1 2013-08-26 Jan Hubicka <jh@suse.cz>
3 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and formating;
5 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
6 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8 (resolve_noninline_speculation): Update callee keys, too.
10 2013-08-26 Jan Hubicka <jh@suse.cz>
12 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
13 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
15 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
17 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
20 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
22 * config/i386/i386.c (ix86_debug_options): Remove prototype.
23 (x86_64_elf_select_section): Ditto.
24 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
26 (ix86_pass_by_reference): Ditto.
27 (output_set_got): Ditto.
28 (ix86_unary_operator_ok): Ditto.
29 (ix86_expand_builtin): Ditto.
31 2013-08-23 Jan Hubicka <jh@suse.cz>
33 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
35 2013-08-23 Jan Hubicka <jh@suse.cz>
37 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
38 (tree_decl_with_vis): Add FINAL field.
40 2013-08-23 Jeff Law <law@redhat.com>
42 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
43 the debugging dump when the expression is fully redundant.
45 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
47 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
48 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
49 * pretty-print.c (pp_formatted_text_data): Likewise.
50 (pp_write_text_to_stream): Likewise.
51 (pp_write_text_as_dot_label_to_stream): Likewise.
52 (pp_append_r): Likewise.
53 (pp_format): Likewise.
55 (pp_clear_output_area): Likewise.
56 (pp_append_text): Likewise.
57 (pp_formatted_text): Likewise.
58 (pp_remaining_character_count_for_line): Likewise.
59 (pp_newline): Likewise.
60 (pp_character): Likewise.
61 (output_buffer::~output_buffer): Define.
62 (pretty_printer::~pretty_printer): Destruct output buffer.
63 * pretty-print.h (output_buffer::~output_buffer): Declare.
64 (pretty_printer::~pretty_printer): Declare virtual.
66 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
69 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
70 HOST_WIDE_INT_M1U): New macros.
71 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
72 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
73 unsigned -1 for lshift.
74 * cse.c (cse_insn): Likewise.
75 * double-int.c (rshift_double, lshift_double): Likewise.
76 * builtins.c (fold_builtin_bitop): Likewise.
77 * combine.c (force_to_mode): Likewise.
78 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
79 * simplify-rtx.c (simplify_const_unary_operation,
80 simplify_const_binary_operation): Likewise.
81 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
82 check_va_list_escapes): Likewise.
83 * rtlanal.c (nonzero_bits1): Likewise.
84 * expmed.c (expand_smod_pow2): Likewise.
85 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
87 2013-08-23 Jan Hubicka <jh@suse.cz>
89 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
90 as having address taken.
92 2013-08-23 Jan Hubicka <jh@suse.cz>
94 * ipa-utils.h (method_class_type): Declare.
95 * ipa-devirt.c (method_class_type): Export.
97 * cgraphunit.c (analyze_functions): Do basic devirtualization;
98 do not walk base classes of anonymous types.
100 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
102 PR rtl-optimization/58220
104 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
105 handle SEQUENCE insns properly.
107 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
109 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
111 (pp_newline_and_indent): Likewise.
112 (pp_separate_with): Likewise.
113 * pretty-print.c (pp_newline_and_flush): Define.
114 (pp_newline_and_indent): Likewise.
115 (pp_separate_with): Likewise.
117 2013-08-23 Jakub Jelinek <jakub@redhat.com>
120 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
121 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
123 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
125 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
126 * gcc/config/i386/i386.md (*movti_internal): Use
127 predicate to determine if EVEX is needed.
128 (*movsi_internal): Ditto.
129 (*movdf_internal): Ditto.
130 (*movsf_internal): Ditto.
131 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
133 2013-08-23 Jakub Jelinek <jakub@redhat.com>
135 PR tree-optimization/58209
136 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
137 (find_tail_calls): Give up for pointer result types if m is non-NULL.
138 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
139 emit POINTER_PLUS_EXPR.
140 (create_tailcall_accumulator): For pointer result type accumulate in
143 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
145 * configure.ac: Add backslashes missing from the last change.
146 * configure: Regenerate.
148 2013-08-22 Jan Hubicka <jh@susue.cz>
150 * ipa.c (function_and_variable_visibility): First remember function
151 was global and then make it local.
153 2013-08-22 Julian Brown <julian@codesourcery.com>
155 * configure.ac: Add aarch64 to list of arches which use "nop" in
157 * configure: Regenerate.
159 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
161 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
162 gnu_libc_has_function.
163 * config/s390/tpf.h: Likewise.
165 2013-08-22 Jan Hubicka <jh@susue.cz>
167 * timevar.c (validate_phases): Add cast.
169 2013-08-22 Jan Hubicka <jh@susue.cz>
171 * timevar.c (validate_phases): Use size_t for memory.
172 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
174 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
176 * pretty-print.h (output_buffer::output_buffer): Declare.
177 (pretty_printer::pretty_printer): Likewise.
178 (pp_construct): Remove.
179 * pretty-print.c (output_buffer::output_buffer): Define.
180 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
181 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
182 (print_gimple_expr): Likewise.
183 (print_gimple_seq): Likewise.
184 (gimple_dump_bb): Likewise.
185 * sched-vis.c (dump_value_slim): Likewise.
186 (dump_insn_slim): Likewise.
187 (dump_rtl_slim): Likewise.
188 (str_pattern_slim): Likewise.
189 * tree-mudflap.c (mf_varname_tree): Likewise.
190 * graph.c (print_graph_cfg): Likewise.
191 (start_graph_dump): Likewise.
192 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
194 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
195 pretty printer initialization.
196 * coretypes.h (diagnostic_context): Remove superflous type alias
198 (pretty_printer): Likewise. Declare directly as a class.
199 (pretty_print_info): Remove declaration as class.
200 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
201 and pp_clear_output_area.
202 (asan_add_global): Likewise.
204 2013-08-22 Jan Hubicka <jh@suse.cz>
206 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
207 * ipa-utils.h (update_type_inheritance_graph): Declare.
208 (possible_polymorphic_call_target_p): Declare.
209 (possible_polymorphic_call_target_p): New.
210 * ipa-devirt.c: Update toplevel comments.
211 (cached_polymorphic_call_targets): Move up.
212 (odr_type_d): Move ID down.
213 (polymorphic_type_binfo_p): Update comment.
214 (odr_hasher::remove): Likewise;
215 (get_odr_type): Set anonymous_namespace.
216 (dump_odr_type): Dump it.
217 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
218 (maybe_record_node): Record node in cached_polymorphic_call_targets.
219 (record_binfo): Add comment.
220 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
222 (devirt_node_removal_hook): Do not iCE when cache is freed.
223 (possible_polymorphic_call_target_p): New predicate.
224 (update_type_inheritance_graph): New function.
226 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
227 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
228 Sergey Lega <sergey.s.lega@intel.com>
229 Anna Tikhonova <anna.tikhonova@intel.com>
230 Ilya Tocar <ilya.tocar@intel.com>
231 Andrey Turetskiy <andrey.turetskiy@intel.com>
232 Ilya Verbin <ilya.verbin@intel.com>
233 Kirill Yukhin <kirill.yukhin@intel.com>
234 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
236 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
237 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
238 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
239 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
240 (OPTION_MASK_ISA_AVX2_UNSET): Update.
241 (OPTION_MASK_ISA_AVX512F_UNSET): New.
242 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
243 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
244 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
245 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
246 OPT_mavx512pf, OPT_mavx512er cases.
247 * config/i386/constraints.md (v): New constraint.
248 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
249 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
251 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
252 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
253 * config/i386/i386-c.c (ix86_target_macros_internal):
254 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
256 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
257 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
258 * config/i386/i386.c (regclass_map, dbx_register_map)
259 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
260 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
261 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
263 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
264 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
265 -mavx512pf options. Fix formatting.
266 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
267 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
268 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
269 -mavx512cd, -mavx512pf options.
270 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
271 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
272 (ix86_preferred_output_reload_class): Replace SSE_REGS with
274 (ix86_hard_regno_mode_ok): Support 512-bit registers.
275 (ix86_set_reg_reg_cost): Ditto.
276 (x86_order_regs_for_local_alloc): Ditto.
277 (MAX_VECT_LEN): Extend to 64-byte.
278 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
279 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
280 (TARGET_AVX512ER, TARGET_AVX512CD): New.
281 (BIGGEST_ALIGNMENT): Extend to 512-bits.
282 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
283 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
284 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
285 (SSE_REG_MODE_P): Support new modes.
286 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
287 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
288 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
289 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
290 (REG_CLASS_CONTENTS): Add new registers.
291 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
292 (EXT_REX_SSE_REGNO_P): New.
293 (HI_REGISTER_NAMES): Add new registers.
294 * config/i386/i386.md: Define constants for new registers.
295 (mode): Add new 512-bit modes.
296 (prefix): Support evex prefix.
297 (isa): Support avx512f, noavx512f, fma_avx512f.
298 (ssemodesuffix): Add new 512-bit modes.
300 (*movxi_internal_avx512f): Ditto.
301 (*movdi_internal): Replace constraint "x" with the new constraint "v".
303 (*movsi_internal): Likewise.
304 (*movdf_internal): Likewise.
305 (*movsf_internal): Likewise.
306 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
307 (<code><mode>3): Likewise.
308 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
310 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
311 with the new constraint "v".
312 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
314 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
315 with the new constraint "v".
316 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
317 (<sse2>_storedqu<avxsizesuffix>): Likewise.
318 (*<plusminus_insn><mode>3): Likewise.
319 (<sse>_vm<plusminus_insn><mode>3): Likewise.
320 (*mul<mode>3): Likewise.
321 (<sse>_vmmul<mode>3): Likewise.
322 (<sse>_div<mode>3): Likewise.
323 (<sse>_vmdiv<mode>3): Likewise.
324 (<sse>_sqrt<mode>2): Likewise.
325 (<sse>_vmsqrt<mode>2): Likewise.
326 (*<code><mode>3_finite): Likewise.
327 (*<code><mode>3) <smaxmin>: Likewise.
328 (<sse>_vm<code><mode>3): Likewise.
329 (*<code><mode>3) <any_logic>: Likewise.
330 (*fma_fmadd_<mode>): Likewise.
331 (*fma_fmsub_<mode>): Likewise.
332 (*fma_fnmadd_<mode>): Likewise.
333 (*fma_fnmsub_<mode>): Likewise.
334 (*fma_fmaddsub_<mode>): Likewise.
335 (*fma_fmsubadd_<mode>): Likewise.
336 (*fmai_fmadd_<mode>): Likewise.
337 (*fmai_fmsub_<mode>): Likewise.
338 (*fmai_fnmadd_<mode>): Likewise.
339 (*fmai_fnmsub_<mode>): Likewise.
340 (sse_cvtsi2ss): Likewise.
341 (sse_cvtsi2ssq): Likewise.
342 (sse_cvtss2si): Likewise.
343 (sse_cvtss2si_2): Likewise.
344 (sse_cvtss2siq): Likewise.
345 (sse_cvtss2siq_2): Likewise.
346 (sse_cvttss2si): Likewise.
347 (sse_cvtss2siq_2): Likewise.
348 (float<sseintvecmodelower><mode>2): Likewise.
349 (sse2_cvtsd2si_2): Likewise.
350 (sse2_cvtsd2siq_2): Likewise.
351 (*<plusminus_insn><mode>3): Likewise.
352 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
353 (*<sse4_1_avx2>_mul<mode>3): Likewise.
354 (ashr<mode>3): Likewise.
355 (<shift_insn><mode>3): Likewise.
356 (avx2_<code><mode>3): Likewise.
357 (*avx2_<code><mode>3): Likewise.
358 (*andnot<mode>3): Likewise.
359 (*<code><mode>3) <any_logic>: Likewise.
360 (abs<mode>2): Likewise.
361 (avx2_permvar<mode>): Likewise.
362 (avx2_perm<mode>_1): Likewise.
363 (*avx_vpermilp<mode>): Likewise.
364 (avx_vpermilvar<mode>3): Likewise.
365 (avx2_ashrv<mode>): Likewise.
366 (avx2_<shift_insn>v<mode>): Likewise.
367 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
369 * doc/rtl.texi: Document XImode.
371 2013-08-21 Jeff Law <law@redhat.com>
373 * tree-flow.h (register_jump_thread): Pass vector of edges
374 instead of each important edge.
375 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
376 thread path into a vector and pass that to register_jump_thread.
377 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
378 passed in edge vector to the current 3-edge form.
381 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
383 * dce.c (fini_dce): Call df_analyze again just in case
384 delete_unmarked_insns removed anything.
386 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
388 * reload.h (struct reg_equivs): Rename to ..
389 (struct reg_equivs_s): .. this.
391 2013-08-20 Martin Liska <marxin.liska@gmail.com>
393 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
395 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
397 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
399 2013-08-21 Jeff Law <law@redhat.com>
401 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
402 simplify assignments too. If the RHS collapses to a singleton
403 range, then return the value for the range.
405 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
407 * config/i386/sse.md (V16): Rename to...
409 (mov<mode>): Update iterator name.
410 (*mov<mode>_internal): Ditto.
411 (push<mode>1): Ditto.
412 (movmisalign<mode>): Ditto.
414 2013-08-20 Jan Hubicka <jh@suse.cz>
417 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
418 entry for direct edges.
419 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
421 2013-08-20 David Malcolm <dmalcolm@redhat.com>
423 Revert my last two changes, r201865 and r201864:
426 2013-08-20 David Malcolm <dmalcolm@redhat.com>
428 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
429 instances can own GC refs.
431 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
432 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
433 (gcc::context::gt_pch_nx): Likewise.
434 (gcc::context::gt_pch_nx): Likewise.
435 * ggc.h (gt_ggc_mx <T>): New.
436 (gt_pch_nx_with_op <T>): New.
437 (gt_pch_nx <T>): New.
438 * passes.c (opt_pass::gt_ggc_mx): New.
439 (opt_pass::gt_pch_nx): New.
440 (opt_pass::gt_pch_nx_with_op): New.
441 (pass_manager::gt_ggc_mx): New.
442 (pass_manager::gt_pch_nx): New.
443 (pass_manager::gt_pch_nx_with_op): New.
444 (pass_manager::operator new): Use
445 ggc_internal_cleared_alloc_stat rather than xcalloc.
446 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
447 (pass_manager::gt_ggc_mx): New.
448 (pass_manager::gt_pch_nx): New.
449 (pass_manager::gt_pch_nx_with_op): New.
450 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
451 (opt_pass::operator new): New.
452 (opt_pass::gt_ggc_mx): New.
453 (opt_pass::gt_pch_nx): New.
454 (opt_pass::gt_pch_nx_with_op): New.
457 2013-08-20 David Malcolm <dmalcolm@redhat.com>
459 * Makefile.in (GTFILES): Add context.h.
460 * context.c (gcc::context::operator new): New.
461 (gcc::context::gt_ggc_mx): New.
462 (gcc::context::gt_pch_nx): New.
463 (gcc::context::gt_pch_nx): New.
464 * context.h (gcc::context): Add GTY((user)) marking.
465 (gcc::context::operator new): New.
466 (gcc::context::gt_ggc_mx): New.
467 (gcc::context::gt_pch_nx): New.
468 (gcc::context::gt_pch_nx): New.
469 (g): Add GTY marking.
470 (gt_ggc_mx (gcc::context *)): New.
471 (gt_pch_nx (gcc::context *)): New.
472 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
474 * gengtype.c (open_base_files) <ifiles>: Add context.h.
476 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
478 * dce.c (fini_dce): Call df_analyze again just in case
479 delete_unmarked_insns removed anything.
481 2013-08-20 Teresa Johnson <tejohnson@google.com>
483 PR rtl-optimizations/57451
484 * final.c (reemit_insn_block_notes): Prevent lexical blocks
485 from crossing split section boundaries.
487 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
489 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
490 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
491 with MULTLIB_DEFAULTS.
493 2013-08-20 Nick Clifton <nickc@redhat.com>
495 * target.def (narrow_volatile_bitfield): Note that the default
496 value is false, not !TARGET_STRICT_ALIGN.
497 * doc/tm.texi: Regenerate.
499 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
501 Fix LIB_SPEC for systems without libpthread.
503 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
504 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
506 * config/i386/linux-common.h: Likewise.
507 * config/mips/linux-common.h: Likewise.
509 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
511 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
514 2013-08-20 David Malcolm <dmalcolm@redhat.com>
516 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
517 instances can own GC refs.
519 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
520 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
521 (gcc::context::gt_pch_nx): Likewise.
522 (gcc::context::gt_pch_nx): Likewise.
523 * ggc.h (gt_ggc_mx <T>): New.
524 (gt_pch_nx_with_op <T>): New.
525 (gt_pch_nx <T>): New.
526 * passes.c (opt_pass::gt_ggc_mx): New.
527 (opt_pass::gt_pch_nx): New.
528 (opt_pass::gt_pch_nx_with_op): New.
529 (pass_manager::gt_ggc_mx): New.
530 (pass_manager::gt_pch_nx): New.
531 (pass_manager::gt_pch_nx_with_op): New.
532 (pass_manager::operator new): Use
533 ggc_internal_cleared_alloc_stat rather than xcalloc.
534 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
535 (pass_manager::gt_ggc_mx): New.
536 (pass_manager::gt_pch_nx): New.
537 (pass_manager::gt_pch_nx_with_op): New.
538 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
539 (opt_pass::operator new): New.
540 (opt_pass::gt_ggc_mx): New.
541 (opt_pass::gt_pch_nx): New.
542 (opt_pass::gt_pch_nx_with_op): New.
544 2013-08-20 David Malcolm <dmalcolm@redhat.com>
546 * Makefile.in (GTFILES): Add context.h.
547 * context.c (gcc::context::operator new): New.
548 (gcc::context::gt_ggc_mx): New.
549 (gcc::context::gt_pch_nx): New.
550 (gcc::context::gt_pch_nx): New.
551 * context.h (gcc::context): Add GTY((user)) marking.
552 (gcc::context::operator new): New.
553 (gcc::context::gt_ggc_mx): New.
554 (gcc::context::gt_pch_nx): New.
555 (gcc::context::gt_pch_nx): New.
556 (g): Add GTY marking.
557 (gt_ggc_mx (gcc::context *)): New.
558 (gt_pch_nx (gcc::context *)): New.
559 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
561 * gengtype.c (open_base_files) <ifiles>: Add context.h.
563 2013-08-20 Alan Modra <amodra@gmail.com>
566 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
567 (rs6000_emit_epilogue): Likewise.
569 2013-08-19 Dehao Chen (dehao@google.com)
571 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
573 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
574 Jakub Jelinek <jakub@redhat.com>
576 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
577 (BUILT_IN_FABSD64): Likewise.
578 (BUILT_IN_FABSD128): Likewise.
579 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
580 (fold_builtin_1): Likewise.
581 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
582 destination and source operands.
583 (*abstd2_fpr): Likewise.
584 (*nabstd2_fpr): Likewise.
586 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
588 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
591 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
593 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
596 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
598 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
599 * builtins.c (default_libc_has_function): New.
600 (gnu_libc_has_function): Ditto.
601 (no_c99_libc_has_function): Ditto.
602 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
603 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
604 (fold_builtin_sincos): Likewise.
605 (fold_builtin_cexp): Likewise.
606 * builtins.def (DEF_C94_BUILTIN): Likewise.
607 (DEF_C99_BUILTIN): Likewise.
608 (DEF_C99_C90RES_BUILTIN): Likewise.
609 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
610 definitions to using this define.
611 * config/darwin-protos.h (darwin_libc_has_function): New.
612 * config/darwin.c (darwin_libc_has_function): Ditto.
613 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
614 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
615 * config/darwin.h: Ditto.
616 * config/elfos.h: Ditto.
617 * config/freebsd.h: Ditto.
618 * config/i386/cygming.h: Ditto.
619 * config/i386/djgpp.h: Ditto.
620 * config/i386/i386-interix.h: Ditto.
621 * config/microblaze/microblaze.h: Ditto.
622 * config/mmix/mmix.h: Ditto.
623 * config/gnu-user.h: Ditto.
624 * config/ia64/hpux.h: Ditto.
625 * config/pa/pa-hpux.h: Ditto.
626 * config/pdp11/pdp11.h: Ditto.
627 * config/picochip/picochip.h: Ditto.
628 * config/linux.h: Ditto.
629 * config/netbsd.h: Ditto.
630 * config/openbsd.h: Ditto.
631 * config/rs6000/aix43.h: Ditto.
632 * config/rs6000/aix51.h: Ditto.
633 * config/rs6000/aix52.h: Ditto.
634 * config/rs6000/aix53.h: Ditto.
635 * config/rs6000/aix61.h: Ditto.
636 * config/rs6000/darwin.h: Ditto.
637 * config/rs6000/linux.h: Ditto.
638 * config/rs6000/linux64.h: Ditto.
639 * config/s390/tpf.h: Ditto.
640 * config/sol2-10.h: Ditto.
641 * config/sol2.h: Ditto.
642 * config/vms/vms.h: Ditto.
643 * config/vxworks.h: Ditto.
644 * config/linux-android.c (linux_android_libc_has_function):
645 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
646 * config/linux-protos.h (linux_android_libc_has_function):
648 * config/i386/i386.c (ix86_libc_has_function): New.
649 * config/i386/i386-protos.h
650 (ix86_libc_has_function): New declaration.
651 * config/i386/i386.md
652 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
653 ("isinf<mode>2): Likewise.
654 * convert.c (convert_to_integer): Using new target hook
655 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
656 TARGET_C99_FUNCTIONS.
657 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
658 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
659 * coretypes.h (function_class): New enum for different
660 classes of functions.
661 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
662 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
663 (TARGET_HAS_SINCOS): Likewise.
664 (TARGET_LIBC_HAS_FUNCTION): New.
665 * doc/tm.texi: Regenerated.
666 * targhooks.h (default_libc_has_function): New declaration.
667 (no_c99_libc_has_function): Ditto.
668 (gnu_libc_has_function): Ditto.
669 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
670 and TARGET_HAS_SINCOS.
672 2013-08-18 Jan Hubicka <jh@suse.cz>
674 * Makeifle-in (ipa-devirt.o): New.
675 (GTFILES): Add ipa-utils.h and ipa-devirt.c
676 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
677 (analyze_functions): Look into possible targets of polymorphic call.
678 * dumpfile.c (dump_files): Add type-inheritance dump.
679 * dumpfile.h (TDI_inheritance): New.
680 * ipa-devirt.c: New file.
681 * ipa-utils.h (odr_type_d): Forward declare.
682 (odr_type): New type.
683 (build_type_inheritance_graph): Declare.
684 (possible_polymorphic_call_targets): Declare and introduce inline
685 variant when only edge is pased.
686 (dump_possible_polymorphic_call_targets): Likewise.
687 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
688 * tree.c (type_in_anonymous_namespace_p): Break out from ...
689 (types_same_for_odr): ... here.
690 * tree.h (type_in_anonymous_namespace_p): Declare.
692 2013-08-18 Jakub Jelinek <jakub@redhat.com>
694 PR tree-optimization/58006
695 * tree-parloops.c (take_address_of): Don't ICE if get_name
697 (eliminate_local_variables_stmt): Remove clobber stmts.
699 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
701 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
702 error message is issued for an alias to undefined symbol.
704 2013-08-18 Jan Hubicka <jh@suse.cz>
706 * cgraph.c (cgraph_create_indirect_edge): Discover
707 polymorphic calls and record basic info into indirect_info.
708 * gimple-fold.c (gimple_fold_call): When doing BINFO based
709 devirtualization, ignore objc function calls.
710 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
711 call with no parm index info.
712 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
713 * tree.c (virtual_method_call_p): New function.
714 * tree.h (virtual_method_call_p): Declare.
716 2013-08-16 Jan Hubicka <jh@suse.cz>
719 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
721 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
723 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
726 2013-08-16 David Malcolm <dmalcolm@redhat.com>
728 * gengtype.c (type_for_name): Add special-case support for
729 locating types within the "gcc::" namespace.
730 (open_base_files): Emit a "using namespace gcc" directive.
732 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
735 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
736 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
738 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
739 array instead of each individual operand as a separate argument.
740 (emit_fusion_gpr_load): Likewise.
741 (expand_fusion_gpr_load): Add new function declaration.
743 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
744 signature to have the operands passed as an array, instead of as
745 separate arguments. Allow ZERO_EXTEND to be in the memory
746 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
747 depend on the register live/dead flags when peepholes are run.
748 (expand_fusion_gpr_load): New function to be called from the
749 peephole2 pass, to change the register that addis sets to be the
751 (emit_fusion_gpr_load): Change the calling signature to have the
752 operands passed as an array, instead of as separate arguments.
753 Allow ZERO_EXTEND to be in the memory address, and also
754 SIGN_EXTEND if -mpower8-fusion-sign.
756 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
758 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
759 adjust the register addis loads up in the peephole2 pass. Do not
760 depend on the register live/dead state when the peephole pass is done.
762 2013-08-16 David Malcolm <dmalcolm@redhat.com>
764 * gengtype.c (create_user_defined_type): Ensure that the kind
765 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
766 declaration is seen before the GTY((user)) marking.
768 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
771 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
773 2013-08-16 Jan Hubicka <jh@suse.cz>
775 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
776 arugment expected_type.
777 (gimple_fold_call): Use it.
778 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
779 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
780 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
781 (try_make_edge_direct_virtual_call): Likewise.
782 * tree.c (obj_type_ref_class): New.
783 * tree.h (obj_type_ref_class): Use it.
785 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
787 * sched-vis.c (rtl_slim_pp_initialized): Remove.
788 (rtl_slim_pp): Likewise.
789 (init_rtl_slim_pretty_print): Likewise.
790 (dump_value_slim): Don't call it. Use local pretty printer.
791 (dump_insn_slim): Likewise.
792 (dump_rtl_slim): Likewise.
793 (str_pattern_slim): Likewise.
794 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
797 2013-08-16 Jakub Jelinek <jakub@redhat.com>
799 PR tree-optimization/58164
800 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
801 walk gimple_goto_dest of GIMPLE_GOTO.
803 PR tree-optimization/58165
804 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
805 bi_call must be the last stmt in a bb, don't split_block, instead
806 use fallthru edge from it and give up if there is none.
807 Release conds vector when returning early.
809 2013-08-14 Xinliang David Li <davidxl@google.com>
811 * config/i386/i386.c (ix86_option_override_internal):
812 Remove unused variable and field.
814 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
817 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
818 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
819 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
820 and Linux, correct BLKmode alignment when 128-bit alignment is
821 required and compatibility flag is not set.
822 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
823 for zero-size arguments when compatibility flag is not set.
825 2013-08-14 Jakub Jelinek <jakub@redhat.com>
827 PR tree-optimization/58145
828 * tree-sra.c (build_ref_for_offset): If prev_base has
829 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
831 2013-08-14 Xinliang David Li <davidxl@google.com>
833 * config/i386/i386.c (ix86_option_override_internal):
834 Fix uninitialized variable error.
836 2013-08-14 Xinliang David Li <davidxl@google.com>
838 * config/i386/i386.opt: Define two new options.
839 * config/i386/x86-tune.def: Add arch selector field in macros.
840 * config/i386/i386.h: Adjust macro definition.
841 * config/i386/i386.c (ix86_option_override_internal):
843 (parse_mtune_ctrl_str): New function.
844 (set_ix86_tune_features): New function.
845 (ix86_function_specific_restore): Call the new helper function.
847 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
849 PR rtl-optimization/57662
850 * sel-sched.c (code_motion_process_successors): When the current insn
851 is removed after the recursive traversal, break from the loop.
852 Add comments and debug printouts.
854 2013-08-14 Jakub Jelinek <jakub@redhat.com>
855 Alexandre Oliva <aoliva@redhat.com>
858 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
859 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
860 there also UNSPEC_PLTOFF.
862 2013-08-14 Marek Polacek <polacek@redhat.com>
864 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
865 NUM_CONDITIONS bit positions.
867 2013-08-13 Cary Coutant <ccoutant@google.com>
869 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
870 (attr_checksum): Hash vector contents instead of pointer.
871 (attr_checksum_ordered): Likewise.
873 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
875 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
876 when Pmode != word_mode. Add length_address attribute.
877 (sse3_monitor_<mode>): Merge from sse3_monitor and
878 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
879 Pmode != word_mode. Update insn length attribute.
880 * config/i386/i386.c (ix86_option_override_internal): Update
881 ix86_gen_monitor selection for merged sse3_monitor insn.
883 2013-08-13 Julian Brown <julian@codesourcery.com>
885 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
886 perform invalid legitimization on greater-than-word-size modes for
889 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
891 * ira.c (setup_class_translate_array): Use aclass instead of cl
892 for classes not fully covered by allocno classes.
894 2013-08-13 Jakub Jelinek <jakub@redhat.com>
896 PR tree-optimization/57661
897 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
898 * tree-inline.c (tree_function_versioning): Initialize it.
899 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
900 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
901 that is not being copied.
904 * asan.c (instrument_strlen_call): Fix typo in comment.
905 Use char * type even for the lhs of POINTER_PLUS_EXPR.
907 2013-08-13 Steve Ellcey <sellcey@mips.com>
909 * config/mips/mips.md (prefetch): Use lw instead of ld on
910 loongson in 32bit mode.
912 2013-08-13 Nick Clifton <nickc@redhat.com>
914 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
916 2013-08-13 Jan Hubicka <jh@suse.cz>
918 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
919 introduced edge; fix typo in sanity check.
920 (cgraph_resolve_speculation): Export; improve diagnostic.
921 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
922 speculation at type mismatch.
923 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
924 (cgraph_resolve_speculation): Declare.
925 (symtab_can_be_discarded): New function.
926 * value-prof.c (gimple_ic_transform): Remove actual transform code.
927 * ipa-inline-transform.c (speculation_removed): New global var.
928 (clone_inlined_nodes): See if speculation can be removed.
929 (inline_call): If speculations was removed, we growths may not match.
930 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
931 (speculation_useful_p): New function.
932 (resolve_noninline_speculation): New function.
933 (inline_small_functions): Resolve useless speculations.
934 * ipa-inline.h (speculation_useful_p): Declare
935 * ipa.c (can_replace_by_local_alias): Simplify.
936 (ipa_profile): Produce speculative calls in non-lto, too;
937 add simple cost model; produce local aliases.
939 2013-08-13 David Malcolm <dmalcolm@redhat.com>
941 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
944 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
946 * config/i386/i386.c (ix86_function_versions): Use error + inform.
948 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
950 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
951 iterator instead of X87MODEF.
953 2013-08-12 Perez Read <netfirewall@gmail.com>
956 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
957 operand 0 for intel asm alternative.
958 (*movabs<mode>_2): Ditto for operand 1.
960 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
962 * config/aarch64/arm_none.h
963 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
965 2013-08-12 Nick Clifton <nickc@redhat.com>
967 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
969 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
971 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
972 expand for QI/HImode operand to produce more effictive code for
973 unsigned char(short) --> float(double) conversion.
975 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
977 * doc/invoke.texi: Mention that -ftls-model does not force the final
980 2013-08-12 Marek Polacek <polacek@redhat.com>
981 Marc Glisse <marc.glisse@inria.fr>
983 PR tree-optimization/57980
984 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
985 when creating -1 constant.
987 2013-08-10 Jan Hubicka <jh@suse.cz>
989 Workaround binutils PR14342.
990 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
991 (gimple_init_edge_profiler): Likewise.
992 (gimple_gen_ic_func_profiler): Likewise.
994 2013-08-09 Jan Hubicka <jh@suse.cz>
996 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
998 2013-08-09 Xinliang David Li <davidxl@google.com>
1000 * config/i386/stringop.def: New file.
1001 * config/i386/stringop.opt: New file.
1002 * config/i386/i386-opts.h: Include stringopt.def.
1003 * config/i386/i386.opt: Include stringopt.opt.
1004 * config/i386/i386.c (ix86_option_override_internal):
1005 Override default size based stringop inline strategies with options.
1006 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1009 2013-08-09 Jan Hubicka <jh@suse.cz>
1011 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1013 2013-08-09 Jan Hubicka <jh@suse.cz>
1015 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1017 (dump_cgraph_node): Dump profile-id.
1018 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1019 and common_target_probability.
1020 * lto-cgraph.c (lto_output_edge): Stream common targets.
1021 (lto_output_node): Stream profile ids.
1022 (input_node): Stream profile ids.
1023 (input_edge): Stream common targets.
1024 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1025 * ipa.c: Include value-prof.h
1026 (ipa_profile_generate_summary): Turn indirect call statement histograms
1027 into common targets.
1028 (ipa_profile): Turn common targets into speculative edges.
1030 2013-08-09 Jan Hubicka <jh@suse.cz>
1032 * cgraph.h (cgraph_node): Add profile_id.
1033 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1034 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1035 (del_node_map): Update.
1036 (find_func_by_funcdef_no): Replace by ...
1037 (find_func_by_profile_id): ... this one.
1038 (gimple_ic_transform): Do not remove useful histograms when
1039 speculation is not done; dump info when indirect call removal
1041 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1042 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1043 (__gcov_indirect_call_profiler_v2): .. this one.
1044 * profile.h (init_node_map): Update.
1045 * coverage.c (coverage_compute_profile_id): New function.
1046 * coverage.h (coverage_compute_profile_id): Declare.
1047 * tree-profile.c (init_ic_make_global_vars): Make
1048 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1049 (gimple_init_edge_profiler): Update prototype of
1050 __gcov_indirect_call_profiler.
1051 (gimple_gen_ic_func_profiler): Simplify.
1052 (tree_profiling): Use init_node_map
1054 2013-08-09 Jan Hubicka <jh@suse.cz>
1056 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1057 non-speculative refs.
1058 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1059 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1060 (cgraph_set_call_stmt): Likewise.
1061 (cgraph_create_edge_1): Fix release checking compilatoin;
1063 (cgraph_free_edge): Free indirect info.
1064 (cgraph_turn_edge_to_speculative): New function.
1065 (cgraph_speculative_call_info): New function.
1066 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1067 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1068 (dump_cgraph_node): Dump speculation.
1069 (verify_edge_count_and_frequency): Accept speculative edges.
1070 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1071 (verify_cgraph_node): Handle speculation.
1072 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1073 (cgraph_set_call_stmt): Update prototype.
1074 (cgraph_make_edge_direct): Update prototype.
1075 (cgraph_speculative_call_info): Declare.
1076 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1077 to change; update call of ipa_find_references.
1078 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1079 lto_stmt_uid and speculative flags.
1080 (ipa_dump_references): Dump speculation.
1081 (ipa_clone_references): Clone speculative flag.
1082 (ipa_clone_referring): Likewise.
1083 (ipa_clone_ref): New function.
1084 (ipa_find_reference): Look into lto_stmt_uids
1085 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1086 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1087 (ipa_find_reference): Update declaration.
1088 (ipa_clone_ref): Declare.
1089 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1090 stream speculative flag.
1091 (lto_output_ref): Stream statements uids and speculation.
1092 (input_ref): Likewise.
1093 (input_edge): Stream speuclation.
1094 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1095 (cgraph_set_call_stmt_including_clones): Handle speculation.
1096 * ipa-inline.c (heap_edge_removal_hook): New function.
1097 (inline_small_functions): Register it.
1098 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1099 also initialize refs.
1100 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1102 (try_make_edge_direct_simple_call): Likewise.
1103 (try_make_edge_direct_simple_call): Likewise.
1104 (update_indirect_edges_after_inlining): Likewise.
1105 (remove_described_reference): Look proper lto_stmt_uid.
1106 (propagate_controlled_uses): Likewise.
1107 (propagate_controlled_uses): Liekwise.
1108 * tree-inline.c (copy_bb): Copy speculative edges.
1109 (redirect_all_calls): New function.
1110 (copy_cfg_body): Do redirection after loop info is updated.
1111 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1113 2013-08-09 Jan Hubicka <jh@suse.cz>
1115 * lto-streamer-out.c (output_function): Renumber PHIs.
1116 * lto-streamer-in.c (input_function): Likewise.
1118 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1120 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1121 (get_lane_unsigned): Likewise.
1122 (dup_lane_scalar): Likewise.
1123 (get_lane): enable for VALL.
1124 * config/aarch64/aarch64-simd.md
1125 (aarch64_dup_lane_scalar<mode>): Remove.
1126 (aarch64_get_lane_signed<mode>): Likewise.
1127 (aarch64_get_lane_unsigned<mode>): Likewise.
1128 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1129 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1130 (aarch64_get_lane<mode>): Enable for all vector modes.
1131 (aarch64_get_lanedi): Remove misleading constraints.
1132 * config/aarch64/arm_neon.h
1133 (__aarch64_vget_lane_any): Define.
1134 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1135 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1136 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1137 * config/aarch64/iterators.md (VDQQH): New.
1141 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1143 * configure.ac: Add GAS check for LEON instructions on SPARC.
1144 * configure: Regenerate.
1145 * config.in: Likewise.
1146 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1148 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1149 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1151 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1152 (AS_LEON_FLAG): New macro.
1153 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1154 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1155 Deal with LEON and LEON3 for the memory model.
1156 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1157 (atomic_compare_and_swap<mode>_1): Likewise.
1158 (*atomic_compare_and_swap<mode>_1): Likewise.
1160 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1162 * config/arm/neon.md (vcond): Fix floating-point vector
1163 comparisons against 0.
1165 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1167 * lra-constraints.c (emit_spill_move): Remove assert.
1168 (process_alt_operands): Add more debugging
1169 output. Increase reject for spilling into memory. Decrease
1170 reject for reloading scratch.
1171 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1173 2013-08-08 Steve Ellcey <sellcey@mips.com>
1175 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1176 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1177 micromips incompatible. Add nan2008.
1178 (MULTILIB_DIRNAMES): Add nan2008.
1179 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1180 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1181 and micromips incompatible. Add nan2008.
1182 (MULTILIB_DIRNAMES): Add nan2008.
1183 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1185 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1187 PR rtl-optimization/58079
1188 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1189 simplify_comparison has widened a comparison with an integer.
1191 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1193 * config/arm/neon.md (movmisalign<mode>): Disable when we
1194 don't allow unaligned accesses.
1195 (*movmisalign<mode>_neon_store): Likewise.
1196 (*movmisalign<mode>_neon_load): Likewise.
1197 (*movmisalign<mode>_neon_store): Likewise.
1198 (*movmisalign<mode>_neon_load): Likewise.
1200 2013-08-08 Jan Hubicka <jh@suse.cz>
1202 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1203 (make_pass_rebuild_cgraph_edges): Also clear references.
1204 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1205 * ipa-inline-transform.c (inline_transform): Remove all references
1207 * cgraphunit.c (expand_function): Remove all references after
1209 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1210 (ipa_find_reference): Rewrite to iterator.
1211 (remove_stmt_references): Likewise.
1212 (ipa_clear_stmts_in_references): New function.
1213 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1214 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1216 * ipa-split.c (split_function): Remove references in split function.
1218 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1221 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1222 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1224 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1227 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
1228 suggested mode for the assignment isn't compatible with the
1231 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1234 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
1236 2013-08-07 Xinliang David Li <davidxl@google.com>
1238 * config/i386/i386.opt: New option -mtune-ctrl=.
1239 * config/i386/x86-tune.def: New file.
1240 * config/i386/i386.h: include x86-tune.def.
1241 * config/i386/i386.c (ix86_option_override_internal):
1242 Parsing -mtune-ctrl= option and set tune features.
1244 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
1247 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
1248 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
1251 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1253 * rtl.h (update_alignments): Declare.
1254 * final.c (grow_label_align): New function extracted from...
1255 (shorten_branches): ...here. Call it.
1256 (update_alignments): New function.
1257 * reorg.c (sibling_labels): New variable.
1258 (get_label_before): Add SIBLING parameter. If it is non-zero, push
1259 the new label along with it onto the sibling_labels vector.
1260 (fill_simple_delay_slots): Adjust call to get_label_before.
1261 (fill_slots_from_thread): Likewise.
1262 (relax_delay_slots): Likewise.
1263 (make_return_insns): Likewise.
1264 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
1266 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1268 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
1269 document its semantics.
1270 (diagnostic_report_diagnostic): Adjust accordingly.
1272 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1274 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
1275 (sparc_option_override): ...and port to new C++ pass API.
1276 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
1278 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
1280 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
1282 2013-08-06 Caroline Tice <cmtice@google.com>
1284 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
1285 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
1286 * tree-pass.h: Add pass_vtable_verify.
1287 * varasm.c (assemble_variable): Add code to properly set the comdat
1288 section and name for the .vtable_map_vars section.
1289 (assemble_vtyv_preinit_initializer): New function.
1290 (default_sectin_type_flags): Make sure .vtable_map_vars section has
1292 * output.h: Add function decl for assemble_vtv_preinit_initializer.
1293 * vtable-verify.c: New file.
1294 * vtable-verify.h: New file.
1295 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
1296 initialiation levels.
1297 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
1298 * passes.def: Insert pass_vtable_verify.
1299 * aclocal.m4: Reorder includes.
1300 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
1301 -fvtv-counts options.
1302 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
1303 as appropriate, if -fvtable-verify=... is used.
1304 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
1305 -fvtable-verify=... is used.
1306 * Makefile.in (OBJS): Add vtable-verify.o to list.
1307 (vtable-verify.o): Add new build rule.
1308 (GTFILES): Add vtable-verify.c to list.
1309 * common.opt (fvtable-verify=): New flag.
1310 (vtv_priority): Values for fvtable-verify= flag.
1311 (fvtv-counts): New flag.
1312 (fvtv-debug): New flag.
1313 * tree.h (save_vtable_map_decl): New extern function decl.
1315 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1317 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
1318 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
1319 (pass_data_rl78_devirt): ...new pass_data instance and...
1320 (make_pass_rl78_devirt): ...new function.
1321 (rl78_asm_file_start): Port pass registration to new C++ API.
1323 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1325 * coretypes.h (rtl_opt_pass): Add.
1326 (gcc::context): Add.
1327 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
1328 (epiphany_init): Port to new C++ pass API.
1329 (epiphany_optimize_mode_switching): Likewise.
1330 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
1331 (pass_manager::get_pass_mode_switching): New.
1332 (pass_manager::get_pass_peephole2): New.
1333 * mode-switching.c (pass_mode_switching): Add clone method.
1334 * recog.c (pass_peephole2): Add clone method.
1335 (pass_split_all_insns): Add clone method.
1337 2013-08-06 David Malcolm <dmalcolm@redhat.com>
1339 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
1340 (mips_option_override): ...here, porting to new C++ API for passes.
1342 2013-08-06 Jan Hubicka <jh@suse.cz>
1344 * cgraph.c (cgraph_get_body): New function based on lto.c
1346 * cgraph.h (cgraph_get_body): Declare.
1347 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
1349 * cgraphunit.c (expand_function): Get body prior expanding.
1350 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
1351 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
1353 * passes.c (do_per_function_toporder): Get body.
1354 * tree-inline.c (expand_call_inline): Get body prior inlining it.
1355 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
1357 2013-08-06 Martin Jambor <mjambor@suse.cz>
1360 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
1361 is not re-finalized. Rename second parameter to no_collect.
1363 2013-08-06 Martin Jambor <mjambor@suse.cz>
1366 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
1367 MEM_REF has proper alignment information.
1369 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
1372 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
1374 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
1375 rtx (*) (rtx, ...) with insn_gen_fn.
1376 * genoutput.c (output_insn_data): Cast gen_? function pointers to
1377 insn_gen_fn::stored_funcptr. Add initializer braces.
1379 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1381 Rewrite how instances of passes are cloned to remove assumptions
1382 about their sizes (thus allowing pass subclasses to have
1383 additional data fields, albeit non-GC-managed ones at this point).
1385 * passes.c (make_pass_instance): Now that passes have clone
1386 methods, rewrite this function to eliminate XNEW and memcpy
1387 calls that used hardcoded sizes. Since this function no longer
1388 creates pass instances, rename it to...
1389 (add_pass_instance): ...this. Document the old way that passes were
1390 numbered and flagged, and rework this function to continue using it.
1391 (next_pass_1): Add an initial_pass argument for use by
1393 (position_pass): When adding multiple instances of a pass, use
1394 the pass's clone method, rather than relying on the XNEW/memcpy
1395 within the former make_pass_instance (now add_pass_instance).
1396 (pass_manager::pass_manager): When invoking next_pass_1, also supply
1397 the initial instance of the current pass within the pass manager.
1399 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1401 This is the automated part of the conversion of passes from C
1402 structs to C++ classes.
1404 Patch autogenerated by refactor_passes.py from
1405 https://github.com/davidmalcolm/gcc-refactoring-scripts
1406 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
1408 * asan.c (pass_asan): Convert from a global struct to a subclass of
1409 gimple_opt_pass along with...
1410 (pass_data_asan): ...new pass_data instance and...
1411 (make_pass_asan): ...new function.
1412 (pass_asan_O0): Convert from a global struct to a subclass of
1413 gimple_opt_pass along with...
1414 (pass_data_asan_O0): ...new pass_data instance and...
1415 (make_pass_asan_O0): ...new function.
1416 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
1417 subclass of rtl_opt_pass along with...
1418 (pass_data_inc_dec): ...new pass_data instance and...
1419 (make_pass_inc_dec): ...new function.
1420 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
1421 a subclass of rtl_opt_pass along with...
1422 (pass_data_reorder_blocks): ...new pass_data instance and...
1423 (make_pass_reorder_blocks): ...new function.
1424 (pass_duplicate_computed_gotos): Convert from a global struct to a
1425 subclass of rtl_opt_pass along with...
1426 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
1427 (make_pass_duplicate_computed_gotos): ...new function.
1428 (pass_partition_blocks): Convert from a global struct to a subclass of
1429 rtl_opt_pass along with...
1430 (pass_data_partition_blocks): ...new pass_data instance and...
1431 (make_pass_partition_blocks): ...new function.
1432 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
1433 struct to a subclass of rtl_opt_pass along with...
1434 (pass_data_branch_target_load_optimize1): ...new pass_data instance
1436 (make_pass_branch_target_load_optimize1): ...new function.
1437 (pass_branch_target_load_optimize2): Convert from a global struct to a
1438 subclass of rtl_opt_pass along with...
1439 (pass_data_branch_target_load_optimize2): ...new pass_data instance
1441 (make_pass_branch_target_load_optimize2): ...new function.
1442 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
1443 of rtl_opt_pass along with...
1444 (pass_data_jump): ...new pass_data instance and...
1445 (make_pass_jump): ...new function.
1446 (pass_jump2): Convert from a global struct to a subclass of
1447 rtl_opt_pass along with...
1448 (pass_data_jump2): ...new pass_data instance and...
1449 (make_pass_jump2): ...new function.
1450 * cfgexpand.c (pass_expand): Convert from a global struct to a
1451 subclass of rtl_opt_pass along with...
1452 (pass_data_expand): ...new pass_data instance and...
1453 (make_pass_expand): ...new function.
1454 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
1455 of rtl_opt_pass along with...
1456 (pass_data_free_cfg): ...new pass_data instance and...
1457 (make_pass_free_cfg): ...new function.
1458 (pass_into_cfg_layout_mode): Convert from a global struct to a
1459 subclass of rtl_opt_pass along with...
1460 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
1461 (make_pass_into_cfg_layout_mode): ...new function.
1462 (pass_outof_cfg_layout_mode): Convert from a global struct to a
1463 subclass of rtl_opt_pass along with...
1464 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
1465 (make_pass_outof_cfg_layout_mode): ...new function.
1466 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
1467 struct to a subclass of gimple_opt_pass along with...
1468 (pass_data_build_cgraph_edges): ...new pass_data instance and...
1469 (make_pass_build_cgraph_edges): ...new function.
1470 (pass_rebuild_cgraph_edges): Convert from a global struct to a
1471 subclass of gimple_opt_pass along with...
1472 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
1473 (make_pass_rebuild_cgraph_edges): ...new function.
1474 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
1475 subclass of gimple_opt_pass along with...
1476 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
1478 (make_pass_remove_cgraph_callee_edges): ...new function.
1479 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
1480 struct to a subclass of rtl_opt_pass along with...
1481 (pass_data_stack_adjustments): ...new pass_data instance and...
1482 (make_pass_stack_adjustments): ...new function.
1483 * combine.c (pass_combine): Convert from a global struct to a subclass
1484 of rtl_opt_pass along with...
1485 (pass_data_combine): ...new pass_data instance and...
1486 (make_pass_combine): ...new function.
1487 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
1488 global struct to a subclass of rtl_opt_pass along with...
1489 (pass_data_compare_elim_after_reload): ...new pass_data instance
1491 (make_pass_compare_elim_after_reload): ...new function.
1492 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1493 of rtl_opt_pass along with...
1494 (pass_data_rtl_cprop): ...new pass_data instance and...
1495 (make_pass_rtl_cprop): ...new function.
1496 * cse.c (pass_cse): Convert from a global struct to a subclass of
1497 rtl_opt_pass along with...
1498 (pass_data_cse): ...new pass_data instance and...
1499 (make_pass_cse): ...new function.
1500 (pass_cse2): Convert from a global struct to a subclass of
1501 rtl_opt_pass along with...
1502 (pass_data_cse2): ...new pass_data instance and...
1503 (make_pass_cse2): ...new function.
1504 (pass_cse_after_global_opts): Convert from a global struct to a
1505 subclass of rtl_opt_pass along with...
1506 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1507 (make_pass_cse_after_global_opts): ...new function.
1508 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1509 of rtl_opt_pass along with...
1510 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1511 (make_pass_ud_rtl_dce): ...new function.
1512 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1513 rtl_opt_pass along with...
1514 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1515 (make_pass_fast_rtl_dce): ...new function.
1516 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1517 a subclass of rtl_opt_pass along with...
1518 (pass_data_df_initialize_opt): ...new pass_data instance and...
1519 (make_pass_df_initialize_opt): ...new function.
1520 (pass_df_initialize_no_opt): Convert from a global struct to a
1521 subclass of rtl_opt_pass along with...
1522 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1523 (make_pass_df_initialize_no_opt): ...new function.
1524 (pass_df_finish): Convert from a global struct to a subclass of
1525 rtl_opt_pass along with...
1526 (pass_data_df_finish): ...new pass_data instance and...
1527 (make_pass_df_finish): ...new function.
1528 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1529 rtl_opt_pass along with...
1530 (pass_data_rtl_dse1): ...new pass_data instance and...
1531 (make_pass_rtl_dse1): ...new function.
1532 (pass_rtl_dse2): Convert from a global struct to a subclass of
1533 rtl_opt_pass along with...
1534 (pass_data_rtl_dse2): ...new pass_data instance and...
1535 (make_pass_rtl_dse2): ...new function.
1536 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1537 subclass of rtl_opt_pass along with...
1538 (pass_data_dwarf2_frame): ...new pass_data instance and...
1539 (make_pass_dwarf2_frame): ...new function.
1540 * except.c (pass_set_nothrow_function_flags): Convert from a global
1541 struct to a subclass of rtl_opt_pass along with...
1542 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1544 (make_pass_set_nothrow_function_flags): ...new function.
1545 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1546 subclass of rtl_opt_pass along with...
1547 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1549 (make_pass_convert_to_eh_region_ranges): ...new function.
1550 * final.c (pass_compute_alignments): Convert from a global struct to a
1551 subclass of rtl_opt_pass along with...
1552 (pass_data_compute_alignments): ...new pass_data instance and...
1553 (make_pass_compute_alignments): ...new function.
1554 (pass_final): Convert from a global struct to a subclass of
1555 rtl_opt_pass along with...
1556 (pass_data_final): ...new pass_data instance and...
1557 (make_pass_final): ...new function.
1558 (pass_shorten_branches): Convert from a global struct to a subclass of
1559 rtl_opt_pass along with...
1560 (pass_data_shorten_branches): ...new pass_data instance and...
1561 (make_pass_shorten_branches): ...new function.
1562 (pass_clean_state): Convert from a global struct to a subclass of
1563 rtl_opt_pass along with...
1564 (pass_data_clean_state): ...new pass_data instance and...
1565 (make_pass_clean_state): ...new function.
1566 * function.c (pass_instantiate_virtual_regs): Convert from a global
1567 struct to a subclass of rtl_opt_pass along with...
1568 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1569 (make_pass_instantiate_virtual_regs): ...new function.
1570 (pass_leaf_regs): Convert from a global struct to a subclass of
1571 rtl_opt_pass along with...
1572 (pass_data_leaf_regs): ...new pass_data instance and...
1573 (make_pass_leaf_regs): ...new function.
1574 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1575 subclass of rtl_opt_pass along with...
1576 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1578 (make_pass_thread_prologue_and_epilogue): ...new function.
1579 (pass_match_asm_constraints): Convert from a global struct to a
1580 subclass of rtl_opt_pass along with...
1581 (pass_data_match_asm_constraints): ...new pass_data instance and...
1582 (make_pass_match_asm_constraints): ...new function.
1583 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1584 subclass of rtl_opt_pass along with...
1585 (pass_data_rtl_fwprop): ...new pass_data instance and...
1586 (make_pass_rtl_fwprop): ...new function.
1587 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1588 rtl_opt_pass along with...
1589 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1590 (make_pass_rtl_fwprop_addr): ...new function.
1591 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1592 rtl_opt_pass along with...
1593 (pass_data_rtl_pre): ...new pass_data instance and...
1594 (make_pass_rtl_pre): ...new function.
1595 (pass_rtl_hoist): Convert from a global struct to a subclass of
1596 rtl_opt_pass along with...
1597 (pass_data_rtl_hoist): ...new pass_data instance and...
1598 (make_pass_rtl_hoist): ...new function.
1599 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1600 subclass of gimple_opt_pass along with...
1601 (pass_data_lower_cf): ...new pass_data instance and...
1602 (make_pass_lower_cf): ...new function.
1603 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1604 from a global struct to a subclass of gimple_opt_pass along with...
1605 (pass_data_strength_reduction): ...new pass_data instance and...
1606 (make_pass_strength_reduction): ...new function.
1607 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1608 of rtl_opt_pass along with...
1609 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1610 (make_pass_rtl_ifcvt): ...new function.
1611 (pass_if_after_combine): Convert from a global struct to a subclass of
1612 rtl_opt_pass along with...
1613 (pass_data_if_after_combine): ...new pass_data instance and...
1614 (make_pass_if_after_combine): ...new function.
1615 (pass_if_after_reload): Convert from a global struct to a subclass of
1616 rtl_opt_pass along with...
1617 (pass_data_if_after_reload): ...new pass_data instance and...
1618 (make_pass_if_after_reload): ...new function.
1619 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1620 a subclass of rtl_opt_pass along with...
1621 (pass_data_initialize_regs): ...new pass_data instance and...
1622 (make_pass_initialize_regs): ...new function.
1623 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1624 of ipa_opt_pass_d along with...
1625 (pass_data_ipa_cp): ...new pass_data instance and...
1626 (make_pass_ipa_cp): ...new function.
1627 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1628 global struct to a subclass of gimple_opt_pass along with...
1629 (pass_data_inline_parameters): ...new pass_data instance and...
1630 (make_pass_inline_parameters): ...new function.
1631 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1632 subclass of gimple_opt_pass along with...
1633 (pass_data_early_inline): ...new pass_data instance and...
1634 (make_pass_early_inline): ...new function.
1635 (pass_ipa_inline): Convert from a global struct to a subclass of
1636 ipa_opt_pass_d along with...
1637 (pass_data_ipa_inline): ...new pass_data instance and...
1638 (make_pass_ipa_inline): ...new function.
1639 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1640 struct to a subclass of gimple_opt_pass along with...
1641 (pass_data_local_pure_const): ...new pass_data instance and...
1642 (make_pass_local_pure_const): ...new function.
1643 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1644 ipa_opt_pass_d along with...
1645 (pass_data_ipa_pure_const): ...new pass_data instance and...
1646 (make_pass_ipa_pure_const): ...new function.
1647 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1648 to a subclass of ipa_opt_pass_d along with...
1649 (pass_data_ipa_reference): ...new pass_data instance and...
1650 (make_pass_ipa_reference): ...new function.
1651 * ipa-split.c (pass_split_functions): Convert from a global struct to
1652 a subclass of gimple_opt_pass along with...
1653 (pass_data_split_functions): ...new pass_data instance and...
1654 (make_pass_split_functions): ...new function.
1655 (pass_feedback_split_functions): Convert from a global struct to a
1656 subclass of gimple_opt_pass along with...
1657 (pass_data_feedback_split_functions): ...new pass_data instance and...
1658 (make_pass_feedback_split_functions): ...new function.
1659 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1660 global struct to a subclass of simple_ipa_opt_pass along with...
1661 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1663 (make_pass_ipa_function_and_variable_visibility): ...new function.
1664 (pass_ipa_free_inline_summary): Convert from a global struct to a
1665 subclass of simple_ipa_opt_pass along with...
1666 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1667 (make_pass_ipa_free_inline_summary): ...new function.
1668 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1669 subclass of ipa_opt_pass_d along with...
1670 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1672 (make_pass_ipa_whole_program_visibility): ...new function.
1673 (pass_ipa_profile): Convert from a global struct to a subclass of
1674 ipa_opt_pass_d along with...
1675 (pass_data_ipa_profile): ...new pass_data instance and...
1676 (make_pass_ipa_profile): ...new function.
1677 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1678 ipa_opt_pass_d along with...
1679 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1680 (make_pass_ipa_cdtor_merge): ...new function.
1681 * ira.c (pass_ira): Convert from a global struct to a subclass of
1682 rtl_opt_pass along with...
1683 (pass_data_ira): ...new pass_data instance and...
1684 (make_pass_ira): ...new function.
1685 (pass_reload): Convert from a global struct to a subclass of
1686 rtl_opt_pass along with...
1687 (pass_data_reload): ...new pass_data instance and...
1688 (make_pass_reload): ...new function.
1689 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1690 subclass of rtl_opt_pass along with...
1691 (pass_data_cleanup_barriers): ...new pass_data instance and...
1692 (make_pass_cleanup_barriers): ...new function.
1693 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1694 of rtl_opt_pass along with...
1695 (pass_data_loop2): ...new pass_data instance and...
1696 (make_pass_loop2): ...new function.
1697 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1698 rtl_opt_pass along with...
1699 (pass_data_rtl_loop_init): ...new pass_data instance and...
1700 (make_pass_rtl_loop_init): ...new function.
1701 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1702 rtl_opt_pass along with...
1703 (pass_data_rtl_loop_done): ...new pass_data instance and...
1704 (make_pass_rtl_loop_done): ...new function.
1705 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1706 subclass of rtl_opt_pass along with...
1707 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1708 (make_pass_rtl_move_loop_invariants): ...new function.
1709 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1710 rtl_opt_pass along with...
1711 (pass_data_rtl_unswitch): ...new pass_data instance and...
1712 (make_pass_rtl_unswitch): ...new function.
1713 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1714 subclass of rtl_opt_pass along with...
1715 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1717 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1718 (pass_rtl_doloop): Convert from a global struct to a subclass of
1719 rtl_opt_pass along with...
1720 (pass_data_rtl_doloop): ...new pass_data instance and...
1721 (make_pass_rtl_doloop): ...new function.
1722 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1723 a subclass of rtl_opt_pass along with...
1724 (pass_data_lower_subreg): ...new pass_data instance and...
1725 (make_pass_lower_subreg): ...new function.
1726 (pass_lower_subreg2): Convert from a global struct to a subclass of
1727 rtl_opt_pass along with...
1728 (pass_data_lower_subreg2): ...new pass_data instance and...
1729 (make_pass_lower_subreg2): ...new function.
1730 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1731 struct to a subclass of ipa_opt_pass_d along with...
1732 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1733 (make_pass_ipa_lto_gimple_out): ...new function.
1734 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1735 of ipa_opt_pass_d along with...
1736 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1737 (make_pass_ipa_lto_finish_out): ...new function.
1738 * mode-switching.c (pass_mode_switching): Convert from a global struct
1739 to a subclass of rtl_opt_pass along with...
1740 (pass_data_mode_switching): ...new pass_data instance and...
1741 (make_pass_mode_switching): ...new function.
1742 * modulo-sched.c (pass_sms): Convert from a global struct to a
1743 subclass of rtl_opt_pass along with...
1744 (pass_data_sms): ...new pass_data instance and...
1745 (make_pass_sms): ...new function.
1746 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1747 subclass of gimple_opt_pass along with...
1748 (pass_data_expand_omp): ...new pass_data instance and...
1749 (make_pass_expand_omp): ...new function.
1750 (pass_lower_omp): Convert from a global struct to a subclass of
1751 gimple_opt_pass along with...
1752 (pass_data_lower_omp): ...new pass_data instance and...
1753 (make_pass_lower_omp): ...new function.
1754 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1755 of gimple_opt_pass along with...
1756 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1757 (make_pass_diagnose_omp_blocks): ...new function.
1758 * passes.c (pass_early_local_passes): Convert from a global struct to
1759 a subclass of simple_ipa_opt_pass along with...
1760 (pass_data_early_local_passes): ...new pass_data instance and...
1761 (make_pass_early_local_passes): ...new function.
1762 (pass_all_early_optimizations): Convert from a global struct to a
1763 subclass of gimple_opt_pass along with...
1764 (pass_data_all_early_optimizations): ...new pass_data instance and...
1765 (make_pass_all_early_optimizations): ...new function.
1766 (pass_all_optimizations): Convert from a global struct to a subclass
1767 of gimple_opt_pass along with...
1768 (pass_data_all_optimizations): ...new pass_data instance and...
1769 (make_pass_all_optimizations): ...new function.
1770 (pass_all_optimizations_g): Convert from a global struct to a subclass
1771 of gimple_opt_pass along with...
1772 (pass_data_all_optimizations_g): ...new pass_data instance and...
1773 (make_pass_all_optimizations_g): ...new function.
1774 (pass_rest_of_compilation): Convert from a global struct to a subclass
1775 of rtl_opt_pass along with...
1776 (pass_data_rest_of_compilation): ...new pass_data instance and...
1777 (make_pass_rest_of_compilation): ...new function.
1778 (pass_postreload): Convert from a global struct to a subclass of
1779 rtl_opt_pass along with...
1780 (pass_data_postreload): ...new pass_data instance and...
1781 (make_pass_postreload): ...new function.
1782 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1783 subclass of rtl_opt_pass along with...
1784 (pass_data_gcse2): ...new pass_data instance and...
1785 (make_pass_gcse2): ...new function.
1786 * postreload.c (pass_postreload_cse): Convert from a global struct to
1787 a subclass of rtl_opt_pass along with...
1788 (pass_data_postreload_cse): ...new pass_data instance and...
1789 (make_pass_postreload_cse): ...new function.
1790 * predict.c (pass_profile): Convert from a global struct to a subclass
1791 of gimple_opt_pass along with...
1792 (pass_data_profile): ...new pass_data instance and...
1793 (make_pass_profile): ...new function.
1794 (pass_strip_predict_hints): Convert from a global struct to a subclass
1795 of gimple_opt_pass along with...
1796 (pass_data_strip_predict_hints): ...new pass_data instance and...
1797 (make_pass_strip_predict_hints): ...new function.
1798 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1799 of rtl_opt_pass along with...
1800 (pass_data_peephole2): ...new pass_data instance and...
1801 (make_pass_peephole2): ...new function.
1802 (pass_split_all_insns): Convert from a global struct to a subclass of
1803 rtl_opt_pass along with...
1804 (pass_data_split_all_insns): ...new pass_data instance and...
1805 (make_pass_split_all_insns): ...new function.
1806 (pass_split_after_reload): Convert from a global struct to a subclass
1807 of rtl_opt_pass along with...
1808 (pass_data_split_after_reload): ...new pass_data instance and...
1809 (make_pass_split_after_reload): ...new function.
1810 (pass_split_before_regstack): Convert from a global struct to a
1811 subclass of rtl_opt_pass along with...
1812 (pass_data_split_before_regstack): ...new pass_data instance and...
1813 (make_pass_split_before_regstack): ...new function.
1814 (pass_split_before_sched2): Convert from a global struct to a subclass
1815 of rtl_opt_pass along with...
1816 (pass_data_split_before_sched2): ...new pass_data instance and...
1817 (make_pass_split_before_sched2): ...new function.
1818 (pass_split_for_shorten_branches): Convert from a global struct to a
1819 subclass of rtl_opt_pass along with...
1820 (pass_data_split_for_shorten_branches): ...new pass_data instance
1822 (make_pass_split_for_shorten_branches): ...new function.
1823 * ree.c (pass_ree): Convert from a global struct to a subclass of
1824 rtl_opt_pass along with...
1825 (pass_data_ree): ...new pass_data instance and...
1826 (make_pass_ree): ...new function.
1827 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1828 subclass of rtl_opt_pass along with...
1829 (pass_data_stack_regs): ...new pass_data instance and...
1830 (make_pass_stack_regs): ...new function.
1831 (pass_stack_regs_run): Convert from a global struct to a subclass of
1832 rtl_opt_pass along with...
1833 (pass_data_stack_regs_run): ...new pass_data instance and...
1834 (make_pass_stack_regs_run): ...new function.
1835 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1836 subclass of rtl_opt_pass along with...
1837 (pass_data_cprop_hardreg): ...new pass_data instance and...
1838 (make_pass_cprop_hardreg): ...new function.
1839 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1840 subclass of rtl_opt_pass along with...
1841 (pass_data_reginfo_init): ...new pass_data instance and...
1842 (make_pass_reginfo_init): ...new function.
1843 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1844 of rtl_opt_pass along with...
1845 (pass_data_regmove): ...new pass_data instance and...
1846 (make_pass_regmove): ...new function.
1847 * regrename.c (pass_regrename): Convert from a global struct to a
1848 subclass of rtl_opt_pass along with...
1849 (pass_data_regrename): ...new pass_data instance and...
1850 (make_pass_regrename): ...new function.
1851 * reorg.c (pass_delay_slots): Convert from a global struct to a
1852 subclass of rtl_opt_pass along with...
1853 (pass_data_delay_slots): ...new pass_data instance and...
1854 (make_pass_delay_slots): ...new function.
1855 (pass_machine_reorg): Convert from a global struct to a subclass of
1856 rtl_opt_pass along with...
1857 (pass_data_machine_reorg): ...new pass_data instance and...
1858 (make_pass_machine_reorg): ...new function.
1859 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1860 of rtl_opt_pass along with...
1861 (pass_data_sched): ...new pass_data instance and...
1862 (make_pass_sched): ...new function.
1863 (pass_sched2): Convert from a global struct to a subclass of
1864 rtl_opt_pass along with...
1865 (pass_data_sched2): ...new pass_data instance and...
1866 (make_pass_sched2): ...new function.
1867 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1868 to a subclass of rtl_opt_pass along with...
1869 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1870 (make_pass_stack_ptr_mod): ...new function.
1871 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1872 to a subclass of rtl_opt_pass along with...
1873 (pass_data_rtl_store_motion): ...new pass_data instance and...
1874 (make_pass_rtl_store_motion): ...new function.
1875 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1876 of gimple_opt_pass along with...
1877 (pass_data_tracer): ...new pass_data instance and...
1878 (make_pass_tracer): ...new function.
1879 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1880 to a subclass of gimple_opt_pass along with...
1881 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1882 (make_pass_diagnose_tm_blocks): ...new function.
1883 (pass_lower_tm): Convert from a global struct to a subclass of
1884 gimple_opt_pass along with...
1885 (pass_data_lower_tm): ...new pass_data instance and...
1886 (make_pass_lower_tm): ...new function.
1887 (pass_tm_init): Convert from a global struct to a subclass of
1888 gimple_opt_pass along with...
1889 (pass_data_tm_init): ...new pass_data instance and...
1890 (make_pass_tm_init): ...new function.
1891 (pass_tm_mark): Convert from a global struct to a subclass of
1892 gimple_opt_pass along with...
1893 (pass_data_tm_mark): ...new pass_data instance and...
1894 (make_pass_tm_mark): ...new function.
1895 (pass_tm_edges): Convert from a global struct to a subclass of
1896 gimple_opt_pass along with...
1897 (pass_data_tm_edges): ...new pass_data instance and...
1898 (make_pass_tm_edges): ...new function.
1899 (pass_tm_memopt): Convert from a global struct to a subclass of
1900 gimple_opt_pass along with...
1901 (pass_data_tm_memopt): ...new pass_data instance and...
1902 (make_pass_tm_memopt): ...new function.
1903 (pass_ipa_tm): Convert from a global struct to a subclass of
1904 simple_ipa_opt_pass along with...
1905 (pass_data_ipa_tm): ...new pass_data instance and...
1906 (make_pass_ipa_tm): ...new function.
1907 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1908 subclass of gimple_opt_pass along with...
1909 (pass_data_call_cdce): ...new pass_data instance and...
1910 (make_pass_call_cdce): ...new function.
1911 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1912 subclass of gimple_opt_pass along with...
1913 (pass_data_build_cfg): ...new pass_data instance and...
1914 (make_pass_build_cfg): ...new function.
1915 (pass_split_crit_edges): Convert from a global struct to a subclass of
1916 gimple_opt_pass along with...
1917 (pass_data_split_crit_edges): ...new pass_data instance and...
1918 (make_pass_split_crit_edges): ...new function.
1919 (pass_warn_function_return): Convert from a global struct to a
1920 subclass of gimple_opt_pass along with...
1921 (pass_data_warn_function_return): ...new pass_data instance and...
1922 (make_pass_warn_function_return): ...new function.
1923 (pass_warn_function_noreturn): Convert from a global struct to a
1924 subclass of gimple_opt_pass along with...
1925 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1926 (make_pass_warn_function_noreturn): ...new function.
1927 (pass_warn_unused_result): Convert from a global struct to a subclass
1928 of gimple_opt_pass along with...
1929 (pass_data_warn_unused_result): ...new pass_data instance and...
1930 (make_pass_warn_unused_result): ...new function.
1931 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1932 a subclass of gimple_opt_pass along with...
1933 (pass_data_merge_phi): ...new pass_data instance and...
1934 (make_pass_merge_phi): ...new function.
1935 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1936 a subclass of gimple_opt_pass along with...
1937 (pass_data_lower_complex): ...new pass_data instance and...
1938 (make_pass_lower_complex): ...new function.
1939 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1940 gimple_opt_pass along with...
1941 (pass_data_lower_complex_O0): ...new pass_data instance and...
1942 (make_pass_lower_complex_O0): ...new function.
1943 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1944 subclass of gimple_opt_pass along with...
1945 (pass_data_lower_eh): ...new pass_data instance and...
1946 (make_pass_lower_eh): ...new function.
1947 (pass_refactor_eh): Convert from a global struct to a subclass of
1948 gimple_opt_pass along with...
1949 (pass_data_refactor_eh): ...new pass_data instance and...
1950 (make_pass_refactor_eh): ...new function.
1951 (pass_lower_resx): Convert from a global struct to a subclass of
1952 gimple_opt_pass along with...
1953 (pass_data_lower_resx): ...new pass_data instance and...
1954 (make_pass_lower_resx): ...new function.
1955 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1956 of gimple_opt_pass along with...
1957 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1958 (make_pass_lower_eh_dispatch): ...new function.
1959 (pass_cleanup_eh): Convert from a global struct to a subclass of
1960 gimple_opt_pass along with...
1961 (pass_data_cleanup_eh): ...new pass_data instance and...
1962 (make_pass_cleanup_eh): ...new function.
1963 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1964 to a subclass of simple_ipa_opt_pass along with...
1965 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1966 (make_pass_ipa_lower_emutls): ...new function.
1967 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1968 a subclass of gimple_opt_pass along with...
1969 (pass_data_if_conversion): ...new pass_data instance and...
1970 (make_pass_if_conversion): ...new function.
1971 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1972 subclass of gimple_opt_pass along with...
1973 (pass_data_build_ssa): ...new pass_data instance and...
1974 (make_pass_build_ssa): ...new function.
1975 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1976 global struct to a subclass of gimple_opt_pass along with...
1977 (pass_data_loop_distribution): ...new pass_data instance and...
1978 (make_pass_loop_distribution): ...new function.
1979 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1980 subclass of gimple_opt_pass along with...
1981 (pass_data_mudflap_1): ...new pass_data instance and...
1982 (make_pass_mudflap_1): ...new function.
1983 (pass_mudflap_2): Convert from a global struct to a subclass of
1984 gimple_opt_pass along with...
1985 (pass_data_mudflap_2): ...new pass_data instance and...
1986 (make_pass_mudflap_2): ...new function.
1987 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1988 subclass of gimple_opt_pass along with...
1989 (pass_data_mudflap_1): ...new pass_data instance and...
1990 (make_pass_mudflap_1): ...new function.
1991 (pass_mudflap_2): Convert from a global struct to a subclass of
1992 gimple_opt_pass along with...
1993 (pass_data_mudflap_2): ...new pass_data instance and...
1994 (make_pass_mudflap_2): ...new function.
1995 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1996 gimple_opt_pass along with...
1997 (pass_data_nrv): ...new pass_data instance and...
1998 (make_pass_nrv): ...new function.
1999 (pass_return_slot): Convert from a global struct to a subclass of
2000 gimple_opt_pass along with...
2001 (pass_data_return_slot): ...new pass_data instance and...
2002 (make_pass_return_slot): ...new function.
2003 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2004 to a subclass of gimple_opt_pass along with...
2005 (pass_data_object_sizes): ...new pass_data instance and...
2006 (make_pass_object_sizes): ...new function.
2007 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2008 global struct to a subclass of gimple_opt_pass along with...
2009 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2011 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2012 (pass_fixup_cfg): Convert from a global struct to a subclass of
2013 gimple_opt_pass along with...
2014 (pass_data_fixup_cfg): ...new pass_data instance and...
2015 (make_pass_fixup_cfg): ...new function.
2016 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2017 (make_pass_mudflap_1): ...new function.
2018 (pass_mudflap_2): Replace declaration with that of...
2019 (make_pass_mudflap_2): ...new function.
2020 (pass_asan): Replace declaration with that of...
2021 (make_pass_asan): ...new function.
2022 (pass_asan_O0): Replace declaration with that of...
2023 (make_pass_asan_O0): ...new function.
2024 (pass_tsan): Replace declaration with that of...
2025 (make_pass_tsan): ...new function.
2026 (pass_tsan_O0): Replace declaration with that of...
2027 (make_pass_tsan_O0): ...new function.
2028 (pass_lower_cf): Replace declaration with that of...
2029 (make_pass_lower_cf): ...new function.
2030 (pass_refactor_eh): Replace declaration with that of...
2031 (make_pass_refactor_eh): ...new function.
2032 (pass_lower_eh): Replace declaration with that of...
2033 (make_pass_lower_eh): ...new function.
2034 (pass_lower_eh_dispatch): Replace declaration with that of...
2035 (make_pass_lower_eh_dispatch): ...new function.
2036 (pass_lower_resx): Replace declaration with that of...
2037 (make_pass_lower_resx): ...new function.
2038 (pass_build_cfg): Replace declaration with that of...
2039 (make_pass_build_cfg): ...new function.
2040 (pass_early_tree_profile): Replace declaration with that of...
2041 (make_pass_early_tree_profile): ...new function.
2042 (pass_cleanup_eh): Replace declaration with that of...
2043 (make_pass_cleanup_eh): ...new function.
2044 (pass_sra): Replace declaration with that of...
2045 (make_pass_sra): ...new function.
2046 (pass_sra_early): Replace declaration with that of...
2047 (make_pass_sra_early): ...new function.
2048 (pass_early_ipa_sra): Replace declaration with that of...
2049 (make_pass_early_ipa_sra): ...new function.
2050 (pass_tail_recursion): Replace declaration with that of...
2051 (make_pass_tail_recursion): ...new function.
2052 (pass_tail_calls): Replace declaration with that of...
2053 (make_pass_tail_calls): ...new function.
2054 (pass_tree_loop): Replace declaration with that of...
2055 (make_pass_tree_loop): ...new function.
2056 (pass_tree_loop_init): Replace declaration with that of...
2057 (make_pass_tree_loop_init): ...new function.
2058 (pass_lim): Replace declaration with that of...
2059 (make_pass_lim): ...new function.
2060 (pass_tree_unswitch): Replace declaration with that of...
2061 (make_pass_tree_unswitch): ...new function.
2062 (pass_predcom): Replace declaration with that of...
2063 (make_pass_predcom): ...new function.
2064 (pass_iv_canon): Replace declaration with that of...
2065 (make_pass_iv_canon): ...new function.
2066 (pass_scev_cprop): Replace declaration with that of...
2067 (make_pass_scev_cprop): ...new function.
2068 (pass_empty_loop): Replace declaration with that of...
2069 (make_pass_empty_loop): ...new function.
2070 (pass_record_bounds): Replace declaration with that of...
2071 (make_pass_record_bounds): ...new function.
2072 (pass_graphite): Replace declaration with that of...
2073 (make_pass_graphite): ...new function.
2074 (pass_graphite_transforms): Replace declaration with that of...
2075 (make_pass_graphite_transforms): ...new function.
2076 (pass_if_conversion): Replace declaration with that of...
2077 (make_pass_if_conversion): ...new function.
2078 (pass_loop_distribution): Replace declaration with that of...
2079 (make_pass_loop_distribution): ...new function.
2080 (pass_vectorize): Replace declaration with that of...
2081 (make_pass_vectorize): ...new function.
2082 (pass_slp_vectorize): Replace declaration with that of...
2083 (make_pass_slp_vectorize): ...new function.
2084 (pass_complete_unroll): Replace declaration with that of...
2085 (make_pass_complete_unroll): ...new function.
2086 (pass_complete_unrolli): Replace declaration with that of...
2087 (make_pass_complete_unrolli): ...new function.
2088 (pass_parallelize_loops): Replace declaration with that of...
2089 (make_pass_parallelize_loops): ...new function.
2090 (pass_loop_prefetch): Replace declaration with that of...
2091 (make_pass_loop_prefetch): ...new function.
2092 (pass_iv_optimize): Replace declaration with that of...
2093 (make_pass_iv_optimize): ...new function.
2094 (pass_tree_loop_done): Replace declaration with that of...
2095 (make_pass_tree_loop_done): ...new function.
2096 (pass_ch): Replace declaration with that of...
2097 (make_pass_ch): ...new function.
2098 (pass_ccp): Replace declaration with that of...
2099 (make_pass_ccp): ...new function.
2100 (pass_phi_only_cprop): Replace declaration with that of...
2101 (make_pass_phi_only_cprop): ...new function.
2102 (pass_build_ssa): Replace declaration with that of...
2103 (make_pass_build_ssa): ...new function.
2104 (pass_build_alias): Replace declaration with that of...
2105 (make_pass_build_alias): ...new function.
2106 (pass_build_ealias): Replace declaration with that of...
2107 (make_pass_build_ealias): ...new function.
2108 (pass_dominator): Replace declaration with that of...
2109 (make_pass_dominator): ...new function.
2110 (pass_dce): Replace declaration with that of...
2111 (make_pass_dce): ...new function.
2112 (pass_dce_loop): Replace declaration with that of...
2113 (make_pass_dce_loop): ...new function.
2114 (pass_cd_dce): Replace declaration with that of...
2115 (make_pass_cd_dce): ...new function.
2116 (pass_call_cdce): Replace declaration with that of...
2117 (make_pass_call_cdce): ...new function.
2118 (pass_merge_phi): Replace declaration with that of...
2119 (make_pass_merge_phi): ...new function.
2120 (pass_split_crit_edges): Replace declaration with that of...
2121 (make_pass_split_crit_edges): ...new function.
2122 (pass_pre): Replace declaration with that of...
2123 (make_pass_pre): ...new function.
2124 (pass_profile): Replace declaration with that of...
2125 (make_pass_profile): ...new function.
2126 (pass_strip_predict_hints): Replace declaration with that of...
2127 (make_pass_strip_predict_hints): ...new function.
2128 (pass_lower_complex_O0): Replace declaration with that of...
2129 (make_pass_lower_complex_O0): ...new function.
2130 (pass_lower_complex): Replace declaration with that of...
2131 (make_pass_lower_complex): ...new function.
2132 (pass_lower_vector): Replace declaration with that of...
2133 (make_pass_lower_vector): ...new function.
2134 (pass_lower_vector_ssa): Replace declaration with that of...
2135 (make_pass_lower_vector_ssa): ...new function.
2136 (pass_lower_omp): Replace declaration with that of...
2137 (make_pass_lower_omp): ...new function.
2138 (pass_diagnose_omp_blocks): Replace declaration with that of...
2139 (make_pass_diagnose_omp_blocks): ...new function.
2140 (pass_expand_omp): Replace declaration with that of...
2141 (make_pass_expand_omp): ...new function.
2142 (pass_expand_omp_ssa): Replace declaration with that of...
2143 (make_pass_expand_omp_ssa): ...new function.
2144 (pass_object_sizes): Replace declaration with that of...
2145 (make_pass_object_sizes): ...new function.
2146 (pass_strlen): Replace declaration with that of...
2147 (make_pass_strlen): ...new function.
2148 (pass_fold_builtins): Replace declaration with that of...
2149 (make_pass_fold_builtins): ...new function.
2150 (pass_stdarg): Replace declaration with that of...
2151 (make_pass_stdarg): ...new function.
2152 (pass_early_warn_uninitialized): Replace declaration with that of...
2153 (make_pass_early_warn_uninitialized): ...new function.
2154 (pass_late_warn_uninitialized): Replace declaration with that of...
2155 (make_pass_late_warn_uninitialized): ...new function.
2156 (pass_cse_reciprocals): Replace declaration with that of...
2157 (make_pass_cse_reciprocals): ...new function.
2158 (pass_cse_sincos): Replace declaration with that of...
2159 (make_pass_cse_sincos): ...new function.
2160 (pass_optimize_bswap): Replace declaration with that of...
2161 (make_pass_optimize_bswap): ...new function.
2162 (pass_optimize_widening_mul): Replace declaration with that of...
2163 (make_pass_optimize_widening_mul): ...new function.
2164 (pass_warn_function_return): Replace declaration with that of...
2165 (make_pass_warn_function_return): ...new function.
2166 (pass_warn_function_noreturn): Replace declaration with that of...
2167 (make_pass_warn_function_noreturn): ...new function.
2168 (pass_cselim): Replace declaration with that of...
2169 (make_pass_cselim): ...new function.
2170 (pass_phiopt): Replace declaration with that of...
2171 (make_pass_phiopt): ...new function.
2172 (pass_forwprop): Replace declaration with that of...
2173 (make_pass_forwprop): ...new function.
2174 (pass_phiprop): Replace declaration with that of...
2175 (make_pass_phiprop): ...new function.
2176 (pass_tree_ifcombine): Replace declaration with that of...
2177 (make_pass_tree_ifcombine): ...new function.
2178 (pass_dse): Replace declaration with that of...
2179 (make_pass_dse): ...new function.
2180 (pass_nrv): Replace declaration with that of...
2181 (make_pass_nrv): ...new function.
2182 (pass_rename_ssa_copies): Replace declaration with that of...
2183 (make_pass_rename_ssa_copies): ...new function.
2184 (pass_sink_code): Replace declaration with that of...
2185 (make_pass_sink_code): ...new function.
2186 (pass_fre): Replace declaration with that of...
2187 (make_pass_fre): ...new function.
2188 (pass_check_data_deps): Replace declaration with that of...
2189 (make_pass_check_data_deps): ...new function.
2190 (pass_copy_prop): Replace declaration with that of...
2191 (make_pass_copy_prop): ...new function.
2192 (pass_vrp): Replace declaration with that of...
2193 (make_pass_vrp): ...new function.
2194 (pass_uncprop): Replace declaration with that of...
2195 (make_pass_uncprop): ...new function.
2196 (pass_return_slot): Replace declaration with that of...
2197 (make_pass_return_slot): ...new function.
2198 (pass_reassoc): Replace declaration with that of...
2199 (make_pass_reassoc): ...new function.
2200 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2201 (make_pass_rebuild_cgraph_edges): ...new function.
2202 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2203 (make_pass_remove_cgraph_callee_edges): ...new function.
2204 (pass_build_cgraph_edges): Replace declaration with that of...
2205 (make_pass_build_cgraph_edges): ...new function.
2206 (pass_local_pure_const): Replace declaration with that of...
2207 (make_pass_local_pure_const): ...new function.
2208 (pass_tracer): Replace declaration with that of...
2209 (make_pass_tracer): ...new function.
2210 (pass_warn_unused_result): Replace declaration with that of...
2211 (make_pass_warn_unused_result): ...new function.
2212 (pass_diagnose_tm_blocks): Replace declaration with that of...
2213 (make_pass_diagnose_tm_blocks): ...new function.
2214 (pass_lower_tm): Replace declaration with that of...
2215 (make_pass_lower_tm): ...new function.
2216 (pass_tm_init): Replace declaration with that of...
2217 (make_pass_tm_init): ...new function.
2218 (pass_tm_mark): Replace declaration with that of...
2219 (make_pass_tm_mark): ...new function.
2220 (pass_tm_memopt): Replace declaration with that of...
2221 (make_pass_tm_memopt): ...new function.
2222 (pass_tm_edges): Replace declaration with that of...
2223 (make_pass_tm_edges): ...new function.
2224 (pass_split_functions): Replace declaration with that of...
2225 (make_pass_split_functions): ...new function.
2226 (pass_feedback_split_functions): Replace declaration with that of...
2227 (make_pass_feedback_split_functions): ...new function.
2228 (pass_strength_reduction): Replace declaration with that of...
2229 (make_pass_strength_reduction): ...new function.
2230 (pass_ipa_lower_emutls): Replace declaration with that of...
2231 (make_pass_ipa_lower_emutls): ...new function.
2232 (pass_ipa_function_and_variable_visibility): Replace declaration with
2234 (make_pass_ipa_function_and_variable_visibility): ...new function.
2235 (pass_ipa_tree_profile): Replace declaration with that of...
2236 (make_pass_ipa_tree_profile): ...new function.
2237 (pass_early_local_passes): Replace declaration with that of...
2238 (make_pass_early_local_passes): ...new function.
2239 (pass_ipa_whole_program_visibility): Replace declaration with that
2241 (make_pass_ipa_whole_program_visibility): ...new function.
2242 (pass_ipa_lto_gimple_out): Replace declaration with that of...
2243 (make_pass_ipa_lto_gimple_out): ...new function.
2244 (pass_ipa_increase_alignment): Replace declaration with that of...
2245 (make_pass_ipa_increase_alignment): ...new function.
2246 (pass_ipa_inline): Replace declaration with that of...
2247 (make_pass_ipa_inline): ...new function.
2248 (pass_ipa_free_lang_data): Replace declaration with that of...
2249 (make_pass_ipa_free_lang_data): ...new function.
2250 (pass_ipa_free_inline_summary): Replace declaration with that of...
2251 (make_pass_ipa_free_inline_summary): ...new function.
2252 (pass_ipa_cp): Replace declaration with that of...
2253 (make_pass_ipa_cp): ...new function.
2254 (pass_ipa_reference): Replace declaration with that of...
2255 (make_pass_ipa_reference): ...new function.
2256 (pass_ipa_pure_const): Replace declaration with that of...
2257 (make_pass_ipa_pure_const): ...new function.
2258 (pass_ipa_pta): Replace declaration with that of...
2259 (make_pass_ipa_pta): ...new function.
2260 (pass_ipa_lto_finish_out): Replace declaration with that of...
2261 (make_pass_ipa_lto_finish_out): ...new function.
2262 (pass_ipa_tm): Replace declaration with that of...
2263 (make_pass_ipa_tm): ...new function.
2264 (pass_ipa_profile): Replace declaration with that of...
2265 (make_pass_ipa_profile): ...new function.
2266 (pass_ipa_cdtor_merge): Replace declaration with that of...
2267 (make_pass_ipa_cdtor_merge): ...new function.
2268 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
2270 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2271 (pass_init_datastructures): Replace declaration with that of...
2272 (make_pass_init_datastructures): ...new function.
2273 (pass_fixup_cfg): Replace declaration with that of...
2274 (make_pass_fixup_cfg): ...new function.
2275 (pass_expand): Replace declaration with that of...
2276 (make_pass_expand): ...new function.
2277 (pass_instantiate_virtual_regs): Replace declaration with that of...
2278 (make_pass_instantiate_virtual_regs): ...new function.
2279 (pass_rtl_fwprop): Replace declaration with that of...
2280 (make_pass_rtl_fwprop): ...new function.
2281 (pass_rtl_fwprop_addr): Replace declaration with that of...
2282 (make_pass_rtl_fwprop_addr): ...new function.
2283 (pass_jump): Replace declaration with that of...
2284 (make_pass_jump): ...new function.
2285 (pass_jump2): Replace declaration with that of...
2286 (make_pass_jump2): ...new function.
2287 (pass_lower_subreg): Replace declaration with that of...
2288 (make_pass_lower_subreg): ...new function.
2289 (pass_cse): Replace declaration with that of...
2290 (make_pass_cse): ...new function.
2291 (pass_fast_rtl_dce): Replace declaration with that of...
2292 (make_pass_fast_rtl_dce): ...new function.
2293 (pass_ud_rtl_dce): Replace declaration with that of...
2294 (make_pass_ud_rtl_dce): ...new function.
2295 (pass_rtl_dce): Replace declaration with that of...
2296 (make_pass_rtl_dce): ...new function.
2297 (pass_rtl_dse1): Replace declaration with that of...
2298 (make_pass_rtl_dse1): ...new function.
2299 (pass_rtl_dse2): Replace declaration with that of...
2300 (make_pass_rtl_dse2): ...new function.
2301 (pass_rtl_dse3): Replace declaration with that of...
2302 (make_pass_rtl_dse3): ...new function.
2303 (pass_rtl_cprop): Replace declaration with that of...
2304 (make_pass_rtl_cprop): ...new function.
2305 (pass_rtl_pre): Replace declaration with that of...
2306 (make_pass_rtl_pre): ...new function.
2307 (pass_rtl_hoist): Replace declaration with that of...
2308 (make_pass_rtl_hoist): ...new function.
2309 (pass_rtl_store_motion): Replace declaration with that of...
2310 (make_pass_rtl_store_motion): ...new function.
2311 (pass_cse_after_global_opts): Replace declaration with that of...
2312 (make_pass_cse_after_global_opts): ...new function.
2313 (pass_rtl_ifcvt): Replace declaration with that of...
2314 (make_pass_rtl_ifcvt): ...new function.
2315 (pass_into_cfg_layout_mode): Replace declaration with that of...
2316 (make_pass_into_cfg_layout_mode): ...new function.
2317 (pass_outof_cfg_layout_mode): Replace declaration with that of...
2318 (make_pass_outof_cfg_layout_mode): ...new function.
2319 (pass_loop2): Replace declaration with that of...
2320 (make_pass_loop2): ...new function.
2321 (pass_rtl_loop_init): Replace declaration with that of...
2322 (make_pass_rtl_loop_init): ...new function.
2323 (pass_rtl_move_loop_invariants): Replace declaration with that of...
2324 (make_pass_rtl_move_loop_invariants): ...new function.
2325 (pass_rtl_unswitch): Replace declaration with that of...
2326 (make_pass_rtl_unswitch): ...new function.
2327 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
2328 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2329 (pass_rtl_doloop): Replace declaration with that of...
2330 (make_pass_rtl_doloop): ...new function.
2331 (pass_rtl_loop_done): Replace declaration with that of...
2332 (make_pass_rtl_loop_done): ...new function.
2333 (pass_web): Replace declaration with that of...
2334 (make_pass_web): ...new function.
2335 (pass_cse2): Replace declaration with that of...
2336 (make_pass_cse2): ...new function.
2337 (pass_df_initialize_opt): Replace declaration with that of...
2338 (make_pass_df_initialize_opt): ...new function.
2339 (pass_df_initialize_no_opt): Replace declaration with that of...
2340 (make_pass_df_initialize_no_opt): ...new function.
2341 (pass_reginfo_init): Replace declaration with that of...
2342 (make_pass_reginfo_init): ...new function.
2343 (pass_inc_dec): Replace declaration with that of...
2344 (make_pass_inc_dec): ...new function.
2345 (pass_stack_ptr_mod): Replace declaration with that of...
2346 (make_pass_stack_ptr_mod): ...new function.
2347 (pass_initialize_regs): Replace declaration with that of...
2348 (make_pass_initialize_regs): ...new function.
2349 (pass_combine): Replace declaration with that of...
2350 (make_pass_combine): ...new function.
2351 (pass_if_after_combine): Replace declaration with that of...
2352 (make_pass_if_after_combine): ...new function.
2353 (pass_ree): Replace declaration with that of...
2354 (make_pass_ree): ...new function.
2355 (pass_partition_blocks): Replace declaration with that of...
2356 (make_pass_partition_blocks): ...new function.
2357 (pass_match_asm_constraints): Replace declaration with that of...
2358 (make_pass_match_asm_constraints): ...new function.
2359 (pass_regmove): Replace declaration with that of...
2360 (make_pass_regmove): ...new function.
2361 (pass_split_all_insns): Replace declaration with that of...
2362 (make_pass_split_all_insns): ...new function.
2363 (pass_fast_rtl_byte_dce): Replace declaration with that of...
2364 (make_pass_fast_rtl_byte_dce): ...new function.
2365 (pass_lower_subreg2): Replace declaration with that of...
2366 (make_pass_lower_subreg2): ...new function.
2367 (pass_mode_switching): Replace declaration with that of...
2368 (make_pass_mode_switching): ...new function.
2369 (pass_sms): Replace declaration with that of...
2370 (make_pass_sms): ...new function.
2371 (pass_sched): Replace declaration with that of...
2372 (make_pass_sched): ...new function.
2373 (pass_ira): Replace declaration with that of...
2374 (make_pass_ira): ...new function.
2375 (pass_reload): Replace declaration with that of...
2376 (make_pass_reload): ...new function.
2377 (pass_clean_state): Replace declaration with that of...
2378 (make_pass_clean_state): ...new function.
2379 (pass_branch_prob): Replace declaration with that of...
2380 (make_pass_branch_prob): ...new function.
2381 (pass_value_profile_transformations): Replace declaration with that
2383 (make_pass_value_profile_transformations): ...new function.
2384 (pass_postreload_cse): Replace declaration with that of...
2385 (make_pass_postreload_cse): ...new function.
2386 (pass_gcse2): Replace declaration with that of...
2387 (make_pass_gcse2): ...new function.
2388 (pass_split_after_reload): Replace declaration with that of...
2389 (make_pass_split_after_reload): ...new function.
2390 (pass_branch_target_load_optimize1): Replace declaration with that
2392 (make_pass_branch_target_load_optimize1): ...new function.
2393 (pass_thread_prologue_and_epilogue): Replace declaration with that
2395 (make_pass_thread_prologue_and_epilogue): ...new function.
2396 (pass_stack_adjustments): Replace declaration with that of...
2397 (make_pass_stack_adjustments): ...new function.
2398 (pass_peephole2): Replace declaration with that of...
2399 (make_pass_peephole2): ...new function.
2400 (pass_if_after_reload): Replace declaration with that of...
2401 (make_pass_if_after_reload): ...new function.
2402 (pass_regrename): Replace declaration with that of...
2403 (make_pass_regrename): ...new function.
2404 (pass_cprop_hardreg): Replace declaration with that of...
2405 (make_pass_cprop_hardreg): ...new function.
2406 (pass_reorder_blocks): Replace declaration with that of...
2407 (make_pass_reorder_blocks): ...new function.
2408 (pass_branch_target_load_optimize2): Replace declaration with that
2410 (make_pass_branch_target_load_optimize2): ...new function.
2411 (pass_leaf_regs): Replace declaration with that of...
2412 (make_pass_leaf_regs): ...new function.
2413 (pass_split_before_sched2): Replace declaration with that of...
2414 (make_pass_split_before_sched2): ...new function.
2415 (pass_compare_elim_after_reload): Replace declaration with that of...
2416 (make_pass_compare_elim_after_reload): ...new function.
2417 (pass_sched2): Replace declaration with that of...
2418 (make_pass_sched2): ...new function.
2419 (pass_stack_regs): Replace declaration with that of...
2420 (make_pass_stack_regs): ...new function.
2421 (pass_stack_regs_run): Replace declaration with that of...
2422 (make_pass_stack_regs_run): ...new function.
2423 (pass_df_finish): Replace declaration with that of...
2424 (make_pass_df_finish): ...new function.
2425 (pass_compute_alignments): Replace declaration with that of...
2426 (make_pass_compute_alignments): ...new function.
2427 (pass_duplicate_computed_gotos): Replace declaration with that of...
2428 (make_pass_duplicate_computed_gotos): ...new function.
2429 (pass_variable_tracking): Replace declaration with that of...
2430 (make_pass_variable_tracking): ...new function.
2431 (pass_free_cfg): Replace declaration with that of...
2432 (make_pass_free_cfg): ...new function.
2433 (pass_machine_reorg): Replace declaration with that of...
2434 (make_pass_machine_reorg): ...new function.
2435 (pass_cleanup_barriers): Replace declaration with that of...
2436 (make_pass_cleanup_barriers): ...new function.
2437 (pass_delay_slots): Replace declaration with that of...
2438 (make_pass_delay_slots): ...new function.
2439 (pass_split_for_shorten_branches): Replace declaration with that of...
2440 (make_pass_split_for_shorten_branches): ...new function.
2441 (pass_split_before_regstack): Replace declaration with that of...
2442 (make_pass_split_before_regstack): ...new function.
2443 (pass_convert_to_eh_region_ranges): Replace declaration with that
2445 (make_pass_convert_to_eh_region_ranges): ...new function.
2446 (pass_shorten_branches): Replace declaration with that of...
2447 (make_pass_shorten_branches): ...new function.
2448 (pass_set_nothrow_function_flags): Replace declaration with that of...
2449 (make_pass_set_nothrow_function_flags): ...new function.
2450 (pass_dwarf2_frame): Replace declaration with that of...
2451 (make_pass_dwarf2_frame): ...new function.
2452 (pass_final): Replace declaration with that of...
2453 (make_pass_final): ...new function.
2454 (pass_rtl_seqabstr): Replace declaration with that of...
2455 (make_pass_rtl_seqabstr): ...new function.
2456 (pass_release_ssa_names): Replace declaration with that of...
2457 (make_pass_release_ssa_names): ...new function.
2458 (pass_early_inline): Replace declaration with that of...
2459 (make_pass_early_inline): ...new function.
2460 (pass_inline_parameters): Replace declaration with that of...
2461 (make_pass_inline_parameters): ...new function.
2462 (pass_update_address_taken): Replace declaration with that of...
2463 (make_pass_update_address_taken): ...new function.
2464 (pass_convert_switch): Replace declaration with that of...
2465 (make_pass_convert_switch): ...new function.
2466 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
2467 to a subclass of simple_ipa_opt_pass along with...
2468 (pass_data_ipa_tree_profile): ...new pass_data instance and...
2469 (make_pass_ipa_tree_profile): ...new function.
2470 * tree-sra.c (pass_sra_early): Convert from a global struct to a
2471 subclass of gimple_opt_pass along with...
2472 (pass_data_sra_early): ...new pass_data instance and...
2473 (make_pass_sra_early): ...new function.
2474 (pass_sra): Convert from a global struct to a subclass of
2475 gimple_opt_pass along with...
2476 (pass_data_sra): ...new pass_data instance and...
2477 (make_pass_sra): ...new function.
2478 (pass_early_ipa_sra): Convert from a global struct to a subclass of
2479 gimple_opt_pass along with...
2480 (pass_data_early_ipa_sra): ...new pass_data instance and...
2481 (make_pass_early_ipa_sra): ...new function.
2482 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
2483 subclass of gimple_opt_pass along with...
2484 (pass_data_ccp): ...new pass_data instance and...
2485 (make_pass_ccp): ...new function.
2486 (pass_fold_builtins): Convert from a global struct to a subclass of
2487 gimple_opt_pass along with...
2488 (pass_data_fold_builtins): ...new pass_data instance and...
2489 (make_pass_fold_builtins): ...new function.
2490 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2491 subclass of gimple_opt_pass along with...
2492 (pass_data_copy_prop): ...new pass_data instance and...
2493 (make_pass_copy_prop): ...new function.
2494 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2495 global struct to a subclass of gimple_opt_pass along with...
2496 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2497 (make_pass_rename_ssa_copies): ...new function.
2498 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2499 subclass of gimple_opt_pass along with...
2500 (pass_data_dce): ...new pass_data instance and...
2501 (make_pass_dce): ...new function.
2502 (pass_dce_loop): Convert from a global struct to a subclass of
2503 gimple_opt_pass along with...
2504 (pass_data_dce_loop): ...new pass_data instance and...
2505 (make_pass_dce_loop): ...new function.
2506 (pass_cd_dce): Convert from a global struct to a subclass of
2507 gimple_opt_pass along with...
2508 (pass_data_cd_dce): ...new pass_data instance and...
2509 (make_pass_cd_dce): ...new function.
2510 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2511 subclass of gimple_opt_pass along with...
2512 (pass_data_dominator): ...new pass_data instance and...
2513 (make_pass_dominator): ...new function.
2514 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2515 gimple_opt_pass along with...
2516 (pass_data_phi_only_cprop): ...new pass_data instance and...
2517 (make_pass_phi_only_cprop): ...new function.
2518 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2519 subclass of gimple_opt_pass along with...
2520 (pass_data_dse): ...new pass_data instance and...
2521 (make_pass_dse): ...new function.
2522 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2523 a subclass of gimple_opt_pass along with...
2524 (pass_data_forwprop): ...new pass_data instance and...
2525 (make_pass_forwprop): ...new function.
2526 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2527 struct to a subclass of gimple_opt_pass along with...
2528 (pass_data_tree_ifcombine): ...new pass_data instance and...
2529 (make_pass_tree_ifcombine): ...new function.
2530 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2531 subclass of gimple_opt_pass along with...
2532 (pass_data_ch): ...new pass_data instance and...
2533 (make_pass_ch): ...new function.
2534 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2535 subclass of gimple_opt_pass along with...
2536 (pass_data_tree_loop): ...new pass_data instance and...
2537 (make_pass_tree_loop): ...new function.
2538 (pass_tree_loop_init): Convert from a global struct to a subclass of
2539 gimple_opt_pass along with...
2540 (pass_data_tree_loop_init): ...new pass_data instance and...
2541 (make_pass_tree_loop_init): ...new function.
2542 (pass_lim): Convert from a global struct to a subclass of
2543 gimple_opt_pass along with...
2544 (pass_data_lim): ...new pass_data instance and...
2545 (make_pass_lim): ...new function.
2546 (pass_tree_unswitch): Convert from a global struct to a subclass of
2547 gimple_opt_pass along with...
2548 (pass_data_tree_unswitch): ...new pass_data instance and...
2549 (make_pass_tree_unswitch): ...new function.
2550 (pass_predcom): Convert from a global struct to a subclass of
2551 gimple_opt_pass along with...
2552 (pass_data_predcom): ...new pass_data instance and...
2553 (make_pass_predcom): ...new function.
2554 (pass_vectorize): Convert from a global struct to a subclass of
2555 gimple_opt_pass along with...
2556 (pass_data_vectorize): ...new pass_data instance and...
2557 (make_pass_vectorize): ...new function.
2558 (pass_graphite): Convert from a global struct to a subclass of
2559 gimple_opt_pass along with...
2560 (pass_data_graphite): ...new pass_data instance and...
2561 (make_pass_graphite): ...new function.
2562 (pass_graphite_transforms): Convert from a global struct to a subclass
2563 of gimple_opt_pass along with...
2564 (pass_data_graphite_transforms): ...new pass_data instance and...
2565 (make_pass_graphite_transforms): ...new function.
2566 (pass_check_data_deps): Convert from a global struct to a subclass of
2567 gimple_opt_pass along with...
2568 (pass_data_check_data_deps): ...new pass_data instance and...
2569 (make_pass_check_data_deps): ...new function.
2570 (pass_iv_canon): Convert from a global struct to a subclass of
2571 gimple_opt_pass along with...
2572 (pass_data_iv_canon): ...new pass_data instance and...
2573 (make_pass_iv_canon): ...new function.
2574 (pass_scev_cprop): Convert from a global struct to a subclass of
2575 gimple_opt_pass along with...
2576 (pass_data_scev_cprop): ...new pass_data instance and...
2577 (make_pass_scev_cprop): ...new function.
2578 (pass_record_bounds): Convert from a global struct to a subclass of
2579 gimple_opt_pass along with...
2580 (pass_data_record_bounds): ...new pass_data instance and...
2581 (make_pass_record_bounds): ...new function.
2582 (pass_complete_unroll): Convert from a global struct to a subclass of
2583 gimple_opt_pass along with...
2584 (pass_data_complete_unroll): ...new pass_data instance and...
2585 (make_pass_complete_unroll): ...new function.
2586 (pass_complete_unrolli): Convert from a global struct to a subclass of
2587 gimple_opt_pass along with...
2588 (pass_data_complete_unrolli): ...new pass_data instance and...
2589 (make_pass_complete_unrolli): ...new function.
2590 (pass_parallelize_loops): Convert from a global struct to a subclass
2591 of gimple_opt_pass along with...
2592 (pass_data_parallelize_loops): ...new pass_data instance and...
2593 (make_pass_parallelize_loops): ...new function.
2594 (pass_loop_prefetch): Convert from a global struct to a subclass of
2595 gimple_opt_pass along with...
2596 (pass_data_loop_prefetch): ...new pass_data instance and...
2597 (make_pass_loop_prefetch): ...new function.
2598 (pass_iv_optimize): Convert from a global struct to a subclass of
2599 gimple_opt_pass along with...
2600 (pass_data_iv_optimize): ...new pass_data instance and...
2601 (make_pass_iv_optimize): ...new function.
2602 (pass_tree_loop_done): Convert from a global struct to a subclass of
2603 gimple_opt_pass along with...
2604 (pass_data_tree_loop_done): ...new pass_data instance and...
2605 (make_pass_tree_loop_done): ...new function.
2606 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2607 struct to a subclass of gimple_opt_pass along with...
2608 (pass_data_cse_reciprocals): ...new pass_data instance and...
2609 (make_pass_cse_reciprocals): ...new function.
2610 (pass_cse_sincos): Convert from a global struct to a subclass of
2611 gimple_opt_pass along with...
2612 (pass_data_cse_sincos): ...new pass_data instance and...
2613 (make_pass_cse_sincos): ...new function.
2614 (pass_optimize_bswap): Convert from a global struct to a subclass of
2615 gimple_opt_pass along with...
2616 (pass_data_optimize_bswap): ...new pass_data instance and...
2617 (make_pass_optimize_bswap): ...new function.
2618 (pass_optimize_widening_mul): Convert from a global struct to a
2619 subclass of gimple_opt_pass along with...
2620 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2621 (make_pass_optimize_widening_mul): ...new function.
2622 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2623 subclass of gimple_opt_pass along with...
2624 (pass_data_phiopt): ...new pass_data instance and...
2625 (make_pass_phiopt): ...new function.
2626 (pass_cselim): Convert from a global struct to a subclass of
2627 gimple_opt_pass along with...
2628 (pass_data_cselim): ...new pass_data instance and...
2629 (make_pass_cselim): ...new function.
2630 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2631 subclass of gimple_opt_pass along with...
2632 (pass_data_phiprop): ...new pass_data instance and...
2633 (make_pass_phiprop): ...new function.
2634 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2635 subclass of gimple_opt_pass along with...
2636 (pass_data_pre): ...new pass_data instance and...
2637 (make_pass_pre): ...new function.
2638 (pass_fre): Convert from a global struct to a subclass of
2639 gimple_opt_pass along with...
2640 (pass_data_fre): ...new pass_data instance and...
2641 (make_pass_fre): ...new function.
2642 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2643 subclass of gimple_opt_pass along with...
2644 (pass_data_reassoc): ...new pass_data instance and...
2645 (make_pass_reassoc): ...new function.
2646 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2647 subclass of gimple_opt_pass along with...
2648 (pass_data_sink_code): ...new pass_data instance and...
2649 (make_pass_sink_code): ...new function.
2650 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2651 subclass of gimple_opt_pass along with...
2652 (pass_data_strlen): ...new pass_data instance and...
2653 (make_pass_strlen): ...new function.
2654 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2655 struct to a subclass of gimple_opt_pass along with...
2656 (pass_data_build_alias): ...new pass_data instance and...
2657 (make_pass_build_alias): ...new function.
2658 (pass_build_ealias): Convert from a global struct to a subclass of
2659 gimple_opt_pass along with...
2660 (pass_data_build_ealias): ...new pass_data instance and...
2661 (make_pass_build_ealias): ...new function.
2662 (pass_ipa_pta): Convert from a global struct to a subclass of
2663 simple_ipa_opt_pass along with...
2664 (pass_data_ipa_pta): ...new pass_data instance and...
2665 (make_pass_ipa_pta): ...new function.
2666 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2667 subclass of gimple_opt_pass along with...
2668 (pass_data_uncprop): ...new pass_data instance and...
2669 (make_pass_uncprop): ...new function.
2670 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2671 global struct to a subclass of gimple_opt_pass along with...
2672 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2673 (make_pass_late_warn_uninitialized): ...new function.
2674 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2675 to a subclass of gimple_opt_pass along with...
2676 (pass_data_init_datastructures): ...new pass_data instance and...
2677 (make_pass_init_datastructures): ...new function.
2678 (pass_early_warn_uninitialized): Convert from a global struct to a
2679 subclass of gimple_opt_pass along with...
2680 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2681 (make_pass_early_warn_uninitialized): ...new function.
2682 (pass_update_address_taken): Convert from a global struct to a
2683 subclass of gimple_opt_pass along with...
2684 (pass_data_update_address_taken): ...new pass_data instance and...
2685 (make_pass_update_address_taken): ...new function.
2686 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2687 struct to a subclass of gimple_opt_pass along with...
2688 (pass_data_release_ssa_names): ...new pass_data instance and...
2689 (make_pass_release_ssa_names): ...new function.
2690 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2691 subclass of gimple_opt_pass along with...
2692 (pass_data_stdarg): ...new pass_data instance and...
2693 (make_pass_stdarg): ...new function.
2694 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2695 global struct to a subclass of gimple_opt_pass along with...
2696 (pass_data_convert_switch): ...new pass_data instance and...
2697 (make_pass_convert_switch): ...new function.
2698 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2699 to a subclass of gimple_opt_pass along with...
2700 (pass_data_tail_recursion): ...new pass_data instance and...
2701 (make_pass_tail_recursion): ...new function.
2702 (pass_tail_calls): Convert from a global struct to a subclass of
2703 gimple_opt_pass along with...
2704 (pass_data_tail_calls): ...new pass_data instance and...
2705 (make_pass_tail_calls): ...new function.
2706 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2707 struct to a subclass of gimple_opt_pass along with...
2708 (pass_data_lower_vector): ...new pass_data instance and...
2709 (make_pass_lower_vector): ...new function.
2710 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2711 gimple_opt_pass along with...
2712 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2713 (make_pass_lower_vector_ssa): ...new function.
2714 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2715 to a subclass of gimple_opt_pass along with...
2716 (pass_data_slp_vectorize): ...new pass_data instance and...
2717 (make_pass_slp_vectorize): ...new function.
2718 (pass_ipa_increase_alignment): Convert from a global struct to a
2719 subclass of simple_ipa_opt_pass along with...
2720 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2721 (make_pass_ipa_increase_alignment): ...new function.
2722 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2723 gimple_opt_pass along with...
2724 (pass_data_vrp): ...new pass_data instance and...
2725 (make_pass_vrp): ...new function.
2726 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2727 subclass of simple_ipa_opt_pass along with...
2728 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2729 (make_pass_ipa_free_lang_data): ...new function.
2730 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2731 gimple_opt_pass along with...
2732 (pass_data_tsan): ...new pass_data instance and...
2733 (make_pass_tsan): ...new function.
2734 (pass_tsan_O0): Convert from a global struct to a subclass of
2735 gimple_opt_pass along with...
2736 (pass_data_tsan_O0): ...new pass_data instance and...
2737 (make_pass_tsan_O0): ...new function.
2738 * var-tracking.c (pass_variable_tracking): Convert from a global
2739 struct to a subclass of rtl_opt_pass along with...
2740 (pass_data_variable_tracking): ...new pass_data instance and...
2741 (make_pass_variable_tracking): ...new function.
2742 * web.c (pass_web): Convert from a global struct to a subclass of
2743 rtl_opt_pass along with...
2744 (pass_data_web): ...new pass_data instance and...
2745 (make_pass_web): ...new function.
2746 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2747 declaration with that of...
2748 (make_pass_mode_switch_use): ...new function.
2749 (pass_resolve_sw_modes): Replace declaration with that of...
2750 (make_pass_resolve_sw_modes): ...new function.
2751 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2752 from a global struct to a subclass of rtl_opt_pass along with...
2753 (pass_data_mode_switch_use): ...new pass_data instance and...
2754 (make_pass_mode_switch_use): ...new function.
2755 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2756 from a global struct to a subclass of rtl_opt_pass along with...
2757 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2758 (make_pass_resolve_sw_modes): ...new function.
2759 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2760 struct to a subclass of rtl_opt_pass along with...
2761 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2762 (make_pass_insert_vzeroupper): ...new function.
2763 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2764 global struct to a subclass of rtl_opt_pass along with...
2765 (pass_data_work_around_errata): ...new pass_data instance and...
2766 (make_pass_work_around_errata): ...new function.
2767 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2768 struct to a subclass of rtl_opt_pass along with...
2769 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2770 (make_pass_mips_machine_reorg2): ...new function.
2772 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2774 * passes.c (pass_manager::operator new): New.
2776 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2778 Handwritten part of conversion of passes to C++ classes.
2780 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2781 (toplev.o): Add dep on PASS_MANAGER_H.
2782 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2783 of early local pases to reflect this moving from a global to a
2784 member of gcc::pass_manager.
2785 (cgraph_add_new_function): Likewise.
2786 * lto-cgraph.c (lto_output_node): Update for conversion of
2787 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2788 * passes.c (opt_pass::clone): New.
2789 (opt_pass::gate): New.
2790 (opt_pass::execute): New.
2791 (opt_pass::opt_pass): New.
2792 (pass_manager::execute_early_local_passes): New.
2793 (pass_manager::execute_pass_mode_switching): new.
2794 (finish_optimization_passes): Convert to...
2795 (pass_manager::finish_optimization_passes): ...this.
2796 (finish_optimization_passes): Update for conversion of passes to
2798 (register_dump_files_1): Use has_gate since we cannot portably
2799 check a vtable entry against NULL.
2800 (dump_one_pass): Likewise.
2801 (ipa_write_summaries_2): Likewise.
2802 (ipa_write_optimization_summaries_1): Likewise.
2803 (ipa_read_summaries_1): Likewise.
2804 (ipa_read_optimization_summaries_1): Likewise.
2805 (execute_ipa_stmt_fixups): Likewise.
2806 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2807 pass-creation functions rather than wiring up globals, and
2808 storing the results in fields of pass_manager generated using
2810 (pass_manager::dump_profile_report): Update for conversion of
2811 passes to C++ classes.
2812 (pass_manager::execute_ipa_summary_passes): Likewise.
2813 (execute_one_ipa_transform_pass): Likewise.
2814 (execute_one_pass): Use has_gate and has_execute since we cannot
2815 portably check a vtable entry against NULL.
2816 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2817 (pass_manager): Use pass-instances.def to add fields for the
2818 various pass instances.
2819 * toplev.c (finalize): Update for move of
2820 finish_optimization_passes to a method of gcc::pass_manager.
2821 * toplev.h (finish_optimization_passes): Move to method of class
2823 * tree-pass.h (struct pass_data): New.
2824 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2825 (opt_pass::gate): Convert to virtual function.
2826 (opt_pass::~opt_pass): New.
2827 (opt_pass::clone): New.
2828 (opt_pass::execute): Convert to virtual function.
2829 (opt_pass::opt_pass): New.
2830 (opt_pass::ctxt_): new.
2831 (gimple_opt_pass): Convert to subclass of opt_pass.
2832 (gimple_opt_pass::gimple_opt_pass): New.
2833 (rtl_opt_pass): Convert to subclass of opt_pass.
2834 (rtl_opt_pass::rtl_opt_pass): New.
2835 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2836 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2837 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2838 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2839 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2840 invocation of pass_mode_switching to reflect this moving from a
2841 global to a member of gcc::pass_manager.
2842 (ix86_option_override): Rework how pass_insert_vzeroupper is
2843 added to the pass_manager to reflect autogenerated changes.
2844 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2846 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2848 PR rtl-optimization/57708
2849 * recog.c (peep2_find_free_register): Validate all regs in a
2852 2013-08-05 Jan Hubicka <jh@suse.cz>
2855 * cgraph.c (verify_cgraph_node): Accept local flags from other
2857 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2858 (function_and_variable_visibility): Likewise.
2859 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2861 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2863 * graph.c (init_graph_slim_pretty_print): Remove.
2864 (print_graph_cfg): Do not call it. Use local pretty printer.
2865 (start_graph_dump): Likewise.
2867 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2869 * gimple-pretty-print.c (buffer): Remove.
2870 (initialized): Likewise.
2871 (maybe_init_pretty_print): Likewise.
2872 (print_gimple_stmt): Do not call it. Use non-static local
2873 pretty_printer variable.
2874 (print_gimple_expr): Likewise.
2875 (print_gimple_seq): Likewise.
2876 (gimple_dump_bb): Likewise.
2878 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2880 * asan.c (asan_pp): Remove.
2881 (asan_pp_initialized): Likewise.
2882 (asan_pp_initialize): Likewise.
2883 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2884 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2885 (asan_add_global): Likewise.
2887 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2889 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2890 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2891 * pretty-print.c (pp_formatted_text_data): Likewise.
2892 (pp_indent): Rename from pp_base_indent.
2893 (pp_format): Rename from pp_base_format.
2894 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2895 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2896 (pp_flush): Rename from pp_base_flush.
2897 (pp_set_line_maximum_length): Rename from
2898 pp_base_set_line_maximum_length.
2899 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2900 (pp_set_prefix): Rename from pp_base_set_prefix.
2901 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2902 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2903 (pp_append_text): Rename from pp_base_append_text.
2904 (pp_formatted_text): Rename from pp_base_formatted_text.
2905 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2906 (pp_remaining_character_count_for_line): Rename from
2907 pp_base_remaining_character_count_for_line.
2908 (pp_newline): Rename from pp_base_newline.
2909 (pp_character): Rename from pp_base_character.
2910 (pp_string): Rename from pp_base_string.
2911 (pp_maybe_space): Rename from pp_base_maybe_space.
2912 * asan.c (asan_pp_string): Adjust.
2913 (asan_emit_stack_protection): Likewise.
2914 (asan_add_global): Likewise.
2915 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2916 * tree-mudflap.c (mf_varname_tree): Likewise.
2917 * tree-pretty-print.c (pp_tree_identifier): Rename from
2918 pp_base_tree_identifier.
2919 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2920 Declare as function.
2922 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2924 * pretty-print.h (pp_bar_bar): New.
2925 (pp_ampersand_ampersand): Likewise.
2926 (pp_less_equal): Likewise.
2927 (pp_greater_equal): Likewise.
2928 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2929 printer functions instead of pp_string or operators and punctuators.
2930 (dump_gimple_call): Likewise.
2931 (dump_gimple_omp_for): Likewise.
2932 (dump_gimple_transaction): Likewise.
2933 (dump_gimple_phi): Likewise.
2934 (pp_gimple_stmt_1): Likewise.
2935 * sched-vis.c (print_insn): Likewise.
2936 * tree-mudflap.c (mf_varname_tree): Likewise.
2937 * tree-pretty-print.c (dump_block_node): Likewise.
2938 (dump_generic_node): Likewise.
2940 2013-08-02 Jan Hubicka <jh@suse.cz>
2942 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2944 * lto-streamer-out.c (tree_is_indexable): Results decls and
2945 parm decls are not indexable.
2946 (DFS_write_tree_body): Do not follow args and results.
2947 (hash_tree): Likewise.
2948 (output_functions): Rearrange so struct function is needed
2949 only when real body is output; be able to also ouptut abstract
2950 functions; output DECL_ARGUMENTS and DECL_RESULT.
2951 (lto_output): When not in WPA, ale store abstract functions.
2952 (write_symbol): Do not care about RESULT_DECL.
2953 (output_symbol_p): Handle correctly sbtract decls.
2954 * lto-streamer-in.c (input_function): Rearrange so struct
2955 function can be NULL at entry; allow streaming of
2956 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2957 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2958 sanity check during LTO.
2959 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2960 RESULT_DECl and DECL_ARGUMENTS.
2961 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2964 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2966 * pretty-print.h (pp_underscore): New.
2968 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2969 printer functions instead of pp_character.
2970 (dump_binary_rhs): Likewise.
2971 (dump_ternary_rhs): Likewise.
2972 (dump_gimple_call_args): Likewise.
2973 (pp_points_to_solution): Likewise.
2974 (dump_gimple_call): Likewise.
2975 (dump_gimple_switch): Likewise.
2976 (dump_gimple_cond): Likewise.
2977 (dump_gimple_bind): Likewise.
2978 (dump_gimple_try): Likewise.
2979 (dump_gimple_omp_for): Likewise.
2980 (dump_gimple_omp_continue): Likewise.
2981 (dump_gimple_omp_single): Likewise.
2982 (dump_gimple_omp_sections): Likewise.
2983 (dump_gimple_omp_block): Likewise.
2984 (dump_gimple_omp_critical): Likewise.
2985 (dump_gimple_transaction): Likewise.
2986 (dump_gimple_asm): Likewise.
2987 (dump_gimple_phi): Likewise.
2988 (dump_gimple_omp_parallel): Likewise.
2989 (dump_gimple_omp_task): Likewise.
2990 (dump_gimple_omp_atomic_load): Likewise.
2991 (dump_gimple_omp_atomic_store): Likewise.
2992 (dump_gimple_mem_ops): Likewise.
2993 (pp_gimple_stmt_1): Likewise.
2994 (pp_cfg_jump): Likewise.
2995 (dump_implicit_edges): Likewise.
2996 (gimple_dump_bb_for_graph): Likewise.
2997 * graph.c (draw_cfg_node): Likewise.
2998 * langhooks.c (lhd_print_error_function): Likewise.
2999 * sched-vis.c (print_exp): Likewise.
3000 (print_value): Likewise.
3001 (print_pattern): Likewise.
3002 (print_insn): Likewise.
3003 (rtl_dump_bb_for_graph): Likewise.
3004 * tree-pretty-print.c (dump_function_declaration): Likewise.
3005 (dump_array_domain): Likewise.
3006 (dump_omp_clause): Likewise.
3007 (dump_location): Likewise.
3008 (dump_generic_node): Likewise.
3009 (print_struct_decl): Likewise.
3010 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3012 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3014 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3015 candidate table when replacing a candidate statement.
3016 (replace_rhs_if_not_dup): Likewise.
3017 (replace_one_candidate): Likewise.
3019 2013-08-02 Jan Hubicka <jh@suse.cz>
3020 Martin Liska <marxin.liska@gmail.com>
3022 * cgraphunit.c (add_new_function): Fix logic when adding from
3024 (assemble_thunk): Rename to ...
3025 (expand_thunk); .. this one; export; get it working with
3026 general functions; make produced gimple valid.
3027 * cgraph.h (expand_thunk): Declare.
3029 2013-08-02 Jan Hubicka <jh@suse.cz>
3031 * ipa-cp.c (gather_context_independent_values): Use
3032 ipa_get_param_move_cost.
3033 (get_replacement_map): Remove PARAM; move parameter folding
3035 (create_specialized_node): Update.
3036 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3037 assert that we have gimple body; update move_cost.
3038 (count_formal_params): Assert that we have gimple body.
3039 (ipa_dump_param): New function.
3040 (ipa_alloc_node_params): Break out from ...
3041 (ipa_initialize_node_params): ... here.
3042 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3043 (ipa_write_node_info): Stream move costs.
3044 (ipa_read_node_info): Read move costs.
3045 (ipa_update_after_lto_read): Do not recompute node params.
3046 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3047 (ipa_get_param): Check we are not in WPA.
3048 (ipa_get_param_move_cost): New.
3049 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3050 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3051 parm numbers to be present.
3053 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3055 PR rtl-optimization/58048
3056 * lra-constraints.c (process_alt_operands): Don't check asm
3057 operand on register.
3059 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3061 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3062 the implied StoreLoad barrier for atomic operations if before.
3064 2013-08-02 Jan Hubicka <jh@suse.cz>
3065 Martin Liska <marxin.liska@gmail.com>
3067 * cgraph.c (cgraph_function_body_availability): Do not check
3069 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3070 symtab_node_availability): Declare.
3071 * ipa.c (can_replace_by_local_alias): New.
3072 (function_and_variable_visibility): Use it.
3073 * symtab.c (symtab_for_node_and_aliases,
3074 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3076 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3078 PR rtl-optimization/57963
3079 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3080 (lra_constraints): Use them.
3082 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3084 * config/arm/types.md (define_attr "type"): Add "load_acq"
3086 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3088 (cortex_a53_store1): Likewise.
3090 2013-08-01 Jan Hubicka <jh@suse.cz>
3092 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3093 partitions are not needed.
3095 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3097 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3098 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3099 MAYBE_NON_Q_CLASS_P where appropriate.
3101 2013-08-01 Jan Hubicka <jh@suse.cz>
3103 * cgraph.h (release_function_body): Declare.
3104 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3105 of unused delcarations.
3107 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3109 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3110 RTL form when subtracting a constant.
3112 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3114 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3115 Generate canonical plus rtx with negated immediate instead of minus
3117 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3119 2013-08-01 Jan Hubicka <jh@suse.cz>
3121 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3122 (cgraph_release_function_body): Likewise.
3123 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3124 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3125 to used_as_abstract_origin.
3126 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3127 symbols used as abstract origins.
3128 * cgraphunit.c (analyze_functions): Update.
3129 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3130 used_as_abstract_origin.
3131 * tree-inline.c (tree_function_versioning): Update
3132 used_as_abstract_origin; be ready for DECL_RESULT and
3133 DECL_ARGUMENTS to be NULL.
3135 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3136 for abstract functions.
3137 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3140 2013-08-01 Jan Hubicka <jh@suse.cz>
3142 * profile.c (compute_value_histograms): Fix thinko.
3144 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3146 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3147 aarch-common-protos.h to extra_headers.
3148 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3149 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3150 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3152 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3154 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3155 Include "../arm/types.md". Define "type" attribute for all patterns.
3156 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3159 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3161 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3162 to support power8 load fusion.
3163 (fusion_gpr_mem_load): Likewise.
3165 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3167 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3168 declarations for power8 load fusion.
3169 (emit_fusion_gpr_load): Likewise.
3171 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3172 tuning for power8, turn on fusion mode by default. Turn on sign
3173 extending fusion mode if normal fusion mode is on, and we are at
3175 (fusion_gpr_load_p): New function, return true if we can fuse an
3176 addis instruction with a dependent load to a GPR.
3177 (emit_fusion_gpr_load): Emit the instructions for power8 load
3180 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3181 (VSX load fusion peepholes): New peepholes to fuse together an
3182 addi instruction with a VSX load instruction.
3184 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3185 peepholes to fuse an addis instruction with a load to a GPR base
3186 register. If we are supporting sign extending fusions, convert
3187 sign extending loads to zero extending loads and add an explicit
3190 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3192 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3193 aarch-common-protos.h to extra_headers.
3194 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3195 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3196 (arm_early_store_addr_dep): Likewise.
3197 (arm_no_early_alu_shift_dep): Likewise.
3198 (arm_no_early_alu_shift_value_dep): Likewise.
3199 (arm_no_early_mul_dep): Likewise.
3200 (arm_no_early_store_addr_dep): Likewise.
3201 (arm_mac_accumulator_is_mul_result): Likewise.
3202 (arm_mac_accumulator_is_result): Likewise.
3203 * config/arm/aarch-common.c: ... here. New file.
3204 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3206 (arm_early_store_addr_dep): Likewise.
3207 (arm_no_early_alu_shift_dep): Likewise.
3208 (arm_no_early_alu_shift_value_dep): Likewise.
3209 (arm_no_early_mul_dep): Likewise.
3210 (arm_no_early_store_addr_dep): Likewise.
3211 (arm_mac_accumulator_is_mul_result): Likewise.
3212 (arm_mac_accumulator_is_result): Likewise.
3213 * config/arm/aarch-common-protos.h: ... here. New file.
3214 * config/arm/t-arm (aarch-common.o): Define.
3216 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3218 * config/arm/arm.md: Include new file "types.md".
3219 (define_attr "type"): Move from here to ...
3220 (define_attr "mul32"): Likewise.
3221 (define_attr "mul64"): Likewise.
3222 * config/arm/types.md: ... here. New file.
3224 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
3226 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
3227 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
3229 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3231 * gen-pass-instances.awk: Fix offset of substr().
3233 2013-07-31 David Malcolm <dmalcolm@redhat.com>
3235 * Makefile.in (pass-instances.def): New.
3236 (passes.o): Replace dependency on passes.def with one on
3239 * gen-pass-instances.awk: New.
3241 * passes.c (pass_manager::pass_manager): Use pass-instances.def
3242 rather than passes.def, updating local definition of NEXT_PASS
3243 macro to add an extra NUM parameter (currently unused).
3245 2013-07-30 David Malcolm <dmalcolm@redhat.com>
3247 * Makefile.in (PASS_MANAGER_H): New.
3248 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
3249 (passes.o): Likewise.
3250 (statistics.o): Likewise.
3251 (cgraphunit.o): Likewise.
3252 (context.o): Depend on PASS_MANAGER_H.
3254 * pass_manager.h: New.
3256 * cgraphunit.c (cgraph_add_new_function): Update for moves
3257 of globals to fields of pass_manager.
3258 (analyze_function): Likewise.
3259 (expand_function): Likewise.
3260 (ipa_passes): Likewise.
3261 (compile): Likewise.
3263 * context.c (context::context): New.
3264 * context.h (context::context): New.
3265 (context::get_passes): New.
3266 (context::passes_): New.
3268 * lto-cgraph.c (input_node): Update for moves of globals to
3269 fields of pass_manager.
3271 * passes.c (all_passes): Remove, in favor of a field of the
3272 same name within the new class pass_manager.
3273 (all_small_ipa_passes): Likewise.
3274 (all_lowering_passes): Likewise.
3275 (all_regular_ipa_passes): Likewise.
3276 (all_late_ipa_passes): Likewise.
3277 (all_lto_gen_passes): Likewise.
3278 (passes_by_id): Likewise.
3279 (passes_by_id_size): Likewise.
3280 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3281 the new class pass_manager.
3282 (set_pass_for_id): Convert to...
3283 (pass_manager::set_pass_for_id): ...method.
3284 (get_pass_for_id): Convert to...
3285 (pass_manager::get_pass_for_id): ...method.
3286 (register_one_dump_file): Move body of implementation into...
3287 (pass_manager::register_one_dump_file): ...here.
3288 (register_dump_files_1): Convert to...
3289 (pass_manager::register_dump_files_1): ...method.
3290 (register_dump_files): Convert to...
3291 (pass_manager::register_dump_files): ...method.
3292 (create_pass_tab): Update for moves of globals to fields of
3294 (dump_passes): Move body of implementation into...
3295 (pass_manager::dump_passes): ...here.
3296 (register_pass): Move body of implementation into...
3297 (pass_manager::register_pass): ...here.
3298 (init_optimization_passes): Convert into...
3299 (pass_manager::pass_manager): ...constructor for new
3300 pass_manager class, and initialize the pass_lists array.
3301 (check_profile_consistency): Update for moves of globals to
3302 fields of pass_manager.
3303 (dump_profile_report): Move body of implementation into...
3304 (pass_manager::dump_profile_report): ...here.
3305 (ipa_write_summaries_1): Update for moves of pass lists from
3306 being globals to fields of pass_manager.
3307 (ipa_write_optimization_summaries): Likewise.
3308 (ipa_read_summaries): Likewise.
3309 (ipa_read_optimization_summaries): Likewise.
3310 (execute_all_ipa_stmt_fixups): Likewise.
3312 * statistics.c (statistics_fini): Update for moves of globals to
3313 fields of pass_manager.
3315 * toplev.c (general_init): Replace call to
3316 init_optimization_passes with construction of the pass_manager
3319 * tree-pass.h (all_passes): Remove, in favor of a field of the
3320 same name within the new class pass_manager.
3321 (all_small_ipa_passes): Likewise.
3322 (all_lowering_passes): Likewise.
3323 (all_regular_ipa_passes): Likewise.
3324 (all_lto_gen_passes): Likewise.
3325 (all_late_ipa_passes): Likewise.
3326 (passes_by_id): Likewise.
3327 (passes_by_id_size): Likewise.
3328 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3329 the new class pass_manager.
3330 (get_pass_for_id): Remove.
3332 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3334 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
3337 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3339 * arm.md (mulhi3): New expand pattern.
3341 2013-07-30 Jan Hubicka <jh@suse.cz>
3342 Martin Liska <marxin.liska@gmail.com>
3344 * profile.c (compute_value_histograms): Do not ICE when
3345 there is mismatch only on some counters.
3347 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3349 PR rtl-optimization/57637
3350 * function.c (move_insn_for_shrink_wrap): Also check the
3351 GEN set of the LIVE problem for the liveness analysis
3352 if it exists, otherwise give up.
3354 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3356 PR tree-optimization/57993
3357 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
3358 replaced statement in the candidate table.
3359 (phi_add_costs): Return infinite cost when the hidden basis does
3360 not dominate all phis on which the candidate is dependent.
3361 (replace_one_candidate): Record replaced statement in the
3364 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
3366 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
3367 (ashlv2si3): New expander.
3368 (*ashlv2si3_i): New define_insn_and_split.
3369 * predicates.md (float_operation): Allow patterns with three
3372 PR rtl-optimization/58021
3373 * mode-switching.c (create_pre_exit): Always split off preceding
3374 insns if we are not at the basic block head.
3376 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
3378 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
3379 (UCLIBC_DYNAMIC_LINKER): New macro.
3380 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
3382 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
3383 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
3385 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
3386 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
3387 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
3388 for SF and DF modes. Use ieee_quad_format for TF mode.
3389 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
3390 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
3391 (mips_option_override): Handle `-mnan=legacy'.
3392 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
3393 `-mabs=2008' and `-mnan=2008'.
3394 (OPTION_DEFAULT_SPECS): Add "nan" default.
3395 (ASM_SPEC): Handle `-mnan='.
3396 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
3397 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
3398 comment accordingly.
3399 (neg<mode>2): Likewise.
3400 * config/mips/mips.opt (mabs, mnan): New options.
3401 * doc/install.texi (Configuration): Document `--with-nan=' option.
3402 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
3404 (MIPS Options): Document them.
3405 * config.gcc <mips*-*-*>: Handle `--with-nan='.
3406 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
3407 * configure: Regenerate.
3408 * config.in: Regenerate.
3410 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3412 * config/i386/i386.md (float post-reload splitters): Do not check
3413 for subregs of SSE registers.
3415 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3416 H.J. Lu <hongjiu.lu@intel.com>
3420 * config/i386/i386.md (post-reload splitter
3421 to avoid partial SSE reg dependency stalls): New pattern.
3423 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
3425 * config/s390/s390.md ("movcc"): Swap load and store instructions.
3427 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
3429 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
3430 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
3432 2013-07-26 Cary Coutant <ccoutant@google.com>
3434 * dwarf2out.c (die_checksum_ordered): Don't include template
3435 instantiations in signature.
3436 (is_template_parameter): New function.
3437 (is_template_instantiation): New function.
3438 (generate_skeleton_bottom_up): Don't include template instantiations
3440 (generate_skeleton): Likewise.
3441 (break_out_comdat_types): Move recursive call to break out nested
3443 (prune_unused_types_mark_generic_parms_dies): Call
3444 is_template_parameter.
3446 2013-07-26 Ian Bolton <ian.bolton@arm.com>
3448 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
3449 uses vector registers.
3450 * config/aarch64/iterators.md: Add attributes rtn and vas.
3452 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3453 Richard Earnshaw <richard.earnshaw@arm.com>
3455 * combine.c (simplify_comparison): Re-canonicalize operands
3457 * config/arm/arm.md (movcond_addsi): New splitter.
3459 2013-07-25 Sterling Augustine <saugustine@google.com>
3461 * dwarf2out.c (size_of_pubnames): Move code to...
3462 (include_pubname_in_output): ...here. New.
3463 (want_pubnames): Rearrange.
3464 (output_pubnames): Call include_pubname_in_output. Move assertion.
3466 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3468 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
3470 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3473 * doc/extend.texi: Remove obsolete builtins. Fix
3474 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
3476 2013-07-25 Jan Hubicka <jh@suse.cz>
3478 * cgraph.c (release_function_body): Break out from ...
3479 (cgraph_release_function_body): ... this one; also release DECL_RESULT
3481 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
3482 old_tree in the map.
3483 (create_specialized_node): Update.
3484 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
3486 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
3487 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
3488 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
3489 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3492 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3494 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3495 addsi3_carryin_alt2_<optab>): Correct output template.
3497 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3499 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3500 Adjust for arm_restrict_it.
3501 Remove trailing whitespace.
3503 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3505 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3506 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3508 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3510 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3512 PR rtl-optimization/57960
3513 * lra-constraints.c (process_alt_operands): Use the right mode
3514 when checking strict_low.
3516 2013-07-25 Jan Hubicka <jh@suse.cz>
3518 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3519 * cgraph.c (cgraph_remove_node): Do not release function body
3520 when in cgraph streaming.
3521 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3522 in other partitions are not considered reachable; fix handling of
3525 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3527 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3529 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3534 * config/arm/arm.md ("*sibcall_insn): Replace use of
3535 Ss with US. Adjust output for v5 and v4t.
3536 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3538 * config/arm/constraints.md ("Ss"): Rename to US.
3540 2013-07-25 Terry Guo <terry.guo@arm.com>
3542 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3543 shift_add/shift_sub0/shift_sub1 RTXs.
3545 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3546 Anton Blanchard <anton@au1.ibm.com>
3548 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3549 (altivec_vpks<VI_char>ss): Likewise.
3550 (altivec_vpks<VI_char>us): Likewise.
3551 (altivec_vpku<VI_char>us): Likewise.
3552 (altivec_vpku<VI_char>um): Likewise.
3554 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3556 Introduce context class.
3558 * Makefile.in (CONTEXT_H): New.
3559 (OBJS): Add context.o.
3560 (toplev.o): Add CONTEXT_H to dependencies.
3563 * toplev.c (general_init): Create the singleton gcc::context instance.
3569 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3571 PR rtl-optimization/57968
3572 * mode-switching.c (create_pre_exit): Allow instructions that
3573 don't set a return register to need a non-exit mode.
3575 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3576 Anton Blanchard <anton@au1.ibm.com>
3578 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3579 operands to vperm for little endian.
3580 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3581 of lvsl to create the control mask for a vperm for little endian.
3583 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3584 Anton Blanchard <anton@au1.ibm.com>
3586 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3587 two operands for little-endian.
3589 2013-07-23 Steve Ellcey <sellcey@mips.com>
3591 * config/mips/mips.c (mips_case_values_threshold): New.
3592 (TARGET_CASE_VALUES_THRESHOLD): Define.
3594 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3595 Anton Blanchard <anton@au1.ibm.com>
3597 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3598 selection of field for vector splat in little endian mode.
3600 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3602 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3603 expanders to rs6000.md.
3604 (ior<mode>3): Likewise.
3605 (and<mode>3): Likewise.
3606 (one_cmpl<mode>2): Likewise.
3607 (nor<mode>3): Likewise.
3608 (andc<mode>3): Likewise.
3609 (eqv<mode>3): Likewise.
3610 (nand<mode>3): Likewise.
3611 (orc<mode>3): Likewise.
3613 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3616 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3617 to split multi-word logical operations.
3618 (rs6000_split_logical_di): Likewise.
3619 (rs6000_split_logical): Likewise.
3621 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3622 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3623 and allow TImode operations in 32-bit.
3624 (vsx_and<mode>3_64bit): Likewise.
3625 (vsx_ior<mode>3_32bit): Likewise.
3626 (vsx_ior<mode>3_64bit): Likewise.
3627 (vsx_xor<mode>3_32bit): Likewise.
3628 (vsx_xor<mode>3_64bit): Likewise.
3629 (vsx_one_cmpl<mode>2_32bit): Likewise.
3630 (vsx_one_cmpl<mode>2_64bit): Likewise.
3631 (vsx_nor<mode>3_32bit): Likewise.
3632 (vsx_nor<mode>3_64bit): Likewise.
3633 (vsx_andc<mode>3_32bit): Likewise.
3634 (vsx_andc<mode>3_64bit): Likewise.
3635 (vsx_eqv<mode>3_32bit): Likewise.
3636 (vsx_eqv<mode>3_64bit): Likewise.
3637 (vsx_nand<mode>3_32bit): Likewise.
3638 (vsx_nand<mode>3_64bit): Likewise.
3639 (vsx_orc<mode>3_32bit): Likewise.
3640 (vsx_orc<mode>3_64bit): Likewise.
3642 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3643 logical types in GPRs.
3645 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3646 logical insns to rs6000.md, and allow TImode operations in
3648 (altivec_ior<mode>3): Likewise.
3649 (altivec_xor<mode>3): Likewise.
3650 (altivec_one_cmpl<mode>2): Likewise.
3651 (altivec_nor<mode>3): Likewise.
3652 (altivec_andc<mode>3): Likewise.
3654 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3655 attributes for moving the 128-bit logical operations into
3657 (BOOL_REGS_OUTPUT): Likewise.
3658 (BOOL_REGS_OP1): Likewise.
3659 (BOOL_REGS_OP2): Likewise.
3660 (BOOL_REGS_UNARY): Likewise.
3661 (BOOL_REGS_AND_CR0): Likewise.
3662 (one_cmpl<mode>2): Add support for DI logical operations on
3663 32-bit, splitting the operations to 32-bit.
3667 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3668 changes to combine the 32/64-bit code, allow logical operations on
3669 TI mode in 32-bit, and to use similar match_operator patterns like
3670 scalar mode uses. Combine the Altivec and VSX code for logical
3671 operations, and move it here.
3672 (ior<mode>3, 128-bit types): Likewise.
3673 (xor<mode>3, 128-bit types): Likewise.
3674 (one_cmpl<mode>3, 128-bit types): Likewise.
3675 (nor<mode>3, 128-bit types): Likewise.
3676 (andc<mode>3, 128-bit types): Likewise.
3677 (eqv<mode>3, 128-bit types): Likewise.
3678 (nand<mode>3, 128-bit types): Likewise.
3679 (orc<mode>3, 128-bit types): Likewise.
3680 (and<mode>3_internal): Likewise.
3681 (bool<mode>3_internal): Likewise.
3682 (boolc<mode>3_internal1): Likewise.
3683 (boolc<mode>3_internal2): Likewise.
3684 (boolcc<mode>3_internal1): Likewise.
3685 (boolcc<mode>3_internal2): Likewise.
3686 (eqv<mode>3_internal1): Likewise.
3687 (eqv<mode>3_internal2): Likewise.
3688 (one_cmpl1<mode>3_internal): Likewise.
3690 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3692 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3693 Rename flag_stack_usage to flag_stack_usage_info.
3695 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3697 * config/microblaze/sync.md: New file.
3698 * config/microblaze/microblaze.md: Include sync.md
3699 * config/microblaze/microblaze.c: Add print_operand 'y'.
3700 * config/microblaze/constraints.md: Add memory_contraint
3701 'Q' which is a single register.
3703 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3705 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3707 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3709 * reload.c (find_reloads): Exit loop once we find this operand
3710 cannot be reloaded somehow for this alternative.
3712 * reload.c (find_reloads): Exit loop once we find a hard register.
3714 * rtlanal.c (computed_jump_p): Exit loop once we find label
3717 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3719 * cfgloopmanip.c (remove_path): Exit loop after setting
3722 * gensupport.c (subst_dup): Avoid loop if code is not
3723 MATCH_DUP nor MATCH_OP_DUP.
3725 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3727 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3729 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3731 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3732 true for SP_REGNUM if mode == ptr_mode.
3733 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3734 with value R0_REGNUM + 31.
3736 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3738 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3739 pad pointer-typed argument downward.
3741 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3743 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3744 and __ILP32__ when the ILP32 model is in use.
3746 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3748 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3749 (aarch64_load_symref_appropriately): In the case of
3750 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3751 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3752 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3753 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3754 if 'mode' doesn't equal to 'ptr_mode'.
3755 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3756 'vcall_offset'; change to call aarch64_emit_move differently depending
3757 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3758 to calculate the upper bound of 'vcall_offset'.
3759 (aarch64_cannot_force_const_mem): Change to also return true if
3761 (aarch64_legitimize_reload_address): In the case of large
3762 displacements, add new local variable 'xmode' and an assertion
3763 based on it; change to use 'xmode' to generate the new rtx and
3765 (aarch64_asm_trampoline_template): Change to generate the template
3766 differently depending on TARGET_ILP32 or not; change to use
3767 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3768 (aarch64_trampoline_size): Removed.
3769 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3770 and replace immediate literals with it. Change to use 'ptr_mode'
3771 instead of 'DImode' and call convert_memory_address if the mode
3772 of 'fnaddr' doesn't equal to 'ptr_mode'.
3773 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3775 (aarch64_elf_asm_destructor): Likewise.
3776 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3777 on TARGET_ILP32 instead of aarch64_trampoline_size.
3778 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3779 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3780 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3781 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3782 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3783 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3784 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3785 depending on the value of 'mode'.
3786 (add_losym_<mode>): New.
3787 (ldr_got_small_<mode>): New, based on ldr_got_small.
3788 (ldr_got_small): Remove.
3789 (ldr_got_small_sidi): New.
3790 * config/aarch64/iterators.md (P): New.
3791 (PTR): Change to 'ptr_mode' in the condition.
3793 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3795 * config.gcc (aarch64*-*-*): Support --with-abi.
3796 (aarch64*-*-elf): Support --with-multilib-list.
3797 (aarch64*-*-linux*): Likewise.
3798 (supported_defaults): Add abi to aarch64*-*-*.
3799 * configure.ac: Mention AArch64 for --with-multilib-list.
3800 * configure: Re-generated.
3801 * config/aarch64/biarchilp32.h: New file.
3802 * config/aarch64/biarchlp64.h: New file.
3803 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3805 (MULTILIB_DEFAULTS): Ditto.
3806 (DRIVER_SELF_SPECS): Ditto.
3807 (ASM_SPEC): Update to also substitute -mabi.
3808 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3809 file whose name depends on -mabi= and -mbig-endian.
3810 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3812 (POINTER_SIZE): New define.
3813 (POINTERS_EXTEND_UNSIGNED): Ditto.
3814 (enum aarch64_abi_type): New enumeration tag.
3815 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3816 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3817 (TARGET_ILP32): New define.
3818 * config/aarch64/aarch64.opt (mabi): New.
3820 (ilp32, lp64): New values for -mabi.
3821 * config/aarch64/t-aarch64 (comma): New define.
3822 (MULTILIB_OPTIONS): Ditto.
3823 (MULTILIB_DIRNAMES): Ditto.
3824 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3825 * doc/invoke.texi: Document -mabi for AArch64.
3827 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3829 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3831 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3832 Anton Blanchard <anton@au1.ibm.com>
3834 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3835 endianness when selecting field to splat.
3837 2013-07-22 Eric Christopher <echristo@gmail.com>
3839 * dwarf2out.c (die_odr_checksum): New function to use
3840 CHECKSUM_ macros and ULEB128 for DIE tag.
3841 (generate_type_signature): Use.
3843 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3845 * config.gcc (sparc*-*-*): Accept leon3 processor.
3846 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3847 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3848 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3849 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3850 (mfix-ut699): Adjust comment.
3851 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3852 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3853 (CPP_CPU_SPEC): Likewise.
3854 (ASM_CPU_SPEC): Likewise.
3855 * config/sparc/sparc.c (leon3_cost): New constant.
3856 (sparc_option_override): Add leon3 support.
3857 (mem_ref): New function.
3858 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3859 (sparc_do_work_around_errata): Look into the instruction in the delay
3860 slot and adjust accordingly. Add fix for the data cache nullify issues
3861 of the UT699. Change insertion position for the NOP.
3862 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3863 (leon3_load): New reservation.
3864 (leon_store): Bump latency to 2.
3865 (grfpu): New automaton.
3866 (grfpu_alu): New unit.
3867 (grfpu_ds): Likewise.
3868 (leon_fp_alu): Adjust.
3869 (leon_fp_mult): Delete.
3870 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3871 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3872 * config/sparc/sparc.md (cpu): Add leon3.
3873 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3875 (atomic_test_and_set): Likewise.
3878 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3880 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3881 default for R5900 targets.
3882 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3883 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3884 * config/mips/mips.c (mips_option_override): Report an error for
3885 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3886 for -march=r5900 -mhard-float.
3888 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3890 * df-problems.c (can_move_insns_across): Exit loop once we
3891 find a non-fixed, non-global register.
3893 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3896 * omega.c (omega_eliminate_red): Break after setting red_found.
3897 (omega_problem_has_red_equations): Similarly after setting found.
3898 (omega_query_variable): Similarly after setting coupled.
3900 2013-07-22 Marek Polacek <polacek@redhat.com>
3902 * gimplify.c: Don't include gimple.h twice.
3904 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3906 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3907 instead of TARGET_THUMB1.
3908 (Pz): New constraint.
3909 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3911 (compare_negsi_si): Likewise.
3912 (compare_addsi2_op0): Likewise.
3913 (compare_addsi2_op1): Likewise.
3914 (addsi3_carryin_<optab>): Likewise.
3915 (addsi3_carryin_alt2_<optab>): Likewise.
3916 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3917 for arm_restrict_it.
3918 (subsi3_carryin): Likewise.
3919 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3920 (minmax_arithsi): Disable for arm_restrict_it.
3921 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3922 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3923 (satsi_<SAT:code>_shift): Likewise.
3924 (arm_shiftsi3): Add alternative for 16-bit encoding.
3925 (arm32_movhf): Disable for arm_restrict_it.
3926 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3927 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3929 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3931 * config/arm/arm.md (attribute "insn"): Delete.
3932 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3933 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3934 (not_shiftsi): Update for attribute change.
3935 (not_shiftsi_compare0): Likewise.
3936 (not_shiftsi_compare0_scratch): Likewise.
3937 (arm_one_cmplsi2): Likewise.
3938 (thumb1_one_cmplsi2): Likewise.
3939 (notsi_compare0): Likewise.
3940 (notsi_compare0_scratch): Likewise.
3941 (thumb1_movdi_insn): Likewise.
3942 (arm_movsi_insn): Likewise.
3943 (movhi_insn_arch4): Likewise.
3944 (movhi_bytes): Likewise.
3945 (arm_movqi_insn): Likewise.
3946 (thumb1_movqi_insn): Likewise.
3947 (arm32_movhf): Likewise.
3948 (thumb1_movhf): Likewise.
3949 (arm_movsf_soft_insn): Likewise.
3950 (thumb1_movsf_insn): Likewise.
3951 (thumb_movdf_insn): Likewise.
3952 (movsicc_insn): Likewise.
3953 (movsfcc_soft_insn): Likewise.
3954 (and_scc): Likewise.
3955 (cond_move): Likewise.
3956 (if_move_not): Likewise.
3957 (if_not_move): Likewise.
3958 (if_shift_move): Likewise.
3959 (if_move_shift): Likewise.
3960 (if_shift_shift): Likewise.
3961 (if_not_arith): Likewise.
3962 (if_arith_not): Likewise.
3963 (cond_move_not): Likewise.
3964 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3965 (neon_mov<mode>): Likewise.
3966 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3967 (thumb2_movsi_vfp): Likewise.
3968 (movsf_vfp): Likewise.
3969 (thumb2_movsf_vfp): Likewise.
3970 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3972 (cortexa7_older_only): Likewise.
3973 (cortexa7_younger): Likewise.
3974 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3975 (1020alu_shift_op): Likewise.
3976 (1020alu_shift_reg_op): Likewise.
3977 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3978 (alu_shift_op): Likewise.
3979 (alu_shift_reg_op): Likewise.
3980 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3981 (11_alu_shift_op): Likewise.
3982 (11_alu_shift_reg_op): Likewise.
3983 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3984 (9_alu_shift_reg_op): Likewise.
3985 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3987 (cortex_a15_alu_shift): Likewise.
3988 (cortex_a15_alu_shift_reg): Likewise.
3989 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3990 (cortex_a5_alu_shift): Likewise.
3991 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3993 (cortex_a53_alu_shift): Likewise.
3994 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3996 (cortex_a7_alu_reg): Likewise.
3997 (cortex_a7_alu_shift): Likewise.
3998 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3999 (cortex_a8_alu_shift): Likewise.
4000 (cortex_a8_alu_shift_reg): Likewise.
4001 (cortex_a8_mov): Likewise.
4002 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4003 (cortex_a9_dp_shift): Likewise.
4004 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4005 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4006 (cortex_r4_mov): Likewise.
4007 (cortex_r4_alu_shift): Likewise.
4008 (cortex_r4_alu_shift_reg): Likewise.
4009 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4010 (526_alu_shift_op): Likewise.
4011 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4012 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4013 (626te_alu_shift_op): Likewise.
4014 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4015 (726te_alu_op): Likewise.
4016 (726te_alu_shift_op): Likewise.
4017 (726te_alu_shift_reg_op): Likewise.
4018 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4019 (mp626_alu_shift_op): Likewise.
4020 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4021 (pj4_alu_e1_conds): Likewise.
4022 (pj4_alu): Likewise.
4023 (pj4_alu_conds): Likewise.
4024 (pj4_shift): Likewise.
4025 (pj4_shift_conds): Likewise.
4026 (pj4_alu_shift): Likewise.
4027 (pj4_alu_shift_conds): Likewise.
4029 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4031 * config/arm/predicates.md (shiftable_operator_strict_it):
4033 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4034 Disable cond_exec version for arm_restrict_it.
4035 (thumb2_smaxsi3): Convert to generate cond_exec.
4036 (thumb2_sminsi3): Likewise.
4037 (thumb32_umaxsi3): Likewise.
4038 (thumb2_uminsi3): Likewise.
4039 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4040 (thumb2_neg_abssi2): Likewise.
4041 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4042 (thumb2_movsicc_insn): Adjust alternatives.
4043 (thumb2_mov_negscc): Disable for arm_restrict_it.
4044 (thumb2_mov_negscc_strict_it): New pattern.
4045 (thumb2_mov_notscc_strict_it): New pattern.
4046 (thumb2_mov_notscc): Disable for arm_restrict_it.
4047 (thumb2_ior_scc): Likewise.
4048 (thumb2_ior_scc_strict_it): New pattern.
4049 (thumb2_cond_move): Adjust for arm_restrict_it.
4050 (thumb2_cond_arith): Disable for arm_restrict_it.
4051 (thumb2_cond_arith_strict_it): New pattern.
4052 (thumb2_cond_sub): Adjust for arm_restrict_it.
4053 (thumb2_movcond): Likewise.
4054 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4055 (thumb2_zero_extendhisi2_v6): Likewise.
4056 (thumb2_zero_extendqisi2_v6): Likewise.
4057 (orsi_notsi_si): Likewise.
4058 (orsi_not_shiftsi_si): Likewise.
4060 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4062 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4063 instruction sequence is 1 byte shorter.
4065 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4067 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4068 it is not needed after split.
4070 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4073 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4074 second label for nonlocal goto receivers. Don't output pic base labels
4075 unless we're producing PIC; mark that action unreachable().
4076 (ix86_save_reg): If the function contains a nonlocal label, save the
4078 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4079 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4080 (update_pic_label_number_if_needed): New.
4081 (machopic_output_function_base_name): Adjust for nonlocal receiver
4083 (machopic_should_output_picbase_label): New.
4084 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4085 (nonlocal_goto_receiver): New insn and split.
4087 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4089 * config/aarch64/aarch64-builtins.c
4090 (aarch64_fold_builtin): Fold abs in all modes.
4091 * config/aarch64/aarch64-simd-builtins.def
4092 (abs): Enable for all modes.
4093 * config/aarch64/arm_neon.h
4094 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4095 (vabs_f64): Add missing intrinsic.
4097 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4099 * config/aarch64/arm_neon.h (vabs_s64): New function
4101 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4104 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4105 * config/avr/avr.md (adjust_len): Add `round'.
4106 * config/avr/avr-protos.h (avr_out_round): New prototype.
4107 (avr_out_plus): Add `out_label' argument.
4108 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4109 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4110 Handle the case where `insn' is just a pattern.
4111 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4112 (avr_out_round): New function.
4113 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4115 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4117 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4118 Add check for flag_stack_usage to handle -fstack-usage support
4120 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4122 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4123 interaction for new Power8 flags and VSX.
4125 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4128 * tree-inline.c (expand_call_inline): Emit errors during
4129 early_inlining only if optimization is not turned on.
4131 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4135 * passes.c (init_optimization_passes): Move the construction of
4136 the pass hierarchy into a new passes.def file.
4138 * Makefile.in (passes.o): Add dependency on passes.def.
4140 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4142 * passes.c (init_optimization_passes): Introduce macros for
4143 constructing the tree of passes (INSERT_PASSES_AFTER,
4144 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4145 TERMINATE_PASS_LIST).
4147 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4148 Wei Mi <wmi@google.com>
4150 PR rtl-optimization/57878
4151 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4153 (reload_pseudo_compare_func): Check nregs first for reload
4156 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4158 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4160 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4162 * read-rtl.c (validate_const_int): Once an invalid character is
4163 seen, quit the loop.
4165 * gengtype.c (write_roots): Similarly once we find the "deletable"
4166 or "if_marked" option.
4168 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4170 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4171 "xtab" and "sat". Move value "clz" from here to ...
4172 (attriubte "type"): ... here.
4173 (satsi_<SAT:code>): Delete "insn" attribute.
4174 (satsi_<SAT:code>_shift): Likewise.
4175 (arm_zero_extendqisi2addsi): Likewise.
4176 (arm_extendqisi2addsi): Likewise.
4177 (clzsi2): Update for attribute changes.
4178 (rbitsi2): Likewise.
4179 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4181 (arm_usatsihi): Likewise.
4182 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4184 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4186 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4187 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4188 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4189 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4190 in alphabetical order.
4191 (attribute "core_cycles"): Update for attribute changes.
4192 (arm_addsi3): Likewise.
4193 (addsi3_compare0): Likewise.
4194 (addsi3_compare0_scratch): Likewise.
4195 (addsi3_compare_op1): Likewise.
4196 (addsi3_compare_op2): Likewise.
4197 (compare_addsi2_op0): Likewise.
4198 (compare_addsi2_op1): Likewise.
4199 (addsi3_carryin_shift_<optab>): Likewise.
4200 (subsi3_carryin_shift): Likewise.
4201 (rsbsi3_carryin_shift): Likewise.
4202 (arm_subsi3_insn): Likewise.
4203 (subsi3_compare0): Likewise.
4204 (subsi3_compare): Likewise.
4205 (arm_andsi3_insn): Likewise.
4206 (thumb1_andsi3_insn): Likewise.
4207 (andsi3_compare0): Likewise.
4208 (andsi3_compare0_scratch): Likewise.
4209 (zeroextractsi_compare0_scratch
4210 (andsi_not_shiftsi_si): Likewise.
4211 (iorsi3_insn): Likewise.
4212 (iorsi3_compare0): Likewise.
4213 (iorsi3_compare0_scratch): Likewise.
4214 (arm_xorsi3): Likewise.
4215 (thumb1_xorsi3_insn): Likewise.
4216 (xorsi3_compare0): Likewise.
4217 (xorsi3_compare0_scratch): Likewise.
4218 (satsi_<SAT:code>_shift): Likewise.
4220 (arm_shiftsi3): Likewise.
4221 (shiftsi3_compare0): Likewise.
4222 (not_shiftsi): Likewise.
4223 (not_shiftsi_compare0): Likewise.
4224 (not_shiftsi_compare0_scratch): Likewise.
4225 (arm_one_cmplsi2): Likewise.
4226 (thumb_one_complsi2): Likewise.
4227 (notsi_compare0): Likewise.
4228 (notsi_compare0_scratch): Likewise.
4229 (thumb1_zero_extendhisi2): Likewise.
4230 (arm_zero_extendhisi2): Likewise.
4231 (arm_zero_extendhisi2_v6): Likewise.
4232 (arm_zero_extendhisi2addsi): Likewise.
4233 (thumb1_zero_extendqisi2): Likewise.
4234 (thumb1_zero_extendqisi2_v6): Likewise.
4235 (arm_zero_extendqisi2): Likewise.
4236 (arm_zero_extendqisi2_v6): Likewise.
4237 (arm_zero_extendqisi2addsi): Likewise.
4238 (thumb1_extendhisi2): Likewise.
4239 (arm_extendhisi2): Likewise.
4240 (arm_extendhisi2_v6): Likewise.
4241 (arm_extendqisi): Likewise.
4242 (arm_extendqisi_v6): Likewise.
4243 (arm_extendqisi2addsi): Likewise.
4244 (thumb1_extendqisi2): Likewise.
4245 (thumb1_movdi_insn): Likewise.
4246 (arm_movsi_insn): Likewise.
4247 (movsi_compare0): Likewise.
4248 (movhi_insn_arch4): Likewise.
4249 (movhi_bytes): Likewise.
4250 (arm_movqi_insn): Likewise.
4251 (thumb1_movqi_insn): Likewise.
4252 (arm32_movhf): Likewise.
4253 (thumb1_movhf): Likewise.
4254 (arm_movsf_soft_insn): Likewise.
4255 (thumb1_movsf_insn): Likewise.
4256 (movdf_soft_insn): Likewise.
4257 (thumb_movdf_insn): Likewise.
4258 (arm_cmpsi_insn): Likewise.
4259 (cmpsi_shiftsi): Likewise.
4260 (cmpsi_shiftsi_swp): Likewise.
4261 (arm_cmpsi_negshiftsi_si): Likewise.
4262 (movsicc_insn): Likewise.
4263 (movsfcc_soft_insn): Likewise.
4264 (arith_shiftsi): Likewise.
4265 (arith_shiftsi_compare0
4266 (arith_shiftsi_compare0_scratch
4267 (sub_shiftsi): Likewise.
4268 (sub_shiftsi_compare0
4269 (sub_shiftsi_compare0_scratch
4270 (and_scc): Likewise.
4271 (cond_move): Likewise.
4272 (if_plus_move): Likewise.
4273 (if_move_plus): Likewise.
4274 (if_move_not): Likewise.
4275 (if_not_move): Likewise.
4276 (if_shift_move): Likewise.
4277 (if_move_shift): Likewise.
4278 (if_shift_shift): Likewise.
4279 (if_not_arith): Likewise.
4280 (if_arith_not): Likewise.
4281 (cond_move_not): Likewise.
4282 (thumb1_ashlsi3): Set type attribute.
4283 (thumb1_ashrsi3): Likewise.
4284 (thumb1_lshrsi3): Likewise.
4285 (thumb1_rotrsi3): Likewise.
4286 (shiftsi3_compare0_scratch): Likewise.
4287 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
4288 (neon_mov<mode>): Likewise.
4289 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
4291 (thumb2_movsi_insn): Likewise.
4292 (thumb2_cmpsi_neg_shiftsi): Likewise.
4293 (thumb2_extendqisi_v6): Likewise.
4294 (thumb2_zero_extendhisi2_v6): Likewise.
4295 (thumb2_zero_extendqisi2_v6): Likewise.
4296 (thumb2_shiftsi3_short): Likewise.
4297 (thumb2_addsi3_compare0_scratch): Likewise.
4298 (orsi_not_shiftsi_si): Likewise.
4299 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4300 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
4302 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4303 (1020alu_shift_op): Likewise.
4304 (1020alu_shift_reg_op): Likewise.
4305 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4306 (alu_shift_op): Likewise.
4307 (alu_shift_reg_op): Likewise.
4308 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4309 (11_alu_shift_op): Likewise.
4310 (11_alu_shift_reg_op): Likewise.
4311 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4312 (9_alu_shift_reg_op): Likewise.
4313 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4315 (cortex_a15_alu_shift): Likewise.
4316 (cortex_a15_alu_shift_reg): Likewise.
4317 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
4319 (cortex_a5_alu_shift): Likewise.
4320 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
4322 (cortex_a53_alu_shift): Likewise.
4323 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4325 (cortex_a7_alu_reg): Likewise.
4326 (cortex_a7_alu_shift): Likewise.
4327 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
4329 (cortex_a8_alu_shift): Likewise.
4330 (cortex_a8_alu_shift_reg): Likewise.
4331 (cortex_a8_mov): Likewise.
4332 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4333 (cortex_a9_dp_shift): Likewise.
4334 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
4336 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
4338 (cortex_r4_mov): Likewise.
4339 (cortex_r4_alu_shift): Likewise.
4340 (cortex_r4_alu_shift_reg): Likewise.
4341 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4342 (526_alu_shift_op): Likewise.
4343 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4344 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4345 (626te_alu_shift_op): Likewise.
4346 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
4347 (726te_alu_op): Likewise.
4348 (726te_alu_shift_op): Likewise.
4349 (726te_alu_shift_reg_op): Likewise.
4350 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4351 (mp626_alu_shift_op): Likewise.
4352 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
4353 (pj4_alu_e1_conds): Likewise.
4354 (pj4_alu): Likewise.
4355 (pj4_alu_conds): Likewise.
4356 (pj4_shift): Likewise.
4357 (pj4_shift_conds): Likewise.
4358 (pj4_alu_shift): Likewise.
4359 (pj4_alu_shift_conds): Likewise.
4360 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4362 (cortexa7_older_only): Likewise.
4363 (cortexa7_younger): Likewise.
4365 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4367 * ipa-pure-const.c (generate_summary): Rename to...
4368 (pure_const_generate_summary): ... this.
4370 2013-07-17 Iain Sandoe <iain@codesourcery.com>
4372 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
4374 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
4377 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
4380 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4382 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
4383 enabled without -march=zEC12.
4384 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
4387 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4389 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
4390 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
4391 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
4392 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
4393 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
4394 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
4395 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
4397 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
4398 and ISA_HAS_NMADD3_NMSUB3.
4399 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
4400 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
4401 (nmsub4<mode>, nmsub3<mode>): Likewise.
4402 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
4404 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4406 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
4407 TARGET_MIPS5400 checking.
4409 2013-07-16 Jakub Jelinek <jakub@redhat.com>
4410 Peter Bergner <bergner@vnet.ibm.com>
4412 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
4413 registers in the comment.
4414 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
4415 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
4416 rather than FIRST_PSEUDO_REGISTERS.
4418 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
4420 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
4421 enable extra ISA flags with TARGET_HTM.
4423 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4425 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
4428 2013-07-15 Cong Hou <congh@google.com>
4430 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
4431 in compare function for sorting.
4433 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
4435 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
4436 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
4437 * config/rs6000/rs6000.opt: Add -mhtm option.
4438 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
4439 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
4440 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
4441 __HTM__ if the HTM instructions are available.
4442 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
4443 htm_spr_reg_operand): New define_predicates.
4444 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
4445 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
4447 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
4448 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
4449 HTM builtin functions.
4450 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
4451 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
4452 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
4453 (rs6000_builtin_mask_calculate): Likewise.
4454 (rs6000_option_override_internal): Likewise.
4455 (bdesc_htm): Add new HTM builtin support.
4456 (htm_spr_num): New function.
4457 (htm_spr_regno): Likewise.
4458 (rs6000_htm_spr_icode): Likewise.
4459 (htm_expand_builtin): Likewise.
4460 (htm_init_builtins): Likewise.
4461 (rs6000_expand_builtin): Add support for HTM builtin functions.
4462 (rs6000_init_builtins): Likewise.
4463 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
4465 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
4466 (TARGET_HTM, MASK_HTM): Define macros.
4467 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
4468 (FIXED_REGISTERS): Likewise.
4469 (CALL_USED_REGISTERS): Likewise.
4470 (CALL_REALLY_USED_REGISTERS): Likewise.
4471 (REG_ALLOC_ORDER): Likewise.
4472 (enum reg_class): Likewise.
4473 (REG_CLASS_NAMES): Likewise.
4474 (REG_CLASS_CONTENTS): Likewise.
4475 (REGISTER_NAMES): Likewise.
4476 (ADDITIONAL_REGISTER_NAMES): Likewise.
4477 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
4478 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
4479 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
4480 * config/rs6000/htm.md: New file.
4481 * config/rs6000/htmintrin.h: New file.
4482 * config/rs6000/htmxlintrin.h: New file.
4484 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
4486 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4487 Define SYMBOL_TINY_GOT, update comment.
4488 * config/aarch64/aarch64.c
4489 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
4490 (aarch64_expand_mov_immediate): Likewise.
4491 (aarch64_print_operand): Likewise.
4492 (aarch64_classify_symbol): Likewise.
4493 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4494 (ldr_got_tiny): Define.
4496 2013-07-13 Tobias Grosser <tobias@grosser.es>
4498 PR tree-optimization/54094
4499 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4500 scheduling dimension for the parallelism check from the polyhedral
4501 information in the AST.
4502 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4505 2013-07-13 Jason Merrill <jason@redhat.com>
4507 * print-tree.c (debug_vec_tree): Use debug_raw.
4508 (debug_raw (vec<tree, va_gc> &)): New.
4509 (debug_raw (vec<tree, va_gc> *)): New.
4510 * tree.h: Declare them.
4512 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4514 * ifcvt.c (ifcvt_after_combine): New static variable.
4515 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4517 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4518 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4519 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4521 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4523 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4525 2013-07-12 Michael Matz <matz@suse.de>
4528 * convert.c (convert_to_real): Reject non-float inner types.
4530 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4532 * config/aarch64/aarch64-protos.h
4533 (aarch64_simd_immediate_valid_for_move): Remove.
4534 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4535 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4537 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4539 2013-07-11 Steve Ellcey <sellcey@mips.com>
4541 * config/mips/mips.c (mips_conditional_register_usage): Do not
4542 use t[0-7] registers in MIPS16 mode when optimizing for size.
4544 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4546 * config/i386/i386.c (dispatch_function_versions): Fix array
4547 indexing of function_version_info to match actual_versions.
4549 2013-07-11 Teresa Johnson <tejohnson@google.com>
4551 * vec.h (struct va_gc): Move release out-of-line.
4552 (va_gc::release): Call ggc_free on released vec.
4554 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4556 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4557 Require GOT register as additional operand in UNSPEC.
4558 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4559 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4560 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4561 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4562 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4563 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4564 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4566 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4569 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4570 name seen by assembler/linker rather if available.
4572 2013-07-11 Andreas Schwab <schwab@suse.de>
4574 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4576 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4578 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4580 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4582 * doc/tm.texi.in: Move hook documentation to ....
4583 * target.def: ... here.
4585 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4587 * doc/tm.texi: Regenerate.
4589 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4592 * doc/invoke.texi: Document -Wconditionally-supported.
4594 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4597 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4600 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4603 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4604 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4605 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4606 Remove duplicate devices.
4607 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4608 * config/avr/t-multilib: Regenerate.
4609 * config/avr/avr-tables.opt: Regenerate.
4610 * doc/avr-mmcu.texi: Regenerate.
4612 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4615 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4617 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4619 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4620 the cost of MULT when optimizing for size.
4622 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4624 * config/cr16/cr16-protos.h: Don't include target.h.
4626 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4628 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4629 adjust register size for TDmode and TFmode for VSX registers.
4631 2013-07-08 Kai Tietz <ktietz@redhat.com>
4634 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4635 hook_bool_const_tree_true.
4637 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4639 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4640 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4641 * config/s390/s390.md: Define FPR*_REGNUM constants.
4642 Fix FPR2_REGNUM constant (18 -> 17).
4643 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4644 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4645 ("trunc<BFP:mode><DFP_ALL:mode>2")
4646 ("trunc<DFP_ALL:mode><BFP:mode>2")
4647 ("*extend<BFP:mode><DFP_ALL:mode>2")
4648 ("*extend<DFP_ALL:mode><BFP:mode>2")
4649 ("extend<BFP:mode><DFP_ALL:mode>2")
4650 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4653 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4655 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4657 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4659 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4660 and cfun_fpr_bit_p to cfun_fpr_save_p.
4661 (s390_frame_area, s390_register_info, s390_frame_info)
4662 (s390_emit_prologue, s390_emit_epilogue)
4663 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4665 * config/s390/s390.h: Define *_REGNUM macros for floating point
4668 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4670 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4672 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4674 PR rtl-optimization/57786
4675 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4676 and break out of the loop when it is set to false.
4678 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4681 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4682 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4683 (const_int 63)) 0)).
4684 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4685 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4686 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4688 PR rtl-optimization/57829
4689 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4690 mask bits outside of mode are just sign-extension from mode to HWI.
4692 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4694 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4695 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4696 adjust_address instead of change_address to keep info about alignment.
4697 (emit_strmov): Remove.
4698 (emit_memmov): New function.
4699 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4700 (expand_movmem_epilogue): Likewise and return updated rtx for
4702 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4703 destination and source.
4704 (decide_alignment): Refactor, handle vector_loop.
4705 (ix86_expand_movmem): Likewise.
4706 (ix86_expand_setmem): Likewise.
4707 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4709 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4711 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4712 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4714 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4716 * config/i386/sse.md (sse_movlhps): Change alternative 3
4717 of operand 2 to "m".
4719 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4722 * config/i386/sse.md (iptr): New mode attribute.
4723 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4724 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4725 (<sse>_vmmul<mode>3): Ditto.
4726 (<sse>_vmdiv<mode>3): Ditto.
4727 (sse_vmrcpv4sf2): Ditto.
4728 (<sse>_vmsqrt<mode>2): Ditto.
4729 (sse_vmrsqrtv4sf2): Ditto.
4730 (<sse>_vm<code><mode>3): Ditto.
4731 (avx_vmcmp<mode>3): Ditto.
4732 (<sse>_vmmaskcmp<mode>3): Ditto.
4733 (<sse>_comi): Ditto.
4734 (<sse>_ucomi): Ditto.
4735 (*xop_vmfrcz_<mode>): Ditto.
4736 (*fmai_fmadd_<mode>): Ditto.
4737 (*fmai_fmsub_<mode>): Ditto.
4738 (*fmai_fnmadd_<mode>): Ditto.
4739 (*fmai_fnmsub_<mode>): Ditto.
4740 (*fma4i_vmfmadd_<mode>): Ditto.
4741 (*fma4i_vmfmsub_<mode>): Ditto.
4742 (*fma4i_vmfnmadd_<mode>): Ditto.
4743 (*fma4i_vmfnmsub_<mode>): Ditto.
4744 (*xop_vmfrcz_<mode>): Ditto.
4745 (sse_cvtps2pi): Ditto.
4746 (sse_cvttps2pi): Ditto.
4747 (sse_cvtss2si): Ditto.
4748 (sse_cvtss2si_2): Ditto.
4749 (sse_cvtss2siq_2): Ditto.
4750 (sse_cvttss2si): Ditto.
4751 (sse_cvttss2siq): Ditto.
4752 (sse_cvtsd2si): Ditto.
4753 (sse_cvtsd2si_2): Ditto.
4754 (sse_cvtsd2siq_2): Ditto.
4755 (sse_cvttsd2si): Ditto.
4756 (sse_cvttsd2siq): Ditto.
4757 (sse_cvtsd2ss): Ditto.
4758 (sse_cvtss2sd): Ditto.
4759 (avx2_pbroadcast<mode>): Ditto.
4760 (avx2_pbroadcast<mode>_1): Ditto.
4761 (*avx_vperm_broadcast_v4sf): Ditto.
4763 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4764 (sse_movlhps): Ditto.
4765 (sse_storehps): Ditto.
4766 (sse_loadhps): Ditto.
4767 (sse_storelps): Ditto.
4768 (sse_loadlps): Ditto.
4769 (*vec_concatv4sf): Ditto.
4770 (*vec_interleave_highv2df): Ditto.
4771 (*vec_interleave_lowv2df): Ditto.
4772 (*vec_extractv2df_1_sse): Ditto.
4773 (*vec_extractv2df_0_sse): Ditto.
4774 (sse2_storelpd): Ditto.
4775 (sse2_loadlpd): Ditto.
4776 (sse2_movsd): Ditto.
4777 (*vec_concatv4si): Ditto.
4778 (vec_concatv2di): Ditto.
4780 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4781 for Intel asm dialect.
4782 (mmx_punpcklwd): Ditto.
4783 (mmx_punpckldq): Ditto.
4785 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4786 for intel assembler dialect.
4788 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4791 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4792 for BUILT_IN_C{LZ,LRSB}*.
4793 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4794 * tree-vrp.c (extract_range_basic): Handle
4795 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4796 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4797 fall thru to code calling set_value*.
4798 * builtins.c (expand_builtin): Remove *IMAX cases.
4799 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4800 if width is bigger than 2*HWI.
4802 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4804 PR rtl-optimization/55342
4805 * lra-int.h (lra_subreg_reload_pseudos): New.
4806 * lra.c: Add undoing optional reloads to the block diagram.
4807 (lra_subreg_reload_pseudos): New.
4808 (lra_optional_reload_pseudos): Change comments.
4809 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4810 lra_optional_reload_pseudos after undo transformations.
4811 * lra-assigns.c (pseudo_prefix_title): New.
4812 (lra_setup_reg_renumber): Use it.
4813 (spill_for): Ditto. Check subreg reload pseudos too.
4814 (assign_by_spills): Consider subreg reload pseudos too.
4815 * lra-constraints.c (simplify_operand_subreg): Use
4816 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4817 (curr_insn_transform): Recognize and do optional reloads.
4818 (undo_optional_reloads): New.
4819 (lra_undo_inheritance): Call undo_optional_reloads.
4821 2013-07-05 Thomas Quinot <quinot@adacore.com>
4823 * tree-complex.c (expand_complex_operations_1): Fix typo.
4825 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4827 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4828 (tune_params): New member 'const vec_costs'.
4829 * config/aarch64/aarch64.c (generic_vector_cost): New.
4830 (generic_tunings): New member 'generic_vector_cost'.
4831 (aarch64_builtin_vectorization_cost): New.
4832 (aarch64_add_stmt_cost): New.
4833 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4834 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4836 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4839 * config/i386/predicates.md (vsib_address_operand): Disallow
4840 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4842 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4845 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4846 expand_builtin_setjmp_receiver.
4847 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4848 with NULL for the label parameter.
4849 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4850 the frame-pointer. Adjust comments.
4851 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4852 only if LABEL is non-NULL.
4854 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4856 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4857 (ARM_ABI_AAPCS64): Ditto.
4859 (ARM_DEFAULT_ABI): Ditto.
4861 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4863 * config/aarch64/aarch64-builtins.c
4864 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4865 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4867 * config/aarch64/aarch64-simd.md
4868 (aarch64_ld1<VALL:mode>): New.
4869 (aarch64_st1<VALL:mode>): Likewise.
4870 * config/aarch64/arm_neon.h
4871 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4873 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4875 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4877 (ix86_option_override_internal):Turn on all -mavx target flags by
4878 default as they are dependent on AVX anyway.
4880 2013-07-02 Cary Coutant <ccoutant@google.com>
4882 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4884 (loc_checksum_ordered): Likewise.
4885 (hash_loc_operands): Remove inline keyword.
4887 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4889 PR tree-optimization/57741
4890 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4891 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4892 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4893 Allow REAL_CST step_exprs if flag_associative_math.
4894 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4896 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4898 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4900 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4902 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4904 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4906 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4908 (iorsi3_insn): Likewise.
4909 (arm_xorsi3): Likewise.
4911 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4913 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4915 (attribute "type"): ... here, and prefix with "wmmx_".
4916 (attribute "core_cycles"): Update for attribute changes.
4917 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4918 (tbcstv4hi): Likewise.
4919 (tbcstv2si): Likewise.
4920 (iwmmxt_iordi3): Likewise.
4921 (iwmmxt_xordi3): Likewise.
4922 (iwmmxt_anddi3): Likewise.
4923 (iwmmxt_nanddi3): Likewise.
4924 (iwmmxt_arm_movdi): Likewise.
4925 (iwmmxt_movsi_insn): Likewise.
4926 (mov<mode>_internal): Likewise.
4927 (and<mode>3_iwmmxt): Likewise.
4928 (ior<mode>3_iwmmxt): Likewise.
4929 (xor<mode>3_iwmmxt): Likewise.
4930 (add<mode>3_iwmmxt): Likewise.
4931 (ssaddv8qi3): Likewise.
4932 (ssaddv4hi3): Likewise.
4933 (ssaddv2si3): Likewise.
4934 (usaddv8qi3): Likewise.
4935 (usaddv4hi3): Likewise.
4936 (usaddv2si3): Likewise.
4937 (sub<mode>3_iwmmxt): Likewise.
4938 (sssubv8qi3): Likewise.
4939 (sssubv4hi3): Likewise.
4940 (sssubv2si3): Likewise.
4941 (ussubv8qi3): Likewise.
4942 (ussubv4hi3): Likewise.
4943 (ussubv2si3): Likewise.
4944 (mulv4hi3_iwmmxt): Likewise.
4945 (smulv4hi3_highpart): Likewise.
4946 (umulv4hi3_highpart): Likewise.
4947 (iwmmxt_wmacs): Likewise.
4948 (iwmmxt_wmacsz): Likewise.
4949 (iwmmxt_wmacu): Likewise.
4950 (iwmmxt_wmacuz): Likewise.
4951 (iwmmxt_clrdi): Likewise.
4952 (iwmmxt_clrv8qi): Likewise.
4953 (iwmmxt_clr4hi): Likewise.
4954 (iwmmxt_clr2si): Likewise.
4955 (iwmmxt_uavgrndv8qi3): Likewise.
4956 (iwmmxt_uavgrndv4hi3): Likewise.
4957 (iwmmxt_uavgv8qi3): Likewise.
4958 (iwmmxt_uavgv4hi3): Likewise.
4959 (iwmmxt_tinsrb): Likewise.
4960 (iwmmxt_tinsrh): Likewise.
4961 (iwmmxt_tinsrw): Likewise.
4962 (iwmmxt_textrmub): Likewise.
4963 (iwmmxt_textrmsb): Likewise.
4964 (iwmmxt_textrmuh): Likewise.
4965 (iwmmxt_textrmsh): Likewise.
4966 (iwmmxt_textrmw): Likewise.
4967 (iwmxxt_wshufh): Likewise.
4968 (eqv8qi3): Likewise.
4969 (eqv4hi3): Likewise.
4970 (eqv2si3): Likewise.
4971 (gtuv8qi3): Likewise.
4972 (gtuv4hi3): Likewise.
4973 (gtuv2si3): Likewise.
4974 (gtv8qi3): Likewise.
4975 (gtv4hi3): Likewise.
4976 (gtv2si3): Likewise.
4977 (smax<mode>3_iwmmxt): Likewise.
4978 (umax<mode>3_iwmmxt): Likewise.
4979 (smin<mode>3_iwmmxt): Likewise.
4980 (umin<mode>3_iwmmxt): Likewise.
4981 (iwmmxt_wpackhss): Likewise.
4982 (iwmmxt_wpackwss): Likewise.
4983 (iwmmxt_wpackdss): Likewise.
4984 (iwmmxt_wpackhus): Likewise.
4985 (iwmmxt_wpackwus): Likewise.
4986 (iwmmxt_wpackdus): Likewise.
4987 (iwmmxt_wunpckihb): Likewise.
4988 (iwmmxt_wunpckihh): Likewise.
4989 (iwmmxt_wunpckihw): Likewise.
4990 (iwmmxt_wunpckilb): Likewise.
4991 (iwmmxt_wunpckilh): Likewise.
4992 (iwmmxt_wunpckilw): Likewise.
4993 (iwmmxt_wunpckehub): Likewise.
4994 (iwmmxt_wunpckehuh): Likewise.
4995 (iwmmxt_wunpckehuw): Likewise.
4996 (iwmmxt_wunpckehsb): Likewise.
4997 (iwmmxt_wunpckehsh): Likewise.
4998 (iwmmxt_wunpckehsw): Likewise.
4999 (iwmmxt_wunpckelub): Likewise.
5000 (iwmmxt_wunpckeluh): Likewise.
5001 (iwmmxt_wunpckeluw): Likewise.
5002 (iwmmxt_wunpckelsb): Likewise.
5003 (iwmmxt_wunpckelsh): Likewise.
5004 (iwmmxt_wunpckelsw): Likewise.
5005 (ror<mode>3): Likewise.
5006 (ashr<mode>3_iwmmxt): Likewise.
5007 (lshr<mode>3_iwmmxt): Likewise.
5008 (ashl<mode>3_iwmmxt): Likewise.
5009 (ror<mode>3_di): Likewise.
5010 (ashr<mode>3_di): Likewise.
5011 (lshr<mode>3_di): Likewise.
5012 (ashl<mode>3_di): Likewise.
5013 (iwmmxt_wmadds): Likewise.
5014 (iwmmxt_wmaddu): Likewise.
5015 (iwmmxt_tmia): Likewise.
5016 (iwmmxt_tmiaph): Likewise.
5017 (iwmmxt_tmiabb): Likewise.
5018 (iwmmxt_tmiatb): Likewise.
5019 (iwmmxt_tmiabt): Likewise.
5020 (iwmmxt_tmiatt): Likewise.
5021 (iwmmxt_tmovmskb): Likewise.
5022 (iwmmxt_tmovmskh): Likewise.
5023 (iwmmxt_tmovmskw): Likewise.
5024 (iwmmxt_waccb): Likewise.
5025 (iwmmxt_wacch): Likewise.
5026 (iwmmxt_waccw): Likewise.
5027 (iwmmxt_waligni): Likewise.
5028 (iwmmxt_walignr): Likewise.
5029 (iwmmxt_walignr0): Likewise.
5030 (iwmmxt_walignr1): Likewise.
5031 (iwmmxt_walignr2): Likewise.
5032 (iwmmxt_walignr3): Likewise.
5033 (iwmmxt_wsadb): Likewise.
5034 (iwmmxt_wsadh): Likewise.
5035 (iwmmxt_wsadbz): Likewise.
5036 (iwmmxt_wsadhz): Likewise.
5037 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5038 (iwmmxt_wabsdiffb): Likewise.
5039 (iwmmxt_wabsdiffh): Likewise.
5040 (iwmmxt_wabsdiffw): Likewise.
5041 (iwmmxt_waddsubhx): Likewise
5042 (iwmmxt_wsubaddhx): Likewise.
5043 (addc<mode>3): Likewise.
5044 (iwmmxt_avg4): Likewise.
5045 (iwmmxt_avg4r): Likewise.
5046 (iwmmxt_wmaddsx): Likewise.
5047 (iwmmxt_wmaddux): Likewise.
5048 (iwmmxt_wmaddsn): Likewise.
5049 (iwmmxt_wmaddun): Likewise.
5050 (iwmmxt_wmulwsm): Likewise.
5051 (iwmmxt_wmulwum): Likewise.
5052 (iwmmxt_wmulsmr): Likewise.
5053 (iwmmxt_wmulumr): Likewise.
5054 (iwmmxt_wmulwsmr): Likewise.
5055 (iwmmxt_wmulwumr): Likewise.
5056 (iwmmxt_wmulwl): Likewise.
5057 (iwmmxt_wqmulm): Likewise.
5058 (iwmmxt_wqmulwm): Likewise.
5059 (iwmmxt_wqmulmr): Likewise.
5060 (iwmmxt_wqmulwmr): Likewise.
5061 (iwmmxt_waddbhusm): Likewise.
5062 (iwmmxt_waddbhusl): Likewise.
5063 (iwmmxt_wqmiabb): Likewise.
5064 (iwmmxt_wqmiabt): Likewise.
5065 (iwmmxt_wqmiatb): Likewise.
5066 (iwmmxt_wqmiatt): Likewise.
5067 (iwmmxt_wqmiabbn): Likewise.
5068 (iwmmxt_wqmiabtn): Likewise.
5069 (iwmmxt_wqmiatbn): Likewise.
5070 (iwmmxt_wqmiattn): Likewise.
5071 (iwmmxt_wmiabb): Likewise.
5072 (iwmmxt_wmiabt): Likewise.
5073 (iwmmxt_wmiatb): Likewise.
5074 (iwmmxt_wmiatt): Likewise.
5075 (iwmmxt_wmiabbn): Likewise.
5076 (iwmmxt_wmiabtn): Likewise.
5077 (iwmmxt_wmiatbn): Likewise.
5078 (iwmmxt_wmiattn): Likewise.
5079 (iwmmxt_wmiawbb): Likewise.
5080 (iwmmxt_wmiawbt): Likewise.
5081 (iwmmxt_wmiawtb): Likewise.
5082 (iwmmxt_wmiawtt): Likewise.
5083 (iwmmxt_wmiawbbn): Likewise.
5084 (iwmmxt_wmiawbtn): Likewise.
5085 (iwmmxt_wmiawtbn): Likewise.
5086 (iwmmxt_wmiawttn): Likewise.
5087 (iwmmxt_wmerge): Likewise.
5088 (iwmmxt_tandc<mode>3): Likewise.
5089 (iwmmxt_torc<mode>3): Likewise.
5090 (iwmmxt_torvsc<mode>3): Likewise.
5091 (iwmmxt_textrc<mode>3): Likewise.
5092 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5093 (wmmxt_pack): Likewise.
5094 (wmmxt_mult_c1): Likewise.
5095 (wmmxt_mult_c2): Likewise.
5096 (wmmxt_alu_c1): Likewise.
5097 (wmmxt_alu_c2): Likewise.
5098 (wmmxt_alu_c3): Likewise.
5099 (wmmxt_transfer_c1): Likewise.
5100 (wmmxt_transfer_c2): Likewise.
5101 (wmmxt_transfer_c3): Likewise.
5102 (marvell_f_iwmmxt_wstr): Likewise.
5103 (marvell_f_iwmmxt_wldr): Likewise.
5105 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5107 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5109 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5112 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5113 * lra-constraints.c (need_for_split_p): Check call used hard regs
5114 living through calls.
5116 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5117 call used regs for call insn.
5119 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5122 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5123 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5125 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5127 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5129 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5131 * lra-constraints.c (need_for_split_p): Check call used hard regs
5132 living through calls.
5134 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5137 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5138 to tie with any other modes. Eliminate Altivec vector mode tests,
5139 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5140 code, to return 0 if testing MODE2 for a condition, if we've
5141 already tested MODE1 for the same condition.
5143 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5145 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5148 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5150 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5151 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5153 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5155 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5157 (aarch64_symbolic_constant_p): Remove.
5158 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5159 static. Fix line length and white space.
5160 (aarch64_symbolic_constant_p): Remove.
5161 * config/aarch64/predicates.md (aarch64_valid_symref):
5162 Use aarch64_classify_symbol_expression.
5164 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5166 * config/arm/constraints.md (Ts): New constraint.
5167 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5169 (compare_scc): Use "Ts" constraint for operand 0.
5170 (ior_scc_scc): Likewise.
5171 (and_scc_scc): Likewise.
5172 (and_scc_scc_nodom): Likewise.
5173 (ior_scc_scc_cmp): Likewise for operand 7.
5174 (and_scc_scc_cmp): Likewise.
5175 * config/arm/thumb2.md (thumb2_movsi_insn):
5176 Add alternatives for 16-bit encodings.
5177 (thumb2_movhi_insn): Likewise.
5178 (thumb2_movsicc_insn): Likewise.
5179 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5180 (thumb2_negscc): Use "Ts" constraint.
5181 Move mvn instruction outside cond_exec block.
5182 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5183 for 16-bit encodings.
5185 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5187 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5189 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5190 (mulsi3subsi): Likewise.
5191 (mulsidi3adddi): Likewise.
5192 (mulsidi3_v6): Likewise.
5193 (umulsidi3_v6): Likewise.
5194 (umulsidi3adddi_v6): Likewise.
5195 (smulsi3_highpart_v6): Likewise.
5196 (umulsi3_highpart_v6): Likewise.
5197 (mulhisi3tb): Likewise.
5198 (mulhisi3bt): Likewise.
5199 (mulhisi3tt): Likewise.
5200 (maddhisi4): Likewise.
5201 (maddhisi4tb): Likewise.
5202 (maddhisi4tt): Likewise.
5203 (maddhidi4): Likewise.
5204 (maddhidi4tb): Likewise.
5205 (maddhidi4tt): Likewise.
5206 (zeroextractsi_compare0_scratch): Likewise.
5207 (insv_zero): Likewise.
5208 (insv_t2): Likewise.
5209 (anddi_notzesidi_di): Likewise.
5210 (anddi_notsesidi_di): Likewise.
5211 (andsi_notsi_si): Likewise.
5212 (iordi_zesidi_di): Likewise.
5213 (xordi_zesidi_di): Likewise.
5214 (andsi_iorsi3_notsi): Likewise.
5216 (smax_m1): Likewise.
5218 (not_shiftsi): Likewise.
5219 (unaligned_loadsi): Likewise.
5220 (unaligned_loadhis): Likewise.
5221 (unaligned_loadhiu): Likewise.
5222 (unaligned_storesi): Likewise.
5223 (unaligned_storehi): Likewise.
5224 (extv_reg): Likewise.
5225 (extzv_t2): Likewise.
5227 (udivsi3): Likewise.
5228 (arm_zero_extendhisi2addsi): Likewise.
5229 (arm_zero_extendqisi2addsi): Likewise.
5230 (compareqi_eq0): Likewise.
5231 (arm_extendhisi2_v6): Likewise.
5232 (arm_extendqisi2addsi): Likewise.
5233 (arm_movt): Likewise.
5234 (thumb2_ldrd): Likewise.
5235 (thumb2_ldrd_base): Likewise.
5236 (thumb2_ldrd_base_neg): Likewise.
5237 (thumb2_strd): Likewise.
5238 (thumb2_strd_base): Likewise.
5239 (thumb2_strd_base_neg): Likewise.
5240 (arm_negsi2): Add alternative for 16-bit encoding.
5241 (arm_one_cmplsi2): Likewise.
5243 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5245 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
5246 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
5247 (movdfcc): Likewise.
5248 * config/arm/vfp.md (*thumb2_movsf_vfp):
5249 Disable predication for arm_restrict_it.
5250 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
5251 (*thumb2_movdfcc_vfp): Likewise.
5252 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
5253 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
5254 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
5255 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
5256 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
5257 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
5258 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
5259 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
5260 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
5261 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
5262 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
5263 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
5264 Disable predication for arm_restrict_it.
5266 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
5268 * config/i386/bmiintrin.h (_bextr_u32): New.
5269 (_bextr_u64): Ditto.
5271 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
5273 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
5274 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
5275 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
5276 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
5277 * config/mips/n32-elf.h: ...this new file.
5279 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
5282 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
5283 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
5285 2013-06-27 Catherine Moore <clm@codesourcery.com>
5287 * config/mips/mips-tables.opt: Regenerate.
5288 * config/mips/mips-cpus.def: Add m14ke and m14kec.
5289 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
5290 * doc/invoke.texi: Add -m14kc.
5292 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5295 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
5296 constraints of operand 1 and 2.
5299 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
5300 to match RTL canonicalization. Swap predicates and
5301 constraints of operand 1 and 2.
5303 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
5305 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
5306 Process OP_INOUT regs for splitting too.
5308 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5310 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
5311 decl before the loop, initialize to NULL.
5312 (vectorizable_load): Initialize ptr_incr to NULL.
5314 2013-06-27 Martin Jambor <mjambor@suse.cz>
5317 * ipa-ref.h (ipa_maybe_record_reference): Declare.
5318 * ipa-ref.c (ipa_maybe_record_reference): New function.
5319 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
5320 * ipa-cp.c (create_specialized_node): Record potential references from
5322 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
5324 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5326 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
5327 parameter 'mode' of type 'enum machine_mode mode'; change to pass
5328 'mode' to force_reg.
5329 (aarch64_add_offset): Update calls to aarch64_force_temporary.
5330 (aarch64_expand_mov_immediate): Likewise.
5332 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5334 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
5335 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
5337 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5339 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
5340 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5341 (struct machine_function): Add tbegin_p.
5342 (s390_canonicalize_comparison): Fold CC mode compares to
5343 conditional jump if possible.
5344 (s390_emit_jump): Return the emitted jump.
5345 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
5346 Handle CCRAWmode compares.
5347 (s390_option_override): Default to -mhtm if available.
5348 (s390_reg_clobbered_rtx): Handle floating point regs as well.
5349 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
5350 FPRs instead of df_regs_ever_live_p.
5351 (s390_optimize_nonescaping_tx): New function.
5352 (s390_init_frame_layout): Extend clobbered_regs array to cover
5354 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
5355 (s390_expand_tbegin): New function.
5356 (enum s390_builtin): New enum definition.
5357 (code_for_builtin): New array definition.
5358 (s390_init_builtins): New function.
5359 (s390_expand_builtin): New function.
5360 (TARGET_INIT_BUILTINS): Define.
5361 (TARGET_EXPAND_BUILTIN): Define.
5362 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5363 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
5364 (s390_alc_comparison): Likewise.
5365 * config/s390/s390-modes.def: Add CCRAWmode.
5366 * config/s390/s390.h (processor_flags): Add PF_TX.
5367 (TARGET_CPU_HTM): Define macro.
5368 (TARGET_HTM): Define macro.
5369 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
5370 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
5371 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5372 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
5373 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5374 (TBEGIN_MASK, TBEGINC_MASK): New constants.
5375 ("*cc_to_int"): Move up.
5376 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
5377 constants other than 0.
5378 ("*ccraw_to_int"): New insn and splitter definition.
5379 ("tbegin", "tbegin_nofloat", "tbegin_retry")
5380 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
5381 ("tx_assist"): New expander.
5382 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
5383 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
5384 * config/s390/s390.opt: Add -mhtm option.
5385 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
5386 * config/s390/htmxlintrin.h: New file.
5387 * config/s390/htmintrin.h: New file.
5388 * config/s390/s390intrin.h: New file.
5389 * doc/extend.texi: Document htm builtins.
5390 * config.gcc: Add the new header files to extra_headers.
5392 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
5394 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
5395 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
5397 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
5398 Pat Haugen <pthaugen@us.ibm.com>
5399 Peter Bergner <bergner@vnet.ibm.com>
5401 * config/rs6000/power8.md: New.
5402 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
5403 setting for power8 entry.
5404 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
5405 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
5406 test for Power4/Power5 only.
5407 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
5409 (force_new_group): Adjust comment.
5410 * config/rs6000/rs6000.md: Include power8.md.
5412 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
5414 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
5415 * config/arm/arm-protos.h (arm_max_conditional_execute): New
5417 (tune_params): Update comment.
5418 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
5419 (arm_max_conditional_execute): New function.
5420 (thumb2_final_prescan_insn): Use max_insn_skipped and
5421 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
5423 2013-06-25 Jakub Jelinek <jakub@redhat.com>
5425 PR tree-optimization/57705
5426 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
5427 SSA_NAME step, provided that it is not defined inside the loop.
5428 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
5429 (get_initial_def_for_induction): Handle SSA_NAME IV step.
5431 2013-06-25 Martin Jambor <mjambor@suse.cz>
5434 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
5435 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
5437 (ipa_note_param_call): Initialize member_ptr flag.
5438 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
5439 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
5440 (ipa_write_indirect_edge_info): Stream member_ptr flag.
5441 (ipa_read_indirect_edge_info): Likewise.
5443 2013-06-25 Richard Biener <rguenther@suse.de>
5446 * passes.c (init_optimization_passes): Move pass_fold_builtins
5447 and pass_dce earlier with -Og.
5449 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
5451 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
5452 <BIT_FIELD_REF>: Remove trailing TAB.
5453 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
5456 2013-06-24 Martin Jambor <mjambor@suse.cz>
5458 PR tree-optimization/57358
5459 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
5460 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
5461 (ipa_analyze_params_uses): Generate pessimistic info when true.
5463 2013-06-24 Martin Jambor <mjambor@suse.cz>
5465 PR tree-optimization/57539
5466 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
5467 global.inlined_to of the new node to it. All callers changed.
5468 * ipa-inline-transform.c (clone_inlined_nodes): New variable
5469 inlining_into, pass it to cgraph_clone_node.
5470 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
5471 ipa_free_edge_args_substructures.
5472 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
5473 rdesc linked list. Do not assert rdesc edges have inlined caller.
5474 Assert we have found an rdesc in the rdesc list.
5476 2013-06-24 Richard Biener <rguenther@suse.de>
5478 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
5479 (pointer_set_lookup): Declare.
5480 (class pointer_map): New template class implementing a
5481 generic pointer to T map.
5482 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
5483 pointer_map<T>::contains, pointer_map<T>::insert,
5484 pointer_map<T>::traverse): New functions.
5485 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
5486 (pointer_set_lookup): New function.
5487 (pointer_set_contains): Use pointer_set_lookup.
5488 (pointer_set_insert): Likewise.
5489 (insert_aux): Remove.
5490 (struct pointer_map_t): Embed a pointer_set_t.
5491 (pointer_map_create): Adjust.
5492 (pointer_map_destroy): Likewise.
5493 (pointer_map_contains): Likewise.
5494 (pointer_map_insert): Likewise.
5495 (pointer_map_traverse): Likewise.
5496 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5497 pointer_map<unsigned> instead of a pointer_map_t.
5498 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5499 (streamer_tree_cache_lookup): Likewise.
5500 (streamer_tree_cache_create): Likewise.
5501 (streamer_tree_cache_delete): Likewise.
5502 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5503 pointer_map<unsigned> instead of a pointer_map_t.
5504 (lto_init_tree_ref_encoder): Adjust.
5505 (lto_destroy_tree_ref_encoder): Likewise.
5506 * lto-section-out.c (lto_output_decl_index): Likewise.
5507 (lto_record_function_out_decl_state): Likewise.
5508 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5510 2013-06-24 Richard Biener <rguenther@suse.de>
5512 PR tree-optimization/57488
5513 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5515 2013-06-24 Alan Modra <amodra@gmail.com>
5517 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5518 (gen_easy_altivec_constant): Likewise.
5519 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5520 easy_vector_constant_msb): Likewise.
5522 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5525 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5526 add missing return true.
5528 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5531 * config/sh/predicates.md (general_extend_operand): Invoke
5532 general_movsrc_operand for memory operands.
5533 (general_movsrc_operand): Allow reg+reg addressing, do not use
5534 general_operand for memory operands.
5536 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5538 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5539 when current target options does not apply.
5540 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5541 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5542 * config/i386/bmiintrin.h: Pass appropriate target
5543 attributes to header.
5544 * config/i386/mmintrin.h: Ditto.
5545 * config/i386/nmmintrin.h: Ditto.
5546 * config/i386/avx2intrin.h: Ditto.
5547 * config/i386/fxsrintrin.h: Ditto.
5548 * config/i386/tbmintrin.h: Ditto.
5549 * config/i386/xsaveintrin.h: Ditto.
5550 * config/i386/f16cintrin.h: Ditto.
5551 * config/i386/xtestintrin.h: Ditto.
5552 * config/i386/xsaveoptintrin.h: Ditto.
5553 * config/i386/bmi2intrin.h: Ditto.
5554 * config/i386/lzcntintrin.h: Ditto.
5555 * config/i386/smmintrin.h: Ditto.
5556 * config/i386/wmmintrin.h: Ditto.
5557 * config/i386/x86intrin.h: Remove all header include guards.
5558 * config/i386/prfchwintrin.h: Ditto.
5559 * config/i386/pmmintrin.h: Ditto.
5560 * config/i386/tmmintrin.h: Ditto.
5561 * config/i386/xmmintrin.h: Ditto.
5562 * config/i386/popcntintrin.h: Ditto.
5563 * config/i386/rdseedintrin.h: Ditto.
5564 * config/i386/ammintrin.h: Ditto.
5565 * config/i386/emmintrin.h: Ditto.
5566 * config/i386/immintrin.h: Remove all header include guards.
5567 * config/i386/fma4intrin.h: Ditto.
5568 * config/i386/lwpintrin.h: Ditto.
5569 * config/i386/xopintrin.h: Ditto.
5570 * config/i386/ia32intrin.h: Ditto.
5571 * config/i386/avxintrin.h: Ditto.
5572 * config/i386/rtmintrin.h: Ditto.
5573 * config/i386/fmaintrin.h: Ditto.
5574 * config/i386/mm3dnow.h: Ditto.
5576 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5578 * common/config/i386/i386-common.c: Handle LZCNT.
5580 2013-06-22 Andi Kleen <ak@linux.intel.com>
5582 * doc/extend.texi: Use __atomic_store_n instead of
5583 __atomic_store in HLE example.
5585 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5587 * config/sh/sh.c: Remove <cstdlib> workaround.
5589 2013-06-21 Andi Kleen <ak@linux.intel.com>
5591 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5593 2013-06-21 Andi Kleen <ak@linux.intel.com>
5595 * doc/extend.texi: Document that __atomic_clear and
5596 __atomic_test_and_set should only be used with bool.
5598 2013-06-20 Jan Hubicka <jh@suse.cz>
5600 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5602 * tree.c (decls_same_for_odr): New function.
5603 (same_for_edr): New function.
5604 (types_same_for_odr): New function.
5605 (get_binfo_at_offset): Use it.
5606 * tree.h (types_same_for_odr): Declare.
5608 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5609 Jason Merrill <jason@redhat.com>
5611 * system.h: Include <cstdlib> as well as <stdlib.h>.
5613 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5616 * config/i386/i386.c (construct_container): Report error if
5617 long double is used with disabled x87 float returns.
5619 2013-06-20 Jan Hubicka <jh@suse.cz>
5621 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5623 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5625 PR rtl-optimization/57425
5626 PR rtl-optimization/57569
5627 * alias.c (write_dependence_p): Remove parameters mem_mode and
5628 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5629 Changed all callers.
5630 (canon_anti_dependence): Get comments and semantics in sync.
5631 Add parameter mem_canonicalized. Changed all callers.
5632 * rtl.h (canon_anti_dependence): Update prototype.
5634 2013-06-20 Richard Biener <rguenther@suse.de>
5636 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5637 byte case, inline streamer_read_uchar and defer section
5640 2013-06-20 Richard Biener <rguenther@suse.de>
5642 PR tree-optimization/57584
5643 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5644 SSA names into the expanded expression that take part in
5645 abnormal coalescing.
5647 2013-06-19 Sharad Singhai <singhai@google.com>
5649 * gcov.c (print_usage): Handle new option.
5650 (process_args): Ditto.
5651 (get_gcov_intermediate_filename): New function.
5652 (output_intermediate_file): New function.
5653 (output_gcov_file): New function
5654 (generate_results): Handle new option.
5655 (release_function): Relase demangled name.
5656 (read_graph_file): Handle demangled name.
5657 (output_lines): Ditto.
5658 * doc/gcov.texi: Document gcov intermediate format.
5660 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5663 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5664 (lra_emit_add): Use the functions. Add comment about Y as an
5667 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5670 * collect2.c (collect_atexit): New.
5671 (collect_exit): Delete.
5672 (main): Register collect_atexit with atexit.
5673 (collect_wait): Change collect_exit to exit.
5675 * collect2.h (collect_exit): Delete.
5676 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5678 2013-06-19 Wei Mi <wmi@google.com>
5680 PR rtl-optimization/57518
5681 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5682 if regno is used in paradoxical subreg.
5683 (update_equiv_regs): Check pdx_subregs[regno] before
5684 set a reg to be equivalent with a mem.
5686 2013-06-19 Matthias Klose <doko@ubuntu.com>
5689 * file-find.h (find_a_file): Add a mode parameter.
5690 * file-find.c (find_a_file): Likewise.
5691 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5692 with X_OK for the executables.
5693 * collect2.c (main): Call find_a_file with X_OK.
5695 2013-06-19 Steve Ellcey <sellcey@mips.com>
5698 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5699 Use NEXT_INSN instead of next_real_insn.
5701 2013-06-19 Jan Hubicka <jh@suse.cz>
5703 * cgraph.h (const_value_known_p): Replace by ...
5704 (ctor_for_folding): .. this one.
5705 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5706 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5707 * expr.c (expand_expr_real_1): Likewise.
5708 (string_constant): Likewise.
5709 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5710 * ipa.c (process_references): Likewise.
5711 (symtab_remove_unreachable_nodes): Likewise.
5712 * ipa-inline-analysis.c (param_change_prob): Likewise.
5713 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5714 (get_base_constructor): Likwise.
5715 * varpool.c (varpool_remove_node): Likewise.
5716 (varpool_remove_initializer): LIkewise.
5717 (dump_varpool_node): LIkwise.
5718 (const_value_known_p): Rewrite to ...
5719 (ctor_for_folding): ... this one.
5721 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5724 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5725 PERSONALITY in $PATH derived prefixes.
5727 2013-06-19 Jeff Law <law@redhat.com>
5729 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5732 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5733 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5736 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5738 * config/arm/vfp.md: Move VFP instruction classification documentation
5740 * config/arm/arm.md: ... here. Update instruction classification
5743 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5745 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5746 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5747 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5749 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5752 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5753 with NEXT_INSN. Use tablejump_p to check for jump table data
5756 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5759 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5760 that now in C++ the value is correct per the C++ standards.
5762 2013-06-19 Richard Biener <rguenther@suse.de>
5764 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5767 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5770 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5773 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5774 with next_active_insn.
5776 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5778 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5779 functions are inlined during failures to flag an error.
5780 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5781 in early inline pass.
5783 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5785 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5788 2013-06-18 Julian Brown <julian@codesourcery.com>
5790 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5791 Permit virtual register pre-reload if !strict.
5792 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5794 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5796 * config/arm/neon.md (movmisalign<mode>): Use
5797 neon_perm_struct_or_reg_operand instead of
5798 neon_struct_or_register_operand.
5799 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5800 neon_permissive_struct_operand instead of neon_struct_operand.
5801 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5802 neon_vector_mem_operand.
5803 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5804 neon_vector_mem_operand.
5805 (neon_permissive_struct_operand): New.
5806 (neon_struct_or_register_operand): Rename to...
5807 (neon_perm_struct_or_reg_operand): This. Adjust call to
5808 neon_vector_mem_operand.
5810 2013-06-18 Richard Biener <rguenther@suse.de>
5812 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5813 * lto-streamer.h: Include pointer-set.h.
5814 (struct lto_decl_slot): Remove.
5815 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5816 Remove next_index entry.
5817 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5818 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5819 (lto_init_tree_ref_encoder): Adjust.
5820 (lto_destroy_tree_ref_encoder): Likewise.
5821 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5822 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5823 (lto_output_decl_index): Adjust.
5824 (lto_new_out_decl_state): Likewise.
5825 (lto_record_function_out_decl_state): Likewise.
5826 * lto-streamer-out.c (copy_function): Likewise.
5828 2013-06-18 Richard Biener <rguenther@suse.de>
5830 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5831 * cgraphunit.c: Include cfgloop.h.
5832 (init_lowered_empty_function): Initialize the loop tree.
5833 (assemble_thunk): Insert new BBs into loops.
5835 2013-06-18 Richard Biener <rguenther@suse.de>
5837 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5838 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5839 the map from cache entry to cache index optional.
5840 (streamer_tree_cache_replace_tree): Adjust accordingly.
5841 (streamer_tree_cache_append): Likewise.
5842 (streamer_tree_cache_delete): Likewise.
5843 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5844 streamer cache map from cache entry to cache index.
5845 * lto-streamer-out.c (create_output_block): Adjust.
5847 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5849 * config/arm/arm.md (attribute "insn"): Move multiplication and
5850 division attributes to...
5851 (attribute "type"): ... here. Remove mult.
5852 (attribute "mul32"): New attribute.
5853 (attribute "mul64"): Add umaal.
5854 (*arm_mulsi3): Update attributes.
5855 (*arm_mulsi3_v6): Likewise.
5856 (*thumb_mulsi3): Likewise.
5857 (*thumb_mulsi3_v6): Likewise.
5858 (*mulsi3_compare0): Likewise.
5859 (*mulsi3_compare0_v6): Likewise.
5860 (*mulsi_compare0_scratch): Likewise.
5861 (*mulsi_compare0_scratch_v6): Likewise.
5862 (*mulsi3addsi): Likewise.
5863 (*mulsi3addsi_v6): Likewise.
5864 (*mulsi3addsi_compare0): Likewise.
5865 (*mulsi3addsi_compare0_v6): Likewise.
5866 (*mulsi3addsi_compare0_scratch): Likewise.
5867 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5868 (*mulsi3subsi): Likewise.
5869 (*mulsidi3adddi): Likewise.
5870 (*mulsi3addsi_v6): Likewise.
5871 (*mulsidi3adddi_v6): Likewise.
5872 (*mulsidi3_nov6): Likewise.
5873 (*mulsidi3_v6): Likewise.
5874 (*umulsidi3_nov6): Likewise.
5875 (*umulsidi3_v6): Likewise.
5876 (*umulsidi3adddi): Likewise.
5877 (*umulsidi3adddi_v6): Likewise.
5878 (*smulsi3_highpart_nov6): Likewise.
5879 (*smulsi3_highpart_v6): Likewise.
5880 (*umulsi3_highpart_nov6): Likewise.
5881 (*umulsi3_highpart_v6): Likewise.
5882 (mulhisi3): Likewise.
5883 (*mulhisi3tb): Likewise.
5884 (*mulhisi3bt): Likewise.
5885 (*mulhisi3tt): Likewise.
5886 (maddhisi4): Likewise.
5887 (*maddhisi4tb): Likewise.
5888 (*maddhisi4tt): Likewise.
5889 (maddhidi4): Likewise.
5890 (*maddhidi4tb): Likewise.
5891 (*maddhidi4tt): Likewise.
5893 (udivsi3): Likewise.
5894 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5895 (thumb2_mulsi_short_compare0): Likewise.
5896 (thumb2_mulsi_short_compare0_scratch): Likewise.
5897 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5898 (1020mult2): Likewise.
5899 (1020mult3): Likewise.
5900 (1020mult4): Likewise.
5901 (1020mult5): Likewise.
5902 (1020mult6): Likewise.
5903 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5905 (cortex_a15_mult64): Likewise.
5906 (cortex_a15_sdiv): Likewise.
5907 (cortex_a15_udiv): Likewise.
5908 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5914 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5915 (pj4_ir_div): Likewise.
5916 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5917 (11_mult2): Likewise.
5918 (11_mult3): Likewise.
5919 (11_mult4): Likewise.
5920 (11_mult5): Likewise.
5921 (11_mult6): Likewise.
5922 (11_mult7): Likewise.
5923 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5924 (cortex_a8_mla): Likewise.
5925 (cortex_a8_mull): Likewise.
5926 (cortex_a8_smulwy): Likewise.
5927 (cortex_a8_smlald): Likewise.
5928 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5929 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5930 (cortex_r4_mul_3): Likewise.
5931 (cortex_r4_mla_4): Likewise.
5932 (cortex_r4_mla_3): Likewise.
5933 (cortex_r4_smlald): Likewise.
5934 (cortex_r4_mull): Likewise.
5935 (cortex_r4_sdiv): Likewise.
5936 (cortex_r4_udiv): Likewise.
5937 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5938 (cortex_a7_idiv): Likewise.
5939 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5940 (9_mult2): Likewise.
5941 (9_mult3): Likewise.
5942 (9_mult4): Likewise.
5943 (9_mult5): Likewise.
5944 (9_mult6): Likewise.
5945 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5946 (cortex_a53_sdiv): Likewise.
5947 (cortex_a53_udiv): Likewise.
5948 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5949 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5950 (mp626_mult2): Likewise.
5951 (mp626_mult3): Likewise.
5952 (mp626_mult4): Likewise.
5953 * config/arm/fa526.md (526_mult1): Update attribute change.
5954 (526_mult2): Likewise.
5955 * config/arm/arm-generic.md (mult): Update attribute change.
5956 (mult_ldsched_strongarm): Likewise.
5957 (mult_ldsched): Likewise.
5958 (multi_cycle): Likewise.
5959 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5960 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5961 (606te_mult2): Likewise.
5962 (606te_mult3): Likewise.
5963 (606te_mult4): Likewise.
5964 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5965 (cortex_a9_mac16): Likewise.
5966 (cortex_a9_multiply): Likewise.
5967 (cortex_a9_mac): Likewise.
5968 (cortex_a9_multiply_long): Likewise.
5969 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5970 (626te_mult2): Likewise.
5971 (626te_mult3): Likewise.
5972 (626te_mult4): Likewise.
5974 2013-06-18 Richard Biener <rguenther@suse.de>
5977 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5979 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5982 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5983 with next_active_insn.
5985 2013-06-18 Alan Modra <amodra@gmail.com>
5987 * config/rs6000/rs6000.h (enum data_align): New.
5988 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5989 (DATA_ABI_ALIGNMENT): Define.
5990 (CONSTANT_ALIGNMENT): Correct comment.
5991 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5992 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5994 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5996 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5997 ATTRIBUTE_UNUSED marking.
5999 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6001 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6002 alternative and update.
6003 (aarch64_dup_lanedi): Delete.
6004 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6005 * config/aarch64/aarch64-simd-builtins.def: Update.
6007 2013-06-17 Richard Biener <rguenther@suse.de>
6009 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6010 (lto_input_scc): Declare.
6011 (lto_input_tree_1): Likewise.
6012 (struct lto_stats_d): Add num_tree_bodies_output and
6013 num_pickle_refs_output.
6014 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6015 (lto_read_tree_1): Split out from ...
6016 (lto_read_tree): ... this.
6017 (lto_input_scc): New function.
6018 (lto_input_tree_1): Split out from ...
6019 (lto_input_tree): ... this. Handle LTO_tree_scc.
6020 (lto_data_in_create): Create the streamer cache without hashes.
6021 * lto-streamer-out.c (create_output_block): Create the streamer
6022 cache with hashes when not doing WPA.
6023 (lto_write_tree_1): Split out from ...
6024 (lto_write_tree): ... this.
6025 (get_symbol_initial_value): New function.
6026 (lto_output_tree_1): Split out from ...
6027 (lto_output_tree): ... this. Write trees as series of SCCs
6028 using a DFS walk via DFS_write_tree.
6029 (struct sccs, struct scc_entry): New types.
6030 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6031 (DFS_write_tree_body): New function.
6032 (DFS_write_tree): Likewise.
6033 (hash_tree): Likewise.
6034 (scc_entry_compare): Likewise.
6035 (hash_scc): Likewise.
6036 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6037 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6038 TREE_CHAIN as regular reference.
6039 (streamer_read_integer_cst): Remove.
6040 (streamer_get_pickled_tree): Adjust.
6041 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6042 of DECL_EXTERNALs in BLOCK_VARS for now.
6043 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6045 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6046 Add hash value argument and record that if hashes are recorded
6048 (streamer_tree_cache_insert_1): Adjust.
6049 (streamer_tree_cache_insert): Likewise.
6050 (streamer_tree_cache_insert_at): Rename to ...
6051 (streamer_tree_cache_replace_tree): ... this and adjust.
6052 (streamer_tree_cache_append): Adjust.
6053 (record_common_node): Likewise.
6054 (streamer_tree_cache_create): Add argument whether to
6055 record hash values together with trees.
6056 (streamer_tree_cache_delete): Adjust.
6057 * tree-streamer.h (struct streamer_tree_cache_d): Add
6059 (streamer_read_integer_cst): Remove.
6060 (streamer_tree_cache_insert): Adjust.
6061 (streamer_tree_cache_append): Likewise.
6062 (streamer_tree_cache_insert_at): Rename to ...
6063 (streamer_tree_cache_replace_tree): ... this and adjust.
6064 (streamer_tree_cache_create): Add argument whether to record hashes.
6065 (streamer_tree_cache_get): Rename to ...
6066 (streamer_tree_cache_get_tree): ... this.
6067 (streamer_tree_cache_get_hash): New function.
6068 * tree.c (cache_integer_cst): New function.
6069 * tree.h (cache_integer_cst): Declare.
6070 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6071 * lto-symtab.c (lto_varpool_replace_node): Only release
6072 DECL_INITIAL of non-prevailing decls.
6073 * varpool.c (varpool_remove_initializer): Do not release
6074 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6076 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6078 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6079 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6080 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6081 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6082 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6083 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6084 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6085 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6086 instead of TARGET_64BIT.
6087 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6088 Require ISA_HAS_<D>DIV.
6090 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6092 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6093 (mips*-*-linux*): Move default with_llsc setting to where other
6095 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6097 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6098 Likewise. Remove default with_tune setting. Move default float
6099 setting to its own block. Handle with_llsc in the same block as above.
6101 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6103 PR rtl-optimization/57425
6104 PR rtl-optimization/57569
6105 * alias.c (write_dependence_p): Add new parameters mem_mode,
6106 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6107 Changed all callers.
6108 (canon_anti_dependence): New function.
6109 * cse.c (check_dependence): Use canon_anti_dependence.
6110 * cselib.c (cselib_invalidate_mem): Likewise.
6111 * rtl.h (canon_anti_dependence): Declare.
6113 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6115 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6116 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6117 ".set mips3" for 64-bit targets.
6119 2013-06-15 Dehao Chen <dehao@google.com>
6121 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6122 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6123 (gimple_check_call_args): Likewise.
6124 * value-prof.c (check_ic_target): Likewise.
6125 * ipa-inline.c (early_inliner): Likewise.
6126 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6127 * cgraph.c (cgraph_create_edge_1): Likewise.
6128 (cgraph_make_edge_direct): Likewise.
6130 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6133 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6134 rs6000_output_move_128bit to handle emitting quad memory
6135 operations. Set attribute length to 8 bytes.
6137 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6139 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6141 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6142 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6143 (aarch64_<su>mlsl<mode>): Likewise.
6145 2013-06-14 Mike Stump <mikestump@comcast.net>
6147 * Makefile.in (TARGET_H): Add insn-codes.h.
6149 2013-06-14 Alan Modra <amodra@gmail.com>
6153 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6154 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6155 bitfield expansion when EXPAND_MEMORY.
6156 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6158 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6160 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6161 test for clearing quad memory on 32-bit later.
6163 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6165 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6166 (fold_negate_expr): Likewise.
6167 (fold_real_zero_addition_p): Handle vectors.
6168 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6170 2013-06-14 Alan Modra <amodra@gmail.com>
6172 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6174 2013-06-13 Jan Hubicka <jh@suse.cz>
6176 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6177 Local comdats are not externally visible.
6178 * symtab.c (dump_symtab_base): Dump externally visible.
6179 (verify_symtab_base): Verify back links in the symtab hash.
6181 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6183 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6184 CONVERT_EXPR as equal nodes.
6186 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6188 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6190 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6192 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6193 Generalize to complex and vector.
6194 * tree.c (build_all_ones_cst): New function.
6195 * tree.h (build_all_ones_cst): Declare it.
6197 2013-06-13 Alan Modra <amodra@gmail.com>
6199 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6200 * config/rs6000/rs6000.md (signbittf2): New insn.
6201 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6202 (abstf2_internal, cmptf_internal2): Likewise.
6203 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6205 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6206 Pat Haugen <pthaugen@us.ibm.com>
6207 Peter Bergner <bergner@vnet.ibm.com>
6209 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6210 power8 byte, half-word, and quad-word atomic instructions.
6211 (emit_store_conditional): Likewise.
6212 (rs6000_expand_atomic_compare_and_swap): Likewise.
6213 (rs6000_expand_atomic_op): Likewise.
6215 * config/rs6000/sync.md (larx): Add new modes for power8.
6217 (AINT): New mode iterator to include TImode as well as normal
6218 integer modes on power8.
6219 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6220 that VSX registers are not considered. Use AINT mode iterator
6221 instead of INT1 to allow inclusion of quad word atomic operations
6223 (load_locked<mode>): Likewise.
6224 (store_conditional<mode>): Likewise.
6225 (atomic_compare_and_swap<mode>): Likewise.
6226 (atomic_exchange<mode>): Likewise.
6227 (atomic_nand<mode>): Likewise.
6228 (atomic_fetch_<fetchop_name><mode>): Likewise.
6229 (atomic_nand_fetch<mode>): Likewise.
6230 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
6232 (ATOMIC): On power8, add QImode, HImode modes.
6233 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
6234 modes that promote to SImode.
6235 (load_lockedti): Convert TImode arguments to PTImode, so that we
6236 get a guaranteed even/odd register pair.
6237 (load_lockedpti): Likewise.
6238 (store_conditionalti): Likewise.
6239 (store_conditionalpti): Likewise.
6241 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
6242 atomic load/store instructions.
6245 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
6247 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
6249 (insn_count): New attribute, with most cases extracted from...
6250 (length): ...here. Redefine most cases in terms of insn_count.
6251 (single_insn): Delete.
6252 (can_delay): Use insn_count to check for single instructions.
6253 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
6254 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
6255 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
6256 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
6257 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
6258 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
6259 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
6260 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
6261 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
6262 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
6263 rather than "length".
6264 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
6265 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
6266 Use "insn_count" rather than "length".
6267 * config/mips/mips-dsp.md
6268 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
6269 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
6272 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
6274 PR tree-optimization/57361
6275 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
6277 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
6279 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
6281 (aarch64_simd_combine<mode>): New instruction expansion.
6282 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
6284 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
6285 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
6287 2013-06-12 Jan Hubicka <jh@suse.cz>
6289 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
6290 decl has when in streaming stage.
6291 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
6292 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
6294 2013-06-12 Roland Stigge <stigge@antcom.de>
6297 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
6299 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6301 PR tree-optimization/57537
6302 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
6303 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
6305 2013-06-12 Richard Biener <rguenther@suse.de>
6307 * data-streamer.h (streamer_write_char_stream): CSE
6308 obs->current_pointer.
6309 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
6310 streamer_write_char_stream manually and optimize the resulting loop.
6311 (streamer_write_hwi_stream): Likewise.
6313 2013-06-12 Jan Hubicka <jh@suse.cz>
6315 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
6316 * cgraph.h (varpool_create_empty_node): Declare.
6317 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
6319 * symtab.c (symtab_unregister_node): Be lax about missin entries
6321 (symtab_get_node): Update comment.
6322 * varpool.c (varpool_create_empty_node): Break out from ...
6323 (varpool_node_for_decl): ... here.
6324 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
6326 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
6328 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
6329 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
6330 part. Use straight-line flow at the end.
6331 <COMPONENT_REF>: Remove superfluous else.
6332 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
6334 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6337 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
6338 target hook even for !TREE_PUBLIC decls. If no resolution info
6339 is available, return false for common and external decls.
6341 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6343 * config/rl78/constraints.md (U): New constraint.
6344 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
6347 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6350 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
6351 to allow returning address to AT_PLATFORM name.
6353 2013-06-11 Jan Hubicka <jh@suse.cz>
6355 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
6356 * cgraph.h (symtab_node_base): Add weakref flag.
6357 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
6358 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
6359 (output_weakrefs): Use weakref flag.
6360 * fold-const.c (simple_operand_p): Handle WEAK.
6361 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
6362 * ipa.c (varpool_externally_visible_p): Drop weakref.
6363 (function_and_variable_visibility): Update comment; fix weakref
6364 sanity checks; do not clear DECL_WEAK on them.
6365 * lto-cgraph.c (lto_output_node): update.
6366 (lto_output_varpool_node): Update.
6367 (input_overwrite_node): Update.
6368 (input_node): Update.
6369 (input_varpool_node): Update.
6370 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
6371 (lto_symtab_merge_symbols): Add sanity check.
6372 (lto_symtab_prevailing_decl): Do not special case weakrefs.
6373 * passes.c (rest_of_decl_compilation): Set static flag, too.
6374 * symtab.c (dump_symtab_base): Dump weakref.
6375 (verify_symtab_base): Sanity check weakrefs.
6376 (symtab_make_decl_local): Remove duplicated code.
6377 (symtab_alias_ultimate_target): Simplify.
6378 * varpool.c (varpool_create_variable_alias): Set weakref flag.
6380 2013-06-11 DJ Delorie <dj@redhat.com>
6382 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
6383 (rl78_unwind_word_mode): New.
6385 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6387 * final.c (debug_prefix_maps): Make static.
6389 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6391 * function.c (initial_trampoline): Remove stray copy.
6393 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
6395 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
6397 2013-06-11 Martin Jambor <mjambor@suse.cz>
6399 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
6400 within bounds at the beginning of the function.
6402 2013-06-11 Alan Modra <amodra@gmail.com>
6404 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
6406 (get_named_section): Don't NULL !DECL_P decl.
6408 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
6410 * doc/invoke.texi (core-avx2): Document.
6412 (atom): Updated with MOVBE.
6414 2013-06-11 Richard Biener <rguenther@suse.de>
6416 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
6418 2013-06-11 Anton Blanchard <anton@samba.org>
6420 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
6421 correct shift value in little-endian mode.
6423 2013-06-11 Jakub Jelinek <jakub@redhat.com>
6426 * varasm.c (get_variable_align): Move #endif to the right place.
6428 2013-06-10 Cary Coutant <ccoutant@google.com>
6430 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
6431 for hash so that hash table traversal order is deterministic.
6433 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
6434 Pat Haugen <pthaugen@us.ibm.com>
6435 Peter Bergner <bergner@vnet.ibm.com>
6437 * config/rs6000/vector.md (GPR move splitter): Do not split moves
6438 of vectors in GPRS if they are direct moves or quad word load or
6441 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
6443 (direct_move_p): Likewise.
6444 (quad_load_store_p): Likewise.
6446 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
6447 classes into bins based on the physical register type.
6448 (reg_class_to_reg_type): Likewise.
6449 (IS_STD_REG_TYPE): Likewise.
6450 (IS_FP_VECT_REG_TYPE): Likewise.
6451 (reload_fpr_gpr): Arrays to determine what insn to use if we can
6452 use direct move instructions.
6453 (reload_gpr_vsx): Likewise.
6454 (reload_vsx_gpr): Likewise.
6455 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
6456 information that is a simplification of register classes. Also
6457 precalculate direct move reload helpers.
6458 (direct_move_p): New function to return true if the operation can
6459 be done as a direct move instruciton.
6460 (quad_load_store_p): New function to return true if the operation
6461 is a quad memory operation.
6462 (rs6000_legitimize_address): If quad memory, only allow register
6463 indirect for TImode addresses.
6464 (rs6000_legitimate_address_p): Likewise.
6465 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
6466 (rs6000_reload_register_type): Likewise.
6467 (register_to_reg_type): Return register type.
6468 (rs6000_secondary_reload_simple_move): New helper function for
6469 secondary reload and secondary memory needed to identify anything
6470 that is a simple move, and does not need reloading.
6471 (rs6000_secondary_reload_direct_move): New helper function for
6472 secondary reload to identify cases that can be done with several
6473 instructions via the direct move instructions.
6474 (rs6000_secondary_reload_move): New helper function for secondary
6475 reload to identify moves between register types that can be done.
6476 (rs6000_secondary_reload): Add support for quad memory operations
6477 and for direct move.
6478 (rs6000_secondary_memory_needed): Likewise.
6479 (rs6000_debug_secondary_memory_needed): Change argument names.
6480 (rs6000_output_move_128bit): New function to return the move to
6481 use for 128-bit moves, including knowing about the various
6482 limitations of quad memory operations.
6484 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
6485 memory operations. call rs6000_output_move_128bit for the actual
6486 instruciton(s) to generate.
6487 (vsx_movti_64bit): Likewise.
6489 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
6490 (UNSPEC_P8V_MTVSRWZ): Likewise.
6491 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6492 (UNSPEC_P8V_MTVSRD): Likewise.
6493 (UNSPEC_P8V_XXPERMDI): Likewise.
6494 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6495 (UNSPEC_FUSION_GPR): Likewise.
6496 (FMOVE128_GPR): New iterator for direct move.
6497 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6500 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6501 loads and direct move instructions.
6502 (zero_extendsidi2_lfiwzx): Likewise.
6503 (extendsidi2_lfiwax): Likewise.
6504 (extendsidi2_nocell): Likewise.
6505 (floatsi<mode>2_lfiwax): Likewise.
6507 (floatunssi<mode>2_lfiwzx): Likewise.
6509 (fix_trunc<mode>_stfiwx): Likewise.
6510 (fixuns_trunc<mode>_stfiwx): Likewise.
6511 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6512 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6513 (parity<mode>2_cmpb): Set length/type attr.
6514 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6515 for 'mr.' to fast_compare.
6516 (bpermd_<mode>): Change type attr to popcnt.
6517 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6518 (p8_mtvsrwz_1): Likewise.
6519 (p8_mtvsrwz_2): Likewise.
6520 (reload_fpr_from_gpr<mode>): Likewise.
6521 (p8_mtvsrd_1): Likewise.
6522 (p8_mtvsrd_2): Likewise.
6523 (p8_xxpermdi_<mode>): Likewise.
6524 (reload_vsx_from_gpr<mode>): Likewise.
6525 (reload_vsx_from_gprsf): Likewise.
6526 (p8_mfvsrd_3_<mode>): LIkewise.
6527 (reload_gpr_from_vsx<mode>): Likewise.
6528 (reload_gpr_from_vsxsf): Likewise.
6529 (p8_mfvsrd_4_disf): Likewise.
6530 (multi-word GPR splits): Do not split direct moves or quad memory
6533 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6535 * tree-into-ssa.c (interesting_blocks): Make static.
6537 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6540 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6541 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6542 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6543 (get_variable_align): New function.
6544 (get_variable_section, emit_bss, emit_common,
6545 assemble_variable_contents, place_block_symbol): Use
6546 get_variable_align instead of DECL_ALIGN.
6547 (assemble_noswitch_variable): Add align argument, use it
6548 instead of DECL_ALIGN.
6549 (assemble_variable): Adjust caller. Use get_variable_align
6550 instead of DECL_ALIGN.
6551 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6553 (DATA_ABI_ALIGNMENT): Define.
6554 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6555 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6556 opt is false, only return the psABI mandated alignment increase.
6557 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6558 (DATA_ABI_ALIGNMENT): ... this.
6559 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6560 (DATA_ABI_ALIGNMENT): ... this.
6561 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6562 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6563 (DATA_ABI_ALIGNMENT): ... this.
6564 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6565 * doc/tm.texi: Regenerated.
6567 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6569 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6570 cmp_code to construct REG_EQUAL note.
6572 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6575 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6576 that operands[2] doesn't overlap with operands[0].
6578 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6579 Jan Hubicka <jh@suse.cz>
6581 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6582 hack to mark symbols as used.
6584 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6586 PR rtl-optimization/57559
6587 * lra-constraints.c (process_alt_operands): Don't discourage
6588 memory with known offset for offsetable memory constraint.
6589 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6591 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6593 * varasm.c (struct oc_local_state): Reorder fields.
6594 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6595 and adjust accordingly.
6596 (output_constructor): Reorder initialization code and adjust call to
6597 output_constructor_bitfield.
6599 2013-06-07 Jan Hubicka <jh@suse.cz>
6601 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6603 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6605 * tree-object-size.c (unknown): Make const.
6607 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6609 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6610 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6611 for last alternative in the cpu_facility attribute.
6613 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6616 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6617 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6618 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6619 * config/arm/constraints.md (Dg): New constraint.
6620 * config/arm/neon.md (xordi3_neon): Remove.
6621 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6622 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6624 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6626 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6627 Clean up alternatives.
6629 2013-06-07 Alan Modra <amodra@gmail.com>
6631 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6634 2013-06-07 Alan Modra <amodra@gmail.com>
6636 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6638 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6640 * config/arm/constraints.md (Df): New constraint.
6641 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6642 Correct length attribute for last two alternatives.
6644 2013-06-07 Alan Modra <amodra@gmail.com>
6646 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6647 override user -mfp-in-toc.
6648 (offsettable_ok_by_alignment): Consider just the current access
6649 rather than the whole object, unless BLKmode. Handle
6650 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6651 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6652 for -mcmodel=medium.
6653 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6654 override user -mfp-in-toc or -msum-in-toc. Default to
6655 -mno-fp-in-toc for -mcmodel=medium.
6657 2013-06-06 DJ Delorie <dj@redhat.com>
6659 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6660 TARGET_VALID_POINTER_MODE.
6662 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6663 Pat Haugen <pthaugen@us.ibm.com>
6664 Peter Bergner <bergner@vnet.ibm.com>
6666 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6667 Document new power8 builtins.
6669 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6670 condition code register, to allow 128-bit logical operations to be
6671 done in the VSX or GPR registers.
6672 (nor<mode>3): Use the canonical form for nor.
6673 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6674 vclz*, and vpopcnt* vector instructions.
6675 (nand<mode>3): Likewise.
6676 (orc<mode>3): Likewise.
6677 (clz<mode>2): LIkewise.
6678 (popcount<mode>2): Likewise.
6680 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6681 that only the GPRs are recognized.
6683 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6684 support for new power8 builtins.
6686 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6688 (xscvdpspn): Likewise.
6694 (vpopcnt): Likewise.
6695 (vpopcntb): Likewise.
6696 (vpopcnth): Likewise.
6697 (vpopcntw): Likewise.
6698 (vpopcntd): Likewise.
6703 (eqv_v16qi3): Likewise.
6704 (eqv_v8hi3): Likewise.
6705 (eqv_v4si3): Likewise.
6706 (eqv_v2di3): Likewise.
6707 (eqv_v4sf3): Likewise.
6708 (eqv_v2df3): Likewise.
6710 (nand_v16qi3): Likewise.
6711 (nand_v8hi3): Likewise.
6712 (nand_v4si3): Likewise.
6713 (nand_v2di3): Likewise.
6714 (nand_v4sf3): Likewise.
6715 (nand_v2df3): Likewise.
6717 (orc_v16qi3): Likewise.
6718 (orc_v8hi3): Likewise.
6719 (orc_v4si3): Likewise.
6720 (orc_v2di3): Likewise.
6721 (orc_v4sf3): Likewise.
6722 (orc_v2df3): Likewise.
6724 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6725 allow power8 quad mode in 64-bit.
6726 (rs6000_builtin_vectorized_function): Add support to vectorize
6727 ISA 2.07 count leading zeros, population count builtins.
6728 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6729 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6730 (builtin_function_type): Add vgbbd builtin function which takes an
6732 (altivec_expand_vec_perm_const): Add support for new power8 merge
6735 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6736 that does not include TImdoe for use with 32-bit.
6737 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6739 (UNSPEC_VSX_CVDPSPN): Likewise.
6740 (vsx_xscvdpspn): Likewise.
6741 (vsx_xscvspdpn): Likewise.
6742 (vsx_xscvdpspn_scalar): Likewise.
6743 (vsx_xscvspdpn_directmove): Likewise.
6744 (vsx_and<mode>3): Split logical operations into 32-bit and
6745 64-bit. Add support to do logical operations on TImode as well as
6746 VSX vector types. Allow logical operations to be done in either
6747 VSX registers or in general purpose registers in 64-bit mode. Add
6748 splitters if GPRs were used. For AND, add clobber of CCmode to
6749 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6751 (vsx_and<mode>3_32bit): Likewise.
6752 (vsx_and<mode>3_64bit): Likewise.
6753 (vsx_ior<mode>3): Likewise.
6754 (vsx_ior<mode>3_32bit): Likewise.
6755 (vsx_ior<mode>3_64bit): Likewise.
6756 (vsx_xor<mode>3): Likewise.
6757 (vsx_xor<mode>3_32bit): Likewise.
6758 (vsx_xor<mode>3_64bit): Likewise.
6759 (vsx_one_cmpl<mode>2): Likewise.
6760 (vsx_one_cmpl<mode>2_32bit): Likewise.
6761 (vsx_one_cmpl<mode>2_64bit): Likewise.
6762 (vsx_nor<mode>3): Likewise.
6763 (vsx_nor<mode>3_32bit): Likewise.
6764 (vsx_nor<mode>3_64bit): Likewise.
6765 (vsx_andc<mode>3): Likewise.
6766 (vsx_andc<mode>3_32bit): Likewise.
6767 (vsx_andc<mode>3_64bit): Likewise.
6768 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6769 and xxlorc instructions.
6770 (vsx_eqv<mode>3_64bit): Likewise.
6771 (vsx_nand<mode>3_32bit): Likewise.
6772 (vsx_nand<mode>3_64bit): Likewise.
6773 (vsx_orc<mode>3_32bit): Likewise.
6774 (vsx_orc<mode>3_64bit): Likewise.
6776 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6778 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6780 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6781 (p8_vmrgow): Likewise.
6782 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6783 GPRs to be split under VSX.
6784 (p8v_clz<mode>2): Add power8 count leading zero support.
6785 (p8v_popcount<mode>2): Add power8 population count support.
6786 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6789 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6792 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6794 (vec_nand): Likewise.
6795 (vec_vclz): Likewise.
6796 (vec_vclzb): Likewise.
6797 (vec_vclzd): Likewise.
6798 (vec_vclzh): Likewise.
6799 (vec_vclzw): Likewise.
6800 (vec_vgbbd): Likewise.
6801 (vec_vmrgew): Likewise.
6802 (vec_vmrgow): Likewise.
6803 (vec_vpopcnt): Likewise.
6804 (vec_vpopcntb): Likewise.
6805 (vec_vpopcntd): Likewise.
6806 (vec_vpopcnth): Likewise.
6807 (vec_vpopcntw): Likewise.
6809 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6811 PR rtl-optimization/57468
6812 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6815 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6817 PR rtl-optimization/57459
6818 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6819 type when setting live regs.
6821 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6823 * config/s390/s390.opt (mlra): New option.
6824 * config/s390/s390.c (s390_decompose_address): Check displacement
6825 for all registers for LRA.
6826 (s390_secondary_reload): Don't used secondary reloads for LRA.
6827 (s390_lra_p): New function.
6828 (TARGET_LRA_P): Define.
6829 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6830 of attribute cpu_facility to zarch for the last alternative.
6831 (*cmpmem_short): Ditto.
6833 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6835 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6836 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6837 (arm_expand_prologue): Likewise.
6839 2013-06-06 Teresa Johnson <tejohnson@google.com>
6842 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6843 as this is now done by redirect_edge_and_branch_force.
6844 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6845 barriers, and fix interaction with splitting.
6846 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6847 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6848 reflect changes made in the routine.
6849 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6850 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6851 since this is called in cfglayout mode, and replace partition fixup
6852 with assert as that is now done by force_nonfallthru_and_redirect.
6853 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6854 already be marked with region crossing note.
6855 (insert_section_boundary_note): Make non-static, gate on flag
6856 has_bb_partition, rewrite to also check for multiple partitions.
6857 (rest_of_handle_reorder_blocks): Remove call to
6858 insert_section_boundary_note, now done later during free_cfg.
6859 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6860 * bb-reorder.h (insert_section_boundary_note): Declare.
6861 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6862 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6863 invoke insert_section_boundary_note.
6864 (try_redirect_by_replacing_jump): Remove unnecessary
6865 check for region crossing note.
6866 (fixup_partition_crossing): New function.
6867 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6868 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6869 in non-cfglayout mode.
6870 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6871 remove old code that tried to do this. Emit barrier correctly
6872 when we are in cfglayout mode.
6873 (last_bb_in_partition): New function.
6874 (rtl_split_edge): Correctly fixup partition boundaries.
6875 (commit_one_edge_insertion): Remove old code that tried to
6876 fixup region crossing edge since this is now handled in
6877 split_block, and set up insertion point correctly since
6878 block may now end in a jump.
6879 (verify_hot_cold_block_grouping): Guard against checking when not in
6880 linearized RTL mode.
6881 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6883 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6884 rtl_verify_flow_info, so not called in cfglayout mode.
6885 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6886 (fixup_reorder_chain): Remove old code that attempted to fixup region
6887 crossing note as this is now handled in force_nonfallthru_and_redirect.
6888 (duplicate_insn_chain): Don't duplicate switch section notes.
6889 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6891 * basic-block.h (emit_barrier_after_bb): Declare.
6893 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6895 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6896 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6897 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6899 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6901 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6903 * config/arm/ldmstm.md: Regenerate.
6905 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6907 * config/arm/sync.md (atomic_loaddi_1):
6908 Disable predication for arm_restrict_it.
6909 (arm_load_exclusive<mode>): Likewise.
6910 (arm_load_exclusivesi): Likewise.
6911 (arm_load_exclusivedi): Likewise.
6912 (arm_load_acquire_exclusive<mode>): Likewise.
6913 (arm_load_acquire_exclusivesi): Likewise.
6914 (arm_load_acquire_exclusivedi): Likewise.
6915 (arm_store_exclusive<mode>): Likewise.
6916 (arm_store_exclusive<mode>): Likewise.
6917 (arm_store_release_exclusivedi): Likewise.
6918 (arm_store_release_exclusive<mode>): Likewise.
6920 2013-06-06 Richard Biener <rguenther@suse.de>
6922 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6924 (lto_tag_is_tree_code_p): Adjust.
6925 (lto_tag_is_gimple_code_p): Likewise.
6926 (lto_gimple_code_to_tag): Likewise.
6927 (lto_tag_to_gimple_code): Likewise.
6928 (lto_tree_code_to_tag): Likewise.
6929 (lto_tag_to_tree_code): Likewise.
6930 * data-streamer.h (streamer_write_hwi_in_range): Use
6931 uhwi streaming to stream the normalized range.
6932 (streamer_read_hwi_in_range): Likewise.
6934 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6936 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6937 (predicable_short_it): Likewise.
6938 (predicated): Likewise.
6939 (enabled): Handle above.
6940 (define_cond_exec): Set predicated attribute to yes.
6942 2013-06-05 Mike Stump <mikestump@comcast.net>
6944 * gdbinit.in (__FUNCTION__): Add.
6946 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6948 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6949 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6951 2013-06-05 Jan Hubicka <jh@suse.cz>
6953 * varasm.c (mark_decl_referenced): Revert the removal until targets
6956 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6958 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6959 instead of mark_decl_referenced.
6961 2013-06-05 Jan Hubicka <jh@suse.cz>
6963 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6964 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6965 and symtab_used_from_object_file_p.
6966 (cgraph_make_node_local_1): Clear forced_by_abi.
6967 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6968 * cgraph.h (symtab_node_base): Add forced_by_abi.
6969 (decide_is_variable_needed): Remove.
6970 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6971 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6972 (decide_is_symbol_needed): ... this one; handle symbols in general;
6973 always analyze virtuals; honnor forced_by_abi.
6974 (cgraph_finalize_function): Update.
6975 (varpool_finalize_decl): Update.
6976 (symbol_defined_and_needed): Remove.
6977 (analyze_functions): Update.
6978 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6979 output_refs, input_overwrite_node): Handle forced_by_abi.
6980 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6981 (address_taken_from_non_vtable_p): ... this one.
6982 (comdat_can_be_unshared_p_1): New function.
6983 (cgraph_comdat_can_be_unshared_p): Rename to ...
6984 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6985 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6986 (function_and_variable_visibility): Clear forced_by_abi as needed.
6987 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6988 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6989 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6990 * varpool.c (decide_is_variable_needed): Remove.
6992 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6994 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6995 (arm_option_override): Override arm_restrict_it where appropriate.
6996 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6997 * config/arm/arm.opt (mrestrict-it): New command-line option.
6998 * doc/invoke.texi: Document -mrestrict-it.
7000 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7002 * tsan.c (tsan_atomic_table): Make const.
7004 2013-06-05 Richard Biener <rguenther@suse.de>
7006 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7007 index associated with the tree we are supposed to replace.
7008 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7009 TREE_ASM_WRITTEN as zero for everything but SSA names.
7011 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7013 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7015 2013-06-04 Jan Hubicka <jh@suse.cz>
7017 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7018 (input_node, input_varpool_node): Handle correctly external same
7020 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7021 nodes at ltrans stage.
7023 2013-06-04 Jan Hubicka <jh@suse.cz>
7025 * ipa-inline.c (update_caller_keys): Fix availability test.
7026 (update_callee_keys): Likewise.
7027 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7028 to follow ELF standard.
7030 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7032 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7033 (mips64r5900el-*-elf*): New configurations.
7034 * config/mips/mips-cpus.def (r5900): New processor.
7035 * config/mips/mips-tables.opt: Regenerate.
7036 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7037 (mips_issue_rate): Handle PROCESSOR_R5900.
7038 (mips_reorg_process_insns): Force reorder mode for the R5900.
7039 * config/mips/mips.h (TARGET_MIPS5900): Define.
7040 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7042 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7044 * config/mips/mips.md (processor): Add r5900.
7045 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7047 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7049 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7050 into function to generate MOVI instruction.
7051 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7052 (aarch64_preferred_simd_mode): Turn into wrapper.
7053 (aarch64_output_scalar_simd_mov_immediate): New function.
7054 * config/aarch64/aarch64-protos.h: Add prototype for above.
7056 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7058 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7059 element_char member.
7060 (sizetochar): Return signed char.
7061 (aarch64_simd_valid_immediate): Remove elchar and other
7062 unnecessary variables.
7063 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7064 Calculate element_char as required.
7065 * config/aarch64/aarch64-protos.h: Update and move prototype
7066 for aarch64_output_simd_mov_immediate.
7067 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7070 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7072 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7073 information completed by aarch64_simd_valid_immediate.
7074 (aarch64_legitimate_constant_p): Update arguments.
7075 (aarch64_simd_valid_immediate): Work with struct rather than many
7077 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7078 (aarch64_simd_make_constant): Update arguments.
7079 (aarch64_output_simd_mov_immediate): Work with struct rather than
7080 many pointers. Output immediate directly rather than as operand.
7081 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7083 * config/aarch64/constraints.md (Dn): Update arguments.
7085 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7087 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7089 (aarch64_simd_immediate_valid_for_move): Remove.
7090 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7091 (aarch64_simd_make_constant): Update call.
7092 (aarch64_output_simd_mov_immediate): Update call.
7093 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7095 * config/aarch64/constraints.md (Dn): Update call.
7097 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7099 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7100 return type to bool for prototype.
7101 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7102 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7103 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7105 2013-06-04 Catherine Moore <clm@codesourcery.com>
7107 * config/mips/mips.opt (meva): New.
7108 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7109 (ASM_SPEC): Handle -meva.
7110 * doc/invoke.texi (meva): Document.
7112 2013-06-04 Alan Modra <amodra@gmail.com>
7114 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7117 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7119 * rtl.def: Add extra fourth optional field to define_cond_exec.
7120 * gensupport.c (process_one_cond_exec): Process attributes from
7122 * doc/md.texi: Document fourth field in define_cond_exec.
7124 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7126 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7127 out the processing order as in store_bit_field_1.
7129 2013-06-04 Jan Hubicka <jh@suse.cz>
7132 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7133 non-VAR_DECL node if it does not exist yet.
7135 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7137 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7138 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7139 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7140 target_cpu_default setting.
7142 2013-06-03 Teresa Johnson <tejohnson@google.com>
7144 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7145 default to -fopt-info=optimized instead of all.
7146 * doc/invoke.texi: Ditto.
7147 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7148 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7149 (execute_vect_slp): Emit BB vectorization success under
7150 MSG_OPTIMIZED_LOCATIONS.
7151 * tree-vect-slp.c (vect_slp_transform_bb): Change
7152 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7153 * tree-vect-loop.c (vect_transform_loop): Ditto.
7155 2013-06-03 Jason Merrill <jason@redhat.com>
7158 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7159 Use TARGET_EXPR for C++.
7161 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7163 PR rtl-optimization/57268
7164 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7165 if DEBUG_INSN_P (insn).
7168 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7170 PR rtl-optimization/57268
7171 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7172 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7174 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7176 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7177 (ix86_avoid_lea_for_addr): Likewise.
7178 (exact_dependency_1): Likewise.
7179 (ix86_adjust_cost): Likewise.
7180 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7182 (do_reorder_for_imul): Fix typo, formatting and
7183 !reload_completed check removed.
7184 (ix86_sched_reorder): Fix typo and formatting.
7185 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7188 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7190 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7192 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7194 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7195 <STRING_CST>: Likewise.
7196 <VECTOR_CST>: Likewise.
7198 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7199 Mikael Morin <mikael@gcc.gnu.org>
7201 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7202 * config.in: Regenerated.
7203 * configure: Regenerated.
7205 2013-06-01 Jan Hubicka <jh@suse.cz>
7208 * cgraphunit.c (compile): When weakref is not supported,
7209 set up transparent aliases before final output pass.
7210 * varasm.c (assemble_alias): Do not try to do it here.
7212 2013-06-01 Jan Hubicka <jh@suse.cz>
7215 * passes.c (for_per_function): Skip unanalyzed functions.
7217 2013-06-01 Jan Hubicka <jh@suse.cz>
7219 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7220 (lto_symtab_merge_symbols_1): ... this one.
7221 (lto_symtab_merge_cgraph_nodes): Rename to ...
7222 (lto_symtab_merge_symbols): ... this one; simplify.
7223 * cgraph.c (same_body_aliases_done): Rename to ...
7224 (cpp_implicit_aliases_done): ... this one.
7225 (cgraph_create_function_alias): Update.
7226 (cgraph_same_body_alias): Update.
7227 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
7228 (verify_edge_corresponds_to_fndecl): Simplify.
7229 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
7230 (cgraph_node): Remove same_body_alias.
7231 (varpool_node): Remove alias_of and extra_name_alias.
7232 (same_body_aliases_done): Rename to ..
7233 (cpp_implicit_aliases_done): ... this one.
7234 (symtab_alias_ultimate_target): Add default parameter.
7235 (symtab_resolve_alias): New function.
7236 (fixup_same_cpp_alias_visibility): Declare.
7237 (cgraph_function_node): Add default parameter.
7238 (cgraph_node_asm_name): Likewise.
7239 (cgraph_function_or_thunk_node): Add default parameter; do
7240 not ICE when it is NULL.
7241 (varpool_variable_node): Likewise.
7242 * tree-emutls.c (create_emultls_var): Update.
7243 (ipa_lower_emutls): Update.
7244 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
7245 (cgraph_reset_node): Reset alias info.
7246 (cgraph_finalize_function): Update.
7247 (fixup_same_cpp_alias_visibility): Move to symtab.c.
7248 (analyze_function): Simplify.
7249 (cgraph_process_same_body_aliases): Simplify.
7250 (analyze_functions): Fixup same body aliases.
7251 (handle_alias_pairs): Simplify.
7252 (assemble_thunk): Update.
7253 (assemble_thunks_and_aliases): Update.
7254 (output_weakrefs): Rewrite.
7255 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
7256 (lto_output_varpool_node): Likewise.
7257 (compute_ltrans_boundary): Remve assert.
7258 (get_alias_symbol): New functoin.
7259 (input_node): Rewrite alias handling.
7260 (input_varpool_node): Likewise.
7261 * ipa-pure-const.c (propagate_pure_const): Fix formating.
7262 * ipa.c (process_references): Handle weakrefs correctly.
7263 (symtab_remove_unreachable_nodes): Likewise.
7264 * trans-mem.c (get_cg_data): Update.
7265 (ipa_tm_create_version_alias): Update.
7266 (ipa_tm_execute): Update.
7267 * symtab.c (dump_symtab_base): Dump aliases.
7268 (verify_symtab_base): Verify aliases.
7269 (symtab_node_availability): New function.
7270 (symtab_alias_ultimate_target): Simplify.
7271 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
7272 handle all the fixup cases.
7273 (symtab_resolve_alias): New function.
7274 * passes.c (ipa_write_summaries): Handle weakrefs.
7275 * varpool.c (varpool_analyze_node): Simplify.
7276 (assemble_aliases): Update.
7277 (varpool_create_variable_alias): Simplify.
7278 (varpool_extra_name_alias): Simplify.
7279 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
7280 (lto_symtab_merge_symbols): ... this one.
7282 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
7285 PR rtl-optimization/57268
7286 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7287 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7289 2013-06-01 Tobias Burnus <burnus@net-b.de>
7292 2013-05-31 Tobias Burnus <burnus@net-b.de>
7295 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
7298 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7300 PR rtl-optimization/57268
7301 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7302 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7304 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7306 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
7307 unordered comparison operators when -fno-trapping-math is in effect
7309 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
7310 and implement unordered comparison operators properly on the e500.
7312 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7314 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
7315 for constant scalar integers.
7316 (simplify_relational_operation_1): Likewise.
7318 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
7320 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
7321 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
7324 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7325 Igor Zamyatin <igor.zamyatin@intel.com>
7327 Silvermont (SLM) architecture performance tuning.
7328 * config/i386/i386.h (enum ix86_tune_indices): Add
7329 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
7330 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
7332 * config/i386/i386.c (initial_ix86_tune_features)
7333 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
7334 (ix86_lea_outperforms): Handle Silvermont tuning.
7335 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
7337 (ix86_use_lea_for_mov): Likewise.
7338 (ix86_avoid_lea_for_addr): Likewise.
7339 (ix86_lea_for_add_ok): Likewise.
7340 (exact_dependency_1): New function.
7341 (exact_store_load_dependency): Likewise.
7342 (ix86_adjust_cost): Handle Silvermont tuning.
7343 (do_reoder_for_imul): Likewise.
7344 (swap_top_of_ready_list): New function.
7345 (ix86_sched_reorder): Changed to handle Silvermont tuning.
7347 * config/i386/i386.md (peepholes that split memory operand in fp
7350 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7352 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
7353 Remove un-necessary braces.
7355 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7357 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7358 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
7360 2013-05-31 Tobias Burnus <burnus@net-b.de>
7363 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
7365 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7368 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
7369 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
7370 * config/arm/neon.md (iordi3_neon): Remove.
7371 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
7372 * config/arm/predicates.md (imm_for_neon_logic_operand):
7373 Move to earlier in the file.
7374 (neon_logic_op2): Likewise.
7375 (arm_iordi_operand_neon): New predicate.
7377 2013-05-31 Richard Biener <rguenther@suse.de>
7379 PR tree-optimization/57478
7380 PR tree-optimization/57453
7381 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
7384 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7386 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
7387 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
7389 2013-05-30 Tobias Burnus <burnus@net-b.de>
7390 Thomas Koenig <tkoenig@gcc.gnu.org>
7393 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
7394 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
7396 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
7398 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
7400 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
7402 * target.def (register_usage_leveling_p): New hook.
7403 * targhooks.c (default_register_usage_leveling_p): New.
7404 * targhooks.h (default_register_usage_leveling_p): New prototype.
7405 * lra-assigns.c (register_usage_leveling_p): Use the hook.
7406 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
7407 * doc/tm.texi: Update.
7408 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
7410 2013-05-30 Ian Bolton <ian.bolton@arm.com>
7412 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
7413 (*insv_reg<mode>): New define_insn.
7415 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
7417 PR rtl-optimization/57439
7418 * postreload.c (move2add_valid_value_p): Check that we have
7419 a zero subreg_regno_offset when accessing the register in
7422 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7423 Igor Zamyatin <igor.zamyatin@intel.com>
7425 Silvermont (SLM) architecture pipeline model, tuning and
7427 * config.gcc: Add slm config options and target.
7429 * config/i386/slm.md: New.
7431 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
7433 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
7435 (ix86_target_macros_internal): Likewise.
7437 * gcc/config/i386/i386.c (slm_cost): New cost.
7438 (m_SLM): New macro flag.
7439 (initial_ix86_tune_features): Set m_SLM.
7440 (x86_accumulate_outgoing_args): Likewise.
7441 (x86_arch_always_fancy_math_387): Likewise.
7442 (processor_target_table): Add slm cost.
7443 (cpu_names): Add slm cpu name.
7444 (x86_option_override_internal): Set SLM ISA.
7445 (ix86_issue_rate): New case PROCESSOR_SLM.
7446 (ia32_multipass_dfa_lookahead): Likewise.
7447 (fold_builtin_cpu): Add slm.
7449 * config/i386/i386.h (TARGET_SLM): New target macro.
7450 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
7451 (processor_type): Add PROCESSOR_SLM.
7453 * config/i386/i386.md (cpu): Add new value "slm".
7454 (slm.md): Include slm.md.
7456 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
7457 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7459 * config/arm/arm-protos.h: Add and update function protos.
7460 * config/arm/arm.c (use_simple_return_p): New added.
7461 (thumb2_expand_return): Check simple_return flag.
7462 * config/arm/arm.md: Add simple_return and conditional simple_return.
7463 * config/arm/iterators.md: Add iterator for return and simple_return.
7465 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7467 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
7468 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
7469 (arm_emit_vfp_multi_reg_pop): Likewise.
7470 (thumb2_emit_ldrd_pop): Likewise.
7471 (arm_expand_epilogue): Add misc REG_CFA notes.
7472 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
7474 2013-05-29 Lawrence Crowl <crowl@google.com>
7476 * config/arm/t-arm: Update for below.
7478 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
7479 Change type to hash_table. Update dependent calls and types.
7481 * config/i386/t-cygming: Update for below.
7483 * config/i386/t-interix: Update for below.
7485 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
7486 Change type to hash_table. Update dependent calls and types.
7487 (i386_find_on_wrapper_list::wrappers): Likewise.
7489 * config/ia64/t-ia64: Update for below.
7491 * config/ia64/ia64.c (bundle_state_table):
7492 Change type to hash_table. Update dependent calls and types.
7494 * config/mips/mips.c (mips_reorg_process_insns::htab):
7495 Change type to hash_table. Update dependent calls and types.
7497 * config/sol2.c (solaris_comdat_htab):
7498 Change type to hash_table. Update dependent calls and types.
7500 * config/t-sol2: Update for above.
7502 2013-05-29 Teresa Johnson <tejohnson@google.com>
7504 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7505 functions are not yet marked as defined.
7507 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7508 Pat Haugen <pthaugen@us.ibm.com>
7509 Peter Bergner <bergner@vnet.ibm.com>
7511 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7515 (vrotl<mode>3): Likewise.
7516 (vashl<mode>3): Likewise.
7517 (vlshr<mode>3): Likewise.
7518 (vashr<mode>3): Likewise.
7520 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7521 support for power8 V2DI builtins.
7523 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7524 power8 V2DI builtins.
7525 (vupkhsw): Likewise.
7526 (vupklsw): Likewise.
7527 (vaddudm): Likewise.
7532 (vpkudum): Likewise.
7533 (vpksdss): Likewise.
7534 (vpkudus): Likewise.
7535 (vpksdus): Likewise.
7540 (vsubudm): Likewise.
7541 (vcmpequd): Likewise.
7542 (vcmpgtsd): Likewise.
7543 (vcmpgtud): Likewise.
7544 (vcmpequd_p): Likewise.
7545 (vcmpgtsd_p): Likewise.
7546 (vcmpgtud_p): Likewise.
7547 (vupkhsw): Likewise.
7548 (vupklsw): Likewise.
7549 (vaddudm): Likewise.
7554 (vpksdss): Likewise.
7555 (vpksdus): Likewise.
7556 (vpkudum): Likewise.
7557 (vpkudus): Likewise.
7562 (vsubudm): Likewise.
7564 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7565 support for power8 V2DI instructions.
7567 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7568 power8 V2DI instructions. Combine pack and unpack insns to use an
7569 iterator for each mode. Check whether a particular mode supports
7570 Altivec instructions instead of just checking TARGET_ALTIVEC.
7571 (UNSPEC_VPKUWUM): Likewise.
7572 (UNSPEC_VPKSHSS): Likewise.
7573 (UNSPEC_VPKSWSS): Likewise.
7574 (UNSPEC_VPKUHUS): Likewise.
7575 (UNSPEC_VPKSHUS): Likewise.
7576 (UNSPEC_VPKUWUS): Likewise.
7577 (UNSPEC_VPKSWUS): Likewise.
7578 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7579 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7580 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7581 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7582 (UNSPEC_VUPKHSB): Likewise.
7583 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7584 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7585 (UNSPEC_VUPKHSH): Likewise.
7586 (UNSPEC_VUPKLSB): Likewise.
7587 (UNSPEC_VUPKLSH): Likewise.
7589 (VI_char): Likewise.
7590 (VI_scalar): Likewise.
7591 (VI_unit): Likewise.
7593 (VP_small): Likewise.
7594 (VP_small_lc): Likewise.
7595 (VU_char): Likewise.
7596 (add<mode>3): Likewise.
7597 (altivec_vaddcuw): Likewise.
7598 (altivec_vaddu<VI_char>s): Likewise.
7599 (altivec_vadds<VI_char>s): Likewise.
7600 (sub<mode>3): Likewise.
7601 (altivec_vsubcuw): Likewise.
7602 (altivec_vsubu<VI_char>s): Likewise.
7603 (altivec_vsubs<VI_char>s): Likewise.
7604 (altivec_vavgs<VI_char>): Likewise.
7605 (altivec_vcmpbfp): Likewise.
7606 (altivec_eq<mode>): Likewise.
7607 (altivec_gt<mode>): Likewise.
7608 (altivec_gtu<mode>): Likewise.
7609 (umax<mode>3): Likewise.
7610 (smax<mode>3): Likewise.
7611 (umin<mode>3): Likewise.
7612 (smin<mode>3): Likewise.
7613 (altivec_vpkuhum): Likewise.
7614 (altivec_vpkuwum): Likewise.
7615 (altivec_vpkshss): Likewise.
7616 (altivec_vpkswss): Likewise.
7617 (altivec_vpkuhus): Likewise.
7618 (altivec_vpkshus): Likewise.
7619 (altivec_vpkuwus): Likewise.
7620 (altivec_vpkswus): Likewise.
7621 (altivec_vpks<VI_char>ss): Likewise.
7622 (altivec_vpks<VI_char>us): Likewise.
7623 (altivec_vpku<VI_char>us): Likewise.
7624 (altivec_vpku<VI_char>um): Likewise.
7625 (altivec_vrl<VI_char>): Likewise.
7626 (altivec_vsl<VI_char>): Likewise.
7627 (altivec_vsr<VI_char>): Likewise.
7628 (altivec_vsra<VI_char>): Likewise.
7629 (altivec_vsldoi_<mode>): Likewise.
7630 (altivec_vupkhsb): Likewise.
7631 (altivec_vupkhs<VU_char>): Likewise.
7632 (altivec_vupkls<VU_char>): Likewise.
7633 (altivec_vupkhsh): Likewise.
7634 (altivec_vupklsb): Likewise.
7635 (altivec_vupklsh): Likewise.
7636 (altivec_vcmpequ<VI_char>_p): Likewise.
7637 (altivec_vcmpgts<VI_char>_p): Likewise.
7638 (altivec_vcmpgtu<VI_char>_p): Likewise.
7639 (abs<mode>2): Likewise.
7640 (vec_unpacks_hi_v16qi): Likewise.
7641 (vec_unpacks_hi_v8hi): Likewise.
7642 (vec_unpacks_lo_v16qi): Likewise.
7643 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7644 (vec_unpacks_lo_v8hi): Likewise.
7645 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7646 (vec_pack_trunc_v8h): Likewise.
7647 (vec_pack_trunc_v4si): Likewise.
7648 (vec_pack_trunc_<mode>): Likewise.
7650 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7652 (vec_vmaxsd): Likewise.
7653 (vec_vmaxud): Likewise.
7654 (vec_vminsd): Likewise.
7655 (vec_vminud): Likewise.
7656 (vec_vpksdss): Likewise.
7657 (vec_vpksdus): Likewise.
7658 (vec_vpkudum): Likewise.
7659 (vec_vpkudus): Likewise.
7660 (vec_vrld): Likewise.
7661 (vec_vsld): Likewise.
7662 (vec_vsrad): Likewise.
7663 (vec_vsrd): Likewise.
7664 (vec_vsubudm): Likewise.
7665 (vec_vupkhsw): Likewise.
7666 (vec_vupklsw): Likewise.
7668 2013-05-29 Jan Hubicka <jh@suse.cz>
7670 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7671 flags; reorder rest of fields in more consistent way.
7672 (varpool_node): Remove analyzed, finalized and alias.
7673 (cgraph_ndoe): Likewise.
7674 (symtab_alias_ultimate_target): New function.
7675 (cgraph_function_node): Move offline.
7676 (cgraph_reset_node): Declare.
7677 (cgraph_comdat_can_be_unshared_p): Remove.
7678 (varpool_remove_initializer): Declare.
7679 (varpool_first_defined_variable, varpool_next_defined_variable
7680 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7681 (cgraph_function_with_gimple_body_p): Update.
7682 (varpool_all_refs_explicit_p): Update.
7683 (symtab_alias_target): New function.
7684 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7685 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7686 (cgraph_function_or_thunk_node): Simplify using
7687 symtab_alias_ultimate_target.
7688 (varpool_variable_node): Likewise.
7689 * cgraph.c (cgraph_create_function_alias): Update.
7690 (cgraph_add_thunk): Update.
7691 (cgraph_remove_node): Update.
7692 (dump_cgraph_node): Do not dump removed flags.
7693 (cgraph_function_body_availability): Update.
7694 (cgraph_propagate_frequency): Update.
7695 (verify_cgraph_node): Check sanity of local flag.
7696 (cgraph_function_node): Move here from cgraph.h; revamp for
7697 cgraph_function_or_thunk_node.
7698 * lto-symtab.c (lto_varpool_replace_node): Update.
7699 (lto_symtab_resolve_can_prevail_p): Update.
7700 (lto_symtab_merge_cgraph_nodes): Update.
7701 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7702 propagate_constants_accross_call, devirtualization_time_bonus,
7703 ipcp_propagate_stage): Update.
7704 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7705 * ipa-inline-transform.c (clone_inlined_nodes,
7706 preserve_function_body_p): Update.
7707 * ipa-reference.c (propagate): Update.
7708 (write_node_summary_p): Update.
7709 * toplev.c (wrapup_global_declaration_2): Update.
7710 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7711 (analyze_function) ... this one.
7712 (cgraph_process_new_functions): Update.
7713 (cgraph_reset_node): Export.
7714 (cgraph_finalize_function): Update.
7715 (cgraph_add_new_function): Update.
7716 (process_function_and_variable_attributes): Update.
7717 (varpool_finalize_decl): Update.
7718 (symbol_finalized): Remove.
7719 (symbol_finalized_and_needed): Rename to ...
7720 (symbol_defined_and_needed): ... update.
7721 (cgraph_analyze_functions): Update.
7722 (handle_alias_pairs): Update.
7723 (mark_functions_to_output): Update.
7724 (assemble_thunk): Update.
7725 (output_in_order): Update.
7726 (output_weakrefs): Update.
7727 (finalize_compilation_unit): Update.
7728 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7729 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7730 input_node, input_varpool_node): Update.
7731 * dbxout.c (dbxout_expand_expr): Update.
7732 * cgraphclones.c (cgraph_clone_node): Update.
7733 (cgraph_copy_node_for_versioning): Update.
7734 (cgraph_materialize_clone): Update.
7735 (cgraph_materialize_all_clones): Update.
7736 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7737 propagate_pure_const, propagate_nothrow): Update.
7738 * lto-streamer-out.c (lto_output, write_symbol): Update.
7739 * ipa-utils.c (ipa_reverse_postorder): Update.
7740 * ipa-inline.c (can_inline_edge_p): Update.
7741 (update_caller_keys, ipa_inline): Update.
7742 * dwarf2out.c (reference_to_unused,
7743 premark_types_used_by_global_vars_helper): Update.
7744 * tree-eh.c (tree_could_trap_p): Update.
7745 * ipa-split.c (consider_split, execute_split_functions): Update.
7746 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7747 has_addr_references_p): Update; move ahead in file for better
7749 (process_references): Simplify.
7750 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7752 (cgraph_comdat_can_be_unshared_p): Make static.
7753 (cgraph_externally_visible_p): Update.
7754 (varpool_externally_visible_p): Update.
7755 (function_and_variable_visibility): Update.
7756 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7757 ipa_tm_mark_force_output_node): Update.
7758 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7759 estimate_edge_devirt_benefit, inline_generate_summary,
7760 inline_write_summary): Update.
7761 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7762 * ipa-prop.c (ipa_compute_jump_functions): Update.
7763 (ipa_print_node_params, ipa_prop_read_section,
7764 ipa_update_after_lto_read, read_replacements_section): Update.
7765 * varasm.c (mark_decl_referenced): Update.
7766 (assemble_alias, dump_tm_clone_pairs): Update.
7767 * tree-inline.c (copy_bb): Update.
7768 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7770 * symtab.c (dump_symtab_base): Print new flags.
7771 (verify_symtab_base): Verify new flags.
7772 (symtab_alias_ultimate_target): New function.
7773 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7774 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7776 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7778 * i386.c (ix86_get_function_versions_dispatcher,
7779 ix86_generate_version_dispatcher_body): Update.
7780 (fold_builtin_cpu): Use varpool_add_new_variable.
7781 * varpool.c (varpool_remove_initializer): Break out from ...
7782 (varpool_remove_node): ... this one.
7783 (dump_varpool_node, varpool_node_for_asm,
7784 cgraph_variable_initializer_availability, varpool_analyze_node,
7785 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7786 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7789 2013-05-29 Jan Hubicka <jh@suse.cz>
7791 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7793 2013-05-29 Easwaran Raman <eraman@google.com>
7795 PR tree-optimization/57442
7796 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7797 when control exits the main loop.
7799 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7801 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7803 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7804 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7805 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7808 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7810 PR tree-optimization/57441
7811 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7812 Don't limit size of incr_vec to number of candidates.
7814 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7816 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7817 and mips16 directories.
7818 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7819 (MULTILIB_DIRNAMES): Ditto.
7820 (MULTILIB_EXCEPTIONS): Add new exceptions.
7821 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7822 (MULTILIB_DIRNAMES): Ditto.
7823 (MULTILIB_EXCEPTIONS): Add new exceptions.
7825 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7826 Marcus Shawcroft <marcus.shawcroft@arm.com>
7828 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7829 SYMBOL_TINY_ABSOLUTE.
7830 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7831 SYMBOL_TINY_ABSOLUTE.
7832 (aarch64_expand_mov_immediate): Likewise.
7833 (aarch64_classify_symbol): Likewise.
7834 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7835 Permit SYMBOL_TINY_ABSOLUTE.
7836 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7838 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7839 Marcus Shawcroft <marcus.shawcroft@arm.com>
7841 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7842 Refactor if/switch. Replace gcc_assert with if.
7844 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7846 * config/i386/i386.c (initial_ix86_tune_features): Enable
7847 FP Reassociation for AMD bdver1 and bdver2.
7849 2013-05-29 Martin Jambor <mjambor@suse.cz>
7851 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7852 and IMAGPART_EXPR do not occur within other handled_components.
7854 2013-05-29 Richard Biener <rguenther@suse.de>
7856 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7857 access on whether the use is in the BB we currently try to
7859 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7860 vectorize to vect_bb_slp_scalar_cost.
7862 2013-05-29 Richard Biener <rguenther@suse.de>
7864 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7865 computing scalar cost offsetted by stmts that are kept live
7867 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7868 for computation of scalar cost.
7870 2013-05-28 Steve Ellcey <sellcey@mips.com>
7872 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7874 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7876 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7877 array notation built-in reduction functions.
7878 * doc/passes.texi (Passes): Added documentation about changes done
7880 * doc/invoke.texi (C Dialect Options): Added documentation about
7881 the -fcilkplus flag.
7882 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7883 (BUILTINS_DEF): Depend on cilkplus.def.
7884 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7885 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7886 * cilkplus.def: New file.
7888 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7890 PR rtl-optimization/57439
7891 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7893 2013-05-28 Easwaran Raman <eraman@google.com>
7895 PR tree-optimization/57337
7896 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7897 (find_insert_point): Correctly identify the insertion point
7898 when two statements with the same UID is compared.
7900 2013-05-28 Richard Biener <rguenther@suse.de>
7902 PR tree-optimization/56787
7903 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7904 from the list of data references.
7905 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7907 (vect_analyze_loop_operations): Likewise.
7908 (vect_transform_loop): Remove clobbers.
7910 2013-05-28 Martin Jambor <mjambor@suse.cz>
7912 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7913 and REALPART_EXPRs have scalar type.
7915 2013-05-28 Richard Biener <rguenther@suse.de>
7917 PR tree-optimization/57411
7918 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7920 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7921 virtual operand propagation.
7923 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7925 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7926 destination register for bmasksi_vis.
7927 (vector_init_bshuffle): Likewise.
7928 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7930 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7932 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7933 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7934 mode if the instruction isn't available in the original mode.
7935 * config/sparc/sparc.opt (mfix-ut699): New option.
7936 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7937 (divdf3): Turn into expander.
7938 (divdf3_nofix): New insn.
7939 (divdf3_fix): Likewise.
7940 (divsf3): Disable if -mfix-ut699.
7941 (sqrtdf2): Turn into expander.
7942 (sqrtdf2_nofix): New insn.
7943 (sqrtdf2_fix): Likewise.
7944 (sqrtsf2): Disable if -mfix-ut699.
7946 2013-05-27 Richard Biener <rguenther@suse.de>
7949 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7950 block for the new loop.
7952 2013-05-27 Richard Biener <rguenther@suse.de>
7954 PR tree-optimization/57343
7955 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7956 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7957 (number_of_iterations_cond): Do not build the folded tree.
7959 2013-05-27 Richard Biener <rguenther@suse.de>
7963 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7964 OEP_CONSTANT_ADDRESS_OF retained.
7966 PR tree-optimization/57417
7967 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7969 (set_ssa_val_to): Compare addresses using
7970 get_addr_base_and_unit_offset.
7972 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7974 PR rtl-optimization/56833
7975 * postreload.c (move2add_record_mode): New function.
7976 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7977 (move2add_use_add2_insn): Use move2add_record_sym_value.
7978 (move2add_use_add3_insn): Likewise.
7979 (reload_cse_move2add): Use move2add_valid_value_p and
7980 move2add_record_mode. Invalidate call-clobbered and REG_INC
7981 affected regs by setting reg_mode to VOIDmode.
7982 (move2add_note_store): Don't pretend the inside of a SUBREG is
7983 the actual destination. Invalidate single/leading registers by
7984 setting reg_mode to VOIDmode.
7985 Use move2add_record_sym_value, move2add_valid_value_p and
7986 move2add_record_mode.
7988 2013-05-27 Richard Biener <rguenther@suse.de>
7990 PR tree-optimization/57396
7991 * tree-affine.c (double_int_constant_multiple_p): Properly
7992 return false for val == 0 and div != 0.
7994 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7996 * config/mips/mips.h: Use #elif in preprocessor conditions.
7998 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8001 * config/mips/constraints.md (kl): New constraint.
8002 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8003 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8004 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8006 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8008 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8011 * config/mips/mips.c (mips_can_inline_p): New function.
8012 (TARGET_CAN_INLINE_P): Define.
8014 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8016 * sched-int.h (ds_t, dw_t): Make unsigned int.
8017 Fix documentation that describes how all the ds_t bits are used.
8018 Reserve the last bit for delayed-branch scheduling.
8019 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8020 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8021 (gen_dep_weak_1): Remove prototype.
8022 * sched-deps.c (get_dep_weak_1): Make static.
8023 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8024 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8025 * doc/tm.texi: Regenerate.
8026 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8028 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8031 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8033 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8034 Sandra Loosemore <sandra@codesourcery.com>
8036 * config.gcc (powerpc-*): Allow native for with-cpu.
8038 2013-05-24 Jeff Law <law@redhat.com>
8040 PR tree-optimization/57124
8041 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8042 conversion feeding a condition if the range has an overflow
8043 if -fstrict-overflow. Add warnings for when we do make the
8046 2013-05-24 Dehao Chen <dehao@google.com>
8048 * tree-cfg.c (locus_discrim_map): Fix the typo.
8049 (locus_discrim_hasher): Likewise.
8050 (locus_discrim_hasher::hash): Likewise.
8051 (locus_discrim_hasher::equal): Likewise.
8053 2013-05-24 Martin Jambor <mjambor@suse.cz>
8055 PR tree-optimization/57294
8056 * cgraph.h (ipa_record_stmt_references): Declare.
8057 * cgraphbuild.c (ipa_record_stmt_references): New function.
8058 (build_cgraph_edges): Use ipa_record_stmt_references.
8059 (rebuild_cgraph_edges): Likewise.
8060 (cgraph_rebuild_references): Likewise.
8061 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8062 associated with the old statement and build references from the
8063 newly built statements.
8064 * ipa-ref.c (ipa_remove_stmt_references): New function.
8065 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8067 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8069 * lra-constraints.c (emit_spill_move): Use smaller mode for
8071 (check_and_process_move): Consider mem-reg moves for secondary
8073 (curr_insn_transform): Don't lose insns emitted before for
8074 secondary memory moves.
8075 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8076 reg set up in the current insn.
8078 2013-05-24 Dehao Chen <dehao@google.com>
8080 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8082 (locus_descrim_hasher::equal): Likewise.
8083 (build_gimple_cfg): New discriminator assignment algorithm.
8084 (make_edges): Likewise.
8085 (next_discriminator_for_locus): Likewise.
8086 (same_line_p): Likewise.
8087 (assign_discriminators): Likewise.
8088 (make_cond_expr_edges): Likewise.
8089 (make_gimple_switch_edges): Likewise.
8090 (make_goto_expr_edges): Likewise.
8091 (make_gimple_asm_edges): Likewise.
8093 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8095 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8096 X format specifier to only display bottom 16 bits.
8097 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8098 immediate to match for operand 2, since it will be masked.
8100 2013-05-24 Richard Biener <rguenther@suse.de>
8102 PR tree-optimization/57287
8103 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8104 all SSA names that occur in abnormal PHIs.
8106 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8109 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8110 that index is not negative.
8112 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8114 PR rtl-optimization/55177
8115 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8116 (simplify_byte_swapping_operation): New.
8117 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8118 (simplify_relational_operation_1): Deal with BSWAP.
8120 2013-05-23 Richard Henderson <rth@redhat.com>
8123 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8124 (ix86_reorg): Call it.
8126 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8129 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8130 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8131 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8133 2013-05-23 Christian Bruel <christian.bruel@st.com>
8136 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8138 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8139 Marcus Shawcroft <marcus.shawcroft@arm.com>
8141 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8142 * config/aarch64/constraints.md (Usa): Remove.
8143 * doc/md.texi (AArch64 Usa): Remove.
8145 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8146 Marcus Shawcroft <marcus.shawcroft@arm.com>
8148 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8149 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8150 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8151 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8153 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8155 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8156 instruction (AdvSIMD).
8157 * config/aarch64/aarch64-builtins.c
8158 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8159 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8161 2013-05-23 Martin Jambor <mjambor@suse.cz>
8164 * tree.h (contains_bitfld_component_ref_p): Declare.
8165 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8166 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8168 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8169 not access a bit-field. Assert all final offsets are byte-aligned.
8171 2013-05-23 Richard Biener <rguenther@suse.de>
8173 PR tree-optimization/57380
8174 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8175 least one invariant or re-used load.
8176 * passes.c (init_optimization_passes): Move pass_phiprop before
8179 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8181 * config/aarch64/aarch64-simd.md
8182 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8184 2013-05-23 Richard Biener <rguenther@suse.de>
8187 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8188 OEP_CONSTANT_ADDRESS_OF retained.
8190 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8193 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8194 don't lower unit. Handle unit not being always BITS_PER_WORD.
8196 2013-05-23 Richard Biener <rguenther@suse.de>
8198 PR rtl-optimization/57341
8199 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8200 instead of true_dependence.
8202 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8204 * bb-reorder.c (branch_threshold): Make const.
8205 (exec_threshold): Ditto.
8207 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8208 Pat Haugen <pthaugen@us.ibm.com>
8209 Peter Bergner <bergner@vnet.ibm.com>
8211 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8212 documentation for the power8 crypto builtins.
8214 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8216 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8217 macros for defining power8 builtin functions.
8218 (BU_P8V_AV_2): Likewise.
8219 (BU_P8V_AV_P): Likewise.
8220 (BU_P8V_VSX_1): Likewise.
8221 (BU_P8V_OVERLOAD_1): Likewise.
8222 (BU_P8V_OVERLOAD_2): Likewise.
8223 (BU_CRYPTO_1): Likewise.
8224 (BU_CRYPTO_2): Likewise.
8225 (BU_CRYPTO_3): Likewise.
8226 (BU_CRYPTO_OVERLOAD_1): Likewise.
8227 (BU_CRYPTO_OVERLOAD_2): Likewise.
8228 (XSCVSPDP): Fix typo, point to the correct instruction.
8229 (VCIPHER): Add power8 crypto builtins.
8230 (VCIPHERLAST): Likewise.
8231 (VNCIPHER): Likewise.
8232 (VNCIPHERLAST): Likewise.
8233 (VPMSUMB): Likewise.
8234 (VPMSUMH): Likewise.
8235 (VPMSUMW): Likewise.
8236 (VPERMXOR_V2DI): Likewise.
8237 (VPERMXOR_V4SI: Likewise.
8238 (VPERMXOR_V8HI: Likewise.
8239 (VPERMXOR_V16QI: Likewise.
8240 (VSHASIGMAW): Likewise.
8241 (VSHASIGMAD): Likewise.
8243 (VPERMXOR): Likewise.
8244 (VSHASIGMA): Likewise.
8246 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8247 __CRYPTO__ if the crypto instructions are available.
8248 (altivec_overloaded_builtins): Add support for overloaded power8
8251 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
8252 support for power8 crypto builtins.
8253 (builtin_function_type): Likewise.
8254 (altivec_init_builtins): Add support for builtins that take vector
8255 long long (V2DI) arguments.
8257 * config/rs6000/crypto.md: New file, define power8 crypto
8260 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8261 Pat Haugen <pthaugen@us.ibm.com>
8262 Peter Bergner <bergner@vnet.ibm.com>
8264 * doc/invoke.texi (Option Summary): Add power8 options.
8265 (RS/6000 and PowerPC Options): Likewise.
8267 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
8268 constraints.md instead of rs6000.h. Reorder w* constraints. Add
8269 wm, wn, wr documentation.
8271 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
8272 registers if direct move instructions are enabled.
8273 (wn): New constraint for no registers.
8274 (wq): New constraint for quad word even GPR registers.
8275 (wr): New constraint if 64-bit instructions are enabled.
8276 (wv): New constraint if power8 vector instructions are enabled.
8277 (wQ): New constraint for quad word memory locations.
8279 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
8280 constraint for 0..15 for crypto instructions.
8281 (gpc_reg_operand): If VSX allow registers in VSX registers as well
8282 as GPR and floating point registers.
8283 (int_reg_operand): New predicate to match only GPR registers.
8284 (base_reg_operand): New predicate to match base registers.
8285 (quad_int_reg_operand): New predicate to match even GPR registers
8286 for quad memory operations.
8287 (vsx_reg_or_cint_operand): New predicate to allow vector logical
8288 operations in both GPR and VSX registers.
8289 (quad_memory_operand): New predicate for quad memory operations.
8290 (reg_or_indexed_operand): New predicate for direct move support.
8292 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
8293 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
8294 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
8295 (POWERPC_MASKS): Add power8 options.
8296 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
8299 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8300 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
8302 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
8303 (-mpower8-fusion): New power8 options.
8304 (-mpower8-fusion-sign): Likewise.
8305 (-mpower8-vector): Likewise.
8306 (-mcrypto): Likewise.
8307 (-mdirect-move): Likewise.
8308 (-mquad-memory): Likewise.
8310 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
8312 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
8314 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
8315 (rs6000_debug_vector_unit): Add p8_vector.
8316 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
8317 definitions. Also print fusion state.
8318 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
8319 (rs6000_builtin_mask_calculate): Add power8 builtin support.
8320 (rs6000_option_override_internal): Add support for power8.
8321 (rs6000_common_init_builtins): Add debugging for skipped builtins
8323 (rs6000_adjust_cost): Add power8 support.
8324 (rs6000_issue_rate): Likewise.
8325 (insn_must_be_first_in_group): Likewise.
8326 (insn_must_be_last_in_group): Likewise.
8327 (force_new_group): Likewise.
8328 (rs6000_register_move_cost): Likewise.
8329 (rs6000_opt_masks): Likewise.
8331 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
8332 power8 capable assembler, default to power7 options.
8333 (TARGET_DIRECT_MOVE): Likewise.
8334 (TARGET_CRYPTO): Likewise.
8335 (TARGET_P8_VECTOR): Likewise.
8336 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
8337 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
8338 (VECTOR_MEM_P8_VECTOR_P): Likewise.
8339 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
8340 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
8341 (TARGET_XSCVDPSPN): Likewise.
8342 (TARGET_XSCVSPDPN): Likewsie.
8343 (TARGET_SYNC_HI_QI): Likewise.
8344 (TARGET_SYNC_TI): Likewise.
8345 (MASK_CRYPTO): Likewise.
8346 (MASK_DIRECT_MOVE): Likewise.
8347 (MASK_P8_FUSION): Likewise.
8348 (MASK_P8_VECTOR): Likewise.
8349 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
8350 temporary used by some of the direct move instructions to get two FP
8351 temporary registers does not force creation of a stack frame.
8352 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
8353 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
8354 that any VSX registers are tieable, even if they are also an
8355 Altivec vector mode.
8356 (r6000_reg_class_enum): Add wm, wr, wv constraints.
8357 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
8358 (RS6000_BTM_CRYPTO): Likewise.
8359 (RS6000_BTM_COMMON): Likewise.
8361 * config/rs6000/rs6000.md (cpu attribute): Add power8.
8362 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
8363 (enum rs6000_vector): Add power8 vector support.
8365 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8369 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
8370 (any_sibcall_could_use_r3): this and handle indirect calls.
8371 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
8373 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8375 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
8377 2013-05-22 Richard Biener <rguenther@suse.de>
8380 * profile.c (branch_prob): Do not split blocks that are
8381 abnormally receiving from ECF_RETURNS_TWICE functions.
8383 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8385 * recog.c (offsettable_address_addr_space_p): Fix calculation of
8386 address mode. Move pointer mode initialization to the same place.
8388 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8390 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
8391 while it has any effect.
8393 2013-05-21 Easwaran Raman <eraman@google.com>
8395 PR tree-optimization/57322
8396 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
8397 UID of the statement added to the BB to be 1.
8399 2013-05-21 Jakub Jelinek <jakub@redhat.com>
8401 PR tree-optimization/57331
8402 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
8403 of conversion from pointer type to integral type with integer.
8405 2013-05-21 Martin Jambor <mjambor@suse.cz>
8408 * ipa-prop.c (ipa_read_node_info): Process param_used and
8409 controlled_uses in the same order as when writing.
8411 2013-05-21 Magnus Granberg <baldrick@free.fr>
8414 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
8416 2013-05-21 Richard Biener <rguenther@suse.de>
8418 PR tree-optimization/57318
8419 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
8420 estimate stmts with side-effects as likely eliminated.
8422 2013-05-21 Richard Biener <rguenther@suse.de>
8424 PR tree-optimization/57330
8425 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
8426 preserve the call stmts fntype.
8428 2013-05-21 Richard Biener <rguenther@suse.de>
8430 PR tree-optimization/57303
8431 * tree-ssa-sink.c (statement_sink_location): Improve killing
8432 stmt detection and properly handle self-assignments.
8434 2013-05-21 Christian Bruel <christian.bruel@st.com>
8436 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
8437 spanning registers. LEAF_REG_REMAP is supported only for contiguous
8438 registers. Set register size out of the PARALLEL loop.
8440 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
8443 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
8444 (*fmasf4, *fmasf4_media): New insns.
8446 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8448 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
8449 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
8450 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
8451 (mips_idiv_insns): Update the comments to say that the returned
8452 instruction counts are in units of BASE_INSN_LENGTH.
8453 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
8454 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
8455 using 2 rather than 4 as the length of indirect MIPS16 and
8456 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
8457 length of a NOP. Don't divide MIPS16 lengths by 2.
8458 (mips16_split_long_branches): Assume a branch is long if the
8459 length is greater than 4 rather than 8.
8460 * config/mips/mips.md (length): Give MIPS16 lengths directly,
8461 rather than multiplying them by 2. Multiply instruction counts
8462 by BASE_INSN_LENGTH rather than 4.
8463 (*jump_mips16, tls_get_tp_mips16_<mode>)
8464 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
8466 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8468 * config/mips/mips.md (extended_mips16): Remove branch case.
8469 (length): Remove duplicated extended_mips16 test.
8471 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8473 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
8475 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
8477 * recog.h (Recog_data): Rename to...
8478 (recog_data_d): ...this.
8479 (recog_data): Update accordingly.
8480 * recog.c (recog_data): Likewise.
8481 * reload.c (save_recog_data): Likewise.
8482 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
8483 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
8485 2013-05-17 Julian Brown <julian@codesourcery.com>
8487 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
8488 found in a REG_EQUAL note, invalidate it.
8490 2013-05-17 Easwaran Raman <eraman@google.com>
8492 * tree-ssa-reassoc.c (find_insert_point): New function.
8493 (insert_stmt_after): Likewise.
8494 (get_def_stmt): Likewise.
8495 (ensure_ops_are_available): Likewise.
8496 (not_dominated_by): Likewise.
8497 (rewrite_expr_tree): Do not move statements beyond what is
8498 necessary. Remove call to swap_ops_for_binary_stmt...
8499 (reassociate_bb): ... and move it here.
8500 (build_and_add_sum): Assign UIDs for new statements.
8501 (linearize_expr): Likewise.
8502 (do_reassoc): Renumber gimple statement UIDs.
8504 2013-05-17 Jan Hubicka <jh@suse.cz>
8506 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8508 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8509 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8511 (output_weakrefs): Update.
8513 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8514 Martin Jambor <mjambor@suse.cz>
8517 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8518 value that corresponds to the given aggval is found in values vector.
8520 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8522 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8523 sse, sse2, sse3, ssse3 and sse4a flags to options.
8525 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8527 * gengtype-state.c: (s_expr_writer): New class, to handle
8528 prettifying of output layout of s-expressions.
8529 (state_writer): New class, to write out gtype.state.
8530 (state_written_type_count): Move this variable into member data of
8532 (s_expr_writer::s_expr_writer): New code: constructor for new class
8533 (state_writer::state_writer(): ditto
8534 (s_expr_writer::write_new_line): New function
8535 (s_expr_writer::write_any_indent): ditto
8536 (s_expr_writer::begin_s_expr): ditto
8537 (s_expr_writer::end_s_expr): ditto
8538 (write_state_fileloc): convert to method of state_writer...
8539 (state_writer:: write_state_fileloc): ...and use methods of
8540 s_expr_writer to write indentation into the gtype.state output file
8541 to visually represent the hierarchical structure of the list
8543 (write_state_fields): ditto, renaming to...
8544 (state_writer::write_state_fields)
8545 (write_state_a_string): ditto, renaming to...
8546 (state_writer::write_state_a_string)
8547 (write_state_string_option): ditto, renaming to...
8548 (state_writer::write_state_string_option)
8549 (write_state_type_option): ditto, renaming to...
8550 (state_writer::write_state_type_option)
8551 (write_state_nested_option): ditto, renaming to...
8552 (state_writer::write_state_nested_option)
8553 (write_state_option): ditto, renaming to...
8554 (state_writer::write_state_option)
8555 (write_state_options): ditto, renaming to...
8556 (state_writer::write_state_options)
8557 (write_state_lang_bitmap): ditto, renaming to...
8558 (state_writer::write_state_lang_bitmap)
8559 (write_state_version): ditto, renaming to...
8560 (state_writer::write_state_version)
8561 (write_state_scalar_type): ditto, renaming to...
8562 (state_writer::write_state_scalar_type)
8563 (write_state_string_type): ditto, renaming to...
8564 (state_writer::write_state_string_type)
8565 (write_state_undefined_type): ditto, renaming to...
8566 (state_writer::write_state_undefined_type)
8567 (write_state_struct_union_type): ditto, renaming to...
8568 (state_writer::write_state_struct_union_type)
8569 (write_state_struct_type): ditto, renaming to...
8570 (state_writer::write_state_struct_type)
8571 (write_state_user_struct_type): ditto, renaming to...
8572 (state_writer::write_state_user_struct_type)
8573 (write_state_lang_struct_type): ditto, renaming to...
8574 (state_writer::write_state_lang_struct_type)
8575 (write_state_param_struct_type): ditto, renaming to...
8576 (state_writer::write_state_param_struct_type)
8577 (write_state_pointer_type): ditto, renaming to...
8578 (state_writer::write_state_pointer_type)
8579 (write_state_array_type): ditto, renaming to...
8580 (state_writer::write_state_array_type)
8581 (write_state_gc_used): ditto, renaming to...
8582 (state_writer::write_state_gc_used)
8583 (write_state_common_type_content): ditto, renaming to...
8584 (state_writer::write_state_common_type_content)
8585 (write_state_type): ditto, renaming to...
8586 (state_writer::write_state_type)
8587 (write_state_pair_list): ditto, renaming to...
8588 (state_writer::write_state_pair_list)
8589 (write_state_pair): ditto, renaming to...
8590 (state_writer::write_state_pair)
8591 (write_state_typedefs): ditto, renaming to...
8592 (state_writer::write_state_typedefs)
8593 (write_state_structures): ditto, renaming to...
8594 (state_writer::write_state_structures)
8595 (write_state_param_structs): ditto, renaming to...
8596 (state_writer::write_state_param_structs)
8597 (write_state_variables): ditto, renaming to...
8598 (state_writer::write_state_variables)
8599 (write_state_srcdir): ditto, renaming to...
8600 (state_writer::write_state_srcdir)
8601 (write_state_files_list): ditto, renaming to...
8602 (state_writer::write_state_files_list)
8603 (write_state_languages): ditto, renaming to...
8604 (state_writer::write_state_languages)
8605 (write_state): create a state_writer instance and use it when
8606 writing out the state file
8608 2013-05-17 Mike Stump <mikestump@comcast.net>
8610 PR rtl-optimization/57304
8611 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8612 accessing DF_REF_REAL_LOC.
8614 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8616 PR rtl-optimization/57281
8617 PR rtl-optimization/57300
8618 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8619 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8620 what the other splitter did if the registers are dead.
8622 2013-05-17 Richard Biener <rguenther@suse.de>
8624 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8627 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8629 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8632 2013-05-17 Marek Polacek <polacek@redhat.com>
8634 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8635 length when doing non-zero store of storing '\0' to '\0'.
8637 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8639 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8640 vect_external_def oprnd1 with loop_vinfo, try to emit
8641 optional cast, negation and and stmts on the loop preheader
8642 edge instead of into the pattern def seq.
8644 PR tree-optimization/57051
8645 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8646 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8648 2013-05-16 Nick Clifton <nickc@redhat.com>
8650 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8651 (rl78_is_naked_func): New function.
8652 (rl78_expand_prologue): Skip prologue generation for naked functions.
8653 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8654 * doc/extend.texi (naked): Add RL78 to the list of processors
8655 that supports this attribute.
8657 2013-05-16 Jeff Law <law@redhat.com>
8659 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8661 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8663 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8664 cache parameters using detect_caches_amd also for CYRIX,
8665 NSC and TM2 signatures.
8667 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8668 Dzianis Kahanovich <mahatma@eu.by>
8672 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8673 VIA/Centaur processors and determine their cache parameters
8674 using detect_caches_amd.
8676 2013-05-16 Teresa Johnson <tejohnson@google.com>
8678 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8679 (rtl_verify_edges): New function.
8680 (rtl_verify_bb_insns): Ditto.
8681 (rtl_verify_bb_pointers): Ditto.
8682 (rtl_verify_bb_insn_chain): Ditto.
8683 (rtl_verify_fallthru): Ditto.
8684 (rtl_verify_bb_layout): Ditto.
8685 (rtl_verify_flow_info_1): Outline checks into new functions.
8686 (rtl_verify_flow_info): Ditto.
8688 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8690 * cfghooks.c (copy_bbs): Add update_dominance argument.
8691 * cfghooks.h (copy_bbs): Update prototype.
8692 * tree-cfg.c (gimple_duplicate_sese_region):
8693 Add update_dominance argument.
8694 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8695 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8696 gimple_duplicate_sese_region call.
8697 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8698 Update copy_bbs call.
8699 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8700 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8702 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8704 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8705 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8706 vect_recog_rotate_pattern.
8707 (vect_recog_rotate_pattern): New function.
8709 2013-05-16 Jason Merrill <jason@redhat.com>
8711 * Makefile.in (LLINKER): New variable.
8712 (mostlyclean): Remove link mutex.
8713 * configure.ac: Handle --enable-link-mutex.
8714 * lock-and-run.sh: New script.
8716 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8719 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8722 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8724 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8726 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8728 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8729 * config/arm/arm.c (next_consecutive_mem): New function.
8730 (gen_movmem_ldrd_strd): Likewise.
8731 * config/arm/arm.md (movmemqi): Update condition and code.
8732 (unaligned_loaddi, unaligned_storedi): New patterns.
8734 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8736 * config.gcc: Obsolete *-*-solaris2.9*.
8737 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8739 2013-05-16 Richard Biener <rguenther@suse.de>
8741 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8742 earlier, after GRAPHITE transforms and IV canonicalization.
8744 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8746 * omp-low.c (extract_omp_for_data): For collapsed loops,
8747 if at least one of the loops is known at compile time to
8748 iterate zero times, set count to 0.
8749 (expand_omp_regimplify_p): New function.
8750 (expand_omp_for_generic): For collapsed loops, if at least
8751 one of the loops isn't known to iterate at least once,
8752 add runtime check with setting count to 0.
8753 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8754 For unsigned types if it isn't known at compile time that
8755 the loop will iterate at least once, add runtime check to bypass
8756 the whole loop if initial condition isn't true.
8758 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8760 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8762 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8765 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8766 transformations to avoid an infinite loop.
8768 2013-05-16 Marek Polacek <polacek@redhat.com>
8770 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8772 2013-05-15 Leif Ekblad <leif@rdos.net>
8774 * config/i386/i386.c (ix86_decompose_address): Use
8775 DEFAULT_TLS_SEG_REG to access TLS segment register.
8776 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8777 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8778 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8780 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8783 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8784 sibling calls to functions that would normally be lazily bound,
8785 unless $gp is call-clobbered.
8787 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8789 * config/i386/i386.c (ix86_option_override_internal): Update
8790 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8791 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8792 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8793 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8795 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8797 2013-05-15 Andreas Schwab <schwab@suse.de>
8799 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8800 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8803 2013-05-15 Teresa Johnson <tejohnson@google.com>
8805 * loop-unroll.c (report_unroll_peel): Check decision before
8806 emitting unroll/peel message.
8808 2013-05-15 Teresa Johnson <tejohnson@google.com>
8810 * function.h (has_bb_partition): New rtl_data flag.
8811 (bb_reorder_complete): Ditto.
8812 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8813 instead of flag_reorder_blocks_and_partition.
8814 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8815 with some enhancements.
8816 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8817 * bb-reorder.c (connect_traces): Check for has_bb_partition
8818 instead of flag_reorder_blocks_and_partition.
8819 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8820 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8821 verify_hot_cold_block_grouping.
8822 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8824 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8827 * config/arm/predicates.md (call_insn_operand): New predicate.
8828 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8829 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8830 if insn is not a tail call.
8831 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8833 * config/arm/arm.h (enum reg_class): New caller save register class.
8834 (REG_CLASS_NAMES): Likewise.
8835 (REG_CLASS_CONTENTS): Likewise.
8836 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8839 2013-05-15 Richard Biener <rguenther@suse.de>
8841 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8842 of MSG_OPTIMIZED_LOCATIONS.
8843 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8844 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8846 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8847 of MSG_OPTIMIZED_LOCATIONS.
8848 (execute_vect_slp): Likewise.
8849 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8850 (vect_create_cond_for_alias_checks): Likewise.
8851 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8852 (vect_recog_widen_mult_pattern): Likewise.
8853 (vect_recog_widen_sum_pattern): Likewise.
8854 (vect_recog_over_widening_pattern): Likewise.
8855 (vect_recog_widen_shift_pattern): Likewise.
8856 (vect_recog_vector_vector_shift_pattern): Likewise.
8857 (vect_recog_divmod_pattern): Likewise.
8858 (vect_recog_mixed_size_cond_pattern): Likewise.
8859 (vect_recog_bool_pattern): Likewise.
8860 (vect_pattern_recog_1): Likewise.
8862 2013-05-15 Martin Jambor <mjambor@suse.cz>
8864 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8865 non-functions to builtin_unreachable.
8866 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8867 correct when new direct edges were discovered.
8869 2013-05-15 Martin Jambor <mjambor@suse.cz>
8871 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8872 header, print symbol order instead of node uid, print more information
8873 about indirect edge targets.
8874 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8876 (ipa_make_edge_direct_to_target): Likewise.
8877 (remove_described_reference): Likewise.
8878 (propagate_controlled_uses): Likewise.
8879 (ipa_print_node_params): Also print symbol order.
8880 (ipcp_transform_function): Print symbol order instead of node uids.
8881 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8882 (cgraph_get_create_real_symbol_node): Likewise.
8883 * ipa-cp.c (print_lattice): Likewise.
8884 (print_all_lattices): Likewise.
8885 (determine_versionability): Likewise.
8886 (initialize_node_lattices): Likewise.
8887 (estimate_local_effects): Likewise.
8888 (update_profiling_info): Likewise.
8889 (create_specialized_node): Likewise.
8890 (perhaps_add_new_callers): Likewise.
8891 (decide_about_value): Likewise.
8892 (decide_whether_version_node): Likewise.
8893 (identify_dead_nodes): Likewise.
8894 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8895 (dump_inline_summary): Likewise.
8896 (estimate_node_size_and_time): Likewise.
8897 (inline_analyze_function): Likewise.
8898 * ipa-inline.c (report_inline_failed_reason): Likewise.
8899 (want_early_inline_function_p): Likewise.
8900 (edge_badness): Likewise.
8901 (update_edge_key): Likewise.
8902 (inline_small_functions): Likewise. Add dumping of order to two other
8904 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8905 instead of node uids.
8906 (propagate_pure_const): Likewise.
8907 (propagate_pure_const): Likewise.
8908 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8909 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8910 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8912 * tree-pretty-print.c (dump_function_header): Likewise.
8913 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8914 Print symbol order instead of node uids.
8916 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8918 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8919 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8921 2013-05-15 Richard Biener <rguenther@suse.de>
8923 PR tree-optimization/57275
8924 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8925 return value for fail to do runtime alias checks for gather loads.
8927 2013-05-15 Jan Hubicka <jh@suse.cz>
8931 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8932 weakrefs are not external.
8933 (lto_symtab_merge_decls): Fix thinko when dealing with
8934 non-lto_symtab decls.
8935 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8936 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8937 * varpool.c (dump_varpool_node): Dump more flags.
8939 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8941 * config/i386/i386.c (processor_alias_table): Add instruction
8942 FSGSBASE for AMD bdver3 architecture.
8944 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8946 * tree.c (warn_deprecated_use): Print file:line using locus color.
8947 * diagnostic.c (diagnostic_report_current_module): Print file:line
8948 and file:line:column using locus color.
8950 2013-05-14 Mike Stump <mikestump@comcast.net>
8952 * gdbinit.in: Add __null.
8954 2013-05-14 Mike Stump <mikestump@comcast.net>
8956 * recog.h: Rename struct recog_data to Recog_data.
8957 * recog.c: Likewise.
8958 * reload.c (can_reload_into): Likewise.
8959 * config/picochip/picochip.c: Likewise.
8961 2013-05-14 Mike Stump <mikestump@comcast.net>
8963 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8965 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8967 * resource.h (struct resources): Remove unch_memory member.
8968 (CLEAR_RESOURCE): Don't clear unch_memory.
8969 * resource.c (mark_referenced_resources): Don't set it.
8970 (mark_set_resources): Likewise.
8971 (mark_target_live_regs): Don't clear it.
8972 (init_resource_info): Likewise.
8973 * reorg.c (resource_conflicts_p): Don't compare it.
8974 (redundant_insn): Don't set it.
8976 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8978 * emit-rtl.c (next_label): Remove unused function.
8979 (skip_consecutive_labels, link_cc0_insns): Move to ...
8980 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8981 only place where these functions are used, and make them static.
8983 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8985 * fold-const.c (fold_negate_expr): Handle vectors.
8986 (fold_truth_not_expr): Make it static.
8987 (fold_invert_truthvalue): New static function.
8988 (invert_truthvalue_loc): Handle vectors. Do not call
8989 fold_truth_not_expr directly.
8990 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8991 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8992 (fold_comparison): Handle vector constants.
8993 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8994 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8995 * tree.h (fold_truth_not_expr): Remove declaration.
8997 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8999 * config/aarch64/aarch64-simd.md
9000 (aarch64_vcond_internal<mode>): Rename to...
9001 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9002 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9003 float modes. Clarify all iterator modes.
9004 (vcond<mode><mode>): Use new name for vcond expanders.
9005 (vcond<v_cmp_result><mode>): Likewise.
9006 (vcondu<mode><mode>: Likewise.
9007 * config/aarch64/iterators.md (VDQF_COND): New.
9009 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9012 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9013 variable for the shift amount. Check that we shift by non-negative
9016 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9019 * config/arm/arm.h (EPILOGUE_USES): Only return true
9020 for LR_REGNUM after epilogue_completed.
9022 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9024 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9027 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9030 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9031 and Solaris 11+/x86 with gld.
9032 * configure: Regenerate.
9034 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9036 * expmed.c (expand_shift_1): Canonicalize rotates by
9037 constant bitsize / 2 to bitsize - 1.
9038 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9039 case ROTATERT>: Likewise.
9042 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9044 * config/i386/i386.md (rotateinv): New code attr.
9045 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9046 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9047 roll $31, %eax, etc.
9049 2013-05-14 Richard Biener <rguenther@suse.de>
9052 * tree-eh.c (sink_clobbers): Give up for successors with
9053 multiple predecessors and no virtual uses.
9055 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9057 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9058 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9060 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9063 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9064 the case when both op0 and op1 have VOIDmode.
9066 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9068 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9069 in multiply-accumulate mode.
9071 2013-05-13 Guozhi Wei <carrot@google.com>
9073 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9075 2013-05-13 Kai Tietz <ktietz@redhat.com>
9078 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9079 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9080 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9081 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9082 (ix86_expand_prologue): Likewise.
9083 (ix86_expand_split_stack_prologue): Likewise.
9084 (legitimate_pic_address_disp_p): Likewise.
9085 (legitimize_pic_address): Likewise.
9086 (legitimize_tls_address): Likewise.
9087 (legitimize_pe_coff_symbol): Likewise.
9088 (output_pic_addr_const): Likewise.
9089 (construct_plt_address): Likewise.
9090 (ix86_expand_call): Likewise.
9091 (x86_output_mi_thunk): Likewise.
9092 (x86_function_profiler): Likewise.
9094 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9096 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9097 similar switch cases.
9098 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9099 (aarch64_simd_mov_to_<mode>low): Delete.
9100 (aarch64_simd_mov_to_<mode>high): Delete.
9101 (move_lo_quad_<mode>): Add w<-r alternative.
9102 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9103 (aarch64_simd_mov_from_*): Update type attribute.
9104 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9107 2013-05-13 Jan Hubicka <jh@suse.cz>
9109 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9110 * config/i386/i386.c (ix86_compute_frame_layout,
9111 ix86_expand_epilogue, emit_i387_cw_initialization,
9112 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9113 ix86_local_alignment): Fix use of size/speed predicates.
9115 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9117 PR tree-optimization/45216
9118 PR tree-optimization/57157
9119 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9120 the (-Y) & (B - 1) variant if OP is |.
9121 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9122 return shifted. Use (-op1) & (prec - 1) as other_amount
9123 instead of prec - op1.
9125 2013-05-13 Martin Jambor <mjambor@suse.cz>
9128 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9129 (ipa_constant_data): New type.
9130 (ipa_jump_func): Use ipa_constant_data to hold information about
9131 constant jump functions.
9132 (ipa_get_jf_constant): Adjust to jump function type changes.
9133 (ipa_get_jf_constant_rdesc): New function.
9134 (ipa_param_descriptor): New field controlled_uses.
9135 (ipa_get_controlled_uses): New function.
9136 (ipa_set_controlled_uses): Likewise.
9137 * ipa-ref.h (ipa_find_reference): Declare.
9138 * ipa-prop.c (ipa_cst_ref_desc): New type.
9139 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9141 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9142 New parameter cs. Adjust all callers.
9143 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9144 (remove_described_reference): New function.
9145 (jfunc_rdesc_usable): Likewise.
9146 (try_make_edge_direct_simple_call): Decrement controlled use count,
9147 attempt to remove reference if it hits zero.
9148 (combine_controlled_uses_counters): New function.
9149 (propagate_controlled_uses): Likewise.
9150 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9151 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9152 (ipa_print_node_params): Print described use counter.
9153 (ipa_write_jump_function): Adjust to jump function type changes.
9154 (ipa_read_jump_function): New parameter CS, pass it to
9155 ipa_set_jf_constant. Adjust caller.
9156 (ipa_write_node_info): Stream controlled use count
9157 (ipa_read_node_info): Likewise.
9158 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9160 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9161 count. Remove cloning-added reference if it reaches zero.
9162 * ipa-ref.c (ipa_find_reference): New function.
9164 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9166 * config/i386/i386.c (processor_target_table): Modified default
9167 alignment values for AMD BD and BT architectures.
9169 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9171 * tree-vect-generic.c (uniform_vector_p): Move ...
9172 * tree.c (uniform_vector_p): ... here.
9173 * tree.h (uniform_vector_p): Declare it.
9174 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9177 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9179 PR tree-optimization/57230
9180 * tree-ssa-strlen.c (handle_char_store): Record length for
9181 array store from STRING_CST.
9183 PR tree-optimization/57230
9184 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9187 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9189 * config/epiphany/epiphany.c (epiphany_init): Check size of
9190 NUM_MODES_FOR_MODE_SWITCHING.
9191 (epiphany_expand_prologue):
9192 Remove CONFIG_REGNUM initial value handling code.
9193 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9194 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9195 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9196 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9197 Don't return 1 for FP_MODE_NONE.
9198 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9199 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9200 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9201 * config/epiphany/epiphany.md (save_config): New pattern.
9203 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9205 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9207 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9209 * config/i386/i386.md (memory): Handle sseishft1.
9210 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9211 (*vec_extractv2di_1): Ditto.
9213 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9215 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9218 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9220 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9221 Add mthumb/march=armv7-a multilib.
9222 Add mthumb/march=armv7-r multilib.
9223 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
9225 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
9227 * config/v850/t-rtems: Add more multilibs.
9229 2013-05-10 Richard Biener <rguenther@suse.de>
9231 PR tree-optimization/57214
9232 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
9233 not propagate from SSA names that occur in abnormal PHI nodes.
9235 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
9237 * stor-layout.c (element_precision): New function.
9238 * machmode.h (element_precision): Declare it.
9239 * tree.c (build_minus_one_cst): New function.
9240 (element_precision): Likewise.
9241 * tree.h (build_minus_one_cst): Declare new function.
9242 (element_precision): Likewise.
9243 * fold-const.c (operand_equal_p): Use element_precision.
9244 (fold_binary_loc): Handle vector types.
9245 * convert.c (convert_to_integer): Use element_precision.
9246 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
9249 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
9251 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
9252 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
9253 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
9254 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9255 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
9256 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
9257 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
9258 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9259 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
9260 (Uuw8): New constraints.
9261 (Usb4): Move into alphabetical order.
9262 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
9263 (sd8_operand, ub8_operand, uw8_operand): New predicates.
9264 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
9265 previously unnamed patterns.
9266 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
9267 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
9268 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
9269 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
9270 of set_attr_alternative/if_then_else. Use extended_mips16 instead
9271 of specific lengths.
9273 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9275 * config/i386/i386.md (rotateinv): New code attr.
9276 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9277 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9278 roll $31, %eax, etc.
9280 PR tree-optimization/45216
9281 PR tree-optimization/57157
9282 * tree-ssa-forwprop.c (simplify_rotate): New function.
9283 (ssa_forward_propagate_and_combine): Call it.
9285 2013-05-10 Richard Biener <rguenther@suse.de>
9287 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
9288 disable peeling when we version for aliasing.
9289 (vector_alignment_reachable_p): Honor explicit user alignment.
9290 (vect_supportable_dr_alignment): Likewise.
9291 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
9292 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
9293 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
9294 then peeling to arrange for the cost-model check to come first.
9296 2013-05-10 Alan Modra <amodra@gmail.com>
9298 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
9299 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
9300 * configure: Regenerate.
9302 2013-05-10 Alan Modra <amodra@gmail.com>
9305 * varasm.c (default_elf_select_section): Move !DECL_P check..
9306 (get_named_section): ..to here before calling get_section_name.
9308 (default_section_type_flags): Add DECL_P check.
9309 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
9310 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
9312 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
9314 * config/epiphany/epiphany.c (epiphany_expand_prologue):
9315 When using gen_stack_adjust_str with a register offset, add a
9316 REG_FRAME_RELATED_EXPR note.
9318 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
9320 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
9321 (*vec_extractv4si_zext_mem): Ditto.
9322 (*vec_extractv2di): Add 0->x and x->x alternatives.
9323 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
9324 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
9326 2013-05-09 Jason Merrill <jason@redhat.com>
9328 N3639 C++1y VLA support
9329 * gimplify.c (gimplify_vla_decl): Don't touch an existing
9332 * tree.c (build_constructor_va): New.
9333 * tree.h: Declare it.
9335 2013-05-09 Martin Jambor <mjambor@suse.cz>
9338 * gimple-fold.c (canonicalize_constructor_val): Call
9339 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
9341 2013-05-09 Jan Hubicka <jh@suse.cz>
9342 Richard Biener <rguenther@suse.de>
9345 * symtab.c (symtab_make_decl_local): Do not add private names.
9347 2013-05-09 Jan Hubicka <jh@suse.cz>
9350 * symtab.c (insert_to_assembler_name_hash): Handle clones.
9351 (unlink_from_assembler_name_hash): Likewise.
9352 (symtab_prevail_in_asm_name_hash, symtab_register_node,
9353 symtab_unregister_node, symtab_initialize_asm_name_hash,
9354 change_decl_assembler_name): Update.
9356 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
9358 * config/aarch64/aarch64.md: New movtf split.
9359 (*movtf_aarch64): Update.
9360 (aarch64_movdi_tilow): Handle TF modes and rename to
9361 aarch64_movdi_<mode>low.
9362 (aarch64_movdi_tihigh): Handle TF modes and rename to
9363 aarch64_movdi_<mode>high
9364 (aarch64_movtihigh_di): Handle TF modes and rename to
9365 aarch64_mov<mode>high_di
9366 (aarch64_movtilow_di): Handle TF modes and rename to
9367 aarch64_mov<mode>low_di
9368 (aarch64_movtilow_tilow): Remove spurious whitespace.
9369 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
9371 (aarch64_print_operand): Update.
9373 2013-05-09 Alan Modra <amodra@gmail.com>
9375 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
9377 * configure: Regenerate.
9379 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9381 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
9382 splitter preparation statements.
9383 * config/i386/sse.md (*vec_extract* splitters): Ditto.
9384 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
9387 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9389 * gimple-ssa-strength-reduction.c (count_candidates): Change
9390 return value to int.
9391 (analyze_candidates_and_replace): Change type of length to int.
9393 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9395 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
9396 (*vec_extract<mode>): Use VI12_128 mode iterator.
9397 (*vec_extract<mode>_mem): Ditto.
9398 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
9401 2013-05-08 Diego Novillo <dnovillo@google.com>
9406 2012-08-17 Diego Novillo <dnovillo@google.com>
9409 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
9410 * config.in: Regenerate.
9411 * configure: Regenerate.
9412 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
9414 2013-05-08 Jan Hubicka <jh@suse.cz>
9417 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
9418 * cgraph.h (symtab_node_base): Add unique_name.
9419 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9420 input_overwrite_node, input_varpool_node): Stream unique_name.
9421 * cgraphclones.c (cgraph_create_virtual_clone,
9422 cgraph_function_versioning): Set unique_name.
9423 * ipa.c (function_and_variable_visibility): Set unique_name.
9425 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9427 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
9428 (alloc_cand_and_find_basis): Restrict conditional candidate
9429 processing to CAND_MULTs.
9431 2013-05-08 Jan Hubicka <jh@suse.cz>
9434 lto-symtab.c (lto_symtab_symbol_p): New function.
9435 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
9436 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
9437 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
9438 Skip static symbols.
9440 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
9442 PR tree-optimization/57200
9443 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
9444 Only call inform if the preceding warning_at returns true.
9446 2013-05-07 Han Shen <shenhan@google.com>
9448 * cfgexpand.c (record_or_union_type_has_array_p): New function.
9449 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
9450 * common.opt (fstack-protector-strong): New option.
9451 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
9452 * doc/invoke.texi (Optimization Options): Document
9453 "-fstack-protector-strong".
9454 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
9456 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9458 * config/mips/mips.c (mips_machine_reorg2): Return 0.
9460 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
9462 * ira.c (update_equiv_regs): Add insn having equiv memory even if
9463 it is not lhs of the insn.
9464 (setup_reg_equiv): Remove insn having equiv memory which it is not
9466 * lra-constraints.c (process_address): Try to improve generation
9467 code for address base + disp.
9468 (lra_constraints): Make correct the code for checking insn setting
9469 up backward equivalence. Remove insn only if it is in the init
9471 * lra-eliminations.c (update_reg_eliminate): Change return value.
9472 (lra_eliminate): Use the result.
9474 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
9476 * config/i386/sse.md (ssescalarnummask): New mode attribute.
9477 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
9478 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
9479 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
9480 register target operands.
9481 (*vec_extractv8hi_sse2): New pattern.
9482 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
9483 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
9484 (*vec_extract<mode>_mem): New insn and split pattern.
9486 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
9488 * config/arm/arm.c (arm_asan_shadow_offset): New function.
9489 (TARGET_ASAN_SHADOW_OFFSET): Define.
9490 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9491 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9493 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9495 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9496 (incr_vec_index): Return -1 if increment not found.
9497 (create_add_on_incoming_edge): Assert if increment not found.
9498 (record_increment): Limit number of increments recorded.
9499 (all_phi_incrs_profitable): Return false if an increment not found.
9500 (replace_profitable_candidates): Don't process increments that were
9502 (analyze_candidates_and_replace): Limit size of incr_vec.
9504 2013-05-07 Richard Biener <rguenther@suse.de>
9506 * calls.c (special_function_p): setjmp-like functions are leaf.
9507 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9508 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9510 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9512 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9513 (aarch64_simd_mov<mode>): New expander.
9514 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9515 (aarch64_simd_mov_to_<mode>high): Likewise.
9516 (aarch64_simd_mov_from_<mode>low): Likewise.
9517 (aarch64_simd_mov_from_<mode>high): Likewise.
9518 (aarch64_dup_lane<mode>): Update.
9519 (aarch64_dup_lanedi): New instruction pattern.
9520 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9521 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9523 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9525 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9526 (replace_mult_candidate): Remove unnecessary argument; remove
9527 unnecessary parameter from call to introduce_cast_before_cand.
9528 (replace_unconditional_candidate): Remove unnecessary parameter
9529 from call to replace_mult_candidate.
9530 (replace_conditional_candidate): Likewise.
9531 (insert_initializers): Use make_temp_ssa_name.
9532 (introduce_cast_before_cand): Remove unnecessary argument; use
9534 (replace_one_candidate): Remove unnecessary argument; remove
9535 unnecessary parameter from calls to introduce_cast_before_cand.
9536 (replace_profitable_candidates): Remove unnecessary parameters
9537 from calls to replace_one_candidate.
9539 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9541 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9542 phi def as possibly hiding a basis for a CAND_ADD whose operands
9543 have been commuted in the analysis.
9544 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9546 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9548 * config/aarch64/aarch64.md
9549 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9550 shift value between 0-4.
9552 2013-05-07 Richard Biener <rguenther@suse.de>
9554 * double-int.h (rshift): New overload.
9555 * double-int.c (rshift): New function.
9556 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9557 (create_reference_ops_from_ref): Remove.
9558 (vn_reference_insert): Use shared ops for constructing the
9559 reference and copy it.
9561 2013-05-07 Richard Biener <rguenther@suse.de>
9564 * tree-eh.c (sink_clobbers): Properly propagate
9565 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9567 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9569 PR tree-optimization/57149
9570 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9571 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9572 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9573 uninit_undefined_value_p instead of ssa_undefined_value_p.
9576 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9577 for modifier == EXPAND_INITIALIZER.
9579 2013-05-07 Anton Blanchard <anton@samba.org>
9581 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9582 for powerpc64 little endian.
9583 * configure: Regenerate.
9585 2013-05-06 Graham Stott <grahams@btinternet.com>
9587 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9588 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9589 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9590 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9592 2013-05-06 Graham Stott <grahams@btinternet.com>
9594 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9595 codes which allow non-lvalues.
9597 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9599 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9600 components are all 1s.
9601 (integer_minus_onep): New function.
9602 * tree.h (integer_minus_onep): Declare it.
9603 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9604 integer_minus_onep instead of integer_all_onesp.
9606 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9609 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9610 variations of these patterns.
9612 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9614 * config/i386/i386.md (isa): Add x64_sse4 member.
9615 (enabled): Handle x64_sse4.
9616 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9617 instruction for 64bit SSE4_1 targets. Update insn attributes.
9618 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9619 instruction for SSE4_1 targets. Update insn attributes.
9620 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9621 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9622 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9624 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9625 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9626 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9628 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9631 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9633 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9635 * final.c (do_assembler_dialects): Don't handle curly braces and
9636 vertical bar escaped by % as dialect delimiters.
9637 (output_asm_insn): Print curly braces and vertical bar if escaped
9638 by % and ASSEMBLER_DIALECT defined.
9639 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9640 * doc/tm.texi: Regenerated.
9642 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9644 * config/mips/mips.c: Include tree-pass.h.
9645 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9646 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9647 (pass_mips_machine_reorg2): New machine specific pass.
9648 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9649 (mips_option_override): Register the new pass.
9650 * rtl.h (cleanup_barriers): Remove prototype.
9651 (dbr_schedule): Likewise.
9652 * jump.c (cleanup_barriers): Make static.
9653 * reorg.c (dbr_schedule): Likewise.
9655 2013-05-06 Richard Biener <rguenther@suse.de>
9657 PR tree-optimization/57185
9658 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9661 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9664 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9665 between "first_arg" and "insn", not "last" and "insn".
9667 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9669 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9670 (find_candidates_in_block): Re-enable slsr_process_phi.
9671 (create_phi_basis): Fix double counting of candidate adjustment.
9673 2013-05-06 Richard Biener <rguenther@suse.de>
9676 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9677 the edge is also fallthru, preserve it and just clear the
9679 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9680 also complex, preserve that and just clear the fallthru flag.
9681 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9682 update virtual operands.
9684 2013-05-06 Alan Modra <amodra@gmail.com>
9686 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9687 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9688 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9689 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9690 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9691 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9693 2013-05-06 Alan Modra <amodra@gmail.com>
9695 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9696 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9697 (DEFAULT_ASM_ENDIAN): Define.
9698 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9699 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9700 Update -K PIC clause from sysv4.h.
9701 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9702 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9704 2013-05-06 Alan Modra <amodra@gmail.com>
9706 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9707 twice for little-endian.
9708 (ashrdi3_no_power, ashrdi3): Support little-endian.
9710 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9713 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9714 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9716 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9717 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9718 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9721 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9722 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9724 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9725 * config/pa/pa.opt: Make mbig-switch a no-op.
9726 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9727 (CASE_VECTOR_MODE): Always return SImode.
9728 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9729 for the !TARGET_BIG_SWITCH case.
9730 * config/pa/pa-linux.h: Likewise.
9731 * config/pa/pa-openbsd.h: Likewise.
9732 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9733 * config/pa/pa.md (short_jump): Remove define_insn.
9734 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9735 (casesi0): Remove define_insn.
9736 (type): Remove btable_branch.
9737 (pa_combine_type): Likewise.
9738 (in_nullified_branch_delay): Likewise.
9739 (in_call_delay): Likewise.
9740 (define_delay): Likewise.
9741 (define_insn_reservation "Z3"): Likewise.
9742 (define_insn_reservation "Z4"): Likewise.
9743 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9744 (pa_adjust_insn_length): Remove adjustment for btable branches.
9745 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9748 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9750 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9751 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9752 Add m->r,x alternatives.
9753 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9754 splitters using SWI48x mode iterator.
9755 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9756 TARGET_64BIT. Add m->x alternative.
9757 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9758 Add o->x alternative. Enable for TARGET_SSE.
9759 (sse_storeq): Remove expander.
9760 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9761 with memory input operand.
9762 (*vec_extractv2di_1 splitter): New.
9763 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9764 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9766 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9768 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9769 (INT_LOWPART): Delete.
9770 (extract_MB): Adjust.
9771 (extract_ME): Adjust.
9772 (print_operand): Adjust.
9774 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9776 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9777 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9778 (reg_or_logical_cint_operand, easy_fp_constant,
9779 logical_const_operand): Delete "CONST_DOUBLE" case.
9780 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9781 "HOST_BITS_PER_WIDE_INT == 64" test.
9782 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9783 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9784 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9785 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9787 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9788 CONST_DOUBLE DImode/VOIDmode case.
9789 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9790 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9791 CONST_DOUBLE VOIDmode case.
9792 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9793 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9794 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9795 Delete CONST_DOUBLE case.
9796 (splitters for mov FMOVE64 const_double): Delete
9797 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9798 "HOST_BITS_PER_WIDE_INT >= 64" test.
9799 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9801 (mov DI const_double): Delete.
9803 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9805 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9806 on op shows all bits zero in mode of a lowpart subreg, return zero.
9808 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9811 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9812 to save TFmode registers and DImode to save TImode registers for
9813 caller save operations.
9814 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9815 mark being partially clobbered since they only use the first
9818 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9819 and TDmode only use the upper 64-bits of each VSX register.
9821 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9823 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9824 (find_candidates_in_block): Disable slsr_process_phi.
9826 2013-05-03 Guozhi Wei <carrot@google.com>
9828 * coverage.c (coverage_obj_init): Move the construction of gcov
9830 (build_init_ctor): ... here.
9832 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9834 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9835 (slsr_cand_d): Redefine def_phi.
9836 (stride_status, phi_adjust_status, count_phis_status): New enums.
9837 (find_phi_def): New.
9838 (find_basis_for_base_expr): New.
9839 (find_basis_for_candidate): Handle hidden bases.
9840 (alloc_cand_and_find_basis): Handle phi candidates.
9841 (slsr_process_phi): New.
9842 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9843 (create_mul_imm_cand): Likewise.
9844 (create_add_ssa_cand): Exclude phi base candidates.
9845 (create_add_imm_cand): Likewise.
9846 (slsr_process_cast): Likewise.
9847 (slsr_process_copy): Likewise.
9848 (find_candidates_in_block): Handle phi candidates.
9849 (dump_candidate): Likewise.
9850 (unconditional_cands): Delete.
9851 (unconditional_cands_with_known_stride_p): Delete.
9852 (phi_dependent_cand_p): New.
9853 (cand_increment): Handle phi-dependent candidates.
9854 (replace_dependent): Delete.
9855 (replace_mult_candidate): New.
9856 (replace_unconditional_candidate): New.
9857 (incr_vec_index): Move to avoid forward reference.
9858 (create_add_on_incoming_edge): New.
9859 (create_phi_basis): New.
9860 (replace_dependents): Delete.
9861 (replace_conditional_candidate): New.
9862 (phi_add_costs): New.
9863 (replace_uncond_cands_and_profitable_phis): New.
9864 (record_increment): Handle phi adjustments.
9865 (record_phi_increments): New.
9866 (record_increments): Handle phi adjustments.
9867 (phi_incr_cost): New.
9868 (lowest_cost_path): Handle phis.
9869 (total_savings): Likewise.
9870 (analyze_increments): Likewise.
9871 (ncd_with_phi): New.
9872 (ncd_of_cand_and_phis): New.
9873 (nearest_common_dominator_for_cands): Handle phi increments.
9874 (all_phi_incrs_profitable): New.
9875 (replace_profitable_candidates): Handle phi-dependent candidates.
9876 (analyze_candidates_and_replace): Likewise.
9878 2013-05-03 Teresa Johnson <tejohnson@google.com>
9881 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9882 do not exceed REG_BR_PROB_BASE.
9884 2013-05-03 Jeff Law <law@redhat.com>
9886 PR tree-optimization/57144
9887 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9888 operand of the condition will bit into the new type when eliminating
9889 a cast feeding a condition.
9891 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9893 PR rtl-optimization/57130
9894 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9895 of COMPARE as in_code to the recursive call if needed.
9897 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9899 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9900 (enabled): Handle new members.
9901 * config/i386/sse.md (*vec_concatv2si): Merge from
9902 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9903 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9905 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9907 PR tree-optimization/57027
9908 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9909 for fnms opportunity, check we got the prerequisite kind
9910 of tree / gimple before using accessor functions.
9912 2013-05-03 Richard Biener <rguenther@suse.de>
9914 * double-int.h (lshift): New overload without precision
9916 (operator *=, operator +=, operator -=): Move ...
9917 * double-int.c (operator *=, operator +=, operator -=): ... here
9918 and implement more efficiently.
9919 (mul_double_with_sign): Remove.
9920 (lshift_double): Adjust to take unsinged shift argument, push
9921 dispatching code to callers.
9922 (mul_double_wide_with_sign): Add early out for callers that
9923 are not interested in high parts or overflow.
9924 (lshift): New function.
9925 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9927 (lrotate, rrotate): Use logical shifts.
9928 * expr.c (get_inner_reference): Use lshift.
9929 * fixed-value.c (do_fixed_divide): Likewise.
9930 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9931 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9932 (indirect_refs_may_alias_p): Likewise.
9933 (stmt_kills_ref_p_1): Likewise.
9935 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9937 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9939 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9941 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9942 scalar form of FABD instruction.
9944 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9946 * lra-constraints.c (process_alt_operands): Add checking alt
9947 number to choose the best alternative.
9949 2013-05-02 Richard Biener <rguenther@suse.de>
9951 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9952 bitmap and its handling.
9953 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9955 2013-05-02 Richard Biener <rguenther@suse.de>
9958 * tree-inline.c (copy_loops): Properly handle removed loops.
9959 (copy_cfg_body): Mark destination loops for fixup if source
9962 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9965 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9966 generating simple_return for naked functions.
9968 2013-05-02 Martin Jambor <mjambor@suse.cz>
9971 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9972 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9974 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9975 ipa_agg_replacement_value structures.
9976 (known_aggs_to_agg_replacement_list): Likewise.
9977 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9978 (read_agg_replacement_chain): Likewise.
9979 (ipcp_transform_function): Also check that by_ref flags match.
9981 2013-05-02 Richard Biener <rguenther@suse.de>
9983 * graphds.h (struct graph): Add obstack member.
9984 * graphds.c (new_graph): Initialize obstack and allocate
9986 (add_edge): Allocate edge from the obstack.
9987 (free_graph): Free the obstack instead of all edges and vertices.
9989 2013-05-02 Teresa Johnson <tejohnson@google.com>
9991 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9993 * cfg.c (update_bb_profile_for_threading): Ditto.
9994 * tree-inline.c (copy_bb): Ditto.
9995 (copy_edges_for_bb): Ditto.
9996 (initialize_cfun): Ditto.
9997 (copy_cfg_body): Ditto.
9998 (expand_call_inline): Ditto.
9999 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10000 (estimate_node_size_and_time): Ditto.
10001 (inline_merge_summary): Ditto.
10002 * cgraphclones.c (cgraph_clone_edge): Ditto.
10003 (cgraph_clone_node): Ditto.
10004 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10005 (compute_trg_info): Ditto.
10007 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10009 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10010 S reg when fp attribute set.
10011 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10013 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10015 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10017 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10018 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10019 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10021 2013-05-02 Richard Biener <rguenther@suse.de>
10023 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10024 (struct instantiate_cache_entry): New type.
10025 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10026 (struct instantiate_cache_type): New type.
10027 (set_instantiated_value, get_instantiated_value): Remove.
10028 (get_instantiated_value_entry): New function.
10029 (instantiate_scev_name): Use the new cache and adjust.
10030 (instantiate_scev_poly): Adjust.
10031 (instantiate_scev_binary): Likewise.
10032 (instantiate_array_ref): Likewise.
10033 (instantiate_scev_convert): Likewise.
10034 (instantiate_scev_not): Likewise.
10035 (instantiate_scev_3): Likewise.
10036 (instantiate_scev_2): Likewise.
10037 (instantiate_scev_r): Likewise.
10038 (instantiate_scev): Likewise.
10039 (resolve_mixers): Likewise.
10041 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10044 * lra-constraints.c (best_small_class_operands_num): Remove.
10045 (process_alt_operands): Remove small_class_operands_num. Take
10046 small classes operands into losers and only if the operand is not
10047 matched. Modify debugging output.
10048 (curr_insn_transform): Remove best_small_class_operands_num.
10051 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10053 * config/aarch64/aarch64-builtins.c
10054 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10055 * config/aarch64/aarch64-simd-builtins.def
10056 (reduc_splus_): Add new modes.
10057 (reduc_uplus_): New.
10058 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10059 (reduc_uplus_v4sf): Likewise.
10060 (reduc_splus_v4sf): Likewise.
10061 (aarch64_addv<mode>): Likewise.
10062 (reduc_uplus_<mode>): Likewise.
10063 (reduc_splus_<mode>): Likewise.
10064 (aarch64_addvv2di): Likewise.
10065 (reduc_uplus_v2di): Likewise.
10066 (reduc_splus_v2di): Likewise.
10067 (aarch64_addvv2si): Likewise.
10068 (reduc_uplus_v2si): Likewise.
10069 (reduc_splus_v2si): Likewise.
10070 (reduc_<sur>plus_<mode>): New.
10071 (reduc_<sur>plus_v2di): Likewise.
10072 (reduc_<sur>plus_v2si): Likewise.
10073 (reduc_<sur>plus_v4sf): Likewise.
10074 (aarch64_addpv4sf): Likewise.
10075 * config/aarch64/arm_neon.h
10076 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10077 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10078 add UNSPEC_SADDV, UNSPEC_UADDV.
10080 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10082 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10084 * config/aarch64/arm_neon.h
10085 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10087 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10089 * config/aarch64/aarch64-builtins
10090 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10092 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10094 * config/aarch64/aarch64-simd-builtins.def
10095 (reduc_smax_): New.
10096 (reduc_smin_): Likewise.
10097 (reduc_umax_): Likewise.
10098 (reduc_umin_): Likewise.
10099 (reduc_smax_nan_): Likewise.
10100 (reduc_smin_nan_): Likewise.
10103 (smax): Update for V2SF, V4SF and V2DF modes.
10106 (smin_nan): Likewise.
10107 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10108 (<su><maxmin><mode>3): ...This, refactor.
10109 (s<maxmin><mode>3): New.
10110 (<maxmin_uns><mode>3): Likewise.
10111 (reduc_<maxmin_uns>_<mode>): Refactor.
10112 (reduc_<maxmin_uns>_v4sf): Likewise.
10113 (reduc_<maxmin_uns>_v2si): Likewise.
10114 (aarch64_<fmaxmin><mode>: Remove.
10115 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10117 (vmin<q>_f<32,64>): Likewise.
10118 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10120 (su): Add mappings for smax, smin, umax, umin.
10122 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10123 (FMAXMIN): Rename as...
10124 (FMAXMIN_UNS): ...This.
10126 (fmaxminv): Likewise.
10127 (fmaxmin): Likewise.
10129 (maxmin_uns_op): Likewise.
10131 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10133 * config/aarch64/arm_neon.h
10134 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10135 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10136 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10138 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10140 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10141 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10143 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10145 * config/aarch64/aarch64-simd.md
10146 (vcond<mode>_internal): Handle special cases for constant masks.
10147 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10148 (vcondu<mode><mode>): Likewise.
10149 (vcond<v_cmp_result><mode>): New.
10151 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10153 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10154 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10155 * config/aarch64/aarch64-simd-builtins.def
10156 (cmeq): Update to BUILTIN_VALLDI.
10161 * config/aarch64/arm_neon.h
10162 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10163 to builtins or C as appropriate.
10165 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10167 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10169 (cmhi): Rename to...
10171 * config/aarch64/aarch64-simd.md
10172 (simd_mode): Add SF.
10173 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10174 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10175 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10176 (cstore<mode>_neg): ...This.
10177 * config/aarch64/iterators.md
10179 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10180 (COMPARISONS): New.
10181 (UCOMPARISONS): Likewise.
10182 (optab): Add missing comparisons.
10188 (VCMP_S): Likewise.
10189 (VCMP_U): Likewise.
10190 (V_cmp_result): Add DF, SF modes.
10191 (v_cmp_result): Likewise.
10193 (vmtype): Likewise.
10194 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10196 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10198 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10199 define_insn to define_insn_and_split.
10200 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10201 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10202 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10203 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10204 (thumb2_negscc): Likewise.
10206 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10208 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10210 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10212 * config/arm/thumb2.md: Remove trailing whitespaces.
10214 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10216 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10217 Use gen_int_mode rather than GEN_INT.
10219 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10221 * value-prof.c (stream_in_histogram_value): Remove the strayed
10224 2013-04-30 Richard Biener <rguenther@suse.de>
10226 PR middle-end/57122
10227 * cfghooks.c (split_edge): Properly check for the loop latch edge.
10229 2013-04-30 Richard Biener <rguenther@suse.de>
10231 PR middle-end/57107
10232 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
10234 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
10236 PR rtl-optimization/56957
10237 PR rtl-optimization/57105
10238 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
10239 variable. Use just INSN_UID for determining whether an insn
10240 should be only disconnected from the insn stream.
10241 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
10243 2013-04-30 Jakub Jelinek <jakub@redhat.com>
10245 PR tree-optimization/57104
10246 * tsan.c (instrument_expr): Don't instrument accesses to
10247 DECL_HARD_REGISTER VAR_DECLs.
10249 2013-04-30 Richard Biener <rguenther@suse.de>
10251 * function.h (loops_for_fn): New inline function.
10252 (set_loops_for_fn): Likewise.
10253 * cfgloop.h (place_new_loop): Add struct function parameter.
10254 (get_loop): Likewise.
10255 (get_loops): Likewise.
10256 (number_of_loops): Likewise.
10257 (fel_next): Adjust.
10258 (fel_init): Likewise.
10259 * cfg.c (get_loop_copy): Adjust.
10260 * cfgloop.c (flow_loops_dump): Likewise.
10261 (record_loop_exits): Likewise.
10262 (verify_loop_structure): Likewise.
10263 * cfgloopanal.c (mark_irreducible_loops): Likewise.
10264 (estimate_reg_pressure_cost): Likewise.
10265 (mark_loop_exit_edges): Likewise.
10266 * cfgloopmanip.c (place_new_loop): Likewise.
10267 (add_loop): Likewise.
10268 (duplicate_loop): Likewise.
10269 * graph.c (draw_cfg_nodes): Likewise.
10270 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
10271 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
10272 (extract_affine_chrec): Likewise.
10273 (build_scop_iteration_domain): Likewise.
10274 * graphite.c (graphite_initialize): Likewise.
10275 * ira-build.c (create_loop_tree_nodes): Likewise.
10276 (more_one_region_p): Likewise.
10277 (rebuild_regno_allocno_maps): Likewise.
10278 (mark_loops_for_removal): Likewise.
10279 (mark_all_loops_for_removal): Likewise.
10280 (remove_unnecessary_regions): Likewise.
10281 (ira_build): Likewise.
10282 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
10283 * loop-init.c (fix_loop_structure): Likewise.
10284 (gate_rtl_move_loop_invariants): Likewise.
10285 (gate_rtl_unswitch): Likewise.
10286 (gate_rtl_unroll_and_peel_loops): Likewise.
10287 (rtl_doloop): Likewise.
10288 * lto-streamer-in.c (input_cfg): Likewise.
10289 * lto-streamer-out.c (output_cfg): Likewise.
10290 * modulo-sched.c (sms_schedule): Likewise.
10291 * predict.c (tree_estimate_probability): Likewise.
10292 (tree_estimate_probability_driver): Likewise.
10293 (estimate_loops): Likewise.
10294 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
10295 (move_sese_region_to_fn): Likewise.
10296 (debug_loop_num): Likewise.
10297 * tree-chrec.c (chrec_evaluate): Likewise.
10298 (hide_evolution_in_other_loops_than_loop): Likewise.
10299 (chrec_component_in_loop_num): Likewise.
10300 (reset_evolution_in_loop): Likewise.
10301 (evolution_function_is_invariant_rec_p): Likewise.
10302 * tree-if-conv.c (main_tree_if_conversion): Likewise.
10303 * tree-inline.c (copy_loops): Likewise.
10304 (copy_cfg_body): Likewise.
10305 (tree_function_versioning): Likewise.
10306 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
10307 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
10309 (add_to_evolution_1): Likewise.
10310 (scev_const_prop): Likewise.
10311 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
10312 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
10313 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
10314 (tree_ssa_lim_initialize): Likewise.
10315 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
10316 (verify_loop_closed_ssa): Likewise.
10317 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
10318 (tree_ssa_loop_im): Likewise.
10319 (tree_ssa_loop_unswitch): Likewise.
10320 (tree_vectorize): Likewise.
10321 (check_data_deps): Likewise.
10322 (tree_ssa_loop_ivcanon): Likewise.
10323 (tree_ssa_loop_bounds): Likewise.
10324 (tree_complete_unroll): Likewise.
10325 (tree_complete_unroll_inner): Likewise.
10326 (tree_parallelize_loops): Likewise.
10327 (tree_ssa_loop_prefetch): Likewise.
10328 (tree_ssa_loop_ivopts): Likewise.
10329 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
10330 * tree-vectorizer.c (vectorize_loops): Likewise.
10332 2013-04-29 Mike Frysinger <vapier@gentoo.org>
10334 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
10335 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
10336 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
10337 with EABI_LINK_SPEC.
10339 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10342 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
10345 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
10348 * lra-constraints.c (process_alt_operands): Discourage a bit more
10349 using memory for pseudos. Print cost dump for alternatives.
10350 Modify cost values for conflicts with early clobbers.
10351 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
10353 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10356 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
10358 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10360 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
10361 from/to S register.
10362 (movdi_aarch64): Support LDR/STR from/to D register.
10364 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10366 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
10367 or higher by default.
10369 2013-04-29 Richard Biener <rguenther@suse.de>
10371 PR middle-end/57075
10372 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
10373 even if not adding abnormal edges for calls that can make
10376 2013-04-29 Richard Biener <rguenther@suse.de>
10378 PR middle-end/57103
10379 * tree-cfg.c (move_stmt_op): Fix condition under which to update
10381 (move_stmt_r): Remove redundant checking.
10383 2013-04-29 Teresa Johnson <tejohnson@google.com>
10386 * basic-block.h (apply_scale): New function.
10387 (apply_probability): Use apply_scale.
10388 * gimple-streamer-in.c (input_bb): Ditto.
10389 * lto-streamer-in.c (input_cfg): Ditto.
10390 * lto-cgraph.c (merge_profile_summaries): Ditto.
10391 * tree-optimize.c (execute_fixup_cfg): Ditto.
10392 * tree-inline.c (copy_bb): Update comment to use apply_scale.
10393 (copy_edges_for_bb): Ditto.
10394 (copy_cfg_body): Ditto.
10396 2013-04-29 Tom de Vries <tom@codesourcery.com>
10398 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
10399 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
10400 (tail_merge_optimize): Handle current_loops == NULL.
10402 2013-04-26 Jeff Law <law@redhat.com>
10404 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
10405 (simplify_cond_using_ranges): Generalize code to simplify
10406 COND_EXPRs where one argument is a constant and the other
10407 is an SSA_NAME created by an integral type conversion.
10409 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10411 * config/arm/arm.md (store_minmaxsi): Use only when
10412 optimize_insn_for_size_p.
10414 2013-04-29 Christian Bruel <christian.bruel@st.com>
10417 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
10419 2013-04-29 Richard Biener <rguenther@suse.de>
10421 PR middle-end/57089
10422 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
10423 loop tree make sure to schedule a fixup for the child as well.
10424 (expand_omp_for_generic): Properly add loops.
10425 (expand_omp_for_static_nochunk): Likewise.
10426 (expand_omp_for_static_chunk): Likewise.
10427 (expand_omp_for): For the degenerate case fixup loops.
10428 (expand_omp_sections): Fix default bb placement in loops.
10429 (expand_omp_atomic_pipeline): Properly add loops.
10431 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10433 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
10435 2013-04-29 Tom de Vries <tom@codesourcery.com>
10437 * tree-ssa-tail-merge.c: Update header comment.
10439 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10441 * config/aarch64/arm_neon.h
10442 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
10443 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
10444 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
10445 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
10446 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
10447 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
10448 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
10449 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
10451 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10453 * config/aarch64/aarch64-simd.md
10454 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
10455 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
10456 fix_trunc, fixuns_trunc.
10457 (ftrunc<VDQF:mode>2): New.
10458 * config/aarch64/iterators.md (optab): Add fix, fixuns.
10459 (fix_trunc_optab): New.
10461 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10463 * config/aarch64/aarch64-builtins.c
10464 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
10465 iceilf, lround, iroundf.
10467 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10470 * config/i386/i386.h (enum ix86_tune_indices)
10471 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
10472 New, split from X86_TUNE_INTER_UNIT_MOVES.
10473 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
10474 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
10475 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
10476 (TARGET_INTER_UNIT_MOVES): Remove.
10477 * config/i386/i386.c (initial_ix86_tune_features): Update.
10478 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
10479 (ix86_expand_convert_uns_didf_sse): Use
10480 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10481 (ix86_expand_vector_init_one_nonzero): Ditto.
10482 (ix86_expand_vector_init_interleave): Ditto.
10483 (inline_secondary_memory_needed): Return true for moves from SSE class
10484 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
10485 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
10486 * config/i386/constraints.md (Yi, Ym): Depend on
10487 TARGET_INTER_UNIT_MOVES_TO_VEC.
10488 (Yj, Yn): New constraints.
10489 * config/i386/i386.md (*movdi_internal): Change constraints of
10490 operand 1 from Yi to Yj and from Ym to Yn.
10491 (*movsi_internal): Ditto.
10492 (*movdf_internal): Ditto.
10493 (*movsf_internal): Ditto.
10494 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10495 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10496 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10497 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10498 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10499 * config/i386/sse.md (movdi_to_sse): Ditto.
10500 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10501 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10502 TARGET_INTER_UNIT_MOVES.
10503 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10504 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10505 instead of TARGET_INTER_UNIT_MOVES.
10506 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10507 operand 1 from Yi to Yj and from Ym to Yn.
10509 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10511 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10512 (float_truncate_hi_): Likewise.
10513 (float_extend_lo_): Likewise.
10514 (float_truncate_lo_): Likewise.
10515 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10516 (aarch64_float_extend_lo_v2df): Likewise.
10517 (vec_unpacks_hi_v4sf): Likewise.
10518 (aarch64_float_truncate_lo_v2sf): Likewise.
10519 (aarch64_float_truncate_hi_v4sf): Likewise.
10520 (vec_pack_trunc_v2df): Likewise.
10521 (vec_pack_trunc_df): Likewise.
10523 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10525 * config/aarch64/aarch64-builtins.c
10526 (aarch64_fold_builtin): Fold float conversions.
10527 * config/aarch64/aarch64-simd-builtins.def
10528 (floatv2si, floatv4si, floatv2di): New.
10529 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10530 * config/aarch64/aarch64-simd.md
10531 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10532 * config/aarch64/iterators.md (FLOATUORS): New.
10533 (optab): Add float, floatuns.
10534 (su_optab): Likewise.
10536 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10538 * config/aarch64/aarch64-builtins.c
10539 (aarch64_builtin_vectorized_function): Use new names for
10541 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10542 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10543 (fcvtzu): Split as...
10544 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10545 (fcvtas): Split as...
10546 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10547 (fcvtau): Split as...
10548 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10549 (fcvtps): Split as...
10550 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10551 (fcvtpu): Split as...
10552 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10553 (fcvtms): Split as...
10554 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10555 (fcvtmu): Split as...
10556 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10557 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10558 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10559 (lfrintnusf, lfrintnudf): Likewise.
10560 * config/aarch64/aarch64-simd.md
10561 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10563 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10564 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10565 (fcvt_pattern): Likewise.
10567 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10569 * config/aarch64/aarch64-simd.md
10570 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10571 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10573 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10575 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10576 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10577 (vrnd<a,m,n,p>_f32): Implement using builtins.
10578 (vrnd<i,x><q>_f<32, 64>): New.
10580 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10582 * config/aarch64/aarch64-builtins.c
10583 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10584 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10585 (frintz): Rename to...
10587 (frintp): Rename to...
10589 (frintm): Rename to...
10591 (frinti): Rename to...
10592 (nearbyint): ...this.
10593 (frintx): Rename to...
10595 (frinta): Rename to...
10597 * config/aarch64/aarch64-simd.md
10598 (aarch64_frint<frint_suffix><mode>): Delete.
10599 (<frint_pattern><mode>2): Convert to insn.
10600 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10601 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10602 (frint_pattern): Likewise.
10603 (frint_suffix): Likewise.
10605 2013-04-29 Richard Biener <rguenther@suse.de>
10607 PR tree-optimization/57081
10608 * loop-init.c: Include tree-flow.h.
10609 (loop_optimizer_finalize): Free number of iteration estimates.
10610 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10612 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10614 PR tree-optimization/57083
10615 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10616 non-singleton shift count range, zero extend low_bound for uns case.
10618 * config/i386/predicates.md (general_vector_operand): New predicate.
10619 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10620 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10621 if they aren't nonimmediate operands. If their original values
10622 satisfy const_vector_equal_evenodd_p, don't shift them.
10623 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10624 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10625 if not nonimmediate_operand.
10626 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10627 instead of register_operand.
10628 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10630 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10632 * stor-layout.c (finalize_size_functions): Allocate a structure and
10633 reset cfun before dumping the functions.
10635 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10637 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10640 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10641 use xop_pmacsdqh if uns_p.
10642 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10643 the immediate rotate count.
10645 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10647 * rtl.h (struct rtx_def): Add comment for field jump.
10648 (LRA_SUBREG_P): New macro.
10649 * recog.c (register_operand): Check LRA_SUBREG_P.
10650 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10652 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10653 (lra_final_code_change): Skip subreg change for operators.
10654 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10655 if there are no operand changes.
10656 * lra-constraints.c (curr_insn_set): New.
10657 (match_reload): Set LRA_SUBREG_P.
10658 (emit_spill_move): Ditto.
10659 (check_and_process_move): Use curr_insn_set. Process only single
10660 set insns. Don't initialize sec_mem_p and change_p.
10661 (simplify_operand_subreg): Use LRA_SUBREG_P.
10662 (reg_in_class_p): New function.
10663 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10664 of #ifdef. Add code to remove cycling.
10665 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10666 non-null disp. Reload inner instead of disp when base and index
10667 are null. Try to put lo_sum into register.
10668 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10669 (check_and_process_move): Move code for move cost check to
10670 simple_move_p. Remove equiv_substitution.
10671 (simple_move_p): New function.
10672 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10673 curr_insn_set. Call check_and_process_move only for single set
10674 insns. Use the new function. Move call of check_and_process_move
10675 after operand equiv substitution and address process.
10677 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10680 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10681 with nonlocal goto receivers or returns twice calls, ignore
10682 unininitialized values from abnormal edges to nl goto receiver
10683 or returns twice call.
10685 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10687 PR tree-optimization/57051
10688 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10689 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10692 2013-04-26 Richard Biener <rguenther@suse.de>
10694 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10695 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10696 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10697 (fixup_loop_arrays_after_move): New function.
10698 (move_sese_region_to_fn): Properly outline the loop tree parts
10699 of the SESE region.
10701 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10703 * config/i386/i386.md (type, unit): Fix long lines.
10705 2013-04-26 Richard Biener <rguenther@suse.de>
10707 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10708 (lto-streamer-out.o): Likewise.
10709 * cfgloop.c (init_loops_structure): Export, add struct function
10710 argument and adjust.
10711 (flow_loops_find): Adjust.
10712 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10713 (init_loops_structure): Declare.
10714 * lto-streamer-in.c: Include cfgloop.h.
10715 (input_cfg): Input the loop tree.
10716 * lto-streamer-out.c: Include cfgloop.h.
10717 (output_cfg): Output the loop tree.
10718 (output_struct_function_base): Do not drop PROP_loops.
10720 2013-03-26 Richard Biener <rguenther@suse.de>
10722 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10723 (pass_build_cfg): Provide PROP_loops.
10724 (move_sese_region_to_fn): Remove loops that are outlined into fn
10726 * tree-inline.c: Include cfgloop.h.
10727 (initialize_cfun): Do not drop PROP_loops.
10728 (copy_loops): New function.
10729 (copy_cfg_body): Copy loop structure.
10730 (tree_function_versioning): Initialize destination loop tree.
10731 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10732 (pass_parallelize_loops): Do IL verification.
10733 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10734 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10735 the CFG make sure we fixup loops as well.
10736 * tree-ssa-tail-merge.c: Include cfgloop.h.
10737 (replace_block_by): When merging loop latches mark loops for fixup.
10738 * lto-streamer-out.c (output_struct_function_base): Drop
10739 PROP_loops for now.
10740 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10741 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10742 * ipa-split.c: Include cfgloop.h.
10743 (split_function): Add the new return block to the loop tree root.
10744 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10745 whether we have removed the forwarder block.
10746 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10747 * cfgloop.h (place_new_loop): Declare.
10748 * cfgloopmanip.c (place_new_loop): Export.
10749 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10750 (tree-switch-conversion.o): Likewise.
10751 (tree-complex.o): Likewise.
10752 (tree-inline.o): Likewise.
10753 (tree-ssa-tailmerge.o): Likewise.
10754 (ipa-split.o): Likewise.
10755 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10756 (tree-ssa-copy.o): Likewise.
10757 * tree-switch-conversion.c: Include cfgloop.h
10758 (process_switch): If we emit a bit-test cascade, schedule loops
10760 * tree-complex.c: Include cfgloop.h.
10761 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10762 * asan.c: Include cfgloop.h.
10763 (create_cond_insert_point): Properly add new basic-blocks to
10764 loops, schedule loop fixup.
10765 * cfgloop.c (verify_loop_structure): Check that looks are not
10767 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10769 (expand_omp_for_generic): Likewise.
10770 (expand_omp_sections): Likewise.
10771 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10772 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10773 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10774 is initialized, not when loops are present.
10775 * tree-parloops.c (parallelize_loops): Remove checking here.
10776 * passes.c (init_optimization_passes): Schedule a copy-propagation
10777 pass before complete unrolling of inner loops.
10779 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10781 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10782 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10783 (colorize_init): Add argument to _WIN32 version.
10784 * toplev.c: Include diagnostic-color.h.
10785 (process_options): Default to -fdiagnostics-color=auto if
10786 GCC_COLORS env var is in the environment.
10787 * common.opt (fdiagnostics-color=): Add Var and Init.
10788 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10789 env var is in the environment, the default is auto rather than never.
10791 * diagnostic.h (file_name_as_prefix): Add context argument.
10792 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10793 the string as locus.
10794 * langhooks.c (lhd_print_error_function): Adjust caller.
10796 2013-04-25 Lawrence Crowl <crowl@google.com>
10798 * var-tracking.c (shared_hash_def::htab):
10799 Change type to hash_table. Update dependent calls and types.
10801 2013-04-25 Lawrence Crowl <crowl@google.com>
10803 * Makefile.in: Update as needed below.
10805 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10806 Move declaration to after the type's method definitons.
10808 * attribs.c (htab_t scoped_attributes::attribute_hash):
10809 Change type to hash_table. Update dependent calls and types.
10811 * bitmap.c (htab_t bitmap_desc_hash):
10812 Change type to hash_table. Update dependent calls and types.
10814 * cselib.c (htab_t cselib_hash_table):
10815 Change type to hash_table. Update dependent calls and types.
10817 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10818 (hash_string_slot_node): Move implementation into lto-streamer.h
10819 struct string_slot_hasher.
10820 (eq_string_slot_node): Likewise.
10822 * data-streamer-out.c: Update output_block::string_hash_table
10823 dependent calls and types.
10825 * dwarf2cfi.c (htab_t trace_index):
10826 Change type to hash_table. Update dependent calls and types.
10828 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10829 Change type to hash_table. Update dependent calls and types.
10830 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10831 (htab_t optimize_external_refs::map): Likewise.
10832 (htab_t output_comp_unit::extern_map): Likewise.
10833 (htab_t output_comdat_type_unit::extern_map): Likewise.
10834 (htab_t output_macinfo::macinfo_htab): Likewise.
10835 (htab_t optimize_location_lists::htab): Likewise.
10836 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10838 * except.c (htab_t ehspec_hash_type):
10839 Change type to hash_table. Update dependent calls and types.
10840 (assign_filter_values::ttypes): Likewise.
10841 (assign_filter_values::ehspec): Likewise.
10842 (sjlj_assign_call_site_values::ar_hash): Likewise.
10843 (convert_to_eh_region_ranges::ar_hash): Likewise.
10845 * gcse.c (htab_t pre_ldst_table):
10846 Change type to hash_table. Update dependent calls and types.
10848 * ggc-common.c (htab_t saving_htab):
10849 Change type to hash_table. Update dependent calls and types.
10850 (htab_t loc_hash): Likewise.
10851 (htab_t ptr_hash): Likewise.
10852 (call_count): Rename ggc_call_count.
10853 (call_alloc): Rename ggc_call_alloc.
10854 (loc_descriptor): Rename make_loc_descriptor.
10855 (add_statistics): Rename ggc_add_statistics.
10857 * ggc-common.c (saving_htab):
10858 Change type to hash_table. Update dependent calls and types.
10860 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10861 (push_gimplify_context): Likewise.
10862 (pop_gimplify_context): Likewise.
10863 (struct gimple_temp_hash_elt): Added.
10864 (struct gimplify_hasher): Likewise.
10865 (struct gimplify_ctx.temp_htab):
10866 Change type to hash_table. Update dependent calls and types.
10868 * gimple-fold.c: Include gimplify-ctx.h.
10870 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10871 Change type to hash_table. Update dependent calls and types.
10872 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10873 avoid potential global name collision.
10875 * gimplify.c: Include gimplify-ctx.h.
10876 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10877 (htab_t gimplify_ctx::temp_htab):
10878 Update dependent calls and types for new type hash_table.
10879 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10880 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10882 * gimplify-ctx.h: New.
10883 (struct gimple_temp_hash_elt): Move from gimplify.c.
10884 (class gimplify_hasher): New.
10885 (struct gimplify_ctx): Move from gimple.h.
10886 (htab_t gimplify_ctx::temp_htab):
10887 Change type to hash_table. Update dependent calls and types.
10889 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10890 (htab_t ivs_params::newivs_index):
10891 Change type to hash_table. Update dependent calls and types.
10892 (htab_t ivs_params::params_index): Likewise.
10893 (htab_t print_generated_program::params_index): Likewise.
10894 (htab_t gloog::newivs_index): Likewise.
10895 (htab_t gloog::params_index): Likewise.
10897 * graphite.c: Include graphite-htab.h.
10898 4htab_t graphite_transform_loops::bb_pbb_mapping):
10899 Change type to hash_table. Update dependent calls and types.
10901 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10902 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10903 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10905 * graphite-dependences.c: Include graphite-htab.h.
10906 (loop_is_parallel_p): Change hash table type of parameter.
10908 * graphite-htab.h: New.
10909 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10910 (extern find_pbb_via_hash): Move from graphite-poly.h.
10911 (extern loop_is_parallel_p): Move from graphite-poly.h.
10912 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10914 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10915 (extern loop_is_parallel_p): Move to graphite-htab.h.
10916 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10918 * haifa-sched.c (htab_t delay_htab):
10919 Change type to hash_table. Update dependent calls and types.
10920 (htab_t delay_htab_i2): Likewise.
10922 * ira-color.c (htab_t allocno_hard_regs_htab):
10923 Change type to hash_table. Update dependent calls and types.
10925 * ira-costs.c (htab_t cost_classes_htab):
10926 Change type to hash_table. Update dependent calls and types.
10928 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10929 Change type to hash_table. Update dependent calls and types.
10931 * loop-iv.c (htab_t bivs):
10932 Change type to hash_table. Update dependent calls and types.
10934 * loop-unroll.c (htab_t opt_info::insns_to_split):
10935 Change type to hash_table. Update dependent calls and types.
10936 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10938 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10939 (struct string_slot_hasher): New.
10940 (htab_t output_block::string_hash_table):
10941 Change type to hash_table. Update dependent calls and types.
10943 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10944 (htab_t file_name_hash_table):
10945 Change type to hash_table. Update dependent calls and types.
10947 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10950 * lto-streamer.c (htab_t tree_htab):
10951 Change type to hash_table. Update dependent calls and types.
10953 * omp-low.c: Include gimplify-ctx.h.
10955 * passes.c (htab_t name_to_pass_map):
10956 Change type to hash_table. Update dependent calls and types.
10957 (pass_traverse): Rename to passes_pass_traverse.
10959 * plugin.c (htab_t event_tab):
10960 Change type to hash_table. Update dependent calls and types.
10962 * postreload-gcse.c (htab_t expr_table):
10963 Change type to hash_table. Update dependent calls and types.
10964 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10966 * sese.c (debug_rename_map_1): Make extern.
10967 (htab_t copy_bb_and_scalar_dependences::rename_map):
10968 Change type to hash_table. Update dependent calls and types.
10970 * sese.h (extern debug_rename_map): Move to .c file.
10972 * store-motion.c (htab_t store_motion_mems_table):
10973 Change type to hash_table. Update dependent calls and types.
10975 * trans-mem.c (htab_t tm_new_mem_hash):
10976 Change type to hash_table. Update dependent calls and types.
10978 * tree-browser.c (htab_t TB_up_ht):
10979 Change type to hash_table. Update dependent calls and types.
10981 * tree-cfg.c (htab_t discriminator_per_locus):
10982 Change type to hash_table. Update dependent calls and types.
10984 * tree-complex.c: Include tree-hasher.h
10985 (htab_t complex_variable_components):
10986 Change type to hash_table. Update dependent calls and types.
10988 * tree-eh.c (htab_t finally_tree):
10989 Change type to hash_table. Update dependent calls and types.
10991 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10992 struct int_tree_hasher.
10993 (extern int_tree_map_eq): Likewise.
10994 (uid_decl_map_hash): Removed.
10995 (extern decl_tree_map_eq): Likewise.
10997 * tree-hasher.h: New.
10998 (struct int_tree_hasher): New.
10999 (typedef int_tree_htab_type): New.
11001 * tree-inline.c: Include gimplify-ctx.h.
11003 * tree-mudflap.c: Include gimplify-ctx.h.
11005 * tree-parloops.c: Include tree-hasher.h.
11006 (htab_t eliminate_local_variables_stmt::decl_address):
11007 Change type to hash_table. Update dependent calls and types.
11008 (htab_t separate_decls_in_region::decl_copies): Likewise.
11010 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11011 Change type to hash_table. Update dependent calls and types.
11013 * tree-sra.c (candidates):
11014 Change type to hash_table. Update dependent calls and types.
11016 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11018 (int_tree_map_hash): Likewise.
11020 * tree-ssa-dom.c (htab_t avail_exprs):
11021 Change type to hash_table. Update dependent calls and types.
11023 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11024 Change type to hash_table. Update dependent calls and types.
11026 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11027 Change type to hash_table. Update dependent calls and types.
11029 * tree-ssa-phiopt.c (seen_ssa_names):
11030 Change type to hash_table. Update dependent calls and types.
11032 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11033 Change type to hash_table. Update dependent calls and types.
11035 * tree-ssa-uncprop.c (equiv):
11036 Change type to hash_table. Update dependent calls and types.
11038 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11040 PR rtl-optimization/57003
11041 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11042 call note_stores with kill_clobbered_value callback again after
11043 killing regs_invalidated_by_call.
11045 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11047 * config/aarch64/aarch64-simd.md
11048 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11049 (aarch64_simd_bsl<mode>): Likewise.
11050 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11052 2013-04-25 Marek Polacek <polacek@redhat.com>
11054 PR tree-optimization/57066
11055 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11057 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11059 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11061 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11063 * config/aarch64/aarch64-builtins.c
11064 (aarch64_fold_builtin): New.
11065 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11066 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11067 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11068 * config/aarch64/arm_neon.h
11069 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11071 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11072 Tejas Belagod <tejas.belagod@arm.com>
11074 * config/aarch64/aarch64-builtins.c
11075 (aarch64_gimple_fold_builtin): New.
11076 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11077 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11078 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11079 (addvv4sf): Update.
11080 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11082 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11084 * config/aarch64/aarch64.md
11085 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11087 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11089 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11090 (*ngcsi_uxtw): New pattern.
11092 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11093 Julian Brown <julian@codesourcery.com>
11095 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11096 (TB_DREG): Add T_V4HF.
11097 (v4hf_UP): New macro.
11098 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11099 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11100 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11102 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11103 (arm_vector_mode_supported_p): Handle V4HF.
11104 (arm_mangle_map): Handle V4HFmode.
11105 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11106 * config/arm/arm_neon_builtins.def: Add entries for
11107 vcvtv4hfv4sf, vcvtv4sfv4hf.
11108 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11109 (neon_vcvtv4hfv4sf): Likewise.
11110 * config/arm/neon-gen.ml: Handle half-precision floating point
11112 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11113 * config/arm/arm_neon.h: Regenerate.
11114 * config/arm/neon.ml (type elts): Add F16.
11115 (type vectype): Add T_float16x4, T_floatHF.
11116 (type vecmode): Add V4HF.
11117 (type features): Add Requires_FP_bit feature.
11118 (elt_width): Handle F16.
11119 (elt_class): Likewise.
11120 (elt_of_class_width): Likewise.
11121 (mode_of_elt): Refactor.
11122 (type_for_elt): Handle F16, fix error messages.
11123 (vectype_size): Handle T_float16x4.
11124 (vcvt_sh): New function.
11125 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11126 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11127 (string_of_mode): Handle V4HF.
11128 * doc/arm-neon-intrinsics.texi: Regenerate.
11130 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11132 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11133 format specifier in 'X' case.
11135 2013-04-25 Alan Modra <amodra@gmail.com>
11138 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11139 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11140 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11141 Repeat for many other rotate/shift and mask patterns using subregs.
11142 Name lshiftrt insns.
11143 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11144 on WORDS_BIG_ENDIAN.
11146 2013-04-25 Alan Modra <amodra@gmail.com>
11148 * config.gcc: Support little-endian powerpc-linux targets.
11149 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11150 (LINK_OS_LINUX_SPEC): Define.
11151 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11152 Preserve MASK_LITTLE_ENDIAN.
11153 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11154 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11155 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11156 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11157 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11158 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11159 Correct fp word order for little-endian. Don't shift toc entries
11160 smaller than a word for little-endian.
11161 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11162 (bswapdi2 splits): Correct low-part subreg for little-endian.
11163 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11164 low/high where such is correct only for be.
11165 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11166 little-endian for -mcall-aixdesc.
11168 2013-04-25 Alan Modra <amodra@gmail.com>
11170 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11171 replace_equiv_address_nv.
11173 2013-04-25 Alan Modra <amodra@gmail.com>
11175 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11177 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11180 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11181 * rtl.h (struct rtx_def): ...
11183 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11185 PR rtl-optimizations/57046
11186 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11187 for multi-reg splits.
11189 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11191 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11193 2013-04-24 Sterling Augustine <saugustine@google.com>
11195 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11196 (comp_dir_string, debug_str_dwo_section): New.
11197 (DEBUG_STR_DWO_SECTION): Rename to ...
11198 (DEBUG_DWO_STR_SECTION): ... this.
11199 (DEBUG_NORM_STR_SECTION): Delete.
11200 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11201 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11202 (find_AT_string): Move most logic to ...
11203 (find_AT_string_in_table): ... here. New.
11204 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11205 add_skeleton_AT_string. Delete logic.
11206 (output_skeleton_debug_sections): Remove call to
11207 add_top_level_skeleton_die_attrs.
11208 (add_comp_dir_attribute): Move logic to comp_dir_string.
11209 (dwarf2out_init): Initialize debug_str_dwo_section.
11210 (output_indirect_string): Call find_string_form.
11211 (output_indirect_strings): Rewrite.
11212 (prune_unused_types): Empty skeleton_debug_str_hash.
11213 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11214 (dwarf2out_finish): Call output_indirect_strings.
11216 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11218 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11220 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11222 * rtl.h (struct rtx_def): Add comment for field jump.
11223 (LRA_SUBREG_P): New macro.
11224 * recog.c (register_operand): Check LRA_SUBREG_P.
11225 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11227 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11228 (lra_final_code_change): Skip subreg change for operators.
11229 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11230 if there are no operand changes.
11231 * lra-constraints.c (curr_insn_set): New.
11232 (match_reload): Set LRA_SUBREG_P.
11233 (emit_spill_move): Ditto.
11234 (check_and_process_move): Use curr_insn_set. Process only single
11235 set insns. Don't initialize sec_mem_p and change_p.
11236 (simplify_operand_subreg): Use LRA_SUBREG_P.
11237 (reg_in_class_p): New function.
11238 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11239 of #ifdef. Add code to remove cycling.
11240 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11241 non-null disp. Reload inner instead of disp when base and index
11242 are null. Try to put lo_sum into register.
11243 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11244 (check_and_process_move): Move code for move cost check to
11245 simple_move_p. Remove equiv_substitution.
11246 (simple_move_p): New function.
11247 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11248 curr_insn_set. Call check_and_process_move only for single set
11249 insns. Use the new function. Move call of check_and_process_move
11250 after operand equiv substitution and address process.
11252 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
11254 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
11255 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
11256 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
11258 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11260 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
11262 2013-04-24 Marek Polacek <polacek@redhat.com>
11264 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
11265 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
11266 (select_loops_exit_conditions): Likewise.
11267 (number_of_iterations_for_all_loops): Likewise.
11268 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
11269 (scev_analysis): Likewise.
11271 2013-04-02 Catherine Moore <clm@codesourcery.com>
11272 Chao-ying Fu <fu@mips.com>
11274 * config/mips/micromips.md (jraddiusp): New pattern.
11275 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
11276 instruction if possible.
11278 2013-04-24 Alan Modra <amodra@gmail.com>
11280 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
11282 2013-04-24 Julian Brown <julian@codesourcery.com>
11283 Chung-Lin Tang <cltang@codesourcery.com>
11285 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
11286 dependency behavior in enumeration type DIE generation. Add TODO note
11287 to comments about future DW_FORM_sdata/udata re-work of related code.
11289 2013-04-23 Lawrence Crowl <crowl@google.com>
11291 * Makefile.in: Update as needed below.
11293 * hash-table.h (class hash_table):
11294 Correct many methods with parameter types compare_type to the correct
11295 value_type. (Correct code was unlikely to notice the change.)
11296 (hash_table::elements_with_deleted) New.
11297 (class hashtable::iterator): New.
11298 (hashtable::begin()): New.
11299 (hashtable::end()): New.
11300 (FOR_EACH_HASH_TABLE_ELEMENT): New.
11302 * statistics.c (statistics_hashes):
11303 Change type to hash_table. Update dependent calls and types.
11305 * tree-into-ssa.c (var_infos):
11306 Change type to hash_table. Update dependent calls and types.
11308 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
11309 Change type to hash_table. Update dependent calls and types.
11311 * tree-ssa-loop-im.c (struct mem_ref.refs):
11312 Change type to hash_table. Update dependent calls and types.
11314 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
11315 Change type to hash_table. Update dependent calls and types.
11317 * tree-ssa-sccvn.c (vn_tables_s::nary):
11318 Change type to hash_table. Update dependent calls and types.
11319 (vn_tables_s::phis): Likewise.
11320 (vn_tables_s::references): Likewise.
11322 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
11323 (vn_reference_eq): Update parameter and return types.
11325 * tree-ssa-structalias.c (pointer_equiv_class_table):
11326 Change type to hash_table. Update dependent calls and types.
11327 (location_equiv_class_table): Likewise.
11329 * tree-vect-data-refs.c: Consequential changes for making
11330 peeling a hash_table.
11332 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
11333 (destroy_loop_vec_info): Dependent hash_table update.
11335 * tree-vectorizer.h (peeling_htab):
11336 Change type to hash_table. Update dependent calls and types.
11338 2013-04-23 Shiva Chen <shiva0217@gmail.com>
11340 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
11341 to check the register content is equal or not.
11342 * lra-constraints.c (match_reload): Use lra_assign_reg_val
11343 to assign register content record.
11344 * lra-eliminations.c (update_reg_eliminate): Use
11345 lra_update_reg_val_offset to update register content offset.
11346 * lra-int.h (struct lra_reg): Add offset member.
11347 (lra_reg_val_equal_p): New static inline function.
11348 (lra_update_reg_val_offset): New static inline function.
11349 (lra_assign_reg_val): New static inline function.
11350 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
11351 to assign register content record.
11352 (initialize_lra_reg_info_element): Initial offset to zero.
11354 2013-04-23 Catherine Moore <clm@codesourcery.com>
11356 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
11357 operands. Record compression.
11359 2013-04-23 Xinliang David Li <davidxl@google.com>
11361 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
11363 2013-04-23 Richard Biener <rguenther@suse.de>
11365 PR middle-end/57036
11366 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
11367 parameter, only add abnormal goto edges from the copied body
11368 if the call could perform abnormal gotos.
11369 (copy_cfg_body): Adjust.
11371 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
11373 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
11375 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
11377 * coretypes.h (gimple_stmt_iterator): Add struct to make
11380 2013-04-23 Richard Biener <rguenther@suse.de>
11382 PR tree-optimization/57026
11383 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
11384 from SSA names occuring in abnormal PHI nodes.
11386 2013-04-22 Andi Kleen <ak@linux.intel.com>
11388 * lto/lto.c (print_lto_report_1): Fix LTO report names.
11390 2013-04-22 Andi Kleen <ak@linux.intel.com>
11392 * lto/lto.c (print_lto_report_1): Declare early.
11393 (read_cgraph_and_symbols): Call print_lto_report_1 early.
11395 2013-04-22 Andi Kleen <ak@linux.intel.com>
11397 * common.opt (-flto-report-wpa): Add.
11398 * doc/invoke.texi (-flto-report-wpa): Add.
11399 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
11402 2013-04-22 Xinliang David Li <davidxl@google.com>
11404 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
11405 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
11406 * Makefile.in: New dependency
11408 David Daney <ddaney.cavm@gmail.com>
11410 * configure.ac (gcc_cv_as_micromips_support): Use the
11411 --fatal-warnings option.
11412 * configure: Regenerate.
11414 2013-04-22 Marek Polacek <polacek@redhat.com>
11417 * tsan.c (instrument_expr): Don't instrument expression
11418 in case its size is zero.
11420 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
11424 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
11426 * config/alpha/alpha.c (TARGET_LRA_P): New define.
11428 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11430 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
11431 (gimple_stmt_iterator): New typedef.
11432 * gimple.h (gimple_stmt_iterator): Rename to...
11433 (gimple_stmt_iterator_d): ... This.
11434 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
11435 trees be valid for GIMPLE and GENERIC.
11436 (TARGET_GIMPLE_FOLD_BUILTIN): New.
11437 * gimple-fold.c (gimple_fold_call): Call target hook
11438 gimple_fold_builtin.
11439 * hooks.c (hook_bool_gsiptr_false): New.
11440 * hooks.h (hook_bool_gsiptr_false): New.
11441 * target.def (fold_stmt): New.
11442 * doc/tm.texi: Regenerate.
11444 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
11447 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
11448 a set sp if no stack realignment.
11450 2013-04-22 Nick Clifton <nickc@redhat.com>
11452 * config.gcc (tilegx-linux): Extend extra_objs rather than
11454 (tilepro-linux): Likewise.
11456 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11458 * config/aarch64/aarch64-builtins.c
11460 (CF0, CF1, CF2, CF3, CF4, CF10): New.
11461 (VAR<1-12>): Add MAP parameter.
11462 (BUILTIN_*): Likewise.
11463 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
11464 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
11465 (aarch64_ushl_n<mode>): Likewise.
11466 (aarch64_sshr_n<mode>): Likewise.
11467 (aarch64_ushr_n<mode>): Likewise.
11468 (aarch64_<maxmin><mode>): Likewise.
11469 (aarch64_sqrt<mode>): Likewise.
11470 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
11471 (vshr<q>_n_*): Likewise.
11473 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11475 * config/aarch64/aarch64-builtins.c
11476 (aarch64_simd_builtin_type_mode): Handle SF types.
11478 (BUILTIN_GPF): Define.
11479 (aarch64_init_simd_builtins): Handle SF types.
11480 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
11481 (frecps): Likewise.
11482 (frecpx): Likewise.
11483 * config/aarch64/aarch64-simd.md
11484 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
11485 (aarch64_frecpe<mode>): New.
11486 (aarch64_frecps<mode>): Likewise.
11487 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
11488 (v8type): Add frecp<esx>.
11489 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
11490 (aarch64_frecps<mode>): Likewise.
11491 * config/aarch64/iterators.md (FRECP): New.
11492 (frecp_suffix): Likewise.
11493 * config/aarch64/arm_neon.h
11494 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11496 2013-04-22 Christian Bruel <christian.bruel@st.com>
11499 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11500 (REG_CLASS_NAMES): Idem.
11501 (REG_CLASS_CONTENTS): Idem.
11502 (REGCLASS_HAS_FP_REG): Idem.
11503 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11504 (sh_conditional_register_usage): Idem.
11506 2013-04-21 Jeff Law <law@redhat.com>
11508 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11509 (ssa_forward_propagate_and_combine): Use it.
11511 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11513 * lra.c: Update the flow chart diagram.
11515 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11517 PR rtl-optimization/56847
11518 * lra-constraints.c (process_alt_operands): Discourage alternative
11519 with non-matche doffsettable memory constraint fro memory with
11522 2013-04-19 Richard Biener <rguenther@suse.de>
11524 PR tree-optimization/56982
11525 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11527 * gimplify.c (gimplify_call_expr): Notice special calls.
11528 (gimplify_modify_expr): Likewise.
11529 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11530 abnormal control flow receivers.
11531 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11532 in the same way as cfun->has_nonlocal_labels.
11533 (gimple_purge_dead_abnormal_call_edges): Likewise.
11534 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11535 receivers start a basic-block.
11537 2013-04-19 Richard Biener <rguenther@suse.de>
11539 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11541 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11542 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11543 (SLP_TREE_LOAD_PERMUTATION): Add.
11544 (vect_transform_slp_perm_load): Adjust prototype.
11545 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11546 (vect_free_slp_instance): Likewise.
11547 (vect_create_new_slp_node): Likewise.
11548 (vect_supported_slp_permutation_p): Remove.
11549 (vect_slp_rearrange_stmts): Adjust.
11550 (vect_supported_load_permutation_p): Likewise. Inline
11551 vect_supported_slp_permutation_p here.
11552 (vect_analyze_slp_instance): Compute load permutations per
11553 slp node instead of per instance.
11554 (vect_get_slp_defs): Adjust.
11555 (vect_transform_slp_perm_load): Likewise.
11556 (vect_schedule_slp_instance): Remove redundant code.
11557 (vect_schedule_slp): Remove hack for PR56270, add it ...
11558 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11559 CSE loads for SLP. Adjust.
11561 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11563 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11564 spelling in two comments.
11566 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11569 * config/arm/arm.c (load_multiple_sequence): Require SP
11570 as base register for loads if SP is in the register list.
11572 2013-04-19 Martin Jambor <mjambor@suse.cz>
11574 PR tree-optimization/56718
11575 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11576 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11577 and made public. Adjusted all callers.
11578 (ipa_intraprocedural_devirtualization): New function.
11579 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11580 (ipa_intraprocedural_devirtualization): Likewise.
11581 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11583 2013-04-19 Richard Biener <rguenther@suse.de>
11585 PR tree-optimization/57000
11586 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11588 2013-04-19 Terry Guo <terry.guo@arm.com>
11590 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11592 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11593 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11594 (cortex_m4_fmacs): Use new reservations.
11595 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11597 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11599 PR rtl-optimization/56999
11600 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11602 (lra_coalesce): Remove split_origin_bitmap and related code.
11603 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11604 ranges if necessary.
11606 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11608 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11610 (ix86_expand_call): Remove clobbered_registers array and use
11611 x86_64_ms_sysv_extra_clobbered_registers instead.
11612 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11614 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11616 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11617 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11618 (*call_value_rex64_ms_sysv): Ditto.
11620 2013-04-18 Cary Coutant <ccoutant@google.com>
11622 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11623 parent instead of die_mark.
11625 2013-04-18 Diego Novillo <dnovillo@google.com>
11627 * gimple.c (create_gimple_tmp): New.
11628 (get_expr_type): New.
11629 (build_assign): New.
11630 (build_type_cast): New.
11631 * gimple.h (enum ssa_mode): Define.
11632 (gimple_seq_set_location): New.
11633 * asan.c (build_check_stmt): Change some gimple_build_* calls
11634 to use build_assign and build_type_cast.
11636 2013-04-18 Richard Biener <rguenther@suse.de>
11638 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11639 handle negative step. Remove redundant checks.
11640 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11641 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11642 for negative step and grouped loads fail to vectorize.
11644 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11646 * emit-rtl.c (reset_insn_used_flags): New function.
11647 (reset_all_used_flags): Use it.
11648 (verify_insn_sharing): New function.
11649 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11651 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11653 PR tree-optimization/56984
11654 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11655 and (x >> M) >= N don't register any assertion if N << M is the
11658 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11660 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11661 do not delete_insn insns that have not yet been emitted, only
11662 unlink them with remove_insn.
11663 * df-scan.c (df_insn_delete): Revert r197492.
11665 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11667 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11668 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11670 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11672 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11673 define_insn_and_split.
11674 (and_scc,ior_scc,negscc): Likewise.
11675 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11677 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11679 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11680 can benefit from using a sequence of LDRD instructions in epilogue
11681 instead of a single LDM instruction.
11683 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11686 * doc/extend.texi: Fix typo.
11688 2013-04-17 Richard Biener <rguenther@suse.de>
11690 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11691 (vect_build_slp_tree): ... here.
11692 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11693 match. Remove special-casing of mismatched complex loads.
11694 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11695 re-try the match with swapped commutative operands.
11696 (vect_supported_load_permutation_p): Remove special-casing of
11697 mismatched complex loads.
11698 (vect_analyze_slp_instance): Adjust.
11700 2013-04-17 Richard Biener <rguenther@suse.de>
11702 PR rtl-optimization/56921
11703 * cfgloop.h (struct loop): Add simple_loop_desc member.
11704 (struct niter_desc): Mark with GTY(()).
11705 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11706 * loop-iv.c (get_simple_loop_desc): Likewise.
11707 (free_simple_loop_desc): Likewise.
11710 2013-04-16 Richard Biener <rguenther@suse.de>
11712 PR rtl-optimization/56921
11713 * loop-init.c (pass_rtl_move_loop_invariants): Add
11714 TODO_do_not_ggc_collect to todo_flags_finish.
11715 (pass_rtl_unswitch): Same.
11716 (pass_rtl_unroll_and_peel_loops): Same.
11717 (pass_rtl_doloop): Same.
11719 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11721 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11722 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11723 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11725 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11726 * tree-streamer.c (record_common_node): Adjust reference in comment.
11728 2013-04-17 Terry Guo <terry.guo@arm.com>
11730 * config/arm/cortex-m4.md: Add a new bypass.
11732 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11734 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11736 (*subs_<optab><mode>_multp2): New pattern.
11737 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11738 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11740 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11742 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11743 (*subs_mul_imm_<mode>): New pattern.
11745 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11748 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11749 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11750 (vsx_movti_32bit): Change j->wa to O->wa.
11752 2013-04-16 Richard Biener <rguenther@suse.de>
11754 PR rtl-optimization/56921
11755 * loop-init.c (pass_rtl_move_loop_invariants): Add
11756 TODO_do_not_ggc_collect to todo_flags_finish.
11757 (pass_rtl_unswitch): Same.
11758 (pass_rtl_unroll_and_peel_loops): Same.
11759 (pass_rtl_doloop): Same.
11761 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11763 * config/arm/arm.c (emit_multi_reg_push): New declaration
11764 for an existing function.
11765 (arm_emit_strd_push): New function.
11766 (arm_expand_prologue): Used here.
11767 (arm_emit_ldrd_pop): New function.
11768 (arm_expand_epilogue): Used here.
11769 (arm_get_frame_offsets): Update condition.
11770 (arm_emit_multi_reg_pop): Add a special case for load of a single
11771 register with writeback.
11773 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11775 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11778 2013-04-16 Richard Biener <rguenther@suse.de>
11780 PR tree-optimization/56756
11781 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11782 (first_mem_ref_loc): New.
11783 (execute_sm): Place the load temporarily before a previous
11784 access instead of in the latch edge to ensure its SSA dependencies
11785 are defined at points dominating the load.
11787 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11789 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11790 correct fix by moving header and footer insn to the footer of
11791 the merged basic block. Clear BB_END of the merged-away block.
11793 PR middle-end/43631
11794 * emit-rtl.c (make_note_raw): New function.
11795 (link_insn_into_chain): New static inline function.
11796 (add_insn): Use it.
11797 (add_insn_before, add_insn_after): Factor insn chain linking code...
11798 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11799 using link_insn_into_chain.
11800 (note_outside_basic_block_p): New helper function for emit_note_after
11801 and emit_note_before.
11802 (emit_note_after): Use nobb variant of add_insn_after if the note
11803 should not be contained in a basic block.
11804 (emit_note_before): Use nobb variant of add_insn_before if the note
11805 should not be contained in a basic block.
11806 (emit_note_copy): Use make_note_raw.
11807 (emit_note): Likewise.
11808 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11809 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11810 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11811 the moved barrier the tail of the basic block it follows.
11812 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11814 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11816 PR tree-optimization/56962
11817 * gimple-ssa-strength-reduction.c (record_increment): Only set
11818 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11819 either rhs1 or rhs2 is equal to c->base_expr.
11821 2013-04-15 Richard Biener <rguenther@suse.de>
11823 PR tree-optimization/56933
11824 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11826 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11827 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11828 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11829 dependence check ...
11830 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11832 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11833 GROUP_READ_WRITE_DEPENDENCE.
11835 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11837 * emit-rtl.c (reset_all_used_flags): New function.
11838 (verify_rtl_sharing): Call reset_all_used_flags before and after
11839 performing the checks.
11841 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11843 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11844 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11845 * config/arm/constraints.md (De): New constraint.
11846 * config/arm/neon.md (anddi3_neon): Delete.
11847 (neon_vand<mode>): Expand to standard anddi3 pattern.
11848 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11849 Move earlier in the file.
11850 (neon_inv_logic_op2): Likewise.
11851 (arm_anddi_operand_neon): New predicate.
11853 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11855 * configure.ac (gcc_cv_ld_as_needed): Set
11856 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11857 Use -z ignore, -z record on *-*-solaris2*.
11858 (HAVE_LD_AS_NEEDED): Update comment.
11859 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11860 * configure: Regenerate.
11861 * config.in: Regenerate.
11862 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11863 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11864 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11865 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11866 equivalents. Fix markup.
11867 * doc/tm.texi: Regenerate.
11869 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11871 * config/i386/i386.opt: New option mstack-protector-guard=.
11872 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11873 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11874 TARGET_SSP_TLS_GUARD.
11875 * config/i386/i386.c (ix86_option_override_internal): Set
11876 ix86_stack_protector_guard.
11877 * config/i386/i386.md (stack_protect_set): Enable for
11878 TARGET_SSP_TLS_GUARD only.
11879 (stack_protect_set_<mode>): Ditto.
11880 (stack_protect_test): Ditto.
11881 (stack_protect_test_<mode>): Ditto.
11882 * doc/invoke.texi (i386 Option): Document.
11884 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11887 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11888 (S_MODES): Set H_MODE bit.
11889 (SF_MODES): Set only S_MODE and SF_MODE bits.
11890 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11891 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11892 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11893 <MODE_FLOAT>: Likewise.
11895 2013-04-15 Joey Ye <joey.ye@arm.com>
11897 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11899 2013-04-15 Joey Ye <joey.ye@arm.com>
11901 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11903 (thumb_far_jump_used_p): Count instruction size and set
11906 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11908 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11909 * resource.c (mark_target_live_regs): Reformat conditional block.
11911 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11913 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11914 notes, they are emitted only just before final.
11915 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11917 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11919 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11920 * cfgrtl.c (delete_insn): Call it here instead.
11921 * lra-spills.c (lra_final_code_change): Use delete_insn.
11922 * haifa-sched.c (sched_remove_insn): Likewise.
11923 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11924 returning to the nop pool.
11925 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11926 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11928 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11930 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11931 * doc/tm.texi: Regenerated.
11933 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11935 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11938 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11940 * df-core.c (df_find_def): Compare register numbers.
11941 (df_find_use): Likewise.
11943 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11946 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11947 lra_in_progress for return.
11949 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11951 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11952 define_insn into define_insn_and_split and emit movsicc patterns.
11954 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11956 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11958 2013-04-12 Richard Biener <rguenther@suse.de>
11960 * tree-pass.h (TODO_do_not_ggc_collect): New.
11961 * passes.c (execute_one_ipa_transform_pass): Honor
11962 TODO_do_not_ggc_collect.
11963 (execute_one_pass): Likewise.
11966 2013-04-10 Richard Biener <rguenther@suse.de>
11968 * passes.c (init_optimization_passes): Remove reload pass.
11969 * ira.c (do_reload): Merge into ...
11971 (rest_of_handle_reload): Remove.
11972 (pass_reload): Likewise.
11973 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11974 of reload for vzeroupper pass placement.
11976 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11978 PR tree-optimization/56918
11979 PR tree-optimization/56920
11980 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11981 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11982 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11983 use wide_mul_with_sign method.
11985 2013-04-12 Richard Biener <rguenther@suse.de>
11987 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11988 not be considered a gimple constant.
11990 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11992 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11993 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11996 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11997 Jakub Jelinek <jakub@redhat.com>
11999 * opts.c: Include diagnostic-color.h.
12000 (common_handle_option): Handle OPT_fdiagnostics_color_.
12001 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12002 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12003 (diagnostic-color.o): New.
12004 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12005 (diagnostic_color_rule): New enum.
12006 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12007 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12008 the location string.
12009 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12010 either NULL, or color kind.
12011 * diagnostic-color.c: New file.
12012 * diagnostic-color.h: New file.
12013 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12015 * doc/invoke.texi (-fdiagnostics-color): Document.
12016 * pretty-print.h (pp_show_color): Define.
12017 (struct pretty_print_info): Add show_color field.
12018 * diagnostic.c: Include diagnostic-color.h.
12019 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12020 macros. Colorize error:, warning: etc. strings and also the location
12022 (diagnostic_show_locus): Colorize the caret line.
12023 * pretty-print.c: Include diagnostic-color.h.
12024 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12025 inside of %< %> quotes or quoted through q format modifier.
12027 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12029 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12031 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12033 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12034 code in CC_NZ mode.
12035 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12038 2013-04-11 Marek Polacek <polacek@redhat.com>
12040 PR tree-optimization/48184
12041 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12043 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12045 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12046 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12047 (skip_simple_arithmetic): Tidy up.
12048 * tree.h (skip_simple_constant_arithmetic): Declare.
12050 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12052 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12054 2013-04-11 Richard Biener <rguenther@suse.de>
12056 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12057 generate vector constants.
12059 2013-04-11 Richard Biener <rguenther@suse.de>
12061 PR tree-optimization/56878
12062 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12063 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12065 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12066 Prefer to align the DR with the most invariant base address.
12068 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12070 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12073 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12075 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12076 floating-point vector comparisons against 0.
12078 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12080 PR tree-optimization/56899
12081 * fold-const.c (extract_muldiv_1): Apply distributive law
12082 only if TYPE_OVERFLOW_WRAPS (ctype).
12084 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12087 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12088 loading has side effect.
12090 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12092 * config/sparc/sparc.c: Include tree-pass.h.
12093 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12094 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12095 head of file. Change return type. Split off gate function.
12096 (sparc_gate_work_around_errata): New function.
12097 (pass_work_around_errata): New pass definition.
12098 (insert_pass_work_around_errata) New pass insert definition to
12099 insert pass_work_around_errata just after delayed-branch scheduling.
12100 (sparc_option_override): Insert the pass.
12101 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12103 2013-04-10 David S. Miller <davem@davemloft.net>
12105 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12106 or -mcpu=hypersparc.
12108 * target.def (cstore_mode): New hook.
12109 * target.h: Include insn-codes.h
12110 * targhooks.c: Likewise.
12111 (default_cstore_mode): New function.
12112 * targhooks.h: Declare it.
12113 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12114 * doc/tm.texi: Rebuild.
12115 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12116 target hook, rather than inspecting the insn_data.
12117 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12118 (TARGET_CSTORE_MODE): Redefine.
12119 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12121 * config/sparc/predicates.md (cstore_result_operand): New special
12123 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12124 Use it for operand 0.
12125 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12126 (*snesi_special): Likewise.
12127 (*snesi_zero): Likewise.
12128 (*seqsi_zero): Likewise.
12129 (*sltu_insn): Likewise.
12130 (*sgeu_insn): Likewise.
12131 (*seqdi_special): Make operand 0 and comparison operation be of
12133 (*snedi_special): Likewise.
12134 (*snedi_special_vis3): Likewise.
12135 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12136 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12137 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12138 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12139 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12140 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12141 (*sltu_extend_sp64): Likewise.
12142 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12143 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12144 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12145 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12146 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12148 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12150 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12151 (aarch64_start_file): Use the new function.
12153 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12154 Jason Merrill <jason@redhat.com>
12156 * common.opt: Add -gdwarf.
12157 * opts.c (common_handle_option): Handle it.
12158 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12160 2013-04-10 Richard Biener <rguenther@suse.de>
12162 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12163 (execute_one_ipa_transform_pass): But unconditionally here.
12164 (execute_one_pass): And here.
12165 (init_optimization_passes): Remove reload pass.
12166 * tree-pass.h (TODO_ggc_collect): Remove.
12167 (pass_reload): Likewise.
12168 * ira.c (do_reload): Merge into ...
12170 (rest_of_handle_reload): Remove.
12171 (pass_reload): Likewise.
12172 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12173 of reload for vzeroupper pass placement.
12174 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12175 and todo_flags_finish of all passes.
12177 2013-04-10 Richard Biener <rguenther@suse.de>
12179 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12180 first_const_oprnd field, rename first_def_type to first_op_type.
12181 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12182 (vect_get_and_check_slp_defs): Always use the type of the
12183 operand. Allow mixed vect_external_def, vect_constant_def types.
12184 (vect_get_constant_vectors): Handle mixed vect_external_def,
12185 vect_constant_def types.
12187 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12189 PR tree-optimization/55524
12190 * tree-ssa-math-opts.c
12191 (convert_mult_to_fma): Don't use an fms construct
12192 when we don't have an fms operation, but fnma, and it looks
12193 likely that we'll be able to use the latter.
12195 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12197 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12199 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12200 inline fail caused by overwritable functions.
12202 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12204 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12205 unnecessary bits in the constant power of two case.
12207 2013-04-10 Richard Biener <rguenther@suse.de>
12209 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12210 broken code swapping operands.
12211 (vect_build_slp_tree): Do not compute load permutations here.
12212 (vect_analyze_slp_instance): Compute load permutations here,
12213 after building the SLP tree.
12215 2013-04-09 Christian Bruel <christian.bruel@st.com>
12217 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12218 of next/prev_real_insn.
12220 2013-04-09 Jan Hubicka <jh@suse.cz>
12222 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12223 Drop aliased parameter.
12224 (function_and_variable_visibility): Do not handle alias pairs.
12225 * cgraph.c (varpool_externally_visible_p): Update prototype.
12226 * varpool.c (varpool_add_new_variable): Update.
12228 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12230 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
12232 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
12234 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
12236 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
12238 2013-04-09 Marek Polacek <polacek@redhat.com>
12240 PR tree-optimization/48762
12241 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
12243 2013-04-09 Richard Biener <rguenther@suse.de>
12245 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
12247 (vect_build_slp_tree): Likewise.
12248 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
12249 calculating the cost of a SLP instance.
12250 (vect_analyze_slp_instance): Use it from here, after building
12253 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12255 PR middle-end/56883
12256 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
12257 expand_omp_for_static_chunk): Use simple_p = true in
12258 force_gimple_operand_gsi calls when assigning to addressable decls.
12260 2013-04-09 Jeff Law <law@redhat.com>
12262 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
12263 when the boolean was created by converting a wider object which
12264 had a boolean range.
12266 2013-04-09 Richard Biener <rguenther@suse.de>
12268 * tree-vectorizer.h (slp_void_p): Remove.
12269 (slp_tree): Typedef before _slp_tree declaration.
12270 (struct _slp_tree): Use a vector of slp_tree as children.
12271 (vect_get_place_in_interleaving_chain): Remove.
12272 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
12274 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
12276 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
12277 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
12278 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
12279 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
12280 Use slp_node instead of slp_void_p and adjust.
12282 2013-04-09 Richard Biener <rguenther@suse.de>
12284 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
12285 work that is not necessary.
12287 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12289 PR tree-optimization/56854
12290 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
12291 forward into clobber stmts if it would change MEM_REF lhs into
12294 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
12296 * tree.c (type_hash_lookup, type_hash_add): Make static.
12297 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
12299 2013-04-09 Richard Biener <rguenther@suse.de>
12301 * tree.h (unsave_expr_now): Remove.
12302 * tree-inline.c (mark_local_for_remap_r): Remove.
12303 (unsave_expr_1): Likewise.
12304 (unsave_r): Likewise.
12305 (unsave_expr_now): Likewise.
12306 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
12307 (propagate_tree_value): Likewise.
12309 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
12311 * doc/rtl.texi (sequence): Rewrite documentation to match the
12312 current use of SEQUENCE rtl objects.
12313 * rtl.def (SEQUENCE): Likewise.
12315 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
12316 Update documentation.
12317 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
12318 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
12320 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
12322 2013-04-08 Teresa Johnson <tejohnson@google.com>
12324 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
12325 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
12327 (estimate_edge_size_and_time): Add comment to suggest using rounding
12329 (estimate_node_size_and_time): Ditto.
12330 (remap_edge_change_prob): Use helper rounding divide methods.
12331 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
12332 (gimple_mod_pow2_value_transform): Ditto.
12333 (gimple_mod_subtract_transform): Ditto.
12334 (gimple_ic_transform): Ditto.
12335 (gimple_stringops_transform): Ditto.
12336 * stmt.c (conditional_probability): Ditto.
12337 (emit_case_dispatch_table): Ditto.
12338 * lto-cgraph.c (merge_profile_summaries): Ditto.
12339 * tree-optimize.c (execute_fixup_cfg): Ditto.
12340 * cfgcleanup.c (try_forward_edges): Ditto.
12341 * cfgloopmanip.c (scale_loop_profile): Ditto.
12343 (duplicate_loop_to_header_edge): Ditto.
12344 (lv_adjust_loop_entry_edge): Ditto.
12345 * tree-vect-loop.c (vect_transform_loop): Ditto.
12346 * profile.c (compute_branch_probabilities): Ditto.
12347 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
12348 * lto-streamer-in.c (input_cfg): Ditto.
12349 * gimple-streamer-in.c (input_bb): Ditto.
12350 * ipa-cp.c (update_profiling_info): Ditto.
12351 (update_specialized_profile): Ditto.
12352 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
12353 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
12355 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12356 (compute_trg_info): Ditto.
12357 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
12358 (purge_dead_edges): Ditto.
12359 * loop-unswitch.c (unswitch_loop): Ditto.
12360 * cgraphclones.c (cgraph_clone_edge): Ditto.
12361 (cgraph_clone_node): Ditto.
12362 * tree-inline.c (copy_bb): Ditto.
12363 (copy_edges_for_bb): Ditto.
12364 (initialize_cfun): Ditto.
12365 (copy_cfg_body): Ditto.
12366 (expand_call_inline): Ditto.
12368 2013-04-08 Kai Tietz <ktietz@redhat.com>
12370 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
12371 TARGET_CYGWIN64 by TARGET_64BIT.
12373 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12375 * config/epiphany/epiphany.md (GPR_1): New constant.
12376 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
12377 * config/epiphany/epiphany.c (gen_compare_reg):
12378 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
12379 is already in place.
12380 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
12381 Don't require being called during rtl expansion; If y operlaps r0,
12383 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
12384 (epiphany_expand_epilogue): Likewise.
12386 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
12387 Don't use CC_FPmode for ORDERED / UNORDERED.
12388 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
12390 * config/epiphany/constraints.md (CnL): New constraint.
12391 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
12392 * config/epiphany/predicates.md (add_operand): Allow 1024.
12394 * config/epiphany/epiphany.md (logical_op): New code iterator.
12395 (op_mnc): New code attribute.
12396 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
12397 (mov_f+1, mov_f+2): New peephole2 patterns.
12399 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
12400 (cstoresi4): Also allow re-use of zero result when doing a NE
12401 comparison to a non-zero operand.
12402 Use (clobber (scratch)) for first insn if the gpr output is not needed.
12404 * config/epiphany/epiphany.md (<insn_opname>v2si3):
12405 Use gen_addsi3_i / gen_subsi3_i.
12407 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12411 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
12412 contain anything but clobbers, at most one __builtin_stack_restore,
12413 optionally debug stmts and final resx, and if it has at least one
12414 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
12415 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
12416 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
12417 which isn't defaut definition, remove them.
12418 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
12419 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
12420 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
12421 with MEM_REF LHS with SSA_NAME address.
12423 2013-04-08 Jeff Law <law@redhat.com>
12425 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
12427 2013-04-08 Richard Biener <rguenther@suse.de>
12429 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
12431 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
12432 determined vector type.
12433 (vect_analyze_data_refs): Likewise.
12434 (vect_get_new_vect_var): Adjust.
12435 (vect_create_destination_var): Preserve SSA name versions.
12436 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
12437 not dump anything here.
12439 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12441 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
12442 Add member lr_slot_known.
12443 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
12445 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12446 Remove code that sets lr_slot_offset according to what a previous
12447 version of epiphany_emit_save_restore used to do.
12448 (epiphany_emit_save_restore): When doing an lr save or restore,
12449 set/verify lr_slot_known and lr_slot_offset.
12451 2013-04-08 Xinyu Qi <xyqi@marvell.com>
12454 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
12457 2013-04-08 Richard Biener <rguenther@suse.de>
12459 * alias.c (find_base_term): Fix thinko in previous change.
12461 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12463 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
12464 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
12465 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
12466 if possible to compute val.
12467 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
12468 For QImode integers don't require anything about precision. Use
12469 const_with_all_bytes_same to find out if the constant doesn't have
12470 repeated bytes in it.
12472 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12474 * config/s390/s390.c (s390_expand_insv): Only accept insertions
12477 2013-04-08 Marek Polacek <polacek@redhat.com>
12479 PR rtl-optimization/48182
12480 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
12483 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12486 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
12487 nuses, make sure we have a label.
12489 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12492 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12493 (rs6000_emit_swdiv_low_precision): Remove.
12494 (rs6000_emit_swdiv): Rewrite to handle between one and four
12495 iterations of Newton-Raphson generally; modify required number of
12496 iterations for some cases.
12497 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12499 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12501 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12502 set-but-unused variable.
12504 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12505 basic blocks of released function bodies garbage-collectable.
12507 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12508 (struct rtl_opt_pass): Add TODO_df_finish.
12510 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12512 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12514 * config/arm/constraints.md (q): New constraint.
12515 * config/arm/ldrdstrd.md: New file.
12516 * config/arm/arm.md (ldrdstrd.md) New include.
12517 (arm_movdi): Use "q" instead of "r" constraint
12518 for double-word memory access.
12519 (movdf_soft_insn): Likewise.
12520 * config/arm/vfp.md (movdi_vfp): Likewise.
12521 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12522 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12523 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12524 (mem_ok_for_ldrd_strd): Likewise.
12525 (output_move_double): Update assertion.
12527 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12529 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12531 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12533 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12534 define_insn_and_split.
12535 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12537 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12539 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12540 define_insn_and_split.
12541 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12542 (shiftsi3_compare): New pattern.
12543 (rrx): New pattern.
12544 * config/arm/unspecs.md (UNSPEC_RRX): New.
12546 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12548 * config/arm/arm.md (negdi_extendsidi): New pattern.
12549 (negdi_zero_extendsidi): Likewise.
12551 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12553 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12554 define_insn_and_split.
12555 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12556 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12558 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12560 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12561 define_insn_and_split.
12562 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12563 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12565 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12567 * config/arm/arm.md (subsi3_carryin): New pattern.
12568 (subsi3_carryin_const): Likewise.
12569 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12570 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12572 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12574 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12576 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12578 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12579 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12581 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12583 * config/arm/arm.c (arm_expand_builtin): Change fcode
12584 type to unsigned int.
12586 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12588 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12590 2013-04-04 Ian Lance Taylor <iant@google.com>
12592 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12595 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12597 PR middle-end/56729
12598 * df-scan.c (df_insn_delete): Disable failing assert.
12600 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12602 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12603 New function prototype.
12604 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12605 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12606 (arm_builtin_vectorized_function): New function.
12608 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12610 * config/arm/arm_neon_builtins.def: New file.
12611 * config/arm/arm.c (neon_builtin_data): Move contents to
12612 arm_neon_builtins.def.
12613 (enum arm_builtins): Include neon builtin definitions.
12614 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12615 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12617 2013-04-04 Marek Polacek <polacek@redhat.com>
12619 PR tree-optimization/48186
12620 * predict.c (maybe_hot_frequency_p): Return false if
12621 HOT_BB_FREQUENCY_FRACTION is 0.
12622 (cgraph_maybe_hot_edge_p): Likewise.
12624 2013-04-04 Richard Biener <rguenther@suse.de>
12626 PR tree-optimization/56826
12627 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12630 2013-04-04 Richard Biener <rguenther@suse.de>
12632 PR tree-optimization/56213
12633 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12634 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12636 2013-04-04 Richard Biener <rguenther@suse.de>
12638 PR tree-optimization/56837
12639 * tree-loop-distribution.c (classify_partition): For non-zero
12640 values require that the value has the same precision as its
12641 mode to be useful as memset value.
12643 2013-04-03 Nick Clifton <nickc@redhat.com>
12645 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12646 (fmssf4): Use fmsf.s on E3V5 architectures.
12647 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12648 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12650 2013-04-03 Jeff Law <law@redhat.com>
12652 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12653 (lra-eliminations.o): Likewise.
12655 2013-04-03 Teresa Johnson <tejohnson@google.com>
12657 * gcov-io.c (compute_working_sets): Moved most of body of old
12658 compute_working_sets here from profile.c.
12659 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12660 (gcov_working_set_t): Moved typedef here from basic-block.h
12661 (compute_working_set): Declare.
12662 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12663 (get_working_sets): Renamed from compute_working_set,
12664 replace most of body with call to new compute_working_sets.
12665 (get_exec_counts): Replace call to compute_working_sets
12666 to get_working_sets.
12667 * profile.h (get_working_sets): Renamed from compute_working_set.
12668 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12669 to get_working_sets.
12670 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12671 * gcov-dump.c (dump_working_sets): New function.
12673 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12675 * hwint.c (sext_hwi, zext_hwi): New functions.
12676 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12677 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12678 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12679 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12680 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12681 (sext_hwi, zext_hwi): New functions.
12683 2013-04-03 Jeff Law <law@redhat.com>
12685 PR tree-optimization/56799
12686 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12687 back test for widening conversion erroneously dropped in prior change.
12689 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12692 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12693 instead of next_real_insn.
12695 2013-04-03 Marek Polacek <polacek@redhat.com>
12698 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12700 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12703 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12705 (thumb1_output_casesi): Likewise.
12706 (thumb2_output_casesi): Likewise.
12708 2013-04-03 Richard Biener <rguenther@suse.de>
12710 PR tree-optimization/56817
12711 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12713 (tree_unroll_loops_completely_1): ... new function to manually
12714 walk the loop tree, properly defering outer loops of unrolled
12715 loops to later iterations.
12717 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12719 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12720 (vectorizable_load): Likewise.
12721 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12722 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12724 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12726 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12729 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12731 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12733 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12735 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12737 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12739 PR tree-optimization/56790
12740 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12743 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12745 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12747 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12748 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12751 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12754 * tree.h (force_folding_builtin_constant_p): New decl.
12755 * builtins.c (force_folding_builtin_constant_p): New variable.
12756 (fold_builtin_constant_p): Fold immediately also if
12757 force_folding_builtin_constant_p.
12759 2013-04-03 Richard Biener <rguenther@suse.de>
12761 PR tree-optimization/56812
12762 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12763 DRs of the same interleaving chain are independent.
12765 2013-04-02 Jason Merrill <jason@redhat.com>
12767 * gdbinit.in (pbb): Use debug fn.
12769 2013-04-02 Lawrence Crowl <crowl@google.com>
12771 * sese.h (struct ivtype_map_elt_s): Remove unused.
12772 (extern debug_ivtype_map): Remove unused.
12773 (extern eq_ivtype_map_elts): Remove unused.
12774 * sese.c (debug_ivtype_map): Removed unused.
12775 (debug_ivtype_map_1): Removed unused.
12776 (debug_ivtype_elt): Remove unused.
12777 (eq_ivtype_map_elts): Remove unused.
12780 2013-04-02 Kai Tietz <ktietz@redhat.com>
12783 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12784 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12785 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12787 (legitimize_pe_coff_symbol): Likewise.
12788 (is_imported_p): New helper-function.
12789 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12790 for Windows x64 targets.
12791 (ix86_expand_prologue): Optimize for pe-coff targets.
12792 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12793 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12794 medium/large code-model.
12795 (legitimize_pic_address): Likewise.
12796 (legitimize_tls_address): Likewise.
12797 (ix86_expand_call): Likewise.
12798 (x86_output_mi_thunk): Likewise.
12799 (get_dllimport_decl): Add new beimport argument.
12800 (construct_plt_address): Don't assert for x64 pe-coff targets.
12801 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12803 (SYMBOL_FLAG_STUBVAR): New macro.
12804 (SYMBOL_REF_STUBVAR_P): Likewise.
12805 * config/i386/winnt.c (stub_list): New structure.
12806 (stub_head): New local variable.
12807 (i386_pe_record_stub): New function.
12808 (i386_pe_file_end): Emit refptr-stubs.
12810 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12812 PR rtl-optimization/56745
12813 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12814 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12817 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12818 and both of them are MEM_REFs, just compare first argument for
12819 equality and attempt to deal even with differing offsets.
12822 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12823 of gimple_clobber_p to be MEM_REF.
12824 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12825 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12826 after gimplification.
12827 * asan.c (get_mem_ref_of_assignment): Don't instrument
12828 gimple_clobber_p stmts.
12829 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12830 gimple_clobber_p stmt if they have MEM_REF lhs and
12831 are dead because of another gimple_clobber_p stmt.
12832 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12833 gimple_clobber_p stmts like debug stmts.
12834 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12835 that refer to unused VAR_DECLs or uninitialized values.
12836 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12837 gimple_clobber_p stmts if they refer to removed parameters.
12838 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12841 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12843 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12844 using SWI48 mode attribute.
12846 2013-04-02 Wei Mi <wmi@google.com>
12848 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12849 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12850 *<rotate_insn><mode>3_mask in i386.md.
12852 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12854 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12856 2013-04-02 Richard Biener <rguenther@suse.de>
12858 PR tree-optimization/56778
12859 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12860 Runtime alias tests are not supported for gather loads.
12861 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12862 stmts referenced from SSA operands before updating SSA form.
12864 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12865 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12867 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12868 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12869 * config/arm/cortex-a53.md: New file.
12870 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12871 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12872 * config/arm/arm.c (arm_issue_rate): Likewise.
12873 * config/arm/arm-tune.md: Regenerate
12874 * config/arm/arm-tables.opt: Regenerate.
12875 * config/arm/arm-cores.def: Add cortex-a53.
12877 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12879 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12882 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12884 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12885 scalar load/store operations using B/H registers.
12886 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12888 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12890 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12892 * config/aarch64/aarch64.c
12893 (aarch64_simd_scalar_immediate_valid_for_move): New.
12894 * config/aarch64/aarch64-protos.h
12895 (aarch64_simd_scalar_immediate_valid_for_move): New.
12896 * config/aarch64/constraints.md (Dh, Dq): New.
12897 * config/aarch64/iterators.md (hq): New.
12899 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12901 * reorg.c (get_branch_condition): Deal with conditional returns.
12902 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12904 2013-04-01 Wei Mi <wmi@google.com>
12906 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12907 Truncate operand 2 using %b asm operand modifier.
12908 (*<shift_insn><mode>3_mask): Ditto.
12909 (*<rotate_insn><mode>3_mask): Ditto.
12911 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12913 PR middle-end/56798
12914 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12916 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12918 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12920 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12922 2013-03-30 Lawrence Crowl <crowl@google.com>
12924 * dse.c (clear_alias_sets): Remove never set.
12925 (disqualified_clear_alias_sets): Remove never set.
12926 (clear_alias_mode_pool): Remove never set.
12927 (dse_step0): Remove condition that is never true.
12928 (canon_address): Remove condition that is never true.
12929 (dse_step7): Remove condition that is never true.
12930 (rest_of_handle_dse): Remove condition that is never true.
12931 (rest_of_handle_dse::did_global): Remove never read from above.
12932 (dse_step2_spill): Remove never called from above.
12933 (dse_step5_spill): Remove never called from above.
12935 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12937 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12938 JUMP_TABLE_DATA changes.
12939 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12940 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12941 (Insns) <jump_table_data>: New entry.
12942 * doc/tm.texi: Regenerate.
12944 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12946 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12947 for table jump at the end of a basic block using tablejump_p.
12948 * targhooks.c (default_invalid_within_doloop): Likewise.
12949 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12950 target hook implementation that is identical to the default hook.
12951 (rs6000_invalid_within_doloop): Remove.
12953 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12954 unused variable from tablejump_p call.
12956 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12957 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12958 (INSN_DELETED_P): Likewise.
12959 (emit_jump_table_data): New prototype.
12960 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12961 after 4th as unused.
12962 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12963 * sched-vis.c (print_insn): Likewise.
12964 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12965 insn for compatibility with back ends that use next_active_insn to
12966 identify jump table data.
12967 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12968 (remove_insn): Likewise.
12969 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12971 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12972 (emit_jump_table_data): New function.
12974 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12975 basic block, a JUMP_TABLE_DATA never is.
12976 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12977 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12978 off from code handling real insns.
12979 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12980 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12981 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12982 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12983 is not a NONDEBUG_INSN_P.
12984 * ira-costs.c (scan_one_insn): Likewise.
12985 * jump.c (mark_all_labels): Likewise.
12986 (mark_jump_label_1): Likewise.
12987 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12988 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12990 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12991 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12992 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12993 JUMP_TABLE_DATA_P insns.
12994 (calculate_elim_costs_all_insns): Likewise.
12995 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12996 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12997 (delete_output_reload): Code style fixups.
12998 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12999 insn flags on this non-insn.
13000 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13001 as scheduling barriers, for pre-change compatibility.
13002 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13003 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13005 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13006 redundant JUMP_TABLE_DATA_P test.
13007 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13008 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13009 (frv_for_each_packet): Likewise.
13010 * config/i386/i386.c (min_insn_size): Likewise.
13011 (ix86_avoid_jump_mispredicts): Likewise.
13012 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13013 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13014 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13015 (mips16_insn_length): Robustify.
13016 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13017 (mips16_split_long_branches): Likewise.
13018 * config/pa/pa.c (pa_combine_instructions): Likewise.
13019 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13020 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13021 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13022 as contributing to pool range lengths.
13023 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13024 Remove redundant JUMP_TABLE_DATA_P test.
13025 (sh_loop_align): Likewise.
13026 (split_branches): Likewise.
13027 (sh_insn_length_adjustment): Likewise.
13028 * config/spu/spu.c (get_branch_target): Likewise.
13030 2013-03-29 Jan Hubicka <jh@suse.cz>
13032 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13033 gcov streaming; stream hot bb threshold to ltrans.
13034 * predict.c (get_hot_bb_threshold): Break out from ....
13035 (maybe_hot_count_p): ... here.
13036 (set_hot_bb_threshold): New function.
13037 * lto-section-in.c (lto_section_name): Add profile.
13038 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13039 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13040 and data-streamer.h
13041 (histogram_entry): New structure.
13042 (histogram, histogram_pool): New global vars.
13043 (histogram_hash): New structure.
13044 (histogram_hash::hash): New method.
13045 (histogram_hash::equal): Likewise.
13046 (account_time_size): New function.
13047 (cmp_counts): New function.
13048 (dump_histogram): New function.
13049 (ipa_profile_generate_summary): New function.
13050 (ipa_profile_write_summary): New function.
13051 (ipa_profile_read_summary): New function.
13052 (ipa_profile): Decide on threshold.
13053 (pass_ipa_profile): Add ipa_profile_write_summary and
13054 ipa_profile_read_summary.
13055 * Makefile.in (ipa.o): Update dependencies.
13056 * lto-streamer.h (LTO_section_ipa_profile): New section.
13058 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13060 * tree.h (VAR_P): New.
13062 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13065 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13067 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13069 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13070 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13071 (control_flow_insn_p): Likewise.
13072 * cfgrtl.c (duplicate_insn_chain): Likewise.
13073 * final.c (get_attr_length_1): Likewise.
13074 (shorten_branches): Likewise.
13075 (final_scan_insn): Likewise.
13076 * function.c (instantiate_virtual_regs): Likewise.
13077 * gcse.c (insert_insn_end_basic_block): Likewise.
13078 * ira-costs.c (scan_one_insn): Likewise.
13079 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13080 * lra.c (check_rtl): Likewise.
13081 * reload1.c (elimination_costs_in_insn): Likewise.
13082 * reorg.c (follow_jumps): Likewise.
13084 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13085 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13086 (thumb_far_jump_used_p): Likewise.
13087 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13088 (workaround_speculation): Likewise.
13089 (add_sched_insns_for_speculation): Likewise.
13090 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13091 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13092 (frv_for_each_packet): Likewise.
13093 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13094 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13095 (final_emit_insn_group_barriers): Likewise.
13096 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13097 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13098 (mips16_insn_length): Likewise.
13099 * config/pa/pa.c (pa_reorg): Likewise.
13100 (pa_combine_instructions): Likewise.
13101 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13102 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13103 (sh_reorg): Likewise.
13104 (split_branches): Likewise.
13105 * config/spu/spu.c (get_branch_target): Likewise.
13107 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13110 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13112 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13113 Fix declaration name.
13115 2013-03-28 Lawrence Crowl <crowl@google.com>
13117 * graphds.h (struct graph.indicies): Remove unused.
13118 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13119 (SCOP_ORIGINAL_PDDRS): Remove unused.
13120 * sese.h (extern insert_loop_close_phis): Removed unused.
13121 (extern insert_guard_phis): Removed unused.
13122 (extern ivtype_map_elt_info): Removed unused.
13123 (new_ivtype_map_elt): Removed unused.
13124 * sese.c (ivtype_map_elt_info): Removed unused.
13126 2013-03-28 Lawrence Crowl <crowl@google.com>
13128 * Makefile.in: Add several missing include dependences.
13130 (test-dump.o): New. This object is not added to any executable,
13131 but is present for ad-hoc testing.
13133 (debug (const bitmap_head_def &)): New.
13134 (debug (const bitmap_head_def *)): New.
13136 (extern debug (const bitmap_head_def &)): New.
13137 (extern debug (const bitmap_head_def *)): New.
13139 (debug (edge_def &)): New.
13140 (debug (edge_def *)): New.
13142 (debug (basic_block_def &)): New.
13143 (debug (basic_block_def *)): New.
13145 (dump_node (const_tree, int, FILE *)): Correct source file.
13147 (debug (die_struct &)): New.
13148 (debug (die_struct *)): New.
13150 (extern debug (die_struct &)): New.
13151 (extern debug (die_struct *)): New.
13152 * gimple-pretty-print.c
13153 (debug (gimple_statement_d &)): New.
13154 (debug (gimple_statement_d *)): New.
13155 * gimple-pretty-print.h
13156 (extern debug (gimple_statement_d &)): New.
13157 (extern debug (gimple_statement_d *)): New.
13159 (debug (ira_allocno_copy &)): New.
13160 (debug (ira_allocno_copy *)): New.
13161 (debug (ira_allocno &)): New.
13162 (debug (ira_allocno *)): New.
13164 (extern debug (ira_allocno_copy &)): New.
13165 (extern debug (ira_allocno_copy *)): New.
13166 (extern debug (ira_allocno &)): New.
13167 (extern debug (ira_allocno *)): New.
13169 (debug (live_range &)): New.
13170 (debug (live_range *)): New.
13172 (debug (lra_live_range &)): New.
13173 (debug (lra_live_range *)): New.
13175 (debug (lra_live_range &)): New.
13176 (debug (lra_live_range *)): New.
13178 (debug (omega_pb_d &)): New.
13179 (debug (omega_pb_d *)): New.
13181 (extern debug (omega_pb_d &)): New.
13182 (extern debug (omega_pb_d *)): New.
13184 (debug (const rtx_def &)): New.
13185 (debug (const rtx_def *)): New.
13187 (debug_tree (tree): Move within file.
13188 (debug_raw (const tree_node &)): New.
13189 (debug_raw (const tree_node *)): New.
13190 (dump_tree_via_hooks (const tree_node *, int)): New.
13191 (debug (const tree_node &)): New.
13192 (debug (const tree_node *)): New.
13193 (debug_verbose (const tree_node &)): New.
13194 (debug_verbose (const tree_node *)): New.
13195 (debug_head (const tree_node &)): New.
13196 (debug_head (const tree_node *)): New.
13197 (debug_body (const tree_node &)): New.
13198 (debug_body (const tree_node *)): New.
13199 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13200 (debug (vec<tree, va_gc> &)): New.
13201 (debug (vec<tree, va_gc> *)): New.
13203 (extern debug (const rtx_def &)): New.
13204 (extern debug (const rtx_def *)): New.
13206 (debug_raw (simple_bitmap_def &)): New.
13207 (debug_raw (simple_bitmap_def *)): New.
13208 (debug (simple_bitmap_def &)): New.
13209 (debug (simple_bitmap_def *)): New.
13211 (extern debug (simple_bitmap_def &)): New.
13212 (extern debug (simple_bitmap_def *)): New.
13213 (extern debug_raw (simple_bitmap_def &)): New.
13214 (extern debug_raw (simple_bitmap_def *)): New.
13216 (debug (vinsn_def &)): New.
13217 (debug (vinsn_def *)): New.
13218 (debug_verbose (vinsn_def &)): New.
13219 (debug_verbose (vinsn_def *)): New.
13220 (debug (expr_def &)): New.
13221 (debug (expr_def *)): New.
13222 (debug_verbose (expr_def &)): New.
13223 (debug_verbose (expr_def *)): New.
13224 (debug (vec<rtx> &)): New.
13225 (debug (vec<rtx> *)): New.
13227 (extern debug (vinsn_def &)): New.
13228 (extern debug (vinsn_def *)): New.
13229 (extern debug_verbose (vinsn_def &)): New.
13230 (extern debug_verbose (vinsn_def *)): New.
13231 (extern debug (expr_def &)): New.
13232 (extern debug (expr_def *)): New.
13233 (extern debug_verbose (expr_def &)): New.
13234 (extern debug_verbose (expr_def *)): New.
13235 (extern debug (vec<rtx> &)): New.
13236 (extern debug (vec<rtx> *)): New.
13238 (_list_iter_cond_expr): Make inline instead of static.
13240 (debug (sreal &)): New.
13241 (debug (sreal *)): New.
13243 (extern debug (sreal &)): New.
13244 (extern debug (sreal *)): New.
13246 (extern debug_raw (const tree_node &)): New.
13247 (extern debug_raw (const tree_node *)): New.
13248 (extern debug (const tree_node &)): New.
13249 (extern debug (const tree_node *)): New.
13250 (extern debug_verbose (const tree_node &)): New.
13251 (extern debug_verbose (const tree_node *)): New.
13252 (extern debug_head (const tree_node &)): New.
13253 (extern debug_head (const tree_node *)): New.
13254 (extern debug_body (const tree_node &)): New.
13255 (extern debug_body (const tree_node *)): New.
13256 (extern debug (vec<tree, va_gc> &)): New.
13257 (extern debug (vec<tree, va_gc> *)): New.
13259 (debug (struct loop &)): New.
13260 (debug (struct loop *)): New.
13261 (debug_verbose (struct loop &)): New.
13262 (debug_verbose (struct loop *)): New.
13263 * tree-dump.c: Add header dependence.
13265 (extern debug (struct loop &)): New.
13266 (extern debug (struct loop *)): New.
13267 (extern debug_verbose (struct loop &)): New.
13268 (extern debug_verbose (struct loop *)): New.
13270 (debug (data_reference &)): New.
13271 (debug (data_reference *)): New.
13272 (debug (vec<data_reference_p> &)): New.
13273 (debug (vec<data_reference_p> *)): New.
13274 (debug (vec<ddr_p> &)): New.
13275 (debug (vec<ddr_p> *)): New.
13277 (extern debug (data_reference &)): New.
13278 (extern debug (data_reference *)): New.
13279 (extern debug (vec<data_reference_p> &)): New.
13280 (extern debug (vec<data_reference_p> *)): New.
13281 (extern debug (vec<ddr_p> &)): New.
13282 (extern debug (vec<ddr_p> *)): New.
13284 (debug (pt_solution &)): New.
13285 (debug (pt_solution *)): New.
13287 (extern debug (pt_solution &)): New.
13288 (extern debug (pt_solution *)): New.
13290 (debug (_var_map &)): New.
13291 (debug (_var_map *)): New.
13292 (debug (tree_live_info_d &)): New.
13293 (debug (tree_live_info_d *)): New.
13295 (extern debug (_var_map &)): New.
13296 (extern debug (_var_map *)): New.
13297 (extern debug (tree_live_info_d &)): New.
13298 (extern debug (tree_live_info_d *)): New.
13300 2013-03-28 Jan Hubicka <jh@suse.cz>
13302 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
13304 2013-03-28 Ian Bolton <ian.bolton@arm.com>
13306 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
13307 record only when desired or required.
13309 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
13311 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
13312 *vec_extractv2di_1_rex64. Use x64 isa attribute.
13314 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13316 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
13317 (*andsi3_compare0_uxtw): New pattern.
13318 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
13319 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
13321 2013-03-28 Jan Hubicka <jh@suse.cz>
13323 * data-streamer-in.c (streamer_read_gcov_count): New function.
13324 * gimple-streamer-out.c: Include value-prof.h.
13325 (output_gimple_stmt): Output histogram.
13326 (output_bb): Use streamer_write_gcov_count.
13327 * value-prof.c: Include data-streamer.h
13328 (dump_histogram_value): Add HIST_TYPE_MAX.
13329 (stream_out_histogram_value): New function.
13330 (stream_in_histogram_value): New function.
13331 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
13332 (stream_out_histogram_value, stream_in_histogram_value): Declare.
13333 * data-streamer-out.c (streamer_write_gcov_count): New function.
13334 (streamer_write_gcov_count_stream): New function.
13335 * lto-cgraph.c (lto_output_edge): Update counter streaming.
13336 (lto_output_node): Likewise.
13337 (input_node, input_edge): Likewise.
13338 * lto-streamer-out.c (output_cfg): Update streaming.
13339 * lto-streamer-in.c (input_cfg): Likewise.
13340 * data-streamer.h (streamer_write_gcov_count,
13341 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
13342 * gimple-streamer-in.c: Include value-prof.h
13343 (input_gimple_stmt): Input histograms.
13344 (input_bb): Update profile streaming.
13346 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13348 * genmodes.c (emit_max_int): New function.
13349 (emit_insn_modes_h): Added call to emit_max_function.
13350 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
13352 * machmode.def: Fixed comment.
13354 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13356 * combine.c (try_combine): Removed useless assert.
13357 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
13359 2013-03-28 Marek Polacek <polacek@redhat.com>
13360 Richard Biener <rguenther@suse.de>
13362 PR tree-optimization/56695
13363 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
13364 build signed result of a vector comparison.
13365 * tree-cfg.c (verify_gimple_comparison): Check that a result
13366 of a vector comparison has signed type.
13368 2013-03-28 Richard Biener <rguenther@suse.de>
13370 PR tree-optimization/37021
13371 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
13372 do not restrict gaps between groups.
13373 * tree-vect-stmts.c (vectorizable_load): Properly account for
13374 a gap between groups.
13376 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
13378 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
13379 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
13382 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
13384 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
13385 * doc/extend.texi (Named Address Spaces): Ditto.
13386 (Variable Attributes): Ditto.
13388 2013-03-27 Kai Tietz <ktietz@redhat.com>
13390 * config.build: Add support for cygwin x64 target.
13391 * config.gcc: Likewise.
13392 * config.host: Likewise.
13393 * configure.ac: Likewise
13394 * configure: Regenerated.
13396 2013-03-27 Kai Tietz <ktietz@redhat.com>
13398 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
13399 * config/i386/t-cygwin-w64: New file.
13400 * config/i386/cygwin-w64.h: New file.
13401 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
13402 and add support for x64-cygwin target.
13403 (CPP_SPEC): Likewise.
13404 (CXX_WRAP_SPEC_LIST): Undefine before define.
13405 (LIBGCJ_SONAME): Use 15 as version.
13407 2013-03-27 Richard Biener <rguenther@suse.de>
13409 PR tree-optimization/56716
13410 * tree-ssa-structalias.c (perform_var_substitution): Adjust
13411 dumping for ref nodes.
13413 2013-03-27 Martin Jambor <mjambor@suse.cz>
13415 PR tree-optimization/55334
13416 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
13417 restricted pointers to arrays.
13419 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
13421 * Makefile.in (.SUFFIXES): Add .cc.
13422 (.c.o): Apply same recipe for implicit rule .cc.o.
13424 2013-03-27 Richard Biener <rguenther@suse.de>
13426 PR tree-optimization/37021
13427 * tree-vect-data-refs.c (vect_check_strided_load): Allow
13428 REALPART/IMAGPART_EXPRs around the supported refs.
13429 * tree-ssa-structalias.c (find_func_aliases): Assume that
13430 floating-point values are not used to transfer pointers.
13432 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
13434 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
13435 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
13436 * gcc/doc/tm.texi: Regenerate.
13437 * gcc/targhooks.h (default_has_ifunc_p): New.
13438 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
13439 * gcc/config/linux-protos.h: New file.
13440 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
13441 this hook for linux which disables support of indirect functions in
13443 * gcc/config/linux-android.c: New file.
13444 * gcc/config/t-linux-android.c: Ditto.
13445 * gcc/config.gcc: Added new object file linux-android.o.
13446 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
13447 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
13448 * gcc/varasm.c (do_assemble_alias): Likewise.
13449 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
13450 doesn't support indirect functions.
13451 * configure: Regenerate.
13453 2013-03-27 Bin Cheng <bin.cheng@arm.com>
13456 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
13457 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
13460 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13462 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
13464 2013-03-27 Terry Guo <terry.guo@arm.com>
13466 * config/arm/arm-cores.def: Added core cortex-r7.
13467 * config/arm/arm-tune.md: Regenerated.
13468 * config/arm/arm-tables.opt: Regenerated.
13469 * doc/invoke.texi: Added entry for core cortex-r7.
13471 2013-03-27 Walter Lee <walt@tilera.com>
13473 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
13474 double-decrement of next_scratch_regno.
13476 2013-03-27 Walter Lee <walt@tilera.com>
13478 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
13480 (insn_v1mulus): Ditto.
13481 (insn_v2muls): Ditto.
13483 2013-03-27 Walter Lee <walt@tilera.com>
13485 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
13486 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
13488 2013-03-27 Walter Lee <walt@tilera.com>
13490 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13491 (*sibcall_value): Ditto.
13493 2013-03-27 Walter Lee <walt@tilera.com>
13495 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13496 (insn_mnz_v8qi): ... this ...
13497 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13499 (insn_v<n>mnz): Replaced by ...
13500 (insn_v1mnz): ... this ...
13501 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13503 (insn_mz_<mode>): Replaced by ...
13504 (insn_mz_v8qi): ... this ...
13505 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13507 (insn_v<n>mz): Replaced by ...
13508 (insn_v1mz): ... this ...
13509 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13512 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13514 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13516 2013-03-26 Roland McGrath <mcgrathr@google.com>
13518 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13519 than fprintf with a non-constant, non-format string.
13521 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13523 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13524 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13525 operand 0 predicate.
13526 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13527 attribute. Use general_x64nomem_operand as operand 1 predicate.
13528 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13529 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13530 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13531 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13532 (mov<mode>_insv_1): Remove expander. Merge insn with
13533 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13534 Use general_x64nomem_operand as operand 1 predicate.
13535 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13536 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13537 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13538 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13539 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13540 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13541 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13542 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13543 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13544 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13545 (general_x64nomem_operand): Ditto.
13547 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13549 * config/rtems.opt: Add -pthread option.
13551 2013-03-26 Richard Biener <rguenther@suse.de>
13553 * alias.c (find_base_term): Avoid redundant and not used recursion.
13554 (base_alias_check): Get the initial base term from the caller.
13555 (true_dependence_1): Compute and pass base terms to base_alias_check.
13556 (write_dependence_p): Likewise.
13557 (may_alias_p): Likewise.
13559 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13561 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13562 PC-relative load in SI modes and above only.
13564 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13566 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13567 * config/arm/iwmmxt.md (WCGR0): Update.
13568 (WCGR1, WCGR2, WCGR3): Likewise.
13570 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13572 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13573 Use x64 and nox64 isa attributes.
13575 2013-03-26 Richard Biener <rguenther@suse.de>
13577 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13578 alignment computations and rely on get_object_alignment_1
13579 for the !TYPE_P case.
13580 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13582 2013-03-26 Walter Lee <walt@tilera.com>
13584 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13585 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13587 2013-03-25 Jeff Law <law@redhat.com>
13589 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13590 check for INTEGRAL_TYPE_P that was missing due to checking in
13591 wrong version of prior patch.
13593 2013-03-25 Walter Lee <walt@tilera.com>
13595 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13596 TILEGX_INSN_SHUFFLEBYTES1.
13597 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13599 (tilegx_builtins): Ditto.
13600 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13602 2013-03-25 Walter Lee <walt@tilera.com>
13604 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13605 (floatunssisf2): New pattern.
13606 (floatsidf2): New pattern.
13607 (floatunssidf2): New pattern.
13609 2013-03-25 Walter Lee <walt@tilera.com>
13611 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13612 tests for constraint J, K, N, P.
13614 2013-03-25 Walter Lee <walt@tilera.com>
13616 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13617 Use indirect/pcrel encoding.
13618 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13621 2013-03-25 Steve Ellcey <sellcey@mips.com>
13623 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13624 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13625 * config/mips/mips.c (mips_option_override): Set IMADD default.
13626 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13627 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13628 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13629 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13630 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13632 2013-03-25 Jeff Law <law@redhat.com>
13634 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13635 slightly to avoid creating and folding useless trees. Simplify
13636 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13638 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13640 * config/i386/i386.md (*zero_extendsidi2): Merge with
13641 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13642 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13643 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13644 targets. Clarify comment.
13646 2013-03-25 Martin Jambor <mjambor@suse.cz>
13648 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13649 pass-through jump functions differently.
13650 (ipa_read_jump_function): Likewise. Also use setter functions to set
13653 2013-03-25 Martin Jambor <mjambor@suse.cz>
13655 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13656 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13658 (ipa_get_indirect_edge_target): New function.
13659 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13660 ipa_get_indirect_edge_target. Update all callers.
13661 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13662 ipa_get_indirect_edge_target_1 instead of calling
13663 ipa_get_indirect_edge_target.
13664 (create_specialized_node): Pass aggvlas to
13665 ipcp_discover_new_direct_edges.
13667 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13669 * config/arm/arm.md (f_sels, f_seld): New types.
13670 (*cmov<mode>): New pattern.
13671 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13674 2013-03-25 Kai Tietz <ktietz@redhat.com>
13676 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13677 POSIX-printf for mingw-hosted builds.
13679 2013-03-25 Richard Biener <rguenther@suse.de>
13681 PR middle-end/56694
13682 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13683 must-not-throw stmt location.
13685 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13687 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13688 Emit load-acquire versions when acq is true.
13689 (arm_emit_store_exclusive): Add rel parameter.
13690 Emit store-release versions when rel is true.
13691 (arm_split_compare_and_swap): Use acquire-release instructions
13693 of barriers when appropriate.
13694 (arm_split_atomic_op): Likewise.
13695 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13696 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13697 (VUNSPEC_SLX): Likewise.
13698 (VUNSPEC_LDA): Likewise.
13699 (VUNSPEC_STL): Likewise.
13700 * config/arm/sync.md (atomic_load<mode>): New pattern.
13701 (atomic_store<mode>): Likewise.
13702 (arm_load_acquire_exclusive<mode>): Likewise.
13703 (arm_load_acquire_exclusivesi): Likewise.
13704 (arm_load_acquire_exclusivedi): Likewise.
13705 (arm_store_release_exclusive<mode>): Likewise.
13707 2013-03-25 Catherine Moore <clm@codesourcery.com>
13709 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13710 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13711 * config/mip/predicates.md (lwsp_swsp_operand,
13712 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13713 sb16_operand, db4_operand, db7_operand, ib3_operand,
13714 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13715 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13716 andi16_operand): New predicates.
13717 * config/mips/mips.md (compression): New attribute.
13718 (enabled): New attribute.
13719 (length): Consider compression in computing length.
13720 (shift_compression): New code attribute.
13721 (*add<mode>3): New operands. Record compression.
13722 (sub<mode>3): Likewise.
13723 (one_cmpl<mode>2): Likewise.
13724 (*and<mode>3): Likewise.
13725 (*ior<mode>3): Likewise.
13726 (unnamed pattern for xor): Likewise.
13727 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13728 (*<optab><mode>3): Likewise.
13729 (*mov<mode>_internal: Likewise.
13730 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13731 (mips_unsigned_immediate_p): New.
13732 (umips_lwsp_swsp_address_p): New.
13733 (m16_based_address_p): New.
13734 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13735 (mips_unsigned_immediate_p): New prototype.
13736 (lwsp_swsp_address_p): New prototype.
13737 (m16_based_address_p): New prototype.
13738 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13739 (mips_signed_immediate_p): New function.
13740 (m16_based_address_p): New function.
13741 (lwsp_swsp_address_p): New function.
13742 (mips_print_operand_punctuation): Recognize short delay slot insns
13743 for microMIPS.add<mode>3"
13745 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13748 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13749 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13751 2013-03-25 Richard Biener <rguenther@suse.de>
13753 PR tree-optimization/56689
13754 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13757 2013-03-25 Richard Biener <rguenther@suse.de>
13759 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13761 (memory_references): Likewise.
13762 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13763 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13764 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13765 (gather_mem_refs_in_loops): Fold into ...
13766 (analyze_memory_references): ... this. Move initialization
13767 to tree_ssa_lim_initialize.
13768 (fill_always_executed_in): Rename to ...
13769 (fill_always_executed_in_1): ... this.
13770 (fill_always_executed_in): Move contains_call computation to
13771 this new function from ...
13772 (tree_ssa_lim_initialize): ... here.
13773 (tree_ssa_lim): Call fill_always_executed_in.
13775 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13777 * postreload.c (reload_combine): Fix code detecting returns.
13779 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13781 * function.c (emit_use_return_register_into_block): On cc0 targets,
13782 do not emit the sequence between cc0 setter and user.
13784 2013-03-25 Kai Tietz <ktietz@redhat.com>
13786 * config/i386/predicates.md (local_symbolic_operand): Interpret
13787 dll-imported symbols as none-local.
13789 2013-03-25 Richard Biener <rguenther@suse.de>
13791 * tree-ssa-loop-im.c (struct depend): Remove.
13792 (struct lim_aux_data): Make depends a vec of gimples.
13793 (free_lim_aux_data): Adjust.
13794 (add_dependency): Likewise.
13795 (set_level): Likewise.
13797 2013-03-25 Richard Biener <rguenther@suse.de>
13799 PR middle-end/56434
13800 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13801 the pointer returned by calls with ECF_MALLOC set.
13803 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13805 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13807 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13809 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13810 using MMXMODE mode iterator.
13811 (*move<mode>_internal): Merge with *movv2sf_internal and
13812 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13814 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13816 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13817 (record_last_mem_set_info): Likewise.
13819 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13820 of XNEWVEC followed by memset.
13821 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13823 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13825 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13826 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13827 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13828 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13829 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13830 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13831 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13832 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13833 BARRIER_P instead of GET_CODE.
13835 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13837 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13838 inaccuracy in the probing code.
13840 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13841 (ctrapdi4): Likewise.
13843 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13845 * calls.c (expand_call): Add missing guard to code handling return
13846 of non-BLKmode structures in MSB.
13847 * function.c (expand_function_end): Likewise.
13849 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13851 * combine.c (try_combine): Adjust comment. Do not add the set of
13852 insn #0 if the destination indirectly is set or dies in insn #2.
13853 Tidy up code to distribute a new note.
13855 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13857 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13858 also for alternatives 16 and 17.
13860 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13862 * config/i386/sse.md (*mov<mode>_internal): Merge with
13863 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13864 Emit insn template depending on type attribute. Use
13865 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13866 movd instead of movq mnemonic for interunit moves. Rewrite mode
13867 attribute calculation. Remove unit attribute calculation.
13868 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13869 Set prefix_data16 attribute for DImode ssemov types.
13870 Use Ym instead of y for SSE-MMX conversion alternatives.
13871 Reorder operand constraints.
13873 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13875 * df.h (df_insn_delete): Adjust prototype.
13876 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13877 and let it decide whether mark the basic block dirty.
13878 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13879 * df-scan.c (df_insn_info_delete): New helper function, split
13880 off from df_insn_delete.
13881 (df_scan_free_bb_info): Use it.
13882 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13884 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13885 that the insn is actually an insn and it has a non-NULL basic block.
13886 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13888 2013-03-22 Richard Biener <rguenther@suse.de>
13890 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13892 (mem_ref_alloc): Do not allocate them.
13893 (refs_independent_p): Do not query or maintain a cache.
13895 2013-03-22 Richard Biener <rguenther@suse.de>
13897 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13898 (gather_mem_refs_in_loops): Do not compute it.
13899 (analyze_memory_references): Do not allocate it.
13900 (tree_ssa_lim_finalize): Do not free it.
13901 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13903 2013-03-22 Richard Biener <rguenther@suse.de>
13905 * is-a.h (as_a): Use gcc_checking_assert.
13907 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13909 * config/aarch64/aarch64.c (aarch64_print_operand): New
13910 format specifier for printing a constant in hex.
13911 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13912 format specifier for printing second operand.
13914 2013-03-22 Richard Biener <rguenther@suse.de>
13916 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13918 (gather_mem_refs_in_loops): Perform store accumulation here.
13919 (create_vop_ref_mapping_loop): Remove.
13920 (create_vop_ref_mapping): Likewise.
13921 (analyze_memory_references): Initialize refs_stored_in_loop.
13922 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13923 (record_indep_loop): Remove.
13924 (record_dep_loop): New function.
13925 (ref_indep_loop_p_1): Adjust to only walk over references
13926 in the loop, not its subloops.
13927 (ref_indep_loop_p): Rename to ...
13928 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13929 maintaining a more fine-grained cache.
13930 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13931 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13933 2013-03-22 Richard Biener <rguenther@suse.de>
13935 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13936 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13937 aggregate mem_ref_loc.
13938 (free_mem_ref_locs): Inline into ...
13939 (memref_free): ... this and adjust.
13940 (mem_ref_alloc): Adjust.
13941 (mem_ref_locs_alloc): Remove.
13942 (record_mem_ref_loc): Adjust.
13943 (get_all_locs_in_loop): Rewrite into ...
13944 (for_all_locs_in_loop): ... this iterator.
13945 (rewrite_mem_ref_loc): New functor.
13946 (rewrite_mem_refs): Use for_all_locs_in_loop.
13947 (sm_set_flag_if_changed): New functor.
13948 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13949 (ref_always_accessed): New functor.
13950 (ref_always_accessed_p): Use for_all_locs_in_loop.
13952 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13954 * tree-pass.h (PROP_gimple_lvec): New.
13955 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13956 (init_optimization_passes): Move pass_lower_vector.
13957 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13959 (pass_lower_vector): Provide PROP_gimple_lvec.
13960 (pass_lower_vector_ssa): Likewise.
13961 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13963 2013-03-21 Mark Wielaard <mjw@redhat.com>
13965 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13967 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13969 * config/i386/i386.md (*movdi_internal): Disparage slightly
13970 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13971 conversion alternatives.
13973 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13975 PR middle-end/48087
13976 * diagnostic.def (DK_WERROR): New kind.
13977 * diagnostic.h (werrorcount): Define.
13978 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13979 promoted to DK_ERROR, increment DK_WERROR counter instead of
13981 * toplev.c (toplev_main): Call print_ignored_options even if
13982 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13983 even if just werrorcount is non-zero.
13986 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13988 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13989 (string_cst_pool_decl): New function.
13990 (optimize_one_addr_into_implicit_ptr): New function.
13991 (resolve_addr_in_expr): Optimize DWARF location expression
13992 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13993 which doesn't live in memory, but has DW_AT_location or
13994 DW_AT_const_value, or refers to a string literal, into
13995 DW_OP_GNU_implicit_pointer.
13996 (optimize_location_into_implicit_ptr): New function.
13997 (resolve_addr): If removing DW_AT_location of a variable because
13998 it was DW_OP_addr of address of the variable, but the variable doesn't
13999 live in memory, try to emit const value attribute for the initializer.
14001 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14003 * tree.h (VECTOR_TYPE_P): New macro.
14004 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14005 TYPE_MODE): Use it.
14006 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14007 VEC_COND_EXPR cannot be lvalues.
14008 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14010 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14012 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14013 Restrict the transformation to equal modes.
14015 2013-03-21 Richard Biener <rguenther@suse.de>
14017 PR tree-optimization/39326
14018 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14019 (MEM_ANALYZABLE): Adjust.
14020 (record_mem_ref_loc): Move bitmap ops ...
14021 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14022 unanalyzable refs, do not record locations for it.
14023 (analyze_memory_references): Allocate ref zero as shared
14025 (refs_independent_p): Do not test for unanalyzed mems here.
14026 (ref_indep_loop_p_1): Special-case disambiguation against
14027 the unanalyzed ref.
14028 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14030 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14032 * config/arm/arm-protos.h (tune_params): Add
14033 prefer_neon_for_64bits field.
14034 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14035 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14036 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14037 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14038 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14039 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14040 (arm_option_override): Handle -mneon-for-64bits new option.
14041 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14042 (prefer_neon_for_64bits): Declare new variable.
14043 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14044 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14045 (arch_enabled): Handle new arch types. Remove support for onlya8
14047 (one_cmpldi2): Use new arch names.
14048 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14049 * config/arm/arm.opt (mneon-for-64bits): Add option.
14050 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14051 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14052 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14054 * doc/invoke.texi (-mneon-for-64bits): Document.
14056 2013-03-21 Richard Biener <rguenther@suse.de>
14058 PR tree-optimization/39326
14059 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14060 (sort_bbs_in_loop_postorder_cmp): New function.
14061 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14063 2013-03-21 Richard Biener <rguenther@suse.de>
14065 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14066 (vect_insert_into_interleaving_chain): Likewise.
14067 (vect_drs_dependent_in_basic_block): Inline ...
14068 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14070 (vect_analyze_data_ref_dependence): ... here. Simplify.
14071 (vect_check_interleaving): Simplify.
14072 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14073 (vect_slp_analyze_data_ref_dependences): ... this new function.
14074 (dr_group_sort_cmp): New function.
14075 (vect_analyze_data_ref_accesses): Compute data-reference groups
14076 here instead of in vect_analyze_data_ref_dependence. Use
14077 a more efficient algorithm.
14078 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14079 vect_slp_analyze_data_ref_dependences. Call
14080 vect_analyze_data_ref_accesses earlier.
14081 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14082 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14083 (vect_slp_analyze_data_ref_dependences): New prototype.
14085 2013-03-21 Richard Biener <rguenther@suse.de>
14087 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14088 ref is stored in the loop.
14089 (find_refs_for_sm): Walk only over all stores.
14090 (store_motion_loop): Allocate from lim_bitmap_obstack.
14091 (store_motion): Likewise.
14093 2013-03-21 Richard Biener <rguenther@suse.de>
14095 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14096 Update virtual SSA form.
14098 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14100 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14101 * configure: Regenerate.
14102 * config.in: Regenerate.
14103 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14104 if !HAVE_LD_EH_FRAME_CIEV3.
14106 2013-03-21 Richard Biener <rguenther@suse.de>
14108 * tree-cfg.c (verify_expr_no_block): New function.
14109 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14110 nor DECL_VALUE_EXPR have locations with associated blocks.
14111 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14112 (clear_unused_block_pointer): Remove code dealing with
14113 blocks in DECL_DEBUG_EXPR locations.
14115 2013-03-21 Richard Biener <rguenther@suse.de>
14117 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14118 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14119 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14120 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14121 instead of DECL_DEBUG_EXPR_IS_FROM.
14122 * gimplify.c (gimplify_modify_expr): Likewise.
14123 * tree-cfg.c (verify_expr_location_1): Likewise.
14124 * tree-complex.c (create_one_component_var): Likewise.
14125 * tree-sra.c (create_access_replacement): Likewise.
14126 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14127 (clear_unused_block_pointer): Likewise.
14128 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14129 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14130 * var-tracking.c (var_debug_decl): Likewise.
14131 (track_expr_p): Likewise.
14132 * tree-inline.c (add_local_variables): Likewise. Set
14133 DECL_HAS_DEBUG_EXPR_P after copying it.
14134 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14135 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14137 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14140 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14141 * configure: Regenerate.
14142 * config.in: Regenerate.
14143 * config/i386/i386.md (*movdf_internal): Use
14144 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14145 movd instead of movq mnemonic for interunit moves.
14146 (*movdi_internal): Ditto.
14148 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14150 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14151 (abd<mode>_3): New pattern.
14152 (aba<mode>_3): New pattern.
14153 (fabd<mode>_3): New pattern.
14155 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14157 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14158 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14159 occurrence of REGISTER_PREFIX as its empty string.
14161 2013-03-20 Jeff Law <law@redhat.com>
14163 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14164 addititional equivalences for equality comparisons between an SSA_NAME
14165 and a constant where the SSA_NAME was set from a widening conversion.
14167 2013-03-20 Walter Lee <walt@tilera.com>
14169 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14171 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14173 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14174 depending on type attribute.
14175 (*movti_internal): Ditto.
14176 (*movtf_internal): Ditto.
14177 (*movxf_internal): Ditto.
14178 (*movdf_internal): Ditto.
14179 (*movsf_internal): Ditto.
14181 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14183 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14184 maybe_vex for sselog1 and ssemov types.
14185 (*movdi_internal): Reorder operand constraints.
14186 (*movsi_internal): Ditto. Set prefix attribute to
14187 maybe_vex for sselog1 and ssemov types.
14188 (*movtf_internal): Set prefix attribute to maybe_vex
14189 for sselog1 and ssemov types.
14190 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14191 DImode ssemov types. Reorder operand constraints.
14192 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14193 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14194 attribute for SImode ssemov types. Reorder operand constraints.
14196 2013-03-20 Martin Jambor <mjambor@suse.cz>
14198 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14199 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14201 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14203 * config/rs6000/predicates.md (indexed_address, update_address_mem
14204 update_indexed_address_mem): New predicates.
14205 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14206 attribute for load/store instructions.
14207 * config/rs6000/dfp.md (movsd_store): Likewise.
14208 (movsd_load): Likewise.
14209 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14210 (unnamed HI->DI extend define_insn): Likewise.
14211 (unnamed SI->DI extend define_insn): Likewise.
14212 (unnamed QI->SI extend define_insn): Likewise.
14213 (unnamed QI->HI extend define_insn): Likewise.
14214 (unnamed HI->SI extend define_insn): Likewise.
14215 (unnamed HI->SI extend define_insn): Likewise.
14216 (extendsfdf2_fpr): Likewise.
14217 (movsi_internal1): Likewise.
14218 (movsi_internal1_single): Likewise.
14219 (movhi_internal): Likewise.
14220 (movqi_internal): Likewise.
14221 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14222 attribute for load/store instructions.
14223 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
14225 (mov<mode>_softfloat): Likewise.
14226 (mov<mode>_hardfloat32): Likewise.
14227 (mov<mode>_hardfloat64): Likewise.
14228 (mov<mode>_softfloat64): Likewise.
14229 (movdi_internal32): Likewise.
14230 (movdi_internal64): Likewise.
14231 (probe_stack_<mode>): Likewise.
14233 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
14235 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
14236 floating point, and decimal floating point to reload iterator.
14238 * config/rs6000/constraints.md (wl constraint): New constraints to
14239 return FLOAT_REGS if certain options are used to reduce the number
14240 of separate patterns that exist in the file.
14241 (wx constraint): Likewise.
14242 (wz constraint): Likewise.
14244 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14245 -mdebug=reg, print wg, wl, wx, and wz constraints.
14246 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14247 Initialize the reload functions for 64-bit binary/decimal floating
14249 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14250 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14251 create the buffer on the stack to overcome not having a 32-bit
14253 (rs6000_emit_move): Likewise.
14254 (rs6000_secondary_memory_needed_rtx): Likewise.
14255 (rs6000_alloc_sdmode_stack_slot): Likewise.
14256 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14257 via xxlxor, just like DFmode 0.0.
14259 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
14260 define as 1 if we are running on a power7 or newer.
14261 (enum r6000_reg_class_enum): Add new constraints.
14263 * config/rs6000/dfp.md (movsd): Delete, combine with binary
14264 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14265 with other moves by using conditional constraits (wg). Use LFIWZX
14266 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
14267 (movsd splitter): Likewise.
14268 (movsd_hardfloat): Likewise.
14269 (movsd_softfloat): Likewise.
14271 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
14272 binary and decimal floating point moves.
14273 (fmove_ok): New attributes to combine binary and decimal floating
14274 point moves, and to combine power6x (mfpgpr) moves along normal
14276 (real_value_to_target): Likewise.
14277 (f32_lr): Likewise.
14278 (f32_lm): Likewise.
14279 (f32_li): Likewise.
14280 (f32_sr): Likewise.
14281 (f32_sm): Likewise.
14282 (f32_si): Likewise.
14283 (movsf): Combine binary and decimal floating point moves. Combine
14284 power6x (mfpgpr) moves with other moves by using conditional
14285 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
14286 (mov<mode> for SFmode/SDmode); Likewise.
14287 (SFmode/SDmode splitters): Likewise.
14288 (movsf_hardfloat): Likewise.
14289 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
14290 (movsf_softfloat): Likewise.
14291 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
14293 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
14294 wx and wz constraints.
14296 * config/rs6000/constraints.md (wg constraint): New constraint to
14297 return FLOAT_REGS if -mmfpgpr (power6x) was used.
14299 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
14302 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14303 -mdebug=reg, print wg, wl, wx, and wz constraints.
14304 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14305 Initialize the reload functions for 64-bit binary/decimal floating
14307 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14308 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14309 create the buffer on the stack to overcome not having a 32-bit
14311 (rs6000_emit_move): Likewise.
14312 (rs6000_secondary_memory_needed_rtx): Likewise.
14313 (rs6000_alloc_sdmode_stack_slot): Likewise.
14314 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14315 via xxlxor, just like DFmode 0.0.
14317 * config/rs6000/dfp.md (movdd): Delete, combine with binary
14318 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14319 with other moves by using conditional constraits (wg). Use LFIWZX
14320 and STFIWX for loading SDmode on power7.
14321 (movdd splitters): Likewise.
14322 (movdd_hardfloat32): Likewise.
14323 (movdd_softfloat32): Likewise.
14324 (movdd_hardfloat64_mfpgpr): Likewise.
14325 (movdd_hardfloat64): Likewise.
14326 (movdd_softfloat64): Likewise.
14328 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
14329 64-bit binary and decimal floating point moves.
14330 (FMOVE64X): Likewise.
14331 (movdf): Combine 64-bit binary and decimal floating point moves.
14332 Combine power6x (mfpgpr) moves with other moves by using
14333 conditional constraits (wg).
14334 (mov<mode> for DFmode/DDmode): Likewise.
14335 (DFmode/DDmode splitters): Likewise.
14336 (movdf_hardfloat32): Likewise.
14337 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
14338 (movdf_softfloat32): Likewise.
14339 (movdf_hardfloat64_mfpgpr): Likewise.
14340 (movdf_hardfloat64): Likewise.
14341 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
14342 (movdf_softfloat64): Likewise.
14343 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
14344 (reload_<mode>_load): Move to later in the file so they aren't in
14345 the middle of the floating point move insns.
14346 (reload_<mode>_store): Likewise.
14348 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
14351 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
14352 constraint if -mdebug=reg.
14353 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
14354 Enable using dd reload support if needed.
14356 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
14357 binary and decimal floating point moves in rs6000.md.
14358 (movtd_internal): Likewise.
14360 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
14361 decimal floating point moves.
14363 (movtf_internal): Likewise.
14364 (mov<mode>_internal, TDmode/TFmode): Likewise.
14365 (movtf_softfloat): Likewise.
14366 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
14368 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
14369 movdi_internal64, using wg constraint for move direct operations.
14370 (movdi_internal64): Likewise.
14372 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
14373 MODES_TIEABLE_P for selected modes. Print the numerical value of
14374 the various virtual registers. Use GPR/FPR first/last values,
14375 instead of hard coding the register numbers. Print which modes
14376 have reload functions registered.
14377 (rs6000_option_override_internal): If -mdebug=reg, trace the options
14378 settings before/after setting cpu, target and subtarget settings.
14379 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
14380 and for secondary reload failures in rs6000_secondary_reload_inner.
14381 (rs6000_secondary_reload_fail): Likewise.
14382 (rs6000_secondary_reload_inner): Likewise.
14384 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
14385 macros for first/last GPR and FPR registers.
14386 (LAST_GPR_REGNO): Likewise.
14387 (FIRST_FPR_REGNO): Likewise.
14388 (LAST_FPR_REGNO): Likewise.
14390 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
14391 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
14392 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
14393 (vcond<mode><mode>): Likewise.
14394 (vcondu<mode><mode>): Likewise.
14395 (vector_gtu<mode>): Likewise.
14396 (vector_gte<mode>): Likewise.
14397 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
14398 to prevent the compiler from converting DImode operations to TImode.
14399 (ior<mode>3): Likewise.
14400 (and<mode>3): Likewise.
14401 (one_cmpl<mode>2): Likewise.
14402 (nor<mode>3): Likewise.
14403 (andc<mode>3): Likewise.
14405 * config/rs6000/constraints.md (wt constraint): New constraint
14406 that returns VSX_REGS if TImode is allowed in VSX registers.
14408 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
14409 constant under VSX.
14411 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
14412 similar to TImode, but it is restricted to being in the GPRs.
14414 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
14415 TImode to occupy a single VSX register.
14417 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
14418 -mvsx-timode for power7/power8.
14419 (power7 cpu): Likewise.
14420 (power8 cpu): Likewise.
14422 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
14423 sure that TFmode/TDmode take up two registers if they are ever
14424 allowed in the upper VSX registers.
14425 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
14427 (rs6000_init_hard_regno_mode_ok): Likewise.
14428 (rs6000_debug_reg_global): Add debugging for PTImode and wt
14429 constraint. Print if LRA is turned on.
14430 (rs6000_option_override_internal): Give an error if -mvsx-timode
14431 and VSX is not enabled.
14432 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
14433 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
14434 to reg+offset addressing. Use PTImode when checking offset
14435 addresses for validity.
14436 (reg_offset_addressing_ok_p): Likewise.
14437 (rs6000_legitimate_offset_address_p): Likewise.
14438 (rs6000_legitimize_address): Likewise.
14439 (rs6000_legitimize_reload_address): Likewise.
14440 (rs6000_legitimate_address_p): Likewise.
14441 (rs6000_eliminate_indexed_memrefs): Likewise.
14442 (rs6000_emit_move): Likewise.
14443 (rs6000_secondary_reload): Likewise.
14444 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
14445 reloads to fpr registers to continue to use reg+offset addressing,
14446 but 64-bit reloads to altivec registers need reg+reg addressing.
14447 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
14448 it. Treat LO_SUM like a PLUS operation.
14449 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
14450 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
14451 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
14452 registers to share a register with a smaller sized type, since VSX
14453 puts scalars in the upper 64-bits.
14454 (print_operand): Add support for PTImode.
14455 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
14456 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
14457 registers, but don't have arithmetic support.
14458 (rs6000_memory_move_cost): Add test for VSX.
14459 (rs6000_opt_masks): Add -mvsx-timode.
14461 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
14464 (VSr): Use wt constraint for TImode.
14465 (VSv): Drop TImode support.
14466 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
14467 (vsx_movti_64bit): Likewise.
14468 (vsx_movti_32bit): Likewise.
14469 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
14470 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
14471 one '?' on the appropriate output constraint. Do not allow TImode
14472 logical operations on 32-bit systems.
14473 (vsx_ior<mode>3): Likewise.
14474 (vsx_xor<mode>3): Likewise.
14475 (vsx_one_cmpl<mode>2): Likewise.
14476 (vsx_nor<mode>3): Likewise.
14477 (vsx_andc<mode>3): Likewise.
14478 (vsx_concat_<mode>): Likewise.
14479 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
14481 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
14482 OPTION_MASK_VSX_TIMODE.
14483 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
14484 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
14486 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
14487 (TI2 iterator): New iterator for TImode, PTImode.
14488 (wd mode attribute): Add values for vector types.
14489 (movti_string): Replace TI move operations with operations for TImode
14490 and PTImode. Add support for TImode being allowed in VSX registers.
14491 (mov<mode>_string, TImode/PTImode): Likewise.
14492 (movti_ppc64): Likewise.
14493 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14494 (TI mode splitters): Likewise.
14496 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14499 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14501 PR tree-optimization/56355
14502 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14503 Also handle integers with undefined overflow.
14505 2013-03-20 Catherine Moore <clm@codesourcery.com>
14506 Maciej W. Rozycki <macro@codesourcery.com>
14507 Tom de Vries <tom@codesourcery.com>
14508 Nathan Sidwell <nathan@codesourcery.com>
14509 Iain Sandoe <iain@codesourcery.com>
14510 Nathan Froyd <froydnj@codesourcery.com>
14511 Chao-ying Fu <fu@mips.com>
14513 * doc/extend.texi: (micromips, nomicromips, nocompression):
14514 Document new function attributes.
14515 * doc/invoke.texi (minterlink-compressed, mmicromips,
14516 m14k, m14ke, m14kec): Document new options.
14517 (minterlink-mips16): Update documentation.
14518 * doc/md.texi (ZC, ZD): Document new constraints.
14519 * configure.ac (gcc_cv_as_micromips): Check if linker
14520 supports the .set micromips directive.
14521 * configure: Regenerate.
14522 * config.in: Regenerate.
14523 * config/mips/mips-tables.opt: Regenerate.
14524 * config/mips/micromips.md: New file.
14525 * constraints.md (ZC, ZD): New constraints.
14526 * config/mips/predicates.md (movep_src_register): New predicate.
14527 (movep_src_operand): New predicate.
14528 (non_volatile_mem_operand): New predicate.
14529 * config/mips/mips.md (multimem): New type.
14530 (length): Differentiate between 17-bit and 18-bit branch offsets.
14531 (MOVEP1, MOVEP2): New mode iterator.
14532 (mov_<load>l): Use ZC constraint.
14533 (mov_<load>r): Likewise.
14534 (mov_<store>l): Likewise.
14535 (mov_<store>r): Likewise.
14536 (*branch_equality<mode>_inverted): Add microMIPS support.
14537 (*branch_equality<mode>): Likewise.
14538 (*jump_absolute): Likewise.
14539 (indirect_jump_<mode>): Likewise.
14540 (tablejump_<mode>): Likewise.
14541 (<optab>_internal): Likewise.
14542 (sibcall_internal): Likewise.
14543 (sibcall_value_internal): Likewise.
14544 (prefetch): Use constraint ZD.
14545 * config/mips/mips.opt (minterlink-compressed): New option.
14546 (minterlink-mips16): Now an alias for minterlink-compressed.
14547 (mmicromips): New option.
14548 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14549 (compare_and_swap_12): Likewise.
14550 (sync_add<mode>): Likewise.
14551 (sync_<optab>_12): Likewise.
14552 (sync_old_<optab>_12): Likewise.
14553 (sync_new_<optab>_12): Likewise.
14554 (sync_nand_12): Likewise.
14555 (sync_old_nand_12): Likewise.
14556 (sync_new_nand_12): Likewise.
14557 (sync_sub<mode>): Likewise.
14558 (sync_old_add<mode>): Likewise.
14559 (sync_old_sub<mode>): Likewise.
14560 (sync_new_add<mode>): Likewise.
14561 (sync_new_sub<mode>): Likewise.
14562 (sync_<optab><mode>): Likewise.
14563 (sync_old_<optab><mode>): Likewise.
14564 (sync_new_<optab><mode>): Likewise.
14565 (sync_nand<mode>): Likewise.
14566 (sync_old_nand<mode>): Likewise.
14567 (sync_new_nand<mode>): Likewise.
14568 (sync_lock_test_and_set<mode>): Likewise.
14569 (test_and_set_12): Likewise.
14570 (atomic_compare_and_swap<mode>): Likewise.
14571 (atomic_exchange<mode>_llsc): Likewise.
14572 (atomic_fetch_add<mode>_llsc): Likewise.
14573 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14574 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14575 (umips_save_restore_pattern_p): Likewise.
14576 (umips_load_store_pair_p): Likewise.
14577 (umips_output_load_store_pair): Likewise.
14578 (umips_movep_target_p): Likewise.
14579 (umips_12bit_offset_address_p): Likewise.
14580 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14581 (mips_base_mips16): Rename this...
14582 (mips_base_compression_flags): ...to this. Update all uses.
14583 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14584 (mips_mips16_decl_p): Delete.
14585 (mips_nomips16_decl_p): Delete.
14586 (mips_get_compress_on_flags): New function.
14587 (mips_get_compress_off_flags): New function.
14588 (mips_get_compress_mode): New function.
14589 (mips_get_compress_on_name): New function.
14590 (mips_get_compress_off_name): New function.
14591 (mips_insert_attributes): Support multiple compression types.
14592 (mips_merge_decl_attributes): Likewise.
14593 (umips_12bit_offset_address_p): New function.
14594 (mips_start_function_definition): Emit .set micromips directive.
14595 (mips_call_may_need_jalx_p): New function.
14596 (mips_function_ok_for_sibcall): Add microMIPS support.
14597 (mips_print_operand_punctuation): Support short delay slots and
14599 (umips_swm_mask, umips_swm_encoding): New.
14600 (umips_build_save_restore): New function.
14601 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14602 (was_mips16_p): Remove.
14603 (old_compression_mode): New.
14604 (mips_set_compression_mode): New function.
14605 (mips_set_current_function): Add microMIPS support.
14606 (mips_option_override): Likewise.
14607 (umips_save_restore_pattern_p): New function.
14608 (umips_output_save_restore): New function.
14609 (umips_load_store_pair_p_1): New function.
14610 (umips_load_store_pair_p): New function.
14611 (umips_output_load_store_pair_1): New function.
14612 (umips_output_load_store_pair): New function.
14613 (umips_movep_target_p) New function.
14614 (mips_prepare_pch_save): Add microMIPS support.
14615 * config/mips/mips.h (TARGET_COMPRESSION): New.
14616 (TARGET_CPU_CPP_BUILTINS): Update macro
14617 to use new compression flags and to support microMIPS.
14618 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14619 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14620 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14621 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14622 (ASM_SPEC): Support mmicromips and mno-micromips.
14623 (M16STORE_REG_P): New macro.
14624 (MIPS_CALL): Support TARGET_MICROMIPS.
14625 (MICROMIPS_J): New macro.
14626 (mips_base_mips16): Rename this...
14627 (mips_base_compression_flags): ...to this.
14628 (UMIPS_12BIT_OFFSET_P): New macro.
14629 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14630 (MULTILIB_DIRNAMES): Likewise.
14631 2013-03-20 Richard Biener <rguenther@suse.de>
14633 PR tree-optimization/56661
14634 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14635 the result does not have to be distinct.
14637 2013-03-20 Richard Biener <rguenther@suse.de>
14639 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14642 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14643 Steven Bosscher <steven@gcc.gnu.org>
14645 PR rtl-optimization/56605
14646 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14648 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14651 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14652 that require movd instead of movq.
14654 2013-03-20 Richard Biener <rguenther@suse.de>
14656 * tree-ssa-structalias.c (struct variable_info): Add pointer
14657 to the first field of an aggregate with sub-vars. Make
14658 this and the pointer to the next subfield its ID.
14659 (vi_next): New function.
14660 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14661 storedanything_id, integer_id): Increment by one.
14662 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14663 get_call_clobber_vi): Adjust.
14664 (solution_set_expand): Simplify and speedup.
14665 (solution_set_add): Inline into ...
14666 (set_union_with_increment): ... this. Adjust accordingly.
14667 (do_sd_constraint): Likewise.
14668 (do_ds_constraint): Likewise.
14669 (do_complex_constraint): Simplify.
14670 (build_pred_graph): Adjust.
14671 (solve_graph): Likewise. Simplify and speedup.
14672 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14673 get_constraint_for_component_ref, get_constraint_for_1,
14674 first_vi_for_offset, first_or_preceding_vi_for_offset,
14675 create_function_info_for, create_variable_info_for_1,
14676 create_variable_info_for, intra_create_variable_infos): Adjust.
14677 (init_base_vars): Push NULL for ID zero.
14678 (compute_points_to_sets): Adjust.
14680 2013-03-20 Richard Biener <rguenther@suse.de>
14682 * cfgloop.c (verify_loop_structure): Streamline and avoid
14683 ICEing on corrupt loop tree.
14684 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14687 2013-03-20 Richard Biener <rguenther@suse.de>
14689 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14690 check whether an SSA update is needed.
14692 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14694 * config/mips/constraints.md (T): Rename to...
14698 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14699 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14701 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14703 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14704 (*subsi3_carryin_uxtw): Likewise.
14706 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14708 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14709 (*rorsi3_insn_uxtw): Likewise.
14711 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14713 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14714 (*extrsi5_insn_uxtw): Likewise.
14716 2013-03-19 Richard Biener <rguenther@suse.de>
14718 PR tree-optimization/56273
14719 * passes.c (init_optimization_passes): Move second VRP after DOM.
14721 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14723 * config/i386/i386.md (*movti_internal): Merge from
14724 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14725 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14726 nox64 isa attributes.
14728 2013-03-18 Richard Biener <rguenther@suse.de>
14730 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14732 (merge_node_constraints): Likewise.
14733 (build_succ_graph): Likewise.
14734 (valid_graph_edge): Inline into single caller.
14735 (unify_nodes): Likewise. Use bitmap_set_bit return value
14737 (scc_visit): Fix formatting and variable use.
14738 (do_sd_constraint): Use gcc_checking_assert.
14739 (do_ds_constraint): Likewise.
14740 (do_complex_constraint): Likewise.
14741 (condense_visit): Likewise. Cleanup.
14742 (dump_pred_graph): New function.
14743 (perform_var_substitution): Dump the pred-graph before
14744 variable substitution.
14745 (find_equivalent_node): Use gcc_checking_assert.
14746 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14748 2013-03-18 Richard Biener <rguenther@suse.de>
14750 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14751 Remove cond_expr_stmt_list argument and do not gimplify the
14753 (vect_loop_versioning): Adjust.
14754 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14755 Cleanup to use less temporaries.
14756 (vect_create_data_ref_ptr): Cleanup.
14758 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14760 PR tree-optimization/56635
14761 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14762 require types_compatible_p types.
14764 2013-03-18 Nick Clifton <nickc@redhat.com>
14766 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14767 spurious backslash.
14769 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14770 Add missing line to comment describing function.
14772 2013-03-18 Richard Biener <rguenther@suse.de>
14774 PR tree-optimization/56210
14775 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14776 Handle string / character search functions.
14777 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14779 2013-03-18 Richard Biener <rguenther@suse.de>
14781 PR middle-end/56483
14782 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14783 and implement properly.
14784 * gimple.h (gimple_cond_single_var_p): Remove.
14786 2013-03-18 Richard Biener <rguenther@suse.de>
14788 * tree-data-ref.h (find_data_references_in_loop): Declare.
14789 * tree-data-ref.c (get_references_in_stmt): Use a stack
14790 vector pre-allocated in the callers.
14791 (find_data_references_in_stmt): Adjust.
14792 (graphite_find_data_references_in_stmt): Likewise.
14793 (create_rdg_vertices): Likewise.
14794 (find_data_references_in_loop): Export.
14795 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14796 Compute dependences here...
14797 (vect_analyze_data_refs): ...not here. When we encounter
14798 a non-vectorizable data reference in basic-block vectorization
14799 truncate the data reference vector. Do not bother to
14800 fixup data-dependence information for gather loads.
14801 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14802 of data references, as reported.
14804 2013-03-18 Richard Biener <rguenther@suse.de>
14806 PR tree-optimization/3713
14807 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14808 has_constants and expr.
14809 (stmt_has_constants): Properly valueize SSA names when deciding
14810 whether the stmt has constants.
14812 2013-03-18 Richard Biener <rguenther@suse.de>
14814 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14815 whole function when there is nothing to do.
14816 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14817 * tree-vectorizer.c (vectorize_loops): Update virtual and
14818 loop-closed SSA once.
14819 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14821 2013-03-18 Richard Biener <rguenther@suse.de>
14823 PR middle-end/56113
14824 * domwalk.c (bb_postorder): New global static.
14825 (cmp_bb_postorder): New function.
14826 (walk_dominator_tree): Replace scheme imposing an order for
14827 visiting dominator sons by one sorting them at the time they
14828 are pushed on the stack.
14830 2013-03-18 Richard Biener <rguenther@suse.de>
14832 PR tree-optimization/39326
14833 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14834 (struct mem_ref): Replace mem member with ao_ref typed member.
14835 (MEM_ANALYZABLE): Adjust.
14836 (memref_eq): Likewise.
14837 (mem_ref_alloc): Likewise.
14838 (gather_mem_refs_stmt): Likewise.
14839 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14840 (execute_sm_if_changed_flag_set): Adjust.
14841 (execute_sm): Likewise.
14842 (ref_always_accessed_p): Likewise.
14843 (refs_independent_p): Likewise.
14844 (can_sm_ref_p): Likewise.
14846 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14849 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14850 return 1 even for !unsignedp.
14852 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14854 * config/i386/i386.md (isa): Add x64 and nox64.
14855 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14856 (*pushtf): Enable *roF alternative for x64 isa only.
14857 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14858 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14859 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14860 (*movtf_internal): Merge from *movtf_internal_rex64 and
14861 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14862 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14863 nox64 isa attributes.
14864 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14865 nox64 isa attributes.
14866 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14868 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14870 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14872 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14875 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14876 class names. Remove trailing comma after "ALL_REGS".
14878 2013-03-16 Jan Hubicka <jh@suse.cz>
14880 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14881 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14882 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14883 of cgraph_get_create_node.
14884 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14886 2013-03-16 Jason Merrill <jason@redhat.com>
14889 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14890 with DW_AT_default_value.
14892 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14894 * BASE-VER: Set to 4.9.0.
14896 2013-03-14 Andi Kleen <ak@linux.intel.com>
14899 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14900 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14901 Document _x* TSX intrinsics.
14903 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14904 David Holsgrove <david.holsgrove@xilinx.com>
14906 * configure.ac: Add MicroBlaze TLS support detection.
14907 * configure: Regenerate.
14908 * config/microblaze/microblaze-protos.h
14909 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14910 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14911 * config/microblaze/microblaze.c (microblaze_address_type): Add
14912 ADDRESS_TLS and tls_reloc address types.
14913 (microblaze_address_info): Add tls_reloc.
14914 (TARGET_HAVE_TLS): Define.
14915 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14916 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14917 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14918 load_tls_operand, microblaze_call_tls_get_addr,
14919 microblaze_legitimize_tls_address): New functions.
14920 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14921 (get_base_reg): Use microblaze_tls_symbol_p.
14922 (microblaze_classify_address): Handle TLS.
14923 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14924 label_mentioned_p and microblaze_tls_referenced_p.
14925 (microblaze_legitimize_address): Handle TLS.
14926 (microblaze_address_insns): Handle ADDRESS_TLS.
14927 (pic_address_needs_scratch): Handle TLS.
14928 (print_operand_address): Handle TLS.
14929 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14930 (microblaze_expand_move): Handle TLS.
14931 (microblaze_legitimate_constant_p): Check
14932 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14933 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14934 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14935 (PIC_OFFSET_TABLE_REGNUM): Set.
14936 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14937 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14938 (addsi3, movsi_internal2, movdf_internal): Update constraints
14939 * config/microblaze/predicates.md (arith_plus_operand): Define
14940 (move_operand): Redefine as move_src_operand,
14941 check microblaze_tls_referenced_p.
14943 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14945 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14946 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14948 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14950 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14953 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14955 PR tree-optimization/53265
14956 * common.opt (Waggressive-loop-optimizations): New option.
14957 * tree-ssa-loop-niter.c: Include tree-pass.h.
14958 (do_warn_aggressive_loop_optimizations): New function.
14959 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14960 if number_of_latch_executions returned constant.
14961 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14962 early. If number_of_latch_executions returned constant, set
14963 nb_iterations_upper_bound back to it.
14964 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14966 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14967 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14969 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14970 (MULTILIB_OSDIRNAMES): Set.
14971 * genmultilib: If defaultosdirname doesn't start with :: , set
14972 defaultosdirname2 instead, clear it and emit two . multilib_raw
14973 entries instead of just one.
14975 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14977 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14978 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14979 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14980 (SUBTARGET_OVERRIDE_OPTIONS): New.
14982 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14985 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14986 (musermode): Convert to Var(TARGET_USERMODE).
14987 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14988 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14989 * config/sh/sh.c (sh_option_override): Use
14990 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14991 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14993 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14995 (udivsi3_i4_single, divsi3_i4_single): Use
14996 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14998 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15000 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15003 2013-03-13 Richard Biener <rguenther@suse.de>
15005 PR tree-optimization/56608
15006 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15007 calls when vectorizing basic-blocks.
15009 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15012 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15015 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15017 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15019 2013-03-11 Jan Hubicka <jh@suse.cz>
15022 * lto-streamer-out.c (output_symbol_p): Skip references from
15023 constructors of external variables.
15025 2013-03-11 Jan Hubicka <jh@suse.cz>
15027 PR middle-end/56571
15028 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15030 * emit-rtl.c (verify_rtx_sharing): Likewise.
15031 (copy_insn_1): Likewise.
15032 * rtl.c (copy_rtx): Likewise.
15034 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15037 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15038 output_operand_lossage message.
15040 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15043 * arm.c (shift_op): Validate RTL pattern on the fly.
15044 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15047 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15050 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15051 new call patterns that clobber %fr12.
15052 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15053 split and postreload patterns.
15054 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15055 registers %fr12 and %fr12R as call used.
15057 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15059 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15060 (canon_address, record_store, replace_read, check_mem_read_rtx,
15061 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15062 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15063 rest_of_handle_dse): Likewise.
15065 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15067 PR middle-end/56524
15068 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15070 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15071 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15072 (save_optabs_if_changed): Replace with...
15073 (init_tree_optimization_optabs): ...this.
15074 * optabs.c (save_optabs_if_changed): Rename to...
15075 (init_tree_optimization_optabs): ...this. Take the optimization node
15076 as argument. Do nothing if the base optabs are already correct.
15077 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15078 to recompute optabs.
15079 * function.h (function): Remove optabs field.
15080 * function.c (invoke_set_current_function_hook): Call
15081 init_tree_optimization_optabs. Use the result to initialize
15084 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15086 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15087 if GTMA_HAS_NO_INSTRUMENTATION.
15088 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15089 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15090 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15091 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15092 GTMA_HAS_NO_INSTRUMENTATION.
15094 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15096 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15099 2013-03-08 Marek Polacek <polacek@redhat.com>
15100 Jakub Jelinek <jakub@redhat.com>
15102 PR tree-optimization/56478
15103 * predict.c (is_comparison_with_loop_invariant_p): Change the
15104 type of loop_step to tree.
15105 (predict_loops): Adjust.
15106 (predict_iv_comparison): Perform the computations on double_ints.
15108 2013-03-08 Richard Biener <rguenther@suse.de>
15110 PR tree-optimization/56570
15111 * tree-cfg.c (verify_expr_location_1): Verify locations for
15113 * tree-sra.c (create_access_replacement): Strip locations
15114 from DECL_DEBUG_EXPRs.
15116 2013-03-08 Richard Biener <rguenther@suse.de>
15118 * tree-inline.c (expand_call_inline): Do not associate
15119 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15120 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15122 2013-03-08 Richard Biener <rguenther@suse.de>
15124 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15125 or block changes with -Og. Fix for location / block encoding
15126 changes and PHI arguments with locations.
15128 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15130 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15132 (struct output_info): Likewise.
15133 (register_overhead): Remove bad gcc_assert.
15134 (bitmap_find_bit): If there is only a single bitmap element, do not
15135 count a miss as a search.
15136 (print_statistics): Update for counter type changes.
15137 (dump_bitmap_statistics): Likewise. Print headers such that they
15138 are properly lined up with the printed counters.
15140 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15142 PR tree-optimization/56559
15143 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15144 check that it has only a single use.
15146 2013-03-07 Richard Biener <rguenther@suse.de>
15148 * doc/invoke.texi (fwhole-program): Discourage use in combination
15151 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15153 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15155 PR tree-optimization/56539
15156 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15157 instead of GSI_CONTINUE_LINKING as last argument to
15158 force_gimple_operand_gsi. Adjust function comment.
15160 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15163 PR middle-end/56548
15164 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15165 promoted mode, convert the result back to the original mode.
15167 2013-03-06 Richard Biener <rguenther@suse.de>
15169 PR middle-end/56294
15170 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15171 (insert_updated_phi_nodes_compare_uids): New function.
15172 (update_ssa): Sort symbols_to_rename after UID before
15173 traversing it to insert PHI nodes.
15175 2013-03-06 Richard Biener <rguenther@suse.de>
15177 PR middle-end/50494
15178 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15179 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15182 2013-02-13 Richard Biener <rguenther@suse.de>
15185 * varasm.c (output_constant_def_1): Get the decl representing
15186 the constant as argument.
15187 (output_constant_def): Wrap output_constant_def_1.
15188 (make_decl_rtl): Use output_constant_def_1 with the decl
15189 representing the constant.
15190 (build_constant_desc): Optionally re-use a decl already
15191 representing the constant.
15192 (tree_output_constant_def): Adjust.
15194 2013-03-06 Joey Ye <joey.ye@arm.com>
15197 * gcc.c (convert_white_space): New function.
15198 (main): Handles white space in function name.
15200 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15203 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15204 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15205 to SH_DIV_CALL_TABLE for TARGET_SH2.
15206 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15208 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15209 call-table options.
15211 2013-03-05 Sterling Augustine <saugustine@google.com>
15212 Cary Coutant <ccoutant@google.com>
15215 * dwarf2out.c (resolve_addr): Don't call
15216 remove_loc_list_addr_table_entries a second time for the same
15219 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15222 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15223 (avoid_complex_debug_insns): New function.
15224 (expand_debug_locations): Call it.
15226 PR rtl-optimization/56484
15227 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
15228 lifetimes of hard registers on small register class machines.
15230 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15232 * config/microblaze/microblaze-protos.h: Rename
15233 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
15234 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
15236 (microblaze_fast_interrupt_function_p): New function.
15237 (microblaze_is_interrupt_handler): Rename to
15238 microblaze_is_interrupt_variant and add fast_interrupt check.
15239 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
15240 (save_restore_insns): Likewise.
15241 (compute_frame_size): Likewise.
15242 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
15243 (microblaze_globalize_label): Likewise.
15244 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
15245 * config/microblaze/microblaze.md: Use wrapper
15246 microblaze_is_interrupt_variant.
15248 2013-03-05 Kai Tietz <ktietz@redhat.com>
15250 * sdbout.c (sdbout_one_type): Switch to current function's section
15251 supporting cold/hot.
15253 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15255 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
15258 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15260 PR middle-end/56461
15261 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
15262 if VALGRIND_GET_VBITS is defined, temporarily make object
15263 memory all defined, and restore previous valgrind addressability
15264 and definability afterwards. Free this_object at the end.
15266 PR middle-end/56461
15267 * lra.c (lra): Call lra_clear_live_ranges if live_p,
15268 right before calling lra_create_live_ranges, also call it
15269 when clearing live_p. Only call lra_clear_live_ranges
15270 at the end if live_p.
15272 PR middle-end/56461
15273 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
15275 2013-03-05 Richard Biener <rguenther@suse.de>
15277 PR tree-optimization/56521
15278 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
15281 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
15284 * except.h (remove_unreachable_eh_regions): New prototype.
15285 * except.c (remove_eh_handler_splicer): New function, split out
15286 of remove_eh_handler.
15287 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
15288 warning about running it on many EH regions one at a time.
15289 (remove_unreachable_eh_regions_worker): New function, walk the
15290 EH tree in depth-first order and remove non-marked regions.
15291 (remove_unreachable_eh_regions): New function.
15292 * tree-eh.c (mark_reachable_handlers): New function, split out
15293 from remove_unreachable_handlers.
15294 (remove_unreachable_handlers): Use mark_reachable_handlers and
15295 remove_unreachable_eh_regions.
15296 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
15297 and remove_unreachable_eh_regions.
15299 2013-03-05 Richard Biener <rguenther@suse.de>
15301 PR middle-end/56525
15302 * loop-init.c (fix_loop_structure): Remove loops in two stages,
15303 not freeing them until the end.
15305 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15307 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
15309 2013-03-05 Richard Biener <rguenther@suse.de>
15311 PR tree-optimization/56270
15312 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
15313 of loads after scheduling an SLP instance.
15315 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15317 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
15319 (check_gcc_parallelize): Run guality.exp as a separate job from
15320 vect.exp with unsorted.exp and $(dg_target_exps) separately from
15321 struct-layout-1.exp with stackalign.exp.
15323 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
15325 PR middle-end/56461
15326 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
15327 load_index sbitmap even if some bit in it isn't set.
15329 PR middle-end/56461
15330 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
15331 (discover_iteration_bound_by_body_walk): Change queues to
15332 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
15333 spelling in comment. Call safe_push on queues[bound_index] directly.
15334 Release queues[queue_index] in every iteration unconditionally.
15335 Release bounds vector.
15337 PR middle-end/56461
15338 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
15339 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
15340 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
15343 2013-03-05 Richard Biener <rguenther@suse.de>
15346 * tree-inline.c (remap_blocks_to_null): New function.
15347 (expand_call_inline): When expanding a call stmt without
15348 an associated block inline remap all callee blocks to NULL.
15350 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15352 PR rtl-optimization/56494
15353 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
15354 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
15355 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
15357 PR middle-end/56461
15358 * sel-sched-ir.c (free_sched_pools): Release
15359 succs_info_pool.stack[succs_info_pool.max_top] vectors too
15360 if succs_info_pool.max_top isn't -1.
15363 * opts.c (opts_obstack, opts_concat): Moved to...
15364 * opts-common.c (opts_obstack, opts_concat): ... here.
15366 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15368 PR middle-end/56461
15369 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
15371 2013-03-04 Martin Jambor <mjambor@suse.cz>
15373 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
15374 all appropriate places.
15376 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
15378 PR tree-optimization/56424
15379 * ipa-split.c (split_function): Do not set the RSO flag if result is
15380 not by reference and its type is a register type.
15382 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
15384 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
15385 (microblaze_legitimate_pic_operand): Likewise
15386 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
15387 new function microblaze_legitimate_pic_operand
15388 * config/microblaze/microblaze-protos.h
15389 (microblaze_legitimate_pic_operand): Declare.
15391 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15393 * config/microblaze/predicates.md (call_insn_simple_operand):
15394 New predicate for supported rtx code types.
15395 * config/microblaze/microblaze.md (call_internal1): Use
15396 call_insn_simple_operand predicate.
15398 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15400 PR middle-end/56461
15401 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
15402 partitions.ordered_remove.
15404 PR middle-end/56461
15405 * tree-vect-stmts.c (vectorizable_conversion): Don't call
15406 vec_oprnds0.create (1) for modifier == NONE.
15408 PR middle-end/56461
15409 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
15410 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
15411 vec_oprnds1 right before pushing anything to it for
15414 PR middle-end/56461
15415 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
15416 set nbbs to 0 instead of having separate code path.
15417 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
15418 instead of false as last argument if returning NULL.
15420 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
15422 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
15423 the attribute is now called "target" instead of "option".
15424 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
15425 * doc/tm.texi.in (Target Attributes): Likewise document the correct
15426 attribute/pragma name for TARGET_OPTION_VALID_P and
15427 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
15428 * doc/tm.texi: Regenerated.
15430 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
15432 * config/microblaze/microblaze.c:
15433 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
15434 * config/microblaze/microblaze.h: Add -mxl-reorder to
15436 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
15437 instructions emitted if TARGET_REORDER.
15438 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
15439 or 0 for -m/-mno case, but initialises as 2 to detect default use case
15442 2013-03-01 Xinliang David Li <davidxl@google.com>
15444 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
15447 2013-03-01 Jakub Jelinek <jakub@redhat.com>
15449 PR middle-end/56461
15450 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
15451 vector even when returning true. Fix up function comment formatting.
15453 PR middle-end/56461
15454 * ira-build.c (ira_loop_nodes_count): New variable.
15455 (create_loop_tree_nodes): Initialize it.
15456 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
15458 PR middle-end/56461
15459 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
15460 method on dr_chain and result_chain.
15461 * tree-vect-stmts.c (vectorizable_store): Only call
15462 result_chain.create if j == 0.
15464 PR middle-end/56461
15465 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
15466 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
15467 before overwriting it.
15469 2013-03-01 Tobias Burnus <burnus@net-b.de>
15471 * doc/extended.texi (C Extensions): Change order in @menu
15473 (Other MIPS Built-in Functions): Move last MIPS entry before
15474 "picoChip Built-in Functions".
15475 (SH Built-in Functions): Move after RX Built-in Functions.
15476 * doc/gcc.texi (Introduction): Change order in @menu
15478 * doc/md.texi (Constraints): Ditto.
15479 * gty.texi (Type Information): Ditto.
15480 (User-provided marking routines for template types): Make
15482 * doc/invoke.texi (AArch64 Options): Move before
15483 "Adapteva Epiphany Options".
15485 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
15486 Jakub Jelinek <jakub@redhat.com>
15489 * asan.c (gate_asan): Lookup no_sanitize_address instead of
15490 no_address_safety_analysis attribute.
15491 * doc/extend.texi (no_address_safety_attribute): Rename to
15492 no_sanitize_address attribute, mention no_address_safety_analysis
15493 attribute as deprecated alias.
15495 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15497 PR middle-end/56461
15498 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15499 type to vec<vec<tree> > *.
15500 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15501 to be vec<tree> instead of vec<tree> *, set vec_defs
15502 to vNULL and call vec_defs.create (number_of_vects), adjust other
15504 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15505 vectorizable_condition): Adjust vect_get_slp_defs callers.
15507 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15509 * config/aarch64/aarch64.c
15510 (aarch64_float_const_representable): Remove unused variable.
15512 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15514 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15516 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15518 * config/aarch64/aarch64-builtins.c
15519 (aarch64_init_simd_builtins): Make static.
15521 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15523 * config/aarch64/aarch64.c
15524 (aarch64_simd_make_constant): Make static.
15526 2013-02-28 Martin Jambor <mjambor@suse.cz>
15528 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15529 with no initialization to the RHS of debug statements.
15531 2013-02-28 Martin Jambor <mjambor@suse.cz>
15533 PR tree-optimization/56294
15534 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15536 (get_access_replacement): Do not call create_access_replacement.
15537 Assert a replacement exists.
15538 (get_repl_default_def_ssa_name): Create the replacement declaration
15541 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15543 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15544 final_end_function.
15546 2013-02-28 Marek Polacek <polacek@redhat.com>
15548 PR rtl-optimization/56466
15549 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15550 if we're changing a loop.
15551 (peel_loops_completely): Likewise.
15553 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15556 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15558 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15561 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15562 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15563 INTX_FTYPE_FX, FX_FTYPE_INTX.
15564 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15566 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15568 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15569 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15570 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15571 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15572 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15573 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15574 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15575 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15576 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15577 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15578 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15579 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15580 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15581 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15582 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15583 (avrxmega6): Increase max flash segments from 5 to 6.
15584 * config/avr/t-multilib: Regenerate.
15585 * config/avr/avr-tables.opt: Regenerate.
15586 * doc/avr-mmcu.texi: Regenerate.
15588 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15590 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15591 (avr_device_to_arch): Rename to avr_device_to_ld.
15592 (avr_device_to_as): New prototype.
15593 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15594 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15595 * config/avr/driver-avr.c (avr_device_to_as): New.
15596 (avr_device_to_arch): Rename to avr_device_to_ld.
15598 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15600 PR middle-end/56461
15601 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15602 method on dr_chain and result_chain.
15604 PR middle-end/56461
15605 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15606 pointer_set_destroy on not_executed_last_iteration.
15608 PR middle-end/56461
15609 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15611 PR middle-end/56461
15612 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15613 FOR_EACH_DEFINED_FUNCTION when freeing state.
15615 PR middle-end/56461
15616 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15618 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15621 PR middle-end/56461
15622 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15623 known_aggs[i].items and release known_aggs vector.
15625 PR middle-end/56461
15626 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15628 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15630 * config/microblaze/microblaze.c (microblaze_emit_compare):
15631 Use xor for EQ/NE comparisions.
15632 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15633 (cbranchsf4): Adjust operator to comparison_operator.
15635 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15637 PR middle-end/56461
15638 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15640 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15641 vec_safe_push, always update *slot.
15642 (redirect_edge_var_map_clear): Use vec_free.
15643 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15644 (free_var_map_entry): Use vec_free.
15645 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15646 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15648 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15650 PR middle-end/45472
15651 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15652 when the may_trap_p bit of the exprs being merged differs.
15653 Reorder tests for speculativeness in the logical and operator.
15655 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15657 * incpath.c (add_standard_paths): Use reconcat instead of concat
15658 where appropriate and avoid leaking memory.
15660 * opts.h: Include obstack.h.
15661 (opts_concat): New prototype.
15662 (opts_obstack): New declaration.
15663 * opts.c (opts_concat): New function.
15664 (opts_obstack): New variable.
15665 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15666 (finish_options): Use opts_concat instead of concat
15667 and XOBNEWVEC instead of XNEWVEC.
15668 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15669 generate_option): Likewise.
15670 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15671 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15674 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15675 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15677 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15679 PR middle-end/56461
15680 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15682 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15684 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15685 (arm_block_move_unaligned_straight): Likewise.
15686 (arm_adjust_block_mem): Likewise.
15688 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15691 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15692 temp, cond and label.
15693 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15696 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15697 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15700 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15701 of prologue/epilogue functions.
15704 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15705 Remove unused variable cfa_offset.
15706 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15709 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15710 type promotion to unsigned.
15713 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15714 for HOST_WIDE_INT of 32 bit / same size as int.
15715 (arm_block_move_unaligned_straight): Likewise.
15716 (arm_adjust_block_mem): Likewise.
15719 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15722 2013-02-26 Marek Polacek <polacek@redhat.com>
15724 PR tree-optimization/56426
15725 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15727 2013-02-26 Richard Biener <rguenther@suse.de>
15730 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15731 unused variable loops.
15733 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15735 PR tree-optimization/56448
15736 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15737 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15738 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15739 later operands of the references, or even first operand for
15740 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15742 PR tree-optimization/56443
15743 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15744 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15745 to type_for_mode langhook.
15747 2013-02-25 Matt Turner <mattst88@gmail.com>
15749 * doc/invoke.texi: Document r4700.
15751 2013-02-25 Richard Biener <rguenther@suse.de>
15753 PR tree-optimization/56175
15754 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15756 (simplify_bitwise_binary): ... here. Also guard the conversion
15757 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15759 2013-02-25 Catherine Moore <clm@codesourcery.com>
15762 2013-02-24 Catherine Moore <clm@codesourcery.com>
15763 Maciej W. Rozycki <macro@codesourcery.com>
15764 Tom de Vries <tom@codesourcery.com>
15765 Nathan Sidwell <nathan@codesourcery.com>
15766 Iain Sandoe <iain@codesourcery.com>
15767 Nathan Froyd <froydnj@codesourcery.com>
15768 Chao-ying Fu <fu@mips.com>
15770 * doc/extend.texi: (micromips, nomicromips, nocompression):
15771 Document new function attributes.
15772 * doc/invoke.texi (minterlink-compressed, mmicromips,
15773 m14k, m14ke, m14kec): Document new options.
15774 (minterlink-mips16): Update documentation.
15775 * doc/md.texi (ZC, ZD): Document new constraints.
15776 * configure.ac (gcc_cv_as_micromips): Check if linker
15777 supports the .set micromips directive.
15778 * configure: Regenerate.
15779 * config.in: Regenerate.
15780 * config/mips/mips-tables.opt: Regenerate.
15781 * config/mips/micromips.md: New file.
15782 * constraints.md (ZC, AD): New constraints.
15783 * config/mips/predicates.md (movep_src_register): New predicate.
15784 (movep_src_operand): New predicate.
15785 (non_volatile_mem_operand): New predicate.
15786 * config/mips/mips.md (multimem): New type.
15787 (length): Differentiate between 17-bit and 18-bit branch offsets.
15788 (MOVEP1, MOVEP2): New mode iterator.
15789 (mov_<load>l): Use ZC constraint.
15790 (mov_<load>r): Likewise.
15791 (mov_<store>l): Likewise.
15792 (mov_<store>r): Likewise.
15793 (*branch_equality<mode>_inverted): Add microMIPS support.
15794 (*branch_equality<mode>): Likewise.
15795 (*jump_absolute): Likewise.
15796 (indirect_jump_<mode>): Likewise.
15797 (tablejump_<mode>): Likewise.
15798 (<optab>_internal): Likewise.
15799 (sibcall_internal): Likewise.
15800 (sibcall_value_internal): Likewise.
15801 (prefetch): Use constraint ZD.
15802 * config/mips/mips.opt (minterlink-compressed): New option.
15803 (minterlink-mips16): Now an alias for minterlink-compressed.
15804 (mmicromips): New option.
15805 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15806 (compare_and_swap_12): Likewise.
15807 (sync_add<mode>): Likewise.
15808 (sync_<optab>_12): Likewise.
15809 (sync_old_<optab>_12): Likewise.
15810 (sync_new_<optab>_12): Likewise.
15811 (sync_nand_12): Likewise.
15812 (sync_old_nand_12): Likewise.
15813 (sync_new_nand_12): Likewise.
15814 (sync_sub<mode>): Likewise.
15815 (sync_old_add<mode>): Likewise.
15816 (sync_old_sub<mode>): Likewise.
15817 (sync_new_add<mode>): Likewise.
15818 (sync_new_sub<mode>): Likewise.
15819 (sync_<optab><mode>): Likewise.
15820 (sync_old_<optab><mode>): Likewise.
15821 (sync_new_<optab><mode>): Likewise.
15822 (sync_nand<mode>): Likewise.
15823 (sync_old_nand<mode>): Likewise.
15824 (sync_new_nand<mode>): Likewise.
15825 (sync_lock_test_and_set<mode>): Likewise.
15826 (test_and_set_12): Likewise.
15827 (atomic_compare_and_swap<mode>): Likewise.
15828 (atomic_exchange<mode>_llsc): Likewise.
15829 (atomic_fetch_add<mode>_llsc): Likewise.
15830 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15831 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15832 (umips_save_restore_pattern_p): Likewise.
15833 (umips_load_store_pair_p): Likewise.
15834 (umips_output_load_store_pair): Likewise.
15835 (umips_movep_target_p): Likewise.
15836 (umips_12bit_offset_address_p): Likewise.
15837 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15838 (mips_base_mips16): Rename this...
15839 (mips_base_compression_flags): ...to this. Update all uses.
15840 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15841 (mips_mips16_decl_p): Delete.
15842 (mips_nomips16_decl_p): Delete.
15843 (mips_get_compress_on_flags): New function.
15844 (mips_get_compress_off_flags): New function.
15845 (mips_get_compress_mode): New function.
15846 (mips_get_compress_on_name): New function.
15847 (mips_get_compress_off_name): New function.
15848 (mips_insert_attributes): Support multiple compression types.
15849 (mips_merge_decl_attributes): Likewise.
15850 (umips_12bit_offset_address_p): New function.
15851 (mips_start_function_definition): Emit .set micromips directive.
15852 (mips_call_may_need_jalx_p): New function.
15853 (mips_function_ok_for_sibcall): Add microMIPS support.
15854 (mips_print_operand_punctuation): Support short delay slots and
15856 (umips_swm_mask, umips_swm_encoding): New.
15857 (umips_build_save_restore): New function.
15858 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15859 (was_mips16_p): Remove.
15860 (old_compression_mode): New.
15861 (mips_set_compression_mode): New function.
15862 (mips_set_current_function): Add microMIPS support.
15863 (mips_option_override): Likewise.
15864 (umips_save_restore_pattern_p): New function.
15865 (umips_output_save_restore): New function.
15866 (umips_load_store_pair_p_1): New function.
15867 (umips_load_store_pair_p): New function.
15868 (umips_output_load_store_pair_1): New function.
15869 (umips_output_load_store_pair): New function.
15870 (umips_movep_target_p) New function.
15871 (mips_prepare_pch_save): Add microMIPS support.
15872 * config/mips/mips.h (TARGET_COMPRESSION): New.
15873 (TARGET_CPU_CPP_BUILTINS): Update macro
15874 to use new compression flags and to support microMIPS.
15875 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15876 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15877 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15878 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15879 (ASM_SPEC): Support mmicromips and mno-micromips.
15880 (M16STORE_REG_P): New macro.
15881 (MIPS_CALL): Support TARGET_MICROMIPS.
15882 (MICROMIPS_J): New macro.
15883 (mips_base_mips16): Rename this...
15884 (mips_base_compression_flags): ...to this.
15885 (UMIPS_12BIT_OFFSET_P): New macro.
15886 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15887 (MULTILIB_DIRNAMES): Likewise.
15889 2013-02-25 Tom de Vries <tom@codesourcery.com>
15891 PR rtl-optimization/56131
15892 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15893 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15894 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15896 2013-02-25 Tobias Burnus <burnus@net-b.de>
15898 * doc/invoke.texi (-fsanitize=): Move from optimization
15899 to debugging options.
15901 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15903 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15905 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15906 Alexander Monakov <amonakov@ispras.ru>
15908 PR middle-end/56077
15909 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15910 flush pending lists also on non-jumps. Adjust comment.
15912 2013-02-24 Catherine Moore <clm@codesourcery.com>
15913 Maciej W. Rozycki <macro@codesourcery.com>
15914 Tom de Vries <tom@codesourcery.com>
15915 Nathan Sidwell <nathan@codesourcery.com>
15916 Iain Sandoe <iain@codesourcery.com>
15917 Nathan Froyd <froydnj@codesourcery.com>
15918 Chao-ying Fu <fu@mips.com>
15920 * doc/extend.texi: (micromips, nomicromips, nocompression):
15921 Document new function attributes.
15922 * doc/invoke.texi (minterlink-compressed, mmicromips,
15923 m14k, m14ke, m14kec): Document new options.
15924 (minterlink-mips16): Update documentation.
15925 * doc/md.texi (ZC, ZD): Document new constraints.
15926 * configure.ac (gcc_cv_as_micromips): Check if linker
15927 supports the .set micromips directive.
15928 * configure: Regenerate.
15929 * config.in: Regenerate.
15930 * config/mips/mips-tables.opt: Regenerate.
15931 * config/mips/micromips.md: New file.
15932 * constraints.md (ZC, AD): New constraints.
15933 * config/mips/predicates.md (movep_src_register): New predicate.
15934 (movep_src_operand): New predicate.
15935 (non_volatile_mem_operand): New predicate.
15936 * config/mips/mips.md (multimem): New type.
15937 (length): Differentiate between 17-bit and 18-bit branch offsets.
15938 (MOVEP1, MOVEP2): New mode iterator.
15939 (mov_<load>l): Use ZC constraint.
15940 (mov_<load>r): Likewise.
15941 (mov_<store>l): Likewise.
15942 (mov_<store>r): Likewise.
15943 (*branch_equality<mode>_inverted): Add microMIPS support.
15944 (*branch_equality<mode>): Likewise.
15945 (*jump_absolute): Likewise.
15946 (indirect_jump_<mode>): Likewise.
15947 (tablejump_<mode>): Likewise.
15948 (<optab>_internal): Likewise.
15949 (sibcall_internal): Likewise.
15950 (sibcall_value_internal): Likewise.
15951 (prefetch): Use constraint ZD.
15952 * config/mips/mips.opt (minterlink-compressed): New option.
15953 (minterlink-mips16): Now an alias for minterlink-compressed.
15954 (mmicromips): New option.
15955 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15956 (compare_and_swap_12): Likewise.
15957 (sync_add<mode>): Likewise.
15958 (sync_<optab>_12): Likewise.
15959 (sync_old_<optab>_12): Likewise.
15960 (sync_new_<optab>_12): Likewise.
15961 (sync_nand_12): Likewise.
15962 (sync_old_nand_12): Likewise.
15963 (sync_new_nand_12): Likewise.
15964 (sync_sub<mode>): Likewise.
15965 (sync_old_add<mode>): Likewise.
15966 (sync_old_sub<mode>): Likewise.
15967 (sync_new_add<mode>): Likewise.
15968 (sync_new_sub<mode>): Likewise.
15969 (sync_<optab><mode>): Likewise.
15970 (sync_old_<optab><mode>): Likewise.
15971 (sync_new_<optab><mode>): Likewise.
15972 (sync_nand<mode>): Likewise.
15973 (sync_old_nand<mode>): Likewise.
15974 (sync_new_nand<mode>): Likewise.
15975 (sync_lock_test_and_set<mode>): Likewise.
15976 (test_and_set_12): Likewise.
15977 (atomic_compare_and_swap<mode>): Likewise.
15978 (atomic_exchange<mode>_llsc): Likewise.
15979 (atomic_fetch_add<mode>_llsc): Likewise.
15980 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15981 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15982 (umips_save_restore_pattern_p): Likewise.
15983 (umips_load_store_pair_p): Likewise.
15984 (umips_output_load_store_pair): Likewise.
15985 (umips_movep_target_p): Likewise.
15986 (umips_12bit_offset_address_p): Likewise.
15987 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15988 (mips_base_mips16): Rename this...
15989 (mips_base_compression_flags): ...to this. Update all uses.
15990 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15991 (mips_mips16_decl_p): Delete.
15992 (mips_nomips16_decl_p): Delete.
15993 (mips_get_compress_on_flags): New function.
15994 (mips_get_compress_off_flags): New function.
15995 (mips_get_compress_mode): New function.
15996 (mips_get_compress_on_name): New function.
15997 (mips_get_compress_off_name): New function.
15998 (mips_insert_attributes): Support multiple compression types.
15999 (mips_merge_decl_attributes): Likewise.
16000 (umips_12bit_offset_address_p): New function.
16001 (mips_start_function_definition): Emit .set micromips directive.
16002 (mips_call_may_need_jalx_p): New function.
16003 (mips_function_ok_for_sibcall): Add microMIPS support.
16004 (mips_print_operand_punctuation): Support short delay slots and
16006 (umips_swm_mask, umips_swm_encoding): New.
16007 (umips_build_save_restore): New function.
16008 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16009 (was_mips16_p): Remove.
16010 (old_compression_mode): New.
16011 (mips_set_compression_mode): New function.
16012 (mips_set_current_function): Add microMIPS support.
16013 (mips_option_override): Likewise.
16014 (umips_save_restore_pattern_p): New function.
16015 (umips_output_save_restore): New function.
16016 (umips_load_store_pair_p_1): New function.
16017 (umips_load_store_pair_p): New function.
16018 (umips_output_load_store_pair_1): New function.
16019 (umips_output_load_store_pair): New function.
16020 (umips_movep_target_p) New function.
16021 (mips_prepare_pch_save): Add microMIPS support.
16022 * config/mips/mips.h (TARGET_COMPRESSION): New.
16023 (TARGET_CPU_CPP_BUILTINS): Update macro
16024 to use new compression flags and to support microMIPS.
16025 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16026 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16027 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16028 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16029 (ASM_SPEC): Support mmicromips and mno-micromips.
16030 (M16STORE_REG_P): New macro.
16031 (MIPS_CALL): Support TARGET_MICROMIPS.
16032 (MICROMIPS_J): New macro.
16033 (mips_base_mips16): Rename this...
16034 (mips_base_compression_flags): ...to this.
16035 (UMIPS_12BIT_OFFSET_P): New macro.
16036 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16037 (MULTILIB_DIRNAMES): Likewise.
16039 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16042 * target-globals.c (save_target_globals): For init_reg_sets and
16043 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16045 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16047 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16048 * config/aarch64/t-aarch64
16049 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16051 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16053 PR inline-asm/56148
16054 * lra-constraints.c (process_alt_operands): Reload operand
16055 conflicting with earlier clobber only if no more other conflicting
16058 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16061 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16062 if not linking a shared library.
16064 2013-02-22 Seth LaForge <sethml@google.com>
16066 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16068 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16070 * config/arm/arm.md (split for extendsidi): Update condition.
16071 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16072 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16073 (qhs_zextenddi_cstr): Likewise.
16075 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16077 PR middle-end/56420
16078 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16079 avoid signed wrapping.
16080 (expand_mult): Handle properly multiplication by
16081 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16082 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16083 in the compiler if coeff is HOST_WIDE_INT_MIN.
16084 (expand_divmod): Don't make ext_op1 static, change it's type to
16085 uhwi. Avoid undefined behavior in -INTVAL (op1).
16087 PR rtl-optimization/50339
16088 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16090 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16091 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16092 into splitting_ashiftrt field.
16093 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16095 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16098 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16100 PR middle-end/56108
16101 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16102 are sure to go irrevocable.
16104 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16106 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16107 scalars are valid operands.
16109 2013-02-21 Martin Jambor <mjambor@suse.cz>
16111 PR tree-optimization/56310
16112 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16113 only matching indices and non-negative final offsets.
16114 (intersect_aggregates_with_edge): Pass src_idx to
16115 agg_replacements_to_vector. Pass src_idx insstead of index to
16116 intersect_with_agg_replacements.
16118 2013-02-21 Martin Jambor <mjambor@suse.cz>
16120 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16121 instead of hard-wired defaults.
16123 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16125 * doc/invoke.texi (MIPS Options): Update documentation of the
16126 floating-point multiply-accumulate instruction restrictions.
16128 2013-02-21 Kostya Serebryany <kcc@google.com>
16130 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16131 asan_shadow_offset on x86_64 linux.
16133 2013-02-21 Richard Biener <rguenther@suse.de>
16135 PR tree-optimization/56415
16137 2013-02-11 Richard Biener <rguenther@suse.de>
16139 PR tree-optimization/56273
16140 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16143 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16146 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16149 PR inline-asm/56405
16150 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16151 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16153 2013-02-20 Jan Hubicka <jh@suse.cz>
16155 PR tree-optimization/56265
16156 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16157 when target is referenced for first time.
16159 2013-02-20 Richard Biener <rguenther@suse.de>
16161 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16162 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16163 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16164 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16165 not return anything.
16166 (rename_ssa_copies): Do not remove unused locals.
16167 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16168 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16169 * passes.c (execute_function_todo): Do not schedule unused locals
16170 removal if cleanup_tree_cfg did something.
16171 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16172 about the number of removed locals.
16174 2013-02-20 Richard Biener <rguenther@suse.de>
16176 PR tree-optimization/56398
16177 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16179 2013-02-20 Martin Jambor <mjambor@suse.cz>
16181 PR tree-optimization/55334
16182 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16183 restricted pointers to arrays.
16185 2013-02-20 Richard Biener <rguenther@suse.de>
16186 Jakub Jelinek <jakub@redhat.com>
16188 PR tree-optimization/56396
16189 * tree-ssa-ccp.c (n_const_val): New static variable.
16190 (get_value): Return NULL for SSA names we don't have a lattice
16192 (ccp_initialize): Initialize n_const_val.
16193 * tree-ssa-copy.c (n_copy_of): New static variable.
16194 (init_copy_prop): Initialize n_copy_of.
16195 (get_value): Return NULL_TREE for SSA names we don't have a
16198 2013-02-20 Martin Jambor <mjambor@suse.cz>
16200 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16202 2013-02-20 Richard Biener <rguenther@suse.de>
16204 * genpreds.c (write_lookup_constraint): Do not compare first
16205 letter of the constraint again.
16207 2013-02-20 Richard Biener <rguenther@suse.de>
16209 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16211 (get_use_iv_cost): Terminate hashtable walk when coming across
16214 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16216 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16217 reassociation for avx2 targets.
16219 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16221 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16222 Add version check for v8.10.a to enable microblaze_has_clz
16223 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16224 version and TARGET_PATTERN_COMPARE check
16225 * config/microblaze/microblaze.md: New clzsi2 instruction
16227 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16229 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
16230 function before branching.
16232 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16234 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
16236 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
16238 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16240 PR middle-end/55889
16241 * sel-sched.c: Include ira.h.
16242 (implicit_clobber_conflict_p): New function.
16243 (moveup_expr): Use it.
16244 * Makefile.in (sel-sched.o): Depend on ira.h.
16246 2013-02-19 Richard Biener <rguenther@suse.de>
16248 PR tree-optimization/56384
16249 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
16250 (vn_hash_type): Split out from ...
16251 (vn_hash_constant_with_type): ... here.
16252 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
16253 (vn_phi_eq): Compare types from vn_phi_s structure.
16254 (vn_phi_lookup): Populate vn_phi_s type.
16255 (vn_phi_insert): Likewise.
16257 2013-02-19 Jakub Jelinek <jakub@redhat.com>
16259 PR tree-optimization/56350
16260 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
16261 if haven't found reduction or nested cycle operand, rather than
16262 asserting we must find it.
16264 PR tree-optimization/56381
16265 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
16268 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
16269 Jakub Jelinek <jakub@redhat.com>
16272 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
16273 (swap_optab_enable): Same.
16274 (init_all_optabs): Use argument instead of global.
16275 * tree.h (struct tree_optimization_option): New field target_optabs.
16276 * expr.h (init_all_optabs): Add argument to prototype.
16277 (TREE_OPTIMIZATION_OPTABS): New.
16278 (save_optabs_if_changed): Protoize.
16279 * optabs.h: Declare this_fn_optabs.
16280 * optabs.c (save_optabs_if_changed): New.
16281 Declare this_fn_optabs.
16282 (init_optabs): Add argument to init_all_optabs() call.
16283 * function.c (invoke_set_current_function_hook): Handle per
16285 * function.h (struct function): New field optabs.
16286 * config/mips/mips.c (mips_set_mips16_mode): Handle when
16287 optimization_current_node has changed.
16288 * target-globals.h (save_target_globals_default_opts): Protoize.
16289 * target-globals.c (save_target_globals_default_opts): New.
16291 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16294 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
16295 registers %fr12 and %fr12R as call used.
16298 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
16299 and HImode, require all displacements to be an integer multiple of
16301 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
16302 only allow QImode and HImode when reload is in progress and strict is
16303 true. Likewise for symbolic addresses. Use base14_operand to check
16304 displacements in REG+BASE addresses.
16306 2013-02-18 Richard Biener <rguenther@suse.de>
16308 PR tree-optimization/56366
16309 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16310 handle sign-conversion of outer-loop initial induction value.
16312 2013-02-18 Richard Biener <rguenther@suse.de>
16314 PR middle-end/56349
16315 * cfghooks.c (merge_blocks): If we merge a latch into another
16316 block adjust references to it.
16317 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
16318 (verify_loop_structure): Verify that a recorded latch is in fact
16321 2013-02-18 Richard Biener <rguenther@suse.de>
16323 PR tree-optimization/56321
16324 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
16325 order SSA name release and virtual operand unlinking.
16327 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16329 * config/microblaze/microblaze.md (save_stack_block): Define.
16330 (restore_stack_block): Likewise.
16332 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16334 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
16335 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
16336 * config/microblaze/microblaze.c (microblaze_option_override):
16337 Bail out early for PIC modes when target does not support PIC.
16339 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16341 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
16342 Replace with a microblaze version.
16343 (microblaze_trampoline_init): Adapt for microblaze.
16344 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
16347 2013-02-16 Jakub Jelinek <jakub@redhat.com>
16348 Dodji Seketeli <dodji@redhat.com>
16351 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
16352 (instrument_mem_region_access): Do not forget to always put
16353 instrumentation of the of 'base' and 'base + len' in a "if (len !=
16354 0) statement, even for cases where either 'base' or 'base + len'
16355 are not instrumented -- because they have been previously
16356 instrumented. Simplify the logic by putting all the statements
16357 instrument 'base + len' inside a sequence, and then insert that
16358 sequence right before the current insertion point. Then, to
16359 instrument 'base + len', just get an iterator on that statement.
16360 And do not forget to update the pointer to iterator the function
16361 received as argument.
16363 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
16365 PR rtl-optimization/56348
16366 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
16368 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
16370 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
16371 (clean_graph_dump_file): Pass base to start_graph_dump.
16373 2013-02-14 Richard Henderson <rth@redhat.com>
16376 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
16378 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
16380 * collect2-aix.h: Define F_LOADONLY.
16382 2013-02-14 Richard Biener <rguenther@suse.de>
16385 * varasm.c (output_constant_def_1): Get the decl representing
16386 the constant as argument.
16387 (output_constant_def): Wrap output_constant_def_1.
16388 (make_decl_rtl): Use output_constant_def_1 with the decl
16389 representing the constant.
16390 (build_constant_desc): Optionally re-use a decl already
16391 representing the constant.
16392 (tree_output_constant_def): Adjust.
16394 2013-02-14 Dodji Seketeli <dodji@redhat.com>
16397 * asan.c (instrument_builtin_call): Really put the length of the
16398 second source argument into src1_len.
16400 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16402 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
16403 argument. If it is false, don't create edge from then_bb to
16405 (insert_if_then_before_iter): Pass true to it.
16406 (build_check_stmt): Pass false to it.
16407 (transform_statements): Flush hash table only on extended basic
16408 block boundaries, rather than at the beginning of every bb.
16409 Don't flush hash table on nonfreeing_call_p calls.
16410 * tree-flow.h (nonfreeing_call_p): New prototype.
16411 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
16413 2013-02-13 David S. Miller <davem@davemloft.net>
16415 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
16417 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
16420 * ira.c (max_regno_before_ira): Move from ...
16422 (fix_reg_equiv_init): Use max_regno_before_ira instead of
16425 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16427 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
16429 2013-02-13 Richard Biener <rguenther@suse.de>
16432 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
16433 globals in MEM_REFs.
16435 2013-02-13 Richard Biener <rguenther@suse.de>
16437 * loop-init.c (loop_optimizer_init): Clear loop state when
16438 re-initializing preserved loops.
16439 * loop-unswitch.c (unswitch_single_loop): Return whether
16440 we unswitched the loop. Do not verify loop state here.
16441 (unswitch_loops): When we unswitched a loop discover new loops.
16443 2013-02-13 Kostya Serebryany <kcc@google.com>
16445 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
16447 * sanitizer.def: Rename __asan_init to __asan_init_v1.
16449 2013-02-12 Dodji Seketeli <dodji@redhat.com>
16451 Avoid instrumenting duplicated memory access in the same basic block
16452 * Makefile.in (asan.o): Add new dependency on hash-table.h
16453 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
16454 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
16455 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
16456 (free_mem_ref_resources, has_mem_ref_been_instrumented)
16457 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
16458 (get_mem_ref_of_assignment): New functions.
16459 (get_mem_refs_of_builtin_call): Extract from
16460 instrument_builtin_call and tweak a little bit to make it fit with
16462 (instrument_builtin_call): Use the new
16463 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
16464 of is_gimple_builtin_call.
16465 (instrument_derefs, instrument_mem_region_access): Insert the
16466 instrumented memory reference into the hash table.
16467 (maybe_instrument_assignment): Renamed instrument_assignment into
16468 this, and change it to advance the iterator when instrumentation
16469 actually happened and return true in that case. This makes it
16470 homogeneous with maybe_instrument_assignment, and thus give a
16471 chance to callers to be more 'regular'.
16472 (transform_statements): Clear the memory reference hash table
16473 whenever we enter a new BB, when we cross a function call, or when
16474 we are done transforming statements. Use
16475 maybe_instrument_assignment instead of instrumentation. No more
16476 need to special case maybe_instrument_assignment and advance the
16477 iterator after calling it; it's now handled just like
16478 maybe_instrument_call. Update comment.
16480 2013-02-13 Richard Biener <rguenther@suse.de>
16482 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
16483 Fix loop discovery code.
16485 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
16487 PR inline-asm/56148
16488 * lra-constraints.c (process_alt_operands): Match early clobber
16489 operand with itself. Check conflicts with earlyclobber only if
16490 the operand is not reloaded. Prefer to reload conflicting operand
16491 if earlyclobber and matching operands are the same.
16493 2013-02-12 Richard Biener <rguenther@suse.de>
16496 * lto-streamer-out.c (write_symbol): Do not output symbols
16497 for hard register variables.
16499 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16502 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16503 (umulsidi3_insn, mulsidi3_insn): New insns.
16505 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16507 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16508 (struct tune_params): Add vec_costs field.
16509 * config/arm/arm.c (arm_builtin_vectorization_cost)
16510 (arm_add_stmt_cost): New functions.
16511 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16512 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16513 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16514 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16515 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16516 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16517 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16519 2013-02-12 Richard Biener <rguenther@suse.de>
16522 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16523 decls again if possible.
16525 2013-02-12 Richard Biener <rguenther@suse.de>
16527 PR middle-end/56288
16528 * tree-ssa.c (verify_ssa_name): Fix check, move
16529 SSA_NAME_IN_FREE_LIST check up.
16531 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16532 Steven Bosscher <steven@gcc.gnu.org>
16534 PR rtl-optimization/56151
16535 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16536 equal to op0 or op1, and last_insn pattern is CODE operation
16537 with MEM dest and one of the operands matches that MEM.
16539 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16541 * doc/extend.texi: Document Function Multiversioning and "default"
16542 parameter string to target attribute.
16543 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16544 target attribute parameter is "default".
16545 (ix86_compare_version_priority): Remove checks for target attribute.
16546 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16547 Remove check for target attribute equal to NULL. Add assert.
16548 (ix86_generate_version_dispatcher_body): Change error to sorry.
16550 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16551 Jack Howarth <howarth@bromo.med.uc.edu>
16552 Patrick Marlier <patrick.marlier@gmail.com>
16555 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16556 define ENDFILE_SPEC as TM_DESTRUCTOR.
16557 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16559 2013-02-11 Alexander Potapenko <glider@google.com>
16560 Jack Howarth <howarth@bromo.med.uc.edu>
16561 Jakub Jelinek <jakub@redhat.com>
16564 * config/darwin.c (cdtor_record): Rename ctor_record.
16565 (sort_cdtor_records): Rename sort_ctor_records.
16566 (finalize_dtors): New routine to sort destructors by
16567 priority before use in assemble_integer.
16568 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16570 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16572 PR rtl-optimization/56275
16573 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16574 offset is non-negative and less than cmode size before
16575 calling simplify_subreg.
16577 2013-02-11 Richard Biener <rguenther@suse.de>
16579 PR tree-optimization/56264
16580 * cfgloop.h (fix_loop_structure): Adjust prototype.
16581 * loop-init.c (fix_loop_structure): Return the number of
16582 newly discovered loops.
16583 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16584 are discovered, do a full loop-closed SSA rewrite.
16586 2013-02-11 Richard Biener <rguenther@suse.de>
16588 PR tree-optimization/56273
16589 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16591 (check_array_ref): Fix missing newline in dumps.
16592 (search_for_addr_array): Likewise.
16594 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16596 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16598 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16601 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16603 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16605 PR rtl-optimization/56246
16606 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16608 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16609 constraints are satisfied.
16611 2013-02-08 Jeff Law <law@redhat.com>
16614 * emit-rtl.c (reg_is_parm_p): New function.
16615 * regs.h (reg_is_parm_p): New prototype.
16616 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16617 callee-clobbered registers.
16619 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16622 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16623 If there is no implicit builtin declaration, just return NULL.
16625 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16627 * config/i386/sse.md (FMAMODEM): New mode iterator.
16628 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16629 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16631 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16633 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16634 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16635 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16637 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16639 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16640 (microblaze*-*-elf): Likewise.
16641 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16643 * config/microblaze/microblaze-c.c: Add builtin defines for
16644 _LITTLE_ENDIAN and _BIG_ENDIAN.
16645 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16646 add to TARGET_DEFAULT flags.
16647 Expand ASM_SPEC and LINK_SPEC.
16648 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16649 * config/microblaze/microblaze.md: Update extendsidi2 and
16650 movdi_internal instructions to use low-order / high-order reg
16652 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16653 options and inversemask / mask of LITTLE_ENDIAN.
16654 * config/microblaze/t-microblaze: Expand multilib options to
16655 include mlittle-endian (le) and update exceptions patterns.
16657 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16659 PR rtl-optimization/56195
16660 * lra-constraints.c (get_reload_reg): Don't reuse regs
16661 if they have smaller mode than requested, if they have
16662 wider mode than requested, try to return a SUBREG.
16664 PR tree-optimization/56250
16665 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16666 if type is unsigned and code isn't MULT_EXPR.
16668 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16670 PR tree-optimization/56064
16671 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16672 bits according to mode.
16673 * fixed-value.h (fixed_from_double_int)
16674 (const_fixed_from_double_int): Adjust comments.
16676 2013-02-08 Richard Biener <rguenther@suse.de>
16679 * lto-streamer.h (struct data_in): Remove current_file, current_line
16680 and current_col members.
16681 * lto-streamer-out.c (lto_output_location): Stream changed bits
16682 en-block for efficiency.
16683 * lto-streamer-in.c (clear_line_info): Remove.
16684 (lto_input_location): Cache current file, line and column
16685 globally via local statics. Read changed bits en-block.
16686 (input_function): Do not call clear_line_info.
16687 (lto_read_body): Likewise.
16688 (lto_input_toplevel_asms): Likewise.
16690 2013-02-08 Michael Matz <matz@suse.de>
16692 PR tree-optimization/52448
16693 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16694 (nt_call_phase): New static.
16695 (add_or_mark_expr): Only mark accesses with newer phase than any
16697 (nonfreeing_call_p): New.
16698 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16699 (nt_fini_block): Keep blocks marked as visited.
16700 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16702 2013-02-08 Richard Biener <rguenther@suse.de>
16704 * ira.c (ira): Free broken dominator information.
16706 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16708 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16710 2013-02-08 Marek Polacek <polacek@redhat.com>
16712 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16714 2013-02-08 Richard Biener <rguenther@suse.de>
16716 PR middle-end/56181
16717 * cfgloop.h (flow_loops_find): Adjust.
16718 (bb_loop_header_p): Declare.
16719 * cfgloop.c (bb_loop_header_p): New function split out from ...
16720 (flow_loops_find): ... here. Adjust function signature,
16721 support incremental loop structure update.
16722 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16723 * cfgloopmanip.c (fix_loop_structure): Move ...
16724 * loop-init.c (fix_loop_structure): ... here.
16725 (apply_loop_flags): Split out from ...
16726 (loop_optimizer_init): ... here.
16727 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16728 in incremental mode, only remove dead loops here.
16730 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16733 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16734 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16735 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16736 (*round<mode>3.libgcc): New insns for fixed-modes.
16737 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16738 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16739 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16740 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16741 implementations. Define to __builtin_avr_absFX,
16742 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16743 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16744 __builtin_avr_countlsFX, respectively.
16745 * config/avr/avr-c.c (target.h): Include it.
16746 (enum avr_builtin_id): New enum.
16747 (avr_resolve_overloaded_builtin): New static function.
16748 (avr_register_target_pragmas): Use it to set
16749 targetm.resolve_overloaded_builtin.
16750 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16751 tree nodes used by DEF_BUILTIN.
16752 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16753 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16754 <AVR_BUILTIN_xxBITS>: Same.
16756 2013-02-08 Richard Biener <rguenther@suse.de>
16758 * cfgloop.c (verify_loop_structure): Properly handle
16759 a loop exiting to another loop header.
16760 * ira-int.h (ira_loops): Remove.
16761 * ira.c (ira_loops): Remove.
16762 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16763 (do_reload): Use loop_optimizer_finalize.
16764 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16765 number_of_loops to access the loop tree.
16766 (more_one_region_p): Likewise.
16767 (finish_loop_tree_nodes): Likewise.
16768 (rebuild_regno_allocno_maps): Likewise.
16769 (mark_loops_for_removal): Likewise.
16770 (mark_all_loops_for_removal): Likewise.
16771 (remove_unnecessary_regions): Likewise.
16772 (ira_build): Likewise.
16773 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16775 2013-02-08 Richard Biener <rguenther@suse.de>
16777 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16778 * ipa-pure-const.c (analyze_function): Avoid calling
16779 mark_irreducible_loops twice.
16780 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16782 2013-02-07 David S. Miller <davem@davemloft.net>
16784 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16786 * var-tracking.c (vt_add_function_parameter): Test the presence of
16787 HAVE_window_save properly and do not remap argument registers when
16788 we have a leaf function.
16790 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16793 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16795 * config/i386/i386.c (ix86_print_operand): Ditto.
16797 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16799 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16801 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16803 PR rtl-optimization/56225
16804 * lra-constraints.c (process_alt_operands): Check that reload hard
16805 reg can hold value for strict_low_part.
16807 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16810 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16811 dwarf2out_end_function.
16812 (in_first_function_p, maybe_at_text_label_p,
16813 first_loclabel_num_not_at_text_label): New variables.
16814 (dwarf2out_var_location): In the first function find out
16815 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16816 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16819 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16821 PR rtl-optimization/56178
16822 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16823 SUBREG of a register. Tidy up related block of code.
16824 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16825 note if the source is a register or a SUBREG of a register.
16827 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16830 * config/rs6000/rs6000.md (ptrm): New mode attr.
16831 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16832 call_value_indirect_aix<pttrsize>,
16833 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16836 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16838 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16839 if -bnortl. Convert to strcmp and strncmp.
16841 2013-02-07 Alan Modra <amodra@gmail.com>
16844 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16845 addresses won't wrap when offsetting.
16846 (rs6000_secondary_reload): Provide secondary reloads needed for
16847 wrapping LO_SUM addresses.
16849 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16851 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16852 MACH, just __MACH__.
16854 2013-02-06 Richard Biener <rguenther@suse.de>
16856 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16857 instead of calling fix_loop_structure.
16859 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16861 PR middle-end/56217
16862 * omp-low.c (use_pointer_for_field): Return false if
16863 lower_send_shared_vars doesn't generate any copy-out code.
16865 2013-02-06 Tom de Vries <tom@codesourcery.com>
16867 PR rtl-optimization/56131
16868 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16869 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16870 of the label is NULL. Add comment.
16872 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16874 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16877 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16878 (STATIC_LIBTSAN_LIBS): Likewise.
16879 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16880 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16881 is defined, don't add anything else beyond that.
16882 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16883 (LINK_COMMAND_SPEC): Use them.
16885 PR tree-optimization/56205
16886 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16887 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16888 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16890 2013-02-05 Richard Biener <rguenther@suse.de>
16892 PR tree-optimization/53342
16893 PR tree-optimization/53185
16894 * tree-vectorizer.h (vect_check_strided_load): Remove.
16895 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16896 not disallow peeling for vectorized strided loads.
16897 (vect_check_strided_load): Make static and simplify.
16898 (vect_analyze_data_refs): Adjust.
16899 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16900 correctly when vectorizing strided loads.
16902 2013-02-05 Richard Biener <rguenther@suse.de>
16904 * doc/install.texi: Refer to ISL, not PPL.
16906 2013-02-05 Jan Hubicka <jh@suse.cz>
16908 PR tree-optimization/55789
16909 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16911 2013-02-05 Jan Hubicka <jh@suse.cz>
16913 PR tree-optimization/55789
16914 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16915 the dead call anyway.
16917 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16920 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16922 2013-02-04 Alexander Potapenko <glider@google.com>
16923 Jack Howarth <howarth@bromo.med.uc.edu>
16924 Jakub Jelinek <jakub@redhat.com>
16927 * config/darwin.c (sort_ctor_records): Stabilized qsort
16928 on constructor priority by using original position.
16929 (finalize_ctors): New routine to sort constructors by
16930 priority before use in assemble_integer.
16931 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16933 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16936 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16937 about visibility on artificial decls.
16938 * config/sol2.c (solaris_assemble_visibility): Likewise.
16940 2013-02-04 Kai Tietz <ktietz@redhat.com>
16943 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16944 argument and improve checking of return-argument types for 16-byte
16946 (ix86_function_value_1): Add additional valtype argument on call
16947 of function_value_64.
16948 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16949 handling infunction_value_64 function.
16951 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16953 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16955 2013-02-04 Richard Biener <rguenther@suse.de>
16957 PR tree-optimization/56188
16958 * tree-ssa-structalias.c (label_visit): Consider case with
16959 initially non-empty points-to set.
16960 (perform_var_substitution): Dump node mapping and clean up.
16962 2013-02-04 Richard Guenther <rguenther@suse.de>
16965 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16966 node prevail as last resort.
16967 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16968 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16970 2013-02-04 Richard Biener <rguenther@suse.de>
16972 PR tree-optimization/56113
16973 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16975 (equiv_class_lookup_or_add): ... this.
16976 (label_visit): Adjust and fix error in previous patch.
16977 (perform_var_substitution): Adjust.
16979 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16981 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16982 * config/sh/sh4-300.md: Likewise.
16983 * config/sh/sh4a.md: Likewise.
16984 * config/sh/constraints.md: Likewise.
16985 * config/sh/sh.md: Likewise.
16986 * config/sh/netbsd-elf.h: Likewise.
16987 * config/sh/predicates.md: Likewise.
16988 * config/sh/sh-protos.h: Likewise.
16989 * config/sh/ushmedia.h: Likewise.
16990 * config/sh/linux.h: Likewise.
16991 * config/sh/sh.c: Likewise.
16992 * config/sh/superh.h: Likewise.
16993 * config/sh/elf.h: Likewise.
16994 * config/sh/sh4.md: Likewise.
16995 * config/sh/sh.h: Likewise.
16997 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16999 * config/pa/constraints.md: Adjust unused letters. Change "T"
17000 constraint to match_test floating_point_store_memory_operand().
17001 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17002 (base14_operand): New.
17003 (floating_point_store_memory_operand): New.
17004 (integer_store_memory_operand): Revise to use base14_operand and
17005 reg_plus_base_memory_operand.
17006 (move_dest_operand): Allow symbolic_memory_operands.
17007 (symbolic_memory_operand): Check for LO_SOM.
17008 (symbolic_operand): Change default case to break.
17009 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17010 CONST_DOUBLE values to be reloaded by putting them into memory when
17011 the destination is a floating point register.
17012 (movdf): Remove code to handle CONST_DOUBLE.
17014 (reload_indf_r1): New.
17015 (reload_insf_r1): New.
17016 Consistently use "Q" and "T" constraints with integer and floating
17017 point move instructions, respectively.
17018 (movdi): Remove FAIL.
17019 Change predicate for source operand unamed DImode move from
17020 general_operand to move_src_operand.
17021 (umulsidi3): Change predicate for destination operand to
17023 Likewise for similar unamed patterns.
17024 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17025 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17026 (hppa_legitimize_address): Simplify mask calculation.
17027 (pa_emit_move_sequence): Revised handling of secondary reloads from
17028 REG+D addresses for floating point loads and stores. Directly handle
17029 loading CONST0_RTX (mode) to a floating point register.
17030 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17031 to floating point registers. Don't restrict secondary reloads to
17032 floating point registers to integer modes. Revise some comments and
17034 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17035 (pa_legitimate_address_p): New.
17036 (pa_legitimize_reload_address): New.
17037 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17038 (STRICT_REG_OK_FOR_BASE_P): New.
17039 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17040 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17042 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17043 Andrew Dixie <andrewd@gentrack.com>
17045 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17048 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17050 * expmed.c (extract_bit_field_1): Pass the full width of the
17051 structure to get_best_reg_extraction_insn.
17053 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17056 * configure.ac (use_cxa_atexit): Add AIX.
17057 * configure: Regenerate.
17059 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17061 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17064 * final.c (need_profile_function): New variable.
17065 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17066 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17067 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17068 notes, targetm.asm_out.function_prologue doesn't emit anything,
17069 HAVE_prologue and profiler should be emitted before prologue,
17070 set need_profile_function instead of emitting it.
17071 (final_scan_insn): If need_profile_function, emit
17072 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17073 NOTE_INSN_FUNCTION_BEG note.
17075 2013-02-01 Richard Henderson <rth@redhat.com>
17077 * config/rs6000/rs6000.md (smulditi3): New.
17080 * config/alpha/alpha.md (umulditi3): New.
17082 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17084 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17085 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17087 2013-02-01 Richard Biener <rguenther@suse.de>
17089 PR tree-optimization/56113
17090 * tree-ssa-structalias.c (label_visit): Reduce work for
17091 single-predecessor nodes.
17093 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17095 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17096 range isn't testing for zero.
17098 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17100 PR middle-end/56113
17101 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17103 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17104 Nick Clifton <nickc@redhat.com>
17106 * config/v850/constraints.md (Q): Define as a memory constraint.
17107 * config/v850/predicates.md (label_ref_operand): New predicate.
17108 (e3v5_shift_operand): New predicate.
17109 (ior_operator): New predicate.
17110 * config/v850/t-v850: Add e3v5 multilib.
17111 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17112 (v850_gen_movdi): Prototype.
17113 * config/v850/v850.c: Add support for e3v5 architecture.
17114 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17116 (construct_save_jarl): Add e3v5 long JARL support.
17117 (v850_adjust_insn_length): New function. Adjust length of call
17118 insns when using e3v5 instructions.
17119 (v850_gen_movdi): New function: Generate instructions to move a
17121 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17122 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17123 (TARGET_USE_FPU): Enable for e3v5.
17124 (CONST_OK_FOR_W): New macro.
17125 (ADJUST_INSN_LENGTH): Define.
17126 * config/v850/v850.md (UNSPEC_LOOP): Define.
17127 (attr cpu): Add v850e3v5.
17128 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17129 (movdi): New pattern.
17130 (movdi_internal): New pattern.
17131 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17132 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17133 (cstoresf4): Likewise.
17134 (cstoredf4): Likewise.
17135 (insv): New pattern.
17136 (rotlso3_a): New pattern.
17137 (rotlsi3_b): New pattern
17138 (rotlsi3_v850e3v5): New pattern.
17139 (doloop_begin): New pattern.
17140 (fix_loop_counter): New pattern.
17141 (doloop_end): New pattern.
17142 (branch_normal): Add e3v5 long branch support.
17143 (branch_invert): Likewise.
17144 (branch_z_normal): Likewise.
17145 (branch_z_invert): Likewise.
17146 (branch_nz_normal): Likewise.
17147 (branch_nz_invert): Likewise.
17148 (call_internal_short): Add e3v5 register-indirect JARL support.
17149 (call_internal_long): Likewise.
17150 (call_value_internal_short): Likewise.
17151 (call_value_internal_long): Likewise.
17152 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17153 (mloop): New option.
17154 * config.gcc: Add support for configuring v840e3v5 target.
17155 * doc/invoke.texi: Document new v850 specific command line options.
17157 2013-01-31 Paul Koning <ni1d@arrl.net>
17161 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17162 children if parent is a class.
17163 (prune_unused_types_prune): Don't add DW_AT_declaration.
17165 2013-01-31 Richard Biener <rguenther@suse.de>
17167 PR tree-optimization/56157
17168 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17169 match up operand with SLP child.
17171 2013-01-31 Jason Merrill <jason@redhat.com>
17174 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17175 parameters the first time.
17176 (gen_scheduled_generic_parms_dies): Check completeness here.
17178 2013-01-31 Richard Biener <rguenther@suse.de>
17180 PR middle-end/53073
17181 * common.opt (faggressive-loop-optimizations): New flag,
17182 enabled by default.
17183 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17184 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17185 infer_loop_bounds_from_undefined by it.
17187 2013-01-31 Richard Biener <rguenther@suse.de>
17189 PR tree-optimization/56150
17190 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17191 visit virtual operands.
17192 (find_uses_to_rename_bb): Likewise.
17194 2013-01-31 Richard Biener <rguenther@suse.de>
17196 PR tree-optimization/56150
17197 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17198 mixed store non-store stmts.
17200 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17203 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17204 LIBASAN_EARLY_SPEC is defined.
17205 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17206 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17208 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17211 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17212 invalid args instead of ICEing on it.
17213 (ix86_valid_target_attribute_tree): Return error_mark_node if
17214 ix86_valid_target_attribute_inner_p failed.
17215 (ix86_valid_target_attribute_p): Return false only if
17216 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17217 target("default") attribute.
17218 (sorted_attr_string): Change argument from const char * to tree,
17219 merge in all target attribute arguments rather than just one.
17220 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17221 instead of free. Avoid using strcat.
17222 (ix86_mangle_function_version_assembler_name): Mangle
17223 target("default") as if no target attribute is present. Adjust
17224 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
17225 instead of xmalloc and XDELETEVEC instead of free.
17226 (ix86_function_versions): Don't return true if one of the decls
17227 doesn't have target attribute. If they don't and one of the decls
17228 is DECL_FUNCTION_VERSIONED, report an error. Adjust
17229 sorted_attr_string caller. Use XDELETEVEC instead of free.
17230 (ix86_supports_function_versions): Remove.
17231 (make_name): Fix up formatting.
17232 (make_dispatcher_decl): Remove resolver_name and its initialization.
17233 Avoid leaking memory.
17234 (is_function_default_version): Return true if there is
17235 target("default") attribute rather than no target attribute at all.
17236 (make_resolver_func): Avoid leaking memory.
17237 (ix86_generate_version_dispatcher_body): Likewise.
17238 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
17239 * target.def (supports_function_versions): Remove.
17240 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
17241 * doc/tm.texi: Regenerated.
17243 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
17245 PR rtl-optimization/56144
17246 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
17247 for values with side effects.
17249 2013-01-30 Richard Biener <rguenther@suse.de>
17251 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
17252 (sparseset_pop): Likewise.
17253 * cfganal.c (compute_idf): Likewise. Increase work-stack size
17254 to be able to use quick_push in the worker loop.
17256 2013-01-30 Marek Polacek <polacek@redhat.com>
17258 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
17260 2013-01-30 Richard Biener <rguenther@suse.de>
17263 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
17265 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
17267 PR tree-optimization/56064
17268 * fixed-value.c (fixed_from_double_int): New function.
17269 * fixed-value.h (fixed_from_double_int): New prototype.
17270 (const_fixed_from_double_int): New static inline function.
17271 * fold-const.c (native_interpret_fixed): New static function.
17272 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
17273 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
17274 (native_encode_fixed): New static function.
17275 (native_encode_expr) <FIXED_CST>: Use it.
17276 (native_interpret_int): Move double_int worker code to...
17277 * double-int.c (double_int::from_buffer): ...this new static method.
17278 * double-int.h (double_int::from_buffer): Prototype it.
17280 2013-01-30 Richard Biener <rguenther@suse.de>
17282 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
17283 New pointer-map and obstack.
17284 (init_alias_vars): Allocate pointer-map and obstack.
17285 (delete_points_to_sets): Free them.
17286 (find_what_var_points_to): Cache result.
17287 (find_what_p_points_to): Adjust for changed interface of
17288 find_what_var_points_to.
17289 (compute_points_to_sets): Likewise.
17290 (ipa_pta_execute): Likewise.
17292 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17294 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
17295 * configure: Regenerate.
17296 * config.in: Regenerate.
17297 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
17298 #nobits/#progbits if supported.
17300 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
17303 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
17304 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
17305 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
17307 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17309 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
17310 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
17312 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17314 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
17315 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
17317 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17319 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
17321 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
17322 * config/arm/cortex-a7.md: New bypasses using
17323 arm_mac_accumulator_is_result.
17325 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17327 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
17328 (cortex_a7_neon_mla): Likewise.
17329 (cortex_a7_fpfmad): New reservation.
17330 (cortex_a7_fpmacs): Use ffmas and update required units.
17331 (cortex_a7_fpmuld): Update required units and latency.
17332 (cortex_a7_fpmacd): Likewise.
17333 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
17334 (cortex_a7_neon). Likewise.
17335 (bypass) Update participating units.
17337 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17339 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
17340 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
17342 * config/arm/vfp11.md (vfp_farith): Use ffmas.
17343 (vfp_fmul): Use ffmad.
17344 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
17345 (cortex_r4_fmacd): Use ffmad.
17346 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
17347 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
17348 (cortex_a9_fmacd): Use ffmad.
17349 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
17350 (cortex_a8_vfp_macd): Use ffmad.
17351 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
17352 (cortex_a5_fpmacd): Use ffmad.
17353 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
17354 (cortex_a15_vfp_macd): Use ffmad.
17355 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
17357 2013-01-29 Jason Merrill <jason@redhat.com>
17360 * varasm.c (default_assemble_visibility): Don't warn about
17361 visibility on artificial decls.
17363 2013-01-29 Richard Biener <rguenther@suse.de>
17365 PR tree-optimization/56113
17366 * tree-ssa-structalias.c (equiv_class_lookup): Also return
17368 (label_visit): Free duplicate bitmaps and record the leader instead.
17369 (perform_var_substitution): Adjust.
17371 2013-01-29 Richard Biener <rguenther@suse.de>
17373 PR tree-optimization/55270
17374 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
17375 the CFG, schedule loops for fixup.
17377 2013-01-29 Nick Clifton <nickc@redhat.com>
17379 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
17382 2013-01-28 Leif Ekblad <leif@rdos.net>
17384 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
17385 * config/i386/i386.h (TARGET_RDOS): New macro.
17386 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
17387 * config/i386/i386.c (ix86_option_override_internal): For 64bit
17388 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
17389 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
17390 DEFAULT_LARGE_SECTION_THRESHOLD.
17391 * config/i386/i386.md (R14_REG, R15_REG): New constants.
17392 * config/i386/rdos.h: New file.
17393 * config/i386/rdos64.h: New file.
17395 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
17398 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
17401 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17403 PR rtl-optimization/56117
17404 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
17405 call cselib_lookup_from_insn on the MEM before calling
17406 add_insn_mem_dependence.
17408 2013-01-28 Richard Biener <rguenther@suse.de>
17410 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
17411 to a stmt that didn't have one.
17412 (copy_phis_for_bb): Likewise for PHI arguments.
17413 (copy_debug_stmt): Likewise for debug stmts.
17415 2013-01-28 Richard Biener <rguenther@suse.de>
17417 PR tree-optimization/56034
17418 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
17419 (partition_builtin_p): Adjust.
17420 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
17421 it is the last partition.
17422 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
17423 up the vertex for the definition.
17424 (classify_partition): Classify whether a partition is a
17425 PKIND_REDUCTION, thus has uses outside of the loop.
17426 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
17427 Merge all PKIND_REDUCTION partitions into the last partition.
17428 (tree_loop_distribution): Seed partitions from reductions as well.
17430 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17432 PR tree-optimization/56125
17433 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
17434 pow(x,c) into sqrt(x) * powi(x, n/2) or
17435 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
17436 optimizing for size.
17437 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
17438 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
17441 PR tree-optimization/56094
17442 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
17443 to UNKNOWN_LOCATION while gimplifying expr.
17445 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
17448 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
17449 operand 0 in movabs insn template for -masm=intel asm alternative.
17450 (*movabs<mode>_2): Ditto for operand 1.
17452 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
17455 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
17458 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17460 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
17463 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17465 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
17466 Undef to avoid warning.
17468 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17470 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
17471 * configure: Regenerate.
17473 2013-01-25 Jakub Jelinek <jakub@redhat.com>
17475 PR tree-optimization/56098
17476 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
17477 for stmts with volatile ops.
17478 (cond_store_replacement): Don't optimize if assign has volatile ops.
17479 (cond_if_else_store_replacement_1): Don't optimize if either
17480 then_assign or else_assign have volatile ops.
17481 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
17484 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17486 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
17488 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17490 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17491 missing ':' in asm example.
17493 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17495 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17496 entries into lane and laneq entries.
17497 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17498 Remove AdvSIMD scalar modes.
17499 (aarch64_sq<r>dmulh_laneq<mode>): New.
17500 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17502 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17503 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17504 * config/aarch64/iterators.md (VCOND): New.
17507 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17510 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17511 Add NULL LIBNAME argument to existing definitions.
17512 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17513 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17514 * config/avr/avr.c (DEF_BUILTIN): Same.
17515 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17516 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17517 implementation is available (DECL_ASSEMBLER_NAME is set).
17518 (avr_fold_absfx): New static function.
17519 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17520 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17521 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17522 AVR_BUILTIN_ABSLLK.
17523 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17524 (abshk, absk, abslk, absllk): Provide as static inline functions.
17526 2013-01-25 Marek Polacek <polacek@redhat.com>
17528 PR tree-optimization/56035
17529 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17531 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17533 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17534 (*movtf_internal_rex64): Add (!o,C) alternative
17535 (*movxf_internal_rex64): Ditto.
17536 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17538 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17540 * doc/invoke.texi: fix typo.
17541 * doc/objc.texi: fix typo.
17543 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17545 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17546 for the first two alternatives.
17548 2013-01-24 Diego Novillo <dnovillo@google.com>
17550 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17551 (ggc-zone.o): Remove.
17552 * configure.ac: Remove option --with-gc.
17553 * configure: Re-generate.
17554 * doc/install.texi: Remove documentation for --with-gc.
17555 * gengtype.c (write_enum_defn): Remove. Update all users.
17556 (write_Types_process_field): Remove generation of gt_e_* argument.
17557 (output_type_enum): Remove. Update all users.
17558 (write_enum_defn): Remove. Update all users.
17559 (enum alloc_zone): Remove. Update all users.
17560 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17561 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17562 Update all callers.
17563 (struct ptr_data): Remove field TYPE. Update all users.
17564 (gt_pch_note_object): Remove argument TYPE. Update all users.
17565 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17567 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17568 (struct alloc_zone): Remove.
17569 (ggc_internal_alloc_zone_stat): Remove.
17570 (ggc_internal_cleared_alloc_zone_stat): Remove.
17571 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17572 (ggc_pch_count_object): Remove last argument. Update all users.
17573 (ggc_pch_alloc_object): Remove last argument. Update all users.
17574 (struct alloc_zone): Remove.
17575 * ggc-zone.c: Remove.
17576 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17577 (struct alloc_zone): Remove.
17578 (ggc_alloc_typed_stat): Remove.
17579 (ggc_alloc_typed): Remove.
17580 (ggc_splay_alloc): Remove first argument.
17581 (rtl_zone): Remove. Update all users.
17582 (tree_zone): Remove. Update all users.
17583 (tree_id_zone): Remove. Update all users.
17584 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17585 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17586 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17587 * tree-ssanames.c: Remove references to zone allocator in comments.
17589 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17591 * config/avr/avr.c (avr_out_fract): Make register numbers that
17592 might be outside of source operand signed.
17594 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17596 * config/i386/constraints.md (Yf): New constraint.
17597 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17598 of f constraint to conditionaly disable x87 register preferences.
17599 (*movdf_internal): Ditto.
17600 (*movsf_internal): Ditto.
17602 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17604 PR inline-asm/55934
17605 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17606 that have operands with impossible constraints.
17607 Add a FIXME for a speed-up opportunity.
17608 * lra-constraints.c (process_alt_operands): Verify that a class
17609 selected from constraints on asms is valid for the operand mode.
17610 (curr_insn_transform): Remove incorrect comment.
17612 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17614 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17615 TOC operand is a valid symbol ref in the constant pool.
17617 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17619 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17621 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17624 * config/avr/stdfix.h: New file.
17625 * t-avr (stdfix-gcc.h): New rule to build it.
17626 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17628 2013-01-23 Kostya Serebryany <kcc@google.com>
17630 * config/darwin.h: remove dependency on
17631 CoreFoundation (asan on Mac OS).
17633 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17636 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17637 instead of cmpdi_operand for first comparison operand.
17638 Don't assert that comparison operands aren't both constants.
17640 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17642 * doc/install.texi (Downloading the Source): Update references to
17643 downloading separate components.
17645 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17647 * doc/extend.texi (__int128): Improve grammar.
17649 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17652 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17653 alternative to (o,r).
17654 (*movdi_internal_rex64): Remove (!o,n) alternative.
17655 (DImode immediate->memory splitter): Remove.
17656 (DImode immediate->memory peephole2): Remove.
17657 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17658 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17659 alternative to (!o,*r).
17660 (*movtf_internal_sse): New pattern.
17661 (*movxf_internal_rex64): New pattern.
17662 (*movxf_internal): Disable for TARGET_64BIT.
17663 (*movdf_internal_rex64): Remove (!o,F) alternative.
17665 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17667 PR middle-end/56074
17668 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17669 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17670 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17671 stmt locations where LOCATION_LOCUS of the stmt location is
17672 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17675 * config/i386/i386.md (UNSPEC_STOS): New.
17676 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17677 *strsetqi_1): Add UNSPEC_STOS.
17679 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17682 * doc/invoke.texi: Remove left over -Wsynth example.
17684 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17686 PR tree-optimization/56051
17687 * fold-const.c (fold_binary_loc): Don't fold
17688 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17689 a narrowing conversion, or widening conversion from signed
17692 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17694 PR rtl-optimization/56023
17695 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17696 dependent on debug instruction.
17698 2013-01-21 Martin Jambor <mjambor@suse.cz>
17700 PR middle-end/56022
17701 * function.c (allocate_struct_function): Call
17702 invoke_set_current_function_hook earlier.
17704 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17706 * reload1.c (init_reload): Only initialize reload_obstack
17707 during the first call.
17709 2013-01-21 Marek Polacek <polacek@redhat.com>
17711 * cfgloop.c (verify_loop_structure): Fix up grammar.
17713 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17715 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17716 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17718 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17721 * config/arm/marvell-pj4.md: Update copyright year.
17722 Fix up use of alu to alu_reg and simple_alu_imm.
17724 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17726 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17728 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17731 * lra-constraints.c (curr_insn_transform): Don't reuse original
17732 insn for secondary memory move when memory mode should be different.
17734 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17736 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17737 atomic_storedi_1): New patterns.
17739 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17741 btver2 pipeline descriptions.
17742 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17744 * config/i386/i386.md (btver2_decode): New type attributes.
17745 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17747 * config/i386/btver2.md: New file describing btver2 pipelines.
17749 2013-01-19 Andrew Pinski <apinski@cavium.com>
17751 PR tree-optimization/52631
17752 * tree-ssa-sccvn (visit_use): Before looking up the original
17753 statement, try looking up the simplified expression.
17755 2013-01-19 Anthony Green <green@moxielogic.com>
17757 * config/moxie/moxie.c (moxie_expand_prologue): Set
17758 current_function_static_stack_size.
17760 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17762 PR tree-optimization/56029
17763 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17764 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17766 2013-01-18 Sharad Singhai <singhai@google.com>
17768 PR tree-optimization/55995
17769 * dumpfile.c (dump_loc): Print location only if available.
17770 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17772 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17775 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17776 secondary memory move.
17777 (inherit_reload_reg): Use rclass instead of cl for
17778 check_secondary_memory_needed_p.
17780 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17782 PR middle-end/56015
17783 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17784 the case where writing real complex part of target modifies op1.
17786 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17788 * config/aarch64/aarch64-simd.md
17789 (aarch64_vcond_internal<mode>): Handle unordered cases.
17790 * config/aarch64/iterators.md (v_cmp_result): New.
17792 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17793 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17795 * config/arm/marvell-pj4.md: New file.
17796 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17797 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17798 (generic_vfp): Likewise.
17799 * config/arm/arm-cores.def: Add marvell-pj4.
17800 * config/arm/arm-tune.md: Regenerate.
17801 * config/arm/arm-tables.opt: Regenerate.
17802 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17803 * doc/invoke.texi: Document marvell-pj4.
17805 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17807 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17809 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17814 * var-tracking.c (negative_power_of_two_p): New.
17815 (global_get_addr_cache, local_get_addr_cache): New.
17816 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17817 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17819 (vt_stack_offset_p): Remove.
17820 (vt_canon_true_dep): Always canonicalize loc's address.
17821 (clobber_overlapping_mems): Make sure we have a MEM.
17822 (local_get_addr_clear_given_value): New.
17823 (val_reset): Clear local cached entries.
17824 (compute_bb_dataflow): Create and release the local cache.
17825 Disable duplicate MEMs clobbering.
17826 (emit_notes_in_bb): Clobber MEMs likewise.
17827 (vt_emit_notes): Create and release the local cache.
17828 (vt_initialize, vt_finalize): Create and release the global
17829 cache, respectively.
17830 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17832 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17834 PR libmudflap/53359
17835 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17836 not found in the symtab.
17838 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17841 PR rtl-optimization/55547
17842 PR rtl-optimization/53827
17845 * alias.c (offset_overlap_p): New, factored out of...
17846 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17847 the conservative special case for symbolic constants. Don't
17848 adjust zero sizes on alignment.
17850 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17852 PR rtl-optimization/52573
17853 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17854 REG_UNUSED for the same register.
17856 2013-01-17 Richard Biener <rguenther@suse.de>
17857 Marek Polacek <polacek@redhat.com>
17859 PR rtl-optimization/55833
17860 * loop-unswitch.c (unswitch_loops): Move loop verification...
17861 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17862 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17863 Set it to true when we're removing a loop from hierarchy tree in
17864 an irreducible region.
17865 (fix_bb_placements): Adjust caller.
17866 (fix_loop_placements): Likewise.
17868 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17870 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17871 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17872 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17873 Remove ID. Adjust comments.
17874 * config/avr/avr-c.c (avr_builtin_name): Remove.
17875 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17876 * config/avr/avr.c (avr_tolower): New static function.
17877 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17878 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17879 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17882 2013-01-17 Jan Hubicka <jh@suse.cz>
17884 PR tree-optimization/55273
17885 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17887 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17890 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17891 store through atomic_store<mode>_1.
17892 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17894 2013-01-17 Martin Jambor <mjambor@suse.cz>
17896 PR tree-optimizations/55264
17897 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17898 for virtual methods.
17899 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17900 virtual methods before inlining is over.
17901 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17903 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17906 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17908 PR rtl-optimization/56005
17909 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17910 pending reads for prefetch.
17912 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17914 * config/aarch64/aarch64.md
17915 (*cstoresi_neg_uxtw): New pattern.
17916 (*cmovsi_insn_uxtw): New pattern.
17917 (*<optab>si3_uxtw): New pattern.
17918 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17919 (*<optab>si3_insn_uxtw): New pattern.
17920 (*bswapsi2_uxtw): New pattern.
17922 2013-01-16 Richard Biener <rguenther@suse.de>
17924 * tree-inline.c (tree_function_versioning): Remove set but
17925 never used variable.
17927 2013-01-16 Richard Biener <rguenther@suse.de>
17929 PR tree-optimization/55964
17930 * tree-flow.h (rename_variables_in_loop): Remove.
17931 (rename_variables_in_bb): Likewise.
17932 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17933 (copy_loop_before): Adjust and delete update-ssa status.
17934 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17935 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17936 (rename_variables_in_loop): Remove.
17937 (slpeel_update_phis_for_duplicate_loop): Likewise.
17938 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17939 use available cfg machinery instead of duplicating it.
17940 Update PHI nodes and perform poor-mans SSA update here.
17941 (slpeel_tree_peel_loop_to_edge): Adjust.
17943 2013-01-16 Richard Biener <rguenther@suse.de>
17945 PR tree-optimization/54767
17946 PR tree-optimization/53465
17947 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17948 (vrp_visit_phi_node): For PHI arguments coming via backedges
17949 drop all symbolical range information.
17950 (execute_vrp): Compute backedges.
17952 2013-01-16 Richard Biener <rguenther@suse.de>
17954 * doc/install.texi: Update CLooG and ISL requirements to
17957 2013-01-16 Christian Bruel <christian.bruel@st.com>
17960 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17961 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17962 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17964 2013-01-16 DJ Delorie <dj@redhat.com>
17966 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17967 (UNSPECV_SP_SWITCH_E): New.
17968 (sp_switch_1): Change to an unspec.
17969 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17972 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17974 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17975 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17976 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17977 (expand_mem_thread_fence): Ditto.
17978 (expand_mem_signal_fence): Ditto.
17979 (expand_atomic_load): Ditto.
17980 (expand_atomic_store): Ditto.
17982 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17984 PR rtl-optimization/55547
17985 PR rtl-optimization/53827
17988 * alias.c (memrefs_conflict_p): Set sizes to negative after
17991 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17994 * function.c (thread_prologue_and_epilogue_insns): Always
17995 add crtl->drap_reg to set_up_by_prologue.set, even if
17996 stack_realign_drap is false.
17998 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18000 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18001 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18002 *call): Fix indention.
18004 2013-01-15 Tom de Vries <tom@codesourcery.com>
18007 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18010 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18012 PR rtl-optimization/55153
18013 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18015 2013-01-15 Martin Jambor <mjambor@suse.cz>
18017 PR tree-optimization/55920
18018 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18019 accesses as grp_to_be_debug_replaced.
18021 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18023 PR tree-optimization/55920
18024 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18025 there is non-useless type conversion needed from debug rhs to lhs,
18026 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18028 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18029 Mikael Pettersson <mikpe@it.uu.se>
18032 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18034 (ASM_OUTPUT_CASE_LABEL): Remove.
18035 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18036 * final.c (shorten_branches): Update alignment of labels before
18037 jump tables if CASE_VECTOR_SHORTEN_MODE.
18039 2013-01-15 Richard Biener <rguenther@suse.de>
18042 * system.h: Do not include gmp.h for building host tools.
18044 2013-01-15 Richard Biener <rguenther@suse.de>
18046 PR middle-end/55882
18047 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18048 account for bitpos when computing alignment.
18050 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18052 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18053 (ix86_target_macros_internal): Likewise.
18055 * config/i386/i386.c (m_CORE2I7): Removed.
18056 (m_CORE_HASWELL): New macro.
18057 (m_CORE_ALL): Likewise.
18058 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18059 (initial_ix86_arch_features): Likewise.
18060 (processor_target_table): Initializations for Core avx2.
18061 (cpu_names): New names "core-avx2".
18062 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18063 PROCESSOR_CORE_HASWELL.
18064 (ix86_issue_rate): New case.
18065 (ia32_multipass_dfa_lookahead): Likewise.
18066 (ix86_sched_init_global): Likewise.
18068 * config/i386/i386.h (TARGET_HASWELL): New macro.
18069 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18070 (processor_type): New PROCESSOR_HASWELL.
18072 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18074 PR tree-optimization/55955
18075 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18076 *SHIFT_EXPR and *ROTATE_EXPR codes.
18078 PR tree-optimization/48766
18079 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18080 -ftrapv disable -fwrapv.
18082 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18085 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18086 etc. to 1 and not to __flash.
18087 Use LL suffix for __INT24_MAX__ with -mint8.
18088 Use ULL suffix for __UINT24_MAX__ with -mint8.
18090 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18092 * config/avr/avr-arch.h
18093 (struct base_arch_s): Use typedef avr_arch_t instead.
18094 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18095 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18096 * config/avr/avr.c: Same.
18097 * config/avr/avr-devices.c: Same.
18098 * config/avr/driver-avr.c: Same.
18099 * config/avr/gen-avr-mmcu-texi.c: Same.
18100 * config/avr/avr-mcus.def: Adjust comment.
18102 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18104 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18105 * config/aarch64/iterators.md (VALLDI): New.
18107 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18108 Andi Kleen <ak@linux.intel.com>
18111 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18112 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18115 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18117 * config/avr/avr-stdint.h: Remove trailing blanks.
18118 * config/avr/avr-log.h: Same.
18119 * config/avr/avr-arch.h: Same.
18120 * config/avr/avr-devices.c: Same.
18121 * config/avr/avr-dimode.md: Same.
18122 * config/avr/predicates.md: Same.
18123 * config/avr/avr-c.c: Same. And fix typo.
18125 * config/avr/avr-protos.h: Same. And:
18126 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18127 (init_cumulative_args): Rename to avr_init_cumulative_args.
18128 (expand_prologue): Rename to avr_expand_prologue.
18129 (expand_epilogue): Rename to avr_expand_epilogue.
18130 (adjust_insn_length): Rename to avr_adjust_insn_length.
18131 (notice_update_cc): Rename to avr_notice_update_cc.
18132 (final_prescan_insn): Rename to avr_final_prescan_insn.
18133 * config/avr/avr.c: Same.
18134 * config/avr/avr.h: Same.
18135 * config/avr/avr.md: Remove trailing blanks.
18136 (prologue): Use avr_expand_prologue.
18137 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18139 2013-01-14 Richard Biener <rguenther@suse.de>
18141 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18142 verify_location, collect_subblocks): New functions.
18143 (verify_gimple_in_cfg): Verify that locations only reference
18144 BLOCKs in the functions BLOCK tree.
18146 2013-01-14 Richard Biener <rguenther@suse.de>
18148 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18150 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18152 (insert_out_of_ssa_copy_on_edge): Likewise.
18153 (rewrite_close_phi_out_of_ssa): Likewise.
18154 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18156 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18157 propagated constants.
18158 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18161 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18163 * config/avr/avr-modes.def: Add GPL copyright notice.
18165 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18167 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18168 MEMMODEL_MASK to determine memory model.
18169 (atomic_store<mode>): Ditto from operands[2].
18170 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18172 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18175 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18176 (fold_gimple_assign): Don't call unshare_expr here.
18177 (fold_ctor_reference): Call unshare_expr.
18179 2013-01-13 Terry Guo <terry.guo@arm.com>
18181 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18182 * doc/fragments.texi: Document MULTILIB_REUSE.
18183 * gcc.c (multilib_reuse): New internal spec.
18184 (set_multilib_dir): Also search multilib from multilib_reuse.
18185 * genmultilib (tmpmultilib3): Refactor code.
18186 (tmpmultilib4): Ditto.
18187 (multilib_reuse): New multilib argument.
18189 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18191 * Makefile.in: Update copyright.
18193 2013-01-12 Tom de Vries <tom@codesourcery.com>
18195 PR middle-end/55890
18196 * calls.c (expand_call): Check if arg_nr is valid.
18198 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18200 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18201 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18202 documentation. Add missing '__' in front of
18203 __builtin_ia32_packssdw256.
18205 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18208 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18209 NO_REGS for larl operands.
18210 (s390_reload_larl_operand): Use s390_load_address instead of
18213 2013-01-11 Richard Biener <rguenther@suse.de>
18215 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18216 (verify_node_sharing): ... here.
18217 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18219 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18221 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18222 Substitute TREECHECKING.
18223 * configure: Regenerate.
18224 * Makefile.in (TREECHECKING): New.
18226 2013-01-11 Richard Guenther <rguenther@suse.de>
18228 PR tree-optimization/44061
18229 * tree-vrp.c (extract_range_basic): Compute zero as
18230 value-range for __builtin_constant_p of function parameters.
18232 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
18234 Update copyright years.
18236 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
18238 PR rtl-optimization/55672
18239 * lra-eliminations.c (mark_not_eliminable): Permit addition with
18240 const to be eliminable.
18242 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
18244 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
18245 * configure: Regenerate.
18247 2013-01-10 Richard Biener <rguenther@suse.de>
18249 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
18251 2013-01-10 Richard Biener <rguenther@suse.de>
18254 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
18255 locations for virtual PHI arguments.
18256 (rewrite_update_phi_arguments): Likewise.
18258 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
18260 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
18263 2013-01-10 Jakub Jelinek <jakub@redhat.com>
18265 PR tree-optimization/55921
18266 * tree-complex.c (expand_complex_asm): New function.
18267 (expand_complex_operations_1): Call it for GIMPLE_ASM.
18269 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18272 * config/s390/s390.c (s390_symref_operand_p)
18273 (s390_loadrelative_operand_p): Merge the two functions.
18274 (s390_check_qrst_address, print_operand_address): Add parameters
18275 to s390_loadrelative_operand_p invokation.
18276 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
18277 (s390_reload_larl_operand, s390_secondary_reload): Use
18278 s390_loadrelative_operand_p instead of s390_symref_operand_p.
18279 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
18281 2013-01-09 Mike Stump <mikestump@comcast.net>
18283 * dse.c (record_store): Remove unnecessary assert.
18285 2013-01-09 Jan Hubicka <jh@suse.cz>
18287 PR tree-optimization/55569
18288 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
18289 * cfgloop.h (scale_loop_profile): Likewise.
18291 2013-01-09 Jan Hubicka <jh@suse.cz>
18294 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
18296 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
18298 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
18300 PR middle-end/55114
18301 * expr.h (maybe_emit_group_store): Declare.
18302 * expr.c (maybe_emit_group_store): New function.
18303 * builtins.c (expand_builtin_int_roundingfn): Call it.
18304 (expand_builtin_int_roundingfn_2): Likewise.
18306 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
18308 PR rtl-optimization/55829
18309 * lra-constraints.c (match_reload): Add code for absent output.
18310 (curr_insn_transform): Add code for reloads of matched inputs
18313 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
18315 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
18316 attribute of movddup insn to DF.
18317 (*vec_interleave_lowv2df): Ditto.
18318 (vec_dupv2df): Ditto.
18320 2013-01-09 Jan Hubicka <jh@suse.cz>
18322 PR tree-optimiation/55875
18323 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
18324 EVERY_ITERATION parameter.
18325 (number_of_iterations_exit): Check if exit is executed every iteration.
18326 (idx_infer_loop_bounds): Similarly here.
18327 (n_of_executions_at_most): Simplify
18328 to only test for cases where statement is dominated by the
18329 particular bound; handle correctly the "postdominance" test.
18330 (scev_probably_wraps_p): Use max loop iterations info
18331 as a global bound first.
18333 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
18334 Nick Clifton <nickc@redhat.com>
18336 * config/v850/v850.md (cbranchsf4): New pattern.
18337 (cstoresf4): New pattern.
18338 (cbranchdf4): New pattern.
18339 (cstoredf4): New pattern.
18340 (movsicc): Disallow floating point comparisons.
18341 (cmpsf_le_insn): Fix order of operators.
18342 (cmpsf_lt_insn): Likewise.
18343 (cmpsf_eq_insn): Likewise.
18344 (cmpdf_le_insn): Likewise.
18345 (cmpdf_lt_insn): Likewise.
18346 (cmpdf_eq_insn): Likewise.
18347 (cmpsf_ge_insn): Use LE comparison.
18348 (cmpdf_ge_insn): Likewise.
18349 (cmpsf_gt_insn): Use LT comparison.
18350 (cmpdf_gt_insn): Likewise.
18351 (cmpsf_ne_insn): Delete pattern.
18352 (cmpdf_ne_insn): Delete pattern.
18353 * config/v850/v850.c (v850_gen_float_compare): Use
18354 gen_cmpdf_eq_insn for NE comparison.
18355 (v850_float_z_comparison_operator)
18356 (v850_float_nz_comparison_operator): Move from here ...
18357 * config/v850/predicates.md: ... to here. Move GT and GE
18358 comparisons into v850_float_z_comparison_operator.
18359 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
18361 (v850_float_nz_comparison_operator): Likewise.
18363 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18365 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
18366 with calls to gen_insvsi/gen_insvdi.
18368 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18370 * config/i386/i386.c (initial_ix86_tune_features): Set up
18371 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
18373 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
18374 Jakub Jelinek <jakub@redhat.com>
18376 PR tree-optimization/48189
18377 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
18378 If nitercst is 0, don't predict the exit edge.
18380 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18382 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
18383 in asm_fprintf with reg_names.
18384 (aarch64_print_operand_address): Likewise.
18385 (aarch64_return_addr): Likewise.
18386 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
18388 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18390 * config/pa/pa.h (VAL_U6_BITS_P): Define.
18391 (INT_U6_BITS): Likewise.
18392 * config/pa/predicates.md (uint6_operand): New predicate.
18393 (shift5_operand, shift6_operand): Likewise.
18394 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
18396 (lshrdi3): Use shift6_operand.
18397 (shrpsi4, shrpdi4): New insn patterns.
18398 (extzv): Delete expander.
18399 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
18400 predicates in unamed zero extract patterns. Tighten common constraint.
18401 (extv): Delete expander.
18402 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
18403 predicates in unamed sign extract patterns. Tighten common constraint.
18404 (insv): Delete expander.
18405 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
18406 predicates in unamed insert patterns. Tighten common constraint.
18407 Change uint32_operand predicate to uint6_operand predicate in unamed
18408 DImode pattern to insert constant values of type 1...1xxxx.
18410 2013-01-04 Jan Hubicka <jh@suse.cz>
18412 PR tree-optimization/55823
18413 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
18416 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18417 Uros Bizjak <ubizjak@gmail.com>
18419 PR rtl-optimization/55845
18420 * df-problems.c (can_move_insns_across): Stop scanning at
18421 volatile_insn_p source instruction or give up if
18422 across_from .. across_to range contains any volatile_insn_p
18425 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18427 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
18428 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
18430 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
18431 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
18433 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18436 * asan.c (asan_clear_shadow): New function.
18437 (asan_emit_stack_protection): Use it.
18439 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18441 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
18442 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
18443 with tab instead of space.
18445 2013-01-08 Nick Clifton <nickc@redhat.com>
18447 * config/rl78/rl78.c (rl78_expand_prologue): Always select
18448 register bank 0 at the start of an interrupt handler.
18449 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
18452 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18454 * config/aarch64/aarch64-simd.md
18455 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
18456 (aarch64_simd_bsl): Likewise.
18457 (aarch64_vcond_internal<mode>): Likewise.
18458 (vcond<mode><mode>): Likewise.
18459 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
18460 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
18462 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18464 * config/aarch64/aarch64-builtins.c
18465 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
18467 2013-01-08 Martin Jambor <mjambor@suse.cz>
18470 * tree-sra.c (analyze_access_subtree): Return true also after
18471 potentially creating a debug-only replacement.
18473 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18475 PR middle-end/55890
18476 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
18478 PR tree-optimization/54120
18479 * tree-vrp.c (range_fits_type_p): Don't allow
18480 src_precision < precision from signed vr to unsigned_p
18481 if vr->min or vr->max is negative.
18482 (simplify_float_conversion_using_ranges): Test can_float_p
18483 against CODE_FOR_nothing.
18485 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18486 Richard Biener <rguenther@suse.de>
18488 PR middle-end/55851
18489 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
18490 types instead of just INTEGER_TYPE types.
18492 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18494 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18497 2013-01-07 Steve Ellcey <sellcey@mips.com>
18500 * config/mips/mips.opt: Change mad to mmad to match documentation.
18502 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18505 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18506 .progmemx.data now.
18508 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18511 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18512 (avr_addrspace_t): Add .section_name field.
18513 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18515 (avr_addrspace): Same. Initialize .section_name. Remove last
18516 NULL entry. Put __memx into .progmemx.data.
18517 (progmem_section_prefix): Remove.
18518 (avr_asm_init_sections): No need to initialize progmem_section.
18519 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18520 section name prefix.
18521 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18522 retrieve the progmem section.
18523 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18524 boundary to run over avr_addrspace[].
18525 (avr_register_target_pragmas): Ditto.
18527 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18529 * varasm.c (output_constant_def_contents): For asan_protect_global
18530 protected strings, adjust DECL_ALIGN if needed, before testing for
18532 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18533 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18535 (output_object_block): For asan protected decls, emit asan padding
18536 after their contents.
18537 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18538 (asan_finish_file): Test it here instead.
18540 2013-01-07 Nick Clifton <nickc@redhat.com>
18541 Matthias Klose <doko@debian.org>
18542 Doug Kwan <dougkwan@google.com>
18543 H.J. Lu <hongjiu.lu@intel.com>
18546 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18548 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18550 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18552 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18554 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18556 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18559 * doc/install.texi (Cross-Compiler-Specific Options): Document
18562 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18564 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18565 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18566 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18567 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18568 vqmovun_high_s64): Fix source operand number and update copyright.
18570 2013-01-07 Richard Biener <rguenther@suse.de>
18572 PR middle-end/55890
18573 * gimple.h (gimple_call_builtin_p): New overload.
18574 * gimple.c (validate_call): New function.
18575 (gimple_call_builtin_p): Likewise.
18576 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18577 Use gimple_call_builtin_p.
18578 (find_func_clobbers): Likewise.
18579 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18580 (strlen_optimize_stmt): Likewise.
18582 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18584 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18585 (vld1q_dup_*): Likewise.
18586 (vld1_*): Likewise.
18587 (vld1q_*): Likewise.
18588 (vld1_lane_*): Likewise.
18589 (vld1q_lane_*): Likewise.
18591 2013-01-07 Richard Biener <rguenther@suse.de>
18593 * lto-streamer.h (LTO_minor_version): Bump to 2.
18595 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18597 * config/aarch64/aarch64-protos.h
18598 (aarch64_const_double_zero_rtx_p): Rename to...
18599 (aarch64_float_const_zero_rtx_p): ...this.
18600 (aarch64_float_const_representable_p): New.
18601 (aarch64_output_simd_mov_immediate): Likewise.
18602 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18603 move immediate case.
18604 * config/aarch64/aarch64.c
18605 (aarch64_const_double_zero_rtx_p): Rename to...
18606 (aarch64_float_const_zero_rtx_p): ...this.
18607 (aarch64_print_operand): Allow printing of new constants.
18608 (aarch64_valid_floating_const): New.
18609 (aarch64_legitimate_constant_p): Check for valid floating-point
18611 (aarch64_simd_valid_immediate): Likewise.
18612 (aarch64_vect_float_const_representable_p): New.
18613 (aarch64_float_const_representable_p): Likewise.
18614 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18615 (aarch64_output_simd_mov_immediate): New.
18616 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18617 (*movdf_aarch64): Likewise.
18618 * config/aarch64/constraints.md (Ufc): New.
18619 (Y): call aarch64_float_const_zero_rtx.
18620 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18622 2013-01-07 Richard Biener <rguenther@suse.de>
18624 PR tree-optimization/55888
18625 PR tree-optimization/55862
18626 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18627 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18628 not if it is contained therein.
18630 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18632 * config/avr/t-avr: Typo.
18634 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18637 * config/avr/t-avr: Don't automatically rebuild
18638 $(srcdir)/config/avr/t-multilib
18639 $(srcdir)/config/avr/avr-tables.opt
18640 $(srcdir)/doc/avr-mmcu.texi
18641 (avr-mcus): New phony target to build them on request.
18642 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18643 * avr/avr-mcus.def: Adjust comments.
18645 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18647 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18649 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18651 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18653 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18655 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18657 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18659 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18660 to generate profiling.
18661 * config/rs6000/aix64.h (LIB_SPEC): Same.
18663 2013-01-04 Andrew Pinski <apinski@cavium.com>
18665 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18667 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18669 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18671 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18673 (ix86_expand_move): Ditto.
18674 (ix86_zero_extend_to_Pmode): Ditto.
18675 (ix86_expand_call): Ditto.
18676 (ix86_expand_special_args_builtin): Ditto.
18677 (ix86_expand_builtin): Ditto.
18679 2013-01-04 Richard Biener <rguenther@suse.de>
18681 PR tree-optimization/55862
18682 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18683 translating them through PHI nodes.
18685 2013-01-04 Martin Jambor <mjambor@suse.cz>
18687 PR tree-optimization/55755
18688 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18689 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18691 2013-01-04 Marek Polacek <polacek@redhat.com>
18693 PR middle-end/55859
18694 * opts.c (default_options_optimization): Clarify error message.
18696 2013-01-04 Richard Biener <rguenther@suse.de>
18698 PR middle-end/55863
18699 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18702 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18705 * config/pa/pa.md (movsi): Revert previous change.
18706 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18709 2013-01-03 Richard Henderson <rth@redhat.com>
18711 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18712 after eliminating TLS symbols.
18714 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18717 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18718 * graphite-poly.c (debug_gmp_value): Likewise.
18720 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18723 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18724 selected code model, define __code_mode_small__, __code_model_medium__,
18725 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18726 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18727 xchg temporary register with %k. Declare temporary register as
18729 [__x86_64__]: For medium and large code models, preserve %rbx register.
18731 2013-01-03 Richard Biener <rguenther@suse.de>
18733 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18734 (dump_subscript): Adjust.
18735 (finalize_ddr_dependent): Do not dump redundant info.
18736 (analyze_siv_subscript): Adjust.
18737 (subscript_dependence_tester): Likewise.
18738 (compute_affine_dependence): Likewise.
18740 2013-01-03 Richard Biener <rguenther@suse.de>
18743 2013-01-03 Richard Biener <rguenther@suse.de>
18745 PR tree-optimization/55857
18746 * tree-vect-stmts.c (vectorizable_load): Do not setup
18747 re-alignment for invariant loads.
18749 2013-01-02 Richard Biener <rguenther@suse.de>
18751 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18752 invariant load do not generate a vector load from the scalar location.
18754 2013-01-03 Richard Biener <rguenther@suse.de>
18756 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18757 for not vectorizing.
18758 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18759 not build INDIRECT_REFs, call get_name once only.
18760 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18761 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18763 2013-01-03 Richard Biener <rguenther@suse.de>
18765 PR tree-optimization/55857
18766 * tree-vect-stmts.c (vectorizable_load): Do not setup
18767 re-alignment for invariant loads.
18769 2013-01-03 Richard Biener <rguenther@suse.de>
18772 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18773 prefer a built-in decl.
18775 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18777 * gcc.c (process_command): Update copyright notice dates.
18778 * gcov.c (print_version): Likewise.
18779 * gcov-dump.c (print_version): Likewise.
18781 PR rtl-optimization/55838
18782 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18783 iv0.step, iv1.step and step.
18785 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18786 Marc Glisse <marc.glisse@inria.fr>
18788 PR tree-optimization/55832
18789 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18790 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18791 integer_{one,zero}_node.
18793 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18796 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18797 * var-tracking.c (reverse_op): Don't add reverse ops to
18798 VALUEs that have already
18799 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18801 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18803 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18805 2013-01-02 Teresa Johnson <tejohnson@google.com>
18807 * dumpfile.c (dump_loc): Print filename with location.
18808 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18809 new location_t parameter to emit complete unroll message with
18810 new dump framework.
18811 (canonicalize_loop_induction_variables): Compute loops location
18812 and pass to try_unroll_loop_completely.
18813 * loop-unroll.c (report_unroll_peel): New function.
18814 (peel_loops_completely): Use new dump format with location
18815 for main dumpfile message, and invoke report_unroll_peel on success.
18816 (decide_unrolling_and_peeling): Ditto.
18817 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18818 by report_unroll_peel.
18819 (decide_peel_completely): Ditto.
18820 (decide_unroll_constant_iterations): Ditto.
18821 (decide_unroll_runtime_iterations): Ditto.
18822 (decide_peel_simple): Ditto.
18823 (decide_unroll_stupid): Ditto.
18824 * cfgloop.c (get_loop_location): New function.
18825 * cfgloop.h (get_loop_location): Declare.
18827 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18829 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18832 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18834 PR middle-end/55198
18835 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18836 BLKmode objects when EXPAND_MEMORY is specified.
18838 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18840 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18842 (fold_builtin_cpu): Do not share cpu model decls across statements.
18844 2013-01-02 Jason Merrill <jason@redhat.com>
18847 * tree.c (build_array_type_1): Revert earlier change.
18849 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18851 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18853 * config/aarch64/aarch64-tune.md: Re-generate.
18855 2013-01-02 Richard Biener <rguenther@suse.de>
18857 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18858 invariant load do not generate a vector load from the scalar location.
18860 2013-01-02 Richard Biener <rguenther@suse.de>
18863 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18864 * configure: Regenerate.
18866 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18868 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18869 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18870 (expand_builtin_int_roundingfn_2): Keep the original target around
18871 for the fallback case.
18873 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18875 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18876 to be clear for sign changes.
18878 2013-01-01 Jan Hubicka <jh@suse.cz>
18880 * ipa-inline-analysis.c: Fix formatting.
18882 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18884 PR tree-optimization/55831
18885 * tree-vect-loop.c (get_initial_def_for_induction): Use
18886 gsi_after_labels instead of gsi_start_bb.
18888 Copyright (C) 2013 Free Software Foundation, Inc.
18890 Copying and distribution of this file, with or without modification,
18891 are permitted in any medium without royalty provided the copyright
18892 notice and this notice are preserved.