1 2013-08-26 Jan Hubicka <jh@suse.cz>
3 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4 methods can not be called indirectly when their address is not taken.
6 2013-08-26 Jan Hubicka <jh@suse.cz>
8 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use ctor_for_folding.
10 2013-08-26 Jan Hubicka <jh@suse.cz>
12 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
15 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
17 * reload.c (find_valid_class): Allow classes that do not include
18 FIRST_PSEUDO_REGISTER - 1.
20 2013-08-26 Jan Hubicka <jh@suse.cz>
22 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
23 fix edge count/frequency when speculation failed; fix type check
26 2013-08-26 Jan Hubicka <jh@suse.cz>
28 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
30 2013-08-26 Jan Hubicka <jh@suse.cz>
32 * ipa-inline-transform.c (inline_transform): Be ready for basic block
33 to be changed by edge redirection.
35 2013-08-26 Jan Hubicka <jh@suse.cz>
37 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and formating;
39 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
40 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
41 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
42 (resolve_noninline_speculation): Update callee keys, too.
44 2013-08-26 Jan Hubicka <jh@suse.cz>
46 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
47 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
49 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
51 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
54 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
56 * config/i386/i386.c (ix86_debug_options): Remove prototype.
57 (x86_64_elf_select_section): Ditto.
58 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
60 (ix86_pass_by_reference): Ditto.
61 (output_set_got): Ditto.
62 (ix86_unary_operator_ok): Ditto.
63 (ix86_expand_builtin): Ditto.
65 2013-08-23 Jan Hubicka <jh@suse.cz>
67 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
69 2013-08-23 Jan Hubicka <jh@suse.cz>
71 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
72 (tree_decl_with_vis): Add FINAL field.
74 2013-08-23 Jeff Law <law@redhat.com>
76 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
77 the debugging dump when the expression is fully redundant.
79 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
81 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
82 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
83 * pretty-print.c (pp_formatted_text_data): Likewise.
84 (pp_write_text_to_stream): Likewise.
85 (pp_write_text_as_dot_label_to_stream): Likewise.
86 (pp_append_r): Likewise.
87 (pp_format): Likewise.
89 (pp_clear_output_area): Likewise.
90 (pp_append_text): Likewise.
91 (pp_formatted_text): Likewise.
92 (pp_remaining_character_count_for_line): Likewise.
93 (pp_newline): Likewise.
94 (pp_character): Likewise.
95 (output_buffer::~output_buffer): Define.
96 (pretty_printer::~pretty_printer): Destruct output buffer.
97 * pretty-print.h (output_buffer::~output_buffer): Declare.
98 (pretty_printer::~pretty_printer): Declare virtual.
100 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
103 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
104 HOST_WIDE_INT_M1U): New macros.
105 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
106 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
107 unsigned -1 for lshift.
108 * cse.c (cse_insn): Likewise.
109 * double-int.c (rshift_double, lshift_double): Likewise.
110 * builtins.c (fold_builtin_bitop): Likewise.
111 * combine.c (force_to_mode): Likewise.
112 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
113 * simplify-rtx.c (simplify_const_unary_operation,
114 simplify_const_binary_operation): Likewise.
115 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
116 check_va_list_escapes): Likewise.
117 * rtlanal.c (nonzero_bits1): Likewise.
118 * expmed.c (expand_smod_pow2): Likewise.
119 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
121 2013-08-23 Jan Hubicka <jh@suse.cz>
123 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
124 as having address taken.
126 2013-08-23 Jan Hubicka <jh@suse.cz>
128 * ipa-utils.h (method_class_type): Declare.
129 * ipa-devirt.c (method_class_type): Export.
131 * cgraphunit.c (analyze_functions): Do basic devirtualization;
132 do not walk base classes of anonymous types.
134 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
136 PR rtl-optimization/58220
138 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
139 handle SEQUENCE insns properly.
141 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
143 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
145 (pp_newline_and_indent): Likewise.
146 (pp_separate_with): Likewise.
147 * pretty-print.c (pp_newline_and_flush): Define.
148 (pp_newline_and_indent): Likewise.
149 (pp_separate_with): Likewise.
151 2013-08-23 Jakub Jelinek <jakub@redhat.com>
154 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
155 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
157 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
159 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
160 * gcc/config/i386/i386.md (*movti_internal): Use
161 predicate to determine if EVEX is needed.
162 (*movsi_internal): Ditto.
163 (*movdf_internal): Ditto.
164 (*movsf_internal): Ditto.
165 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
167 2013-08-23 Jakub Jelinek <jakub@redhat.com>
169 PR tree-optimization/58209
170 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
171 (find_tail_calls): Give up for pointer result types if m is non-NULL.
172 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
173 emit POINTER_PLUS_EXPR.
174 (create_tailcall_accumulator): For pointer result type accumulate in
177 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
179 * configure.ac: Add backslashes missing from the last change.
180 * configure: Regenerate.
182 2013-08-22 Jan Hubicka <jh@susue.cz>
184 * ipa.c (function_and_variable_visibility): First remember function
185 was global and then make it local.
187 2013-08-22 Julian Brown <julian@codesourcery.com>
189 * configure.ac: Add aarch64 to list of arches which use "nop" in
191 * configure: Regenerate.
193 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
195 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
196 gnu_libc_has_function.
197 * config/s390/tpf.h: Likewise.
199 2013-08-22 Jan Hubicka <jh@susue.cz>
201 * timevar.c (validate_phases): Add cast.
203 2013-08-22 Jan Hubicka <jh@susue.cz>
205 * timevar.c (validate_phases): Use size_t for memory.
206 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
208 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
210 * pretty-print.h (output_buffer::output_buffer): Declare.
211 (pretty_printer::pretty_printer): Likewise.
212 (pp_construct): Remove.
213 * pretty-print.c (output_buffer::output_buffer): Define.
214 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
215 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
216 (print_gimple_expr): Likewise.
217 (print_gimple_seq): Likewise.
218 (gimple_dump_bb): Likewise.
219 * sched-vis.c (dump_value_slim): Likewise.
220 (dump_insn_slim): Likewise.
221 (dump_rtl_slim): Likewise.
222 (str_pattern_slim): Likewise.
223 * tree-mudflap.c (mf_varname_tree): Likewise.
224 * graph.c (print_graph_cfg): Likewise.
225 (start_graph_dump): Likewise.
226 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
228 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
229 pretty printer initialization.
230 * coretypes.h (diagnostic_context): Remove superflous type alias
232 (pretty_printer): Likewise. Declare directly as a class.
233 (pretty_print_info): Remove declaration as class.
234 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
235 and pp_clear_output_area.
236 (asan_add_global): Likewise.
238 2013-08-22 Jan Hubicka <jh@suse.cz>
240 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
241 * ipa-utils.h (update_type_inheritance_graph): Declare.
242 (possible_polymorphic_call_target_p): Declare.
243 (possible_polymorphic_call_target_p): New.
244 * ipa-devirt.c: Update toplevel comments.
245 (cached_polymorphic_call_targets): Move up.
246 (odr_type_d): Move ID down.
247 (polymorphic_type_binfo_p): Update comment.
248 (odr_hasher::remove): Likewise;
249 (get_odr_type): Set anonymous_namespace.
250 (dump_odr_type): Dump it.
251 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
252 (maybe_record_node): Record node in cached_polymorphic_call_targets.
253 (record_binfo): Add comment.
254 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
256 (devirt_node_removal_hook): Do not iCE when cache is freed.
257 (possible_polymorphic_call_target_p): New predicate.
258 (update_type_inheritance_graph): New function.
260 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
261 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
262 Sergey Lega <sergey.s.lega@intel.com>
263 Anna Tikhonova <anna.tikhonova@intel.com>
264 Ilya Tocar <ilya.tocar@intel.com>
265 Andrey Turetskiy <andrey.turetskiy@intel.com>
266 Ilya Verbin <ilya.verbin@intel.com>
267 Kirill Yukhin <kirill.yukhin@intel.com>
268 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
270 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
271 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
272 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
273 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
274 (OPTION_MASK_ISA_AVX2_UNSET): Update.
275 (OPTION_MASK_ISA_AVX512F_UNSET): New.
276 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
277 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
278 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
279 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
280 OPT_mavx512pf, OPT_mavx512er cases.
281 * config/i386/constraints.md (v): New constraint.
282 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
283 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
285 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
286 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
287 * config/i386/i386-c.c (ix86_target_macros_internal):
288 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
290 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
291 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
292 * config/i386/i386.c (regclass_map, dbx_register_map)
293 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
294 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
295 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
297 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
298 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
299 -mavx512pf options. Fix formatting.
300 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
301 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
302 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
303 -mavx512cd, -mavx512pf options.
304 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
305 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
306 (ix86_preferred_output_reload_class): Replace SSE_REGS with
308 (ix86_hard_regno_mode_ok): Support 512-bit registers.
309 (ix86_set_reg_reg_cost): Ditto.
310 (x86_order_regs_for_local_alloc): Ditto.
311 (MAX_VECT_LEN): Extend to 64-byte.
312 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
313 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
314 (TARGET_AVX512ER, TARGET_AVX512CD): New.
315 (BIGGEST_ALIGNMENT): Extend to 512-bits.
316 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
317 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
318 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
319 (SSE_REG_MODE_P): Support new modes.
320 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
321 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
322 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
323 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
324 (REG_CLASS_CONTENTS): Add new registers.
325 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
326 (EXT_REX_SSE_REGNO_P): New.
327 (HI_REGISTER_NAMES): Add new registers.
328 * config/i386/i386.md: Define constants for new registers.
329 (mode): Add new 512-bit modes.
330 (prefix): Support evex prefix.
331 (isa): Support avx512f, noavx512f, fma_avx512f.
332 (ssemodesuffix): Add new 512-bit modes.
334 (*movxi_internal_avx512f): Ditto.
335 (*movdi_internal): Replace constraint "x" with the new constraint "v".
337 (*movsi_internal): Likewise.
338 (*movdf_internal): Likewise.
339 (*movsf_internal): Likewise.
340 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
341 (<code><mode>3): Likewise.
342 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
344 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
345 with the new constraint "v".
346 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
348 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
349 with the new constraint "v".
350 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
351 (<sse2>_storedqu<avxsizesuffix>): Likewise.
352 (*<plusminus_insn><mode>3): Likewise.
353 (<sse>_vm<plusminus_insn><mode>3): Likewise.
354 (*mul<mode>3): Likewise.
355 (<sse>_vmmul<mode>3): Likewise.
356 (<sse>_div<mode>3): Likewise.
357 (<sse>_vmdiv<mode>3): Likewise.
358 (<sse>_sqrt<mode>2): Likewise.
359 (<sse>_vmsqrt<mode>2): Likewise.
360 (*<code><mode>3_finite): Likewise.
361 (*<code><mode>3) <smaxmin>: Likewise.
362 (<sse>_vm<code><mode>3): Likewise.
363 (*<code><mode>3) <any_logic>: Likewise.
364 (*fma_fmadd_<mode>): Likewise.
365 (*fma_fmsub_<mode>): Likewise.
366 (*fma_fnmadd_<mode>): Likewise.
367 (*fma_fnmsub_<mode>): Likewise.
368 (*fma_fmaddsub_<mode>): Likewise.
369 (*fma_fmsubadd_<mode>): Likewise.
370 (*fmai_fmadd_<mode>): Likewise.
371 (*fmai_fmsub_<mode>): Likewise.
372 (*fmai_fnmadd_<mode>): Likewise.
373 (*fmai_fnmsub_<mode>): Likewise.
374 (sse_cvtsi2ss): Likewise.
375 (sse_cvtsi2ssq): Likewise.
376 (sse_cvtss2si): Likewise.
377 (sse_cvtss2si_2): Likewise.
378 (sse_cvtss2siq): Likewise.
379 (sse_cvtss2siq_2): Likewise.
380 (sse_cvttss2si): Likewise.
381 (sse_cvtss2siq_2): Likewise.
382 (float<sseintvecmodelower><mode>2): Likewise.
383 (sse2_cvtsd2si_2): Likewise.
384 (sse2_cvtsd2siq_2): Likewise.
385 (*<plusminus_insn><mode>3): Likewise.
386 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
387 (*<sse4_1_avx2>_mul<mode>3): Likewise.
388 (ashr<mode>3): Likewise.
389 (<shift_insn><mode>3): Likewise.
390 (avx2_<code><mode>3): Likewise.
391 (*avx2_<code><mode>3): Likewise.
392 (*andnot<mode>3): Likewise.
393 (*<code><mode>3) <any_logic>: Likewise.
394 (abs<mode>2): Likewise.
395 (avx2_permvar<mode>): Likewise.
396 (avx2_perm<mode>_1): Likewise.
397 (*avx_vpermilp<mode>): Likewise.
398 (avx_vpermilvar<mode>3): Likewise.
399 (avx2_ashrv<mode>): Likewise.
400 (avx2_<shift_insn>v<mode>): Likewise.
401 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
403 * doc/rtl.texi: Document XImode.
405 2013-08-21 Jeff Law <law@redhat.com>
407 * tree-flow.h (register_jump_thread): Pass vector of edges
408 instead of each important edge.
409 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
410 thread path into a vector and pass that to register_jump_thread.
411 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
412 passed in edge vector to the current 3-edge form.
415 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
417 * dce.c (fini_dce): Call df_analyze again just in case
418 delete_unmarked_insns removed anything.
420 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
422 * reload.h (struct reg_equivs): Rename to ..
423 (struct reg_equivs_s): .. this.
425 2013-08-20 Martin Liska <marxin.liska@gmail.com>
427 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
429 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
431 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
433 2013-08-21 Jeff Law <law@redhat.com>
435 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
436 simplify assignments too. If the RHS collapses to a singleton
437 range, then return the value for the range.
439 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
441 * config/i386/sse.md (V16): Rename to...
443 (mov<mode>): Update iterator name.
444 (*mov<mode>_internal): Ditto.
445 (push<mode>1): Ditto.
446 (movmisalign<mode>): Ditto.
448 2013-08-20 Jan Hubicka <jh@suse.cz>
451 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
452 entry for direct edges.
453 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
455 2013-08-20 David Malcolm <dmalcolm@redhat.com>
457 Revert my last two changes, r201865 and r201864:
460 2013-08-20 David Malcolm <dmalcolm@redhat.com>
462 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
463 instances can own GC refs.
465 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
466 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
467 (gcc::context::gt_pch_nx): Likewise.
468 (gcc::context::gt_pch_nx): Likewise.
469 * ggc.h (gt_ggc_mx <T>): New.
470 (gt_pch_nx_with_op <T>): New.
471 (gt_pch_nx <T>): New.
472 * passes.c (opt_pass::gt_ggc_mx): New.
473 (opt_pass::gt_pch_nx): New.
474 (opt_pass::gt_pch_nx_with_op): New.
475 (pass_manager::gt_ggc_mx): New.
476 (pass_manager::gt_pch_nx): New.
477 (pass_manager::gt_pch_nx_with_op): New.
478 (pass_manager::operator new): Use
479 ggc_internal_cleared_alloc_stat rather than xcalloc.
480 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
481 (pass_manager::gt_ggc_mx): New.
482 (pass_manager::gt_pch_nx): New.
483 (pass_manager::gt_pch_nx_with_op): New.
484 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
485 (opt_pass::operator new): New.
486 (opt_pass::gt_ggc_mx): New.
487 (opt_pass::gt_pch_nx): New.
488 (opt_pass::gt_pch_nx_with_op): New.
491 2013-08-20 David Malcolm <dmalcolm@redhat.com>
493 * Makefile.in (GTFILES): Add context.h.
494 * context.c (gcc::context::operator new): New.
495 (gcc::context::gt_ggc_mx): New.
496 (gcc::context::gt_pch_nx): New.
497 (gcc::context::gt_pch_nx): New.
498 * context.h (gcc::context): Add GTY((user)) marking.
499 (gcc::context::operator new): New.
500 (gcc::context::gt_ggc_mx): New.
501 (gcc::context::gt_pch_nx): New.
502 (gcc::context::gt_pch_nx): New.
503 (g): Add GTY marking.
504 (gt_ggc_mx (gcc::context *)): New.
505 (gt_pch_nx (gcc::context *)): New.
506 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
508 * gengtype.c (open_base_files) <ifiles>: Add context.h.
510 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
512 * dce.c (fini_dce): Call df_analyze again just in case
513 delete_unmarked_insns removed anything.
515 2013-08-20 Teresa Johnson <tejohnson@google.com>
517 PR rtl-optimizations/57451
518 * final.c (reemit_insn_block_notes): Prevent lexical blocks
519 from crossing split section boundaries.
521 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
523 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
524 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
525 with MULTLIB_DEFAULTS.
527 2013-08-20 Nick Clifton <nickc@redhat.com>
529 * target.def (narrow_volatile_bitfield): Note that the default
530 value is false, not !TARGET_STRICT_ALIGN.
531 * doc/tm.texi: Regenerate.
533 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
535 Fix LIB_SPEC for systems without libpthread.
537 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
538 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
540 * config/i386/linux-common.h: Likewise.
541 * config/mips/linux-common.h: Likewise.
543 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
545 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
548 2013-08-20 David Malcolm <dmalcolm@redhat.com>
550 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
551 instances can own GC refs.
553 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
554 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
555 (gcc::context::gt_pch_nx): Likewise.
556 (gcc::context::gt_pch_nx): Likewise.
557 * ggc.h (gt_ggc_mx <T>): New.
558 (gt_pch_nx_with_op <T>): New.
559 (gt_pch_nx <T>): New.
560 * passes.c (opt_pass::gt_ggc_mx): New.
561 (opt_pass::gt_pch_nx): New.
562 (opt_pass::gt_pch_nx_with_op): New.
563 (pass_manager::gt_ggc_mx): New.
564 (pass_manager::gt_pch_nx): New.
565 (pass_manager::gt_pch_nx_with_op): New.
566 (pass_manager::operator new): Use
567 ggc_internal_cleared_alloc_stat rather than xcalloc.
568 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
569 (pass_manager::gt_ggc_mx): New.
570 (pass_manager::gt_pch_nx): New.
571 (pass_manager::gt_pch_nx_with_op): New.
572 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
573 (opt_pass::operator new): New.
574 (opt_pass::gt_ggc_mx): New.
575 (opt_pass::gt_pch_nx): New.
576 (opt_pass::gt_pch_nx_with_op): New.
578 2013-08-20 David Malcolm <dmalcolm@redhat.com>
580 * Makefile.in (GTFILES): Add context.h.
581 * context.c (gcc::context::operator new): New.
582 (gcc::context::gt_ggc_mx): New.
583 (gcc::context::gt_pch_nx): New.
584 (gcc::context::gt_pch_nx): New.
585 * context.h (gcc::context): Add GTY((user)) marking.
586 (gcc::context::operator new): New.
587 (gcc::context::gt_ggc_mx): New.
588 (gcc::context::gt_pch_nx): New.
589 (gcc::context::gt_pch_nx): New.
590 (g): Add GTY marking.
591 (gt_ggc_mx (gcc::context *)): New.
592 (gt_pch_nx (gcc::context *)): New.
593 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
595 * gengtype.c (open_base_files) <ifiles>: Add context.h.
597 2013-08-20 Alan Modra <amodra@gmail.com>
600 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
601 (rs6000_emit_epilogue): Likewise.
603 2013-08-19 Dehao Chen (dehao@google.com)
605 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
607 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
608 Jakub Jelinek <jakub@redhat.com>
610 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
611 (BUILT_IN_FABSD64): Likewise.
612 (BUILT_IN_FABSD128): Likewise.
613 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
614 (fold_builtin_1): Likewise.
615 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
616 destination and source operands.
617 (*abstd2_fpr): Likewise.
618 (*nabstd2_fpr): Likewise.
620 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
622 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
625 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
627 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
630 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
632 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
633 * builtins.c (default_libc_has_function): New.
634 (gnu_libc_has_function): Ditto.
635 (no_c99_libc_has_function): Ditto.
636 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
637 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
638 (fold_builtin_sincos): Likewise.
639 (fold_builtin_cexp): Likewise.
640 * builtins.def (DEF_C94_BUILTIN): Likewise.
641 (DEF_C99_BUILTIN): Likewise.
642 (DEF_C99_C90RES_BUILTIN): Likewise.
643 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
644 definitions to using this define.
645 * config/darwin-protos.h (darwin_libc_has_function): New.
646 * config/darwin.c (darwin_libc_has_function): Ditto.
647 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
648 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
649 * config/darwin.h: Ditto.
650 * config/elfos.h: Ditto.
651 * config/freebsd.h: Ditto.
652 * config/i386/cygming.h: Ditto.
653 * config/i386/djgpp.h: Ditto.
654 * config/i386/i386-interix.h: Ditto.
655 * config/microblaze/microblaze.h: Ditto.
656 * config/mmix/mmix.h: Ditto.
657 * config/gnu-user.h: Ditto.
658 * config/ia64/hpux.h: Ditto.
659 * config/pa/pa-hpux.h: Ditto.
660 * config/pdp11/pdp11.h: Ditto.
661 * config/picochip/picochip.h: Ditto.
662 * config/linux.h: Ditto.
663 * config/netbsd.h: Ditto.
664 * config/openbsd.h: Ditto.
665 * config/rs6000/aix43.h: Ditto.
666 * config/rs6000/aix51.h: Ditto.
667 * config/rs6000/aix52.h: Ditto.
668 * config/rs6000/aix53.h: Ditto.
669 * config/rs6000/aix61.h: Ditto.
670 * config/rs6000/darwin.h: Ditto.
671 * config/rs6000/linux.h: Ditto.
672 * config/rs6000/linux64.h: Ditto.
673 * config/s390/tpf.h: Ditto.
674 * config/sol2-10.h: Ditto.
675 * config/sol2.h: Ditto.
676 * config/vms/vms.h: Ditto.
677 * config/vxworks.h: Ditto.
678 * config/linux-android.c (linux_android_libc_has_function):
679 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
680 * config/linux-protos.h (linux_android_libc_has_function):
682 * config/i386/i386.c (ix86_libc_has_function): New.
683 * config/i386/i386-protos.h
684 (ix86_libc_has_function): New declaration.
685 * config/i386/i386.md
686 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
687 ("isinf<mode>2): Likewise.
688 * convert.c (convert_to_integer): Using new target hook
689 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
690 TARGET_C99_FUNCTIONS.
691 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
692 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
693 * coretypes.h (function_class): New enum for different
694 classes of functions.
695 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
696 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
697 (TARGET_HAS_SINCOS): Likewise.
698 (TARGET_LIBC_HAS_FUNCTION): New.
699 * doc/tm.texi: Regenerated.
700 * targhooks.h (default_libc_has_function): New declaration.
701 (no_c99_libc_has_function): Ditto.
702 (gnu_libc_has_function): Ditto.
703 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
704 and TARGET_HAS_SINCOS.
706 2013-08-18 Jan Hubicka <jh@suse.cz>
708 * Makeifle-in (ipa-devirt.o): New.
709 (GTFILES): Add ipa-utils.h and ipa-devirt.c
710 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
711 (analyze_functions): Look into possible targets of polymorphic call.
712 * dumpfile.c (dump_files): Add type-inheritance dump.
713 * dumpfile.h (TDI_inheritance): New.
714 * ipa-devirt.c: New file.
715 * ipa-utils.h (odr_type_d): Forward declare.
716 (odr_type): New type.
717 (build_type_inheritance_graph): Declare.
718 (possible_polymorphic_call_targets): Declare and introduce inline
719 variant when only edge is pased.
720 (dump_possible_polymorphic_call_targets): Likewise.
721 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
722 * tree.c (type_in_anonymous_namespace_p): Break out from ...
723 (types_same_for_odr): ... here.
724 * tree.h (type_in_anonymous_namespace_p): Declare.
726 2013-08-18 Jakub Jelinek <jakub@redhat.com>
728 PR tree-optimization/58006
729 * tree-parloops.c (take_address_of): Don't ICE if get_name
731 (eliminate_local_variables_stmt): Remove clobber stmts.
733 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
735 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
736 error message is issued for an alias to undefined symbol.
738 2013-08-18 Jan Hubicka <jh@suse.cz>
740 * cgraph.c (cgraph_create_indirect_edge): Discover
741 polymorphic calls and record basic info into indirect_info.
742 * gimple-fold.c (gimple_fold_call): When doing BINFO based
743 devirtualization, ignore objc function calls.
744 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
745 call with no parm index info.
746 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
747 * tree.c (virtual_method_call_p): New function.
748 * tree.h (virtual_method_call_p): Declare.
750 2013-08-16 Jan Hubicka <jh@suse.cz>
753 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
755 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
757 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
760 2013-08-16 David Malcolm <dmalcolm@redhat.com>
762 * gengtype.c (type_for_name): Add special-case support for
763 locating types within the "gcc::" namespace.
764 (open_base_files): Emit a "using namespace gcc" directive.
766 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
769 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
770 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
772 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
773 array instead of each individual operand as a separate argument.
774 (emit_fusion_gpr_load): Likewise.
775 (expand_fusion_gpr_load): Add new function declaration.
777 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
778 signature to have the operands passed as an array, instead of as
779 separate arguments. Allow ZERO_EXTEND to be in the memory
780 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
781 depend on the register live/dead flags when peepholes are run.
782 (expand_fusion_gpr_load): New function to be called from the
783 peephole2 pass, to change the register that addis sets to be the
785 (emit_fusion_gpr_load): Change the calling signature to have the
786 operands passed as an array, instead of as separate arguments.
787 Allow ZERO_EXTEND to be in the memory address, and also
788 SIGN_EXTEND if -mpower8-fusion-sign.
790 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
792 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
793 adjust the register addis loads up in the peephole2 pass. Do not
794 depend on the register live/dead state when the peephole pass is done.
796 2013-08-16 David Malcolm <dmalcolm@redhat.com>
798 * gengtype.c (create_user_defined_type): Ensure that the kind
799 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
800 declaration is seen before the GTY((user)) marking.
802 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
805 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
807 2013-08-16 Jan Hubicka <jh@suse.cz>
809 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
810 arugment expected_type.
811 (gimple_fold_call): Use it.
812 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
813 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
814 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
815 (try_make_edge_direct_virtual_call): Likewise.
816 * tree.c (obj_type_ref_class): New.
817 * tree.h (obj_type_ref_class): Use it.
819 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
821 * sched-vis.c (rtl_slim_pp_initialized): Remove.
822 (rtl_slim_pp): Likewise.
823 (init_rtl_slim_pretty_print): Likewise.
824 (dump_value_slim): Don't call it. Use local pretty printer.
825 (dump_insn_slim): Likewise.
826 (dump_rtl_slim): Likewise.
827 (str_pattern_slim): Likewise.
828 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
831 2013-08-16 Jakub Jelinek <jakub@redhat.com>
833 PR tree-optimization/58164
834 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
835 walk gimple_goto_dest of GIMPLE_GOTO.
837 PR tree-optimization/58165
838 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
839 bi_call must be the last stmt in a bb, don't split_block, instead
840 use fallthru edge from it and give up if there is none.
841 Release conds vector when returning early.
843 2013-08-14 Xinliang David Li <davidxl@google.com>
845 * config/i386/i386.c (ix86_option_override_internal):
846 Remove unused variable and field.
848 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
851 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
852 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
853 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
854 and Linux, correct BLKmode alignment when 128-bit alignment is
855 required and compatibility flag is not set.
856 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
857 for zero-size arguments when compatibility flag is not set.
859 2013-08-14 Jakub Jelinek <jakub@redhat.com>
861 PR tree-optimization/58145
862 * tree-sra.c (build_ref_for_offset): If prev_base has
863 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
865 2013-08-14 Xinliang David Li <davidxl@google.com>
867 * config/i386/i386.c (ix86_option_override_internal):
868 Fix uninitialized variable error.
870 2013-08-14 Xinliang David Li <davidxl@google.com>
872 * config/i386/i386.opt: Define two new options.
873 * config/i386/x86-tune.def: Add arch selector field in macros.
874 * config/i386/i386.h: Adjust macro definition.
875 * config/i386/i386.c (ix86_option_override_internal):
877 (parse_mtune_ctrl_str): New function.
878 (set_ix86_tune_features): New function.
879 (ix86_function_specific_restore): Call the new helper function.
881 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
883 PR rtl-optimization/57662
884 * sel-sched.c (code_motion_process_successors): When the current insn
885 is removed after the recursive traversal, break from the loop.
886 Add comments and debug printouts.
888 2013-08-14 Jakub Jelinek <jakub@redhat.com>
889 Alexandre Oliva <aoliva@redhat.com>
892 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
893 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
894 there also UNSPEC_PLTOFF.
896 2013-08-14 Marek Polacek <polacek@redhat.com>
898 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
899 NUM_CONDITIONS bit positions.
901 2013-08-13 Cary Coutant <ccoutant@google.com>
903 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
904 (attr_checksum): Hash vector contents instead of pointer.
905 (attr_checksum_ordered): Likewise.
907 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
909 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
910 when Pmode != word_mode. Add length_address attribute.
911 (sse3_monitor_<mode>): Merge from sse3_monitor and
912 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
913 Pmode != word_mode. Update insn length attribute.
914 * config/i386/i386.c (ix86_option_override_internal): Update
915 ix86_gen_monitor selection for merged sse3_monitor insn.
917 2013-08-13 Julian Brown <julian@codesourcery.com>
919 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
920 perform invalid legitimization on greater-than-word-size modes for
923 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
925 * ira.c (setup_class_translate_array): Use aclass instead of cl
926 for classes not fully covered by allocno classes.
928 2013-08-13 Jakub Jelinek <jakub@redhat.com>
930 PR tree-optimization/57661
931 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
932 * tree-inline.c (tree_function_versioning): Initialize it.
933 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
934 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
935 that is not being copied.
938 * asan.c (instrument_strlen_call): Fix typo in comment.
939 Use char * type even for the lhs of POINTER_PLUS_EXPR.
941 2013-08-13 Steve Ellcey <sellcey@mips.com>
943 * config/mips/mips.md (prefetch): Use lw instead of ld on
944 loongson in 32bit mode.
946 2013-08-13 Nick Clifton <nickc@redhat.com>
948 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
950 2013-08-13 Jan Hubicka <jh@suse.cz>
952 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
953 introduced edge; fix typo in sanity check.
954 (cgraph_resolve_speculation): Export; improve diagnostic.
955 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
956 speculation at type mismatch.
957 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
958 (cgraph_resolve_speculation): Declare.
959 (symtab_can_be_discarded): New function.
960 * value-prof.c (gimple_ic_transform): Remove actual transform code.
961 * ipa-inline-transform.c (speculation_removed): New global var.
962 (clone_inlined_nodes): See if speculation can be removed.
963 (inline_call): If speculations was removed, we growths may not match.
964 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
965 (speculation_useful_p): New function.
966 (resolve_noninline_speculation): New function.
967 (inline_small_functions): Resolve useless speculations.
968 * ipa-inline.h (speculation_useful_p): Declare
969 * ipa.c (can_replace_by_local_alias): Simplify.
970 (ipa_profile): Produce speculative calls in non-lto, too;
971 add simple cost model; produce local aliases.
973 2013-08-13 David Malcolm <dmalcolm@redhat.com>
975 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
978 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
980 * config/i386/i386.c (ix86_function_versions): Use error + inform.
982 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
984 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
985 iterator instead of X87MODEF.
987 2013-08-12 Perez Read <netfirewall@gmail.com>
990 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
991 operand 0 for intel asm alternative.
992 (*movabs<mode>_2): Ditto for operand 1.
994 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
996 * config/aarch64/arm_none.h
997 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
999 2013-08-12 Nick Clifton <nickc@redhat.com>
1001 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
1003 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1005 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
1006 expand for QI/HImode operand to produce more effictive code for
1007 unsigned char(short) --> float(double) conversion.
1009 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
1011 * doc/invoke.texi: Mention that -ftls-model does not force the final
1014 2013-08-12 Marek Polacek <polacek@redhat.com>
1015 Marc Glisse <marc.glisse@inria.fr>
1017 PR tree-optimization/57980
1018 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
1019 when creating -1 constant.
1021 2013-08-10 Jan Hubicka <jh@suse.cz>
1023 Workaround binutils PR14342.
1024 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
1025 (gimple_init_edge_profiler): Likewise.
1026 (gimple_gen_ic_func_profiler): Likewise.
1028 2013-08-09 Jan Hubicka <jh@suse.cz>
1030 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
1032 2013-08-09 Xinliang David Li <davidxl@google.com>
1034 * config/i386/stringop.def: New file.
1035 * config/i386/stringop.opt: New file.
1036 * config/i386/i386-opts.h: Include stringopt.def.
1037 * config/i386/i386.opt: Include stringopt.opt.
1038 * config/i386/i386.c (ix86_option_override_internal):
1039 Override default size based stringop inline strategies with options.
1040 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1043 2013-08-09 Jan Hubicka <jh@suse.cz>
1045 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1047 2013-08-09 Jan Hubicka <jh@suse.cz>
1049 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1051 (dump_cgraph_node): Dump profile-id.
1052 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1053 and common_target_probability.
1054 * lto-cgraph.c (lto_output_edge): Stream common targets.
1055 (lto_output_node): Stream profile ids.
1056 (input_node): Stream profile ids.
1057 (input_edge): Stream common targets.
1058 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1059 * ipa.c: Include value-prof.h
1060 (ipa_profile_generate_summary): Turn indirect call statement histograms
1061 into common targets.
1062 (ipa_profile): Turn common targets into speculative edges.
1064 2013-08-09 Jan Hubicka <jh@suse.cz>
1066 * cgraph.h (cgraph_node): Add profile_id.
1067 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1068 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1069 (del_node_map): Update.
1070 (find_func_by_funcdef_no): Replace by ...
1071 (find_func_by_profile_id): ... this one.
1072 (gimple_ic_transform): Do not remove useful histograms when
1073 speculation is not done; dump info when indirect call removal
1075 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1076 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1077 (__gcov_indirect_call_profiler_v2): .. this one.
1078 * profile.h (init_node_map): Update.
1079 * coverage.c (coverage_compute_profile_id): New function.
1080 * coverage.h (coverage_compute_profile_id): Declare.
1081 * tree-profile.c (init_ic_make_global_vars): Make
1082 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1083 (gimple_init_edge_profiler): Update prototype of
1084 __gcov_indirect_call_profiler.
1085 (gimple_gen_ic_func_profiler): Simplify.
1086 (tree_profiling): Use init_node_map
1088 2013-08-09 Jan Hubicka <jh@suse.cz>
1090 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1091 non-speculative refs.
1092 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1093 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1094 (cgraph_set_call_stmt): Likewise.
1095 (cgraph_create_edge_1): Fix release checking compilatoin;
1097 (cgraph_free_edge): Free indirect info.
1098 (cgraph_turn_edge_to_speculative): New function.
1099 (cgraph_speculative_call_info): New function.
1100 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1101 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1102 (dump_cgraph_node): Dump speculation.
1103 (verify_edge_count_and_frequency): Accept speculative edges.
1104 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1105 (verify_cgraph_node): Handle speculation.
1106 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1107 (cgraph_set_call_stmt): Update prototype.
1108 (cgraph_make_edge_direct): Update prototype.
1109 (cgraph_speculative_call_info): Declare.
1110 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1111 to change; update call of ipa_find_references.
1112 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1113 lto_stmt_uid and speculative flags.
1114 (ipa_dump_references): Dump speculation.
1115 (ipa_clone_references): Clone speculative flag.
1116 (ipa_clone_referring): Likewise.
1117 (ipa_clone_ref): New function.
1118 (ipa_find_reference): Look into lto_stmt_uids
1119 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1120 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1121 (ipa_find_reference): Update declaration.
1122 (ipa_clone_ref): Declare.
1123 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1124 stream speculative flag.
1125 (lto_output_ref): Stream statements uids and speculation.
1126 (input_ref): Likewise.
1127 (input_edge): Stream speuclation.
1128 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1129 (cgraph_set_call_stmt_including_clones): Handle speculation.
1130 * ipa-inline.c (heap_edge_removal_hook): New function.
1131 (inline_small_functions): Register it.
1132 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1133 also initialize refs.
1134 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1136 (try_make_edge_direct_simple_call): Likewise.
1137 (try_make_edge_direct_simple_call): Likewise.
1138 (update_indirect_edges_after_inlining): Likewise.
1139 (remove_described_reference): Look proper lto_stmt_uid.
1140 (propagate_controlled_uses): Likewise.
1141 (propagate_controlled_uses): Liekwise.
1142 * tree-inline.c (copy_bb): Copy speculative edges.
1143 (redirect_all_calls): New function.
1144 (copy_cfg_body): Do redirection after loop info is updated.
1145 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1147 2013-08-09 Jan Hubicka <jh@suse.cz>
1149 * lto-streamer-out.c (output_function): Renumber PHIs.
1150 * lto-streamer-in.c (input_function): Likewise.
1152 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1154 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1155 (get_lane_unsigned): Likewise.
1156 (dup_lane_scalar): Likewise.
1157 (get_lane): enable for VALL.
1158 * config/aarch64/aarch64-simd.md
1159 (aarch64_dup_lane_scalar<mode>): Remove.
1160 (aarch64_get_lane_signed<mode>): Likewise.
1161 (aarch64_get_lane_unsigned<mode>): Likewise.
1162 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1163 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1164 (aarch64_get_lane<mode>): Enable for all vector modes.
1165 (aarch64_get_lanedi): Remove misleading constraints.
1166 * config/aarch64/arm_neon.h
1167 (__aarch64_vget_lane_any): Define.
1168 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1169 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1170 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1171 * config/aarch64/iterators.md (VDQQH): New.
1175 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1177 * configure.ac: Add GAS check for LEON instructions on SPARC.
1178 * configure: Regenerate.
1179 * config.in: Likewise.
1180 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1182 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1183 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1185 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1186 (AS_LEON_FLAG): New macro.
1187 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1188 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1189 Deal with LEON and LEON3 for the memory model.
1190 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1191 (atomic_compare_and_swap<mode>_1): Likewise.
1192 (*atomic_compare_and_swap<mode>_1): Likewise.
1194 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1196 * config/arm/neon.md (vcond): Fix floating-point vector
1197 comparisons against 0.
1199 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1201 * lra-constraints.c (emit_spill_move): Remove assert.
1202 (process_alt_operands): Add more debugging
1203 output. Increase reject for spilling into memory. Decrease
1204 reject for reloading scratch.
1205 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1207 2013-08-08 Steve Ellcey <sellcey@mips.com>
1209 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1210 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1211 micromips incompatible. Add nan2008.
1212 (MULTILIB_DIRNAMES): Add nan2008.
1213 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1214 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1215 and micromips incompatible. Add nan2008.
1216 (MULTILIB_DIRNAMES): Add nan2008.
1217 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1219 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1221 PR rtl-optimization/58079
1222 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1223 simplify_comparison has widened a comparison with an integer.
1225 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1227 * config/arm/neon.md (movmisalign<mode>): Disable when we
1228 don't allow unaligned accesses.
1229 (*movmisalign<mode>_neon_store): Likewise.
1230 (*movmisalign<mode>_neon_load): Likewise.
1231 (*movmisalign<mode>_neon_store): Likewise.
1232 (*movmisalign<mode>_neon_load): Likewise.
1234 2013-08-08 Jan Hubicka <jh@suse.cz>
1236 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1237 (make_pass_rebuild_cgraph_edges): Also clear references.
1238 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1239 * ipa-inline-transform.c (inline_transform): Remove all references
1241 * cgraphunit.c (expand_function): Remove all references after
1243 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1244 (ipa_find_reference): Rewrite to iterator.
1245 (remove_stmt_references): Likewise.
1246 (ipa_clear_stmts_in_references): New function.
1247 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1248 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1250 * ipa-split.c (split_function): Remove references in split function.
1252 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1255 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1256 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1258 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1261 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
1262 suggested mode for the assignment isn't compatible with the
1265 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1268 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
1270 2013-08-07 Xinliang David Li <davidxl@google.com>
1272 * config/i386/i386.opt: New option -mtune-ctrl=.
1273 * config/i386/x86-tune.def: New file.
1274 * config/i386/i386.h: include x86-tune.def.
1275 * config/i386/i386.c (ix86_option_override_internal):
1276 Parsing -mtune-ctrl= option and set tune features.
1278 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
1281 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
1282 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
1285 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1287 * rtl.h (update_alignments): Declare.
1288 * final.c (grow_label_align): New function extracted from...
1289 (shorten_branches): ...here. Call it.
1290 (update_alignments): New function.
1291 * reorg.c (sibling_labels): New variable.
1292 (get_label_before): Add SIBLING parameter. If it is non-zero, push
1293 the new label along with it onto the sibling_labels vector.
1294 (fill_simple_delay_slots): Adjust call to get_label_before.
1295 (fill_slots_from_thread): Likewise.
1296 (relax_delay_slots): Likewise.
1297 (make_return_insns): Likewise.
1298 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
1300 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1302 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
1303 document its semantics.
1304 (diagnostic_report_diagnostic): Adjust accordingly.
1306 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1308 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
1309 (sparc_option_override): ...and port to new C++ pass API.
1310 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
1312 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
1314 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
1316 2013-08-06 Caroline Tice <cmtice@google.com>
1318 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
1319 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
1320 * tree-pass.h: Add pass_vtable_verify.
1321 * varasm.c (assemble_variable): Add code to properly set the comdat
1322 section and name for the .vtable_map_vars section.
1323 (assemble_vtyv_preinit_initializer): New function.
1324 (default_sectin_type_flags): Make sure .vtable_map_vars section has
1326 * output.h: Add function decl for assemble_vtv_preinit_initializer.
1327 * vtable-verify.c: New file.
1328 * vtable-verify.h: New file.
1329 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
1330 initialiation levels.
1331 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
1332 * passes.def: Insert pass_vtable_verify.
1333 * aclocal.m4: Reorder includes.
1334 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
1335 -fvtv-counts options.
1336 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
1337 as appropriate, if -fvtable-verify=... is used.
1338 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
1339 -fvtable-verify=... is used.
1340 * Makefile.in (OBJS): Add vtable-verify.o to list.
1341 (vtable-verify.o): Add new build rule.
1342 (GTFILES): Add vtable-verify.c to list.
1343 * common.opt (fvtable-verify=): New flag.
1344 (vtv_priority): Values for fvtable-verify= flag.
1345 (fvtv-counts): New flag.
1346 (fvtv-debug): New flag.
1347 * tree.h (save_vtable_map_decl): New extern function decl.
1349 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1351 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
1352 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
1353 (pass_data_rl78_devirt): ...new pass_data instance and...
1354 (make_pass_rl78_devirt): ...new function.
1355 (rl78_asm_file_start): Port pass registration to new C++ API.
1357 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1359 * coretypes.h (rtl_opt_pass): Add.
1360 (gcc::context): Add.
1361 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
1362 (epiphany_init): Port to new C++ pass API.
1363 (epiphany_optimize_mode_switching): Likewise.
1364 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
1365 (pass_manager::get_pass_mode_switching): New.
1366 (pass_manager::get_pass_peephole2): New.
1367 * mode-switching.c (pass_mode_switching): Add clone method.
1368 * recog.c (pass_peephole2): Add clone method.
1369 (pass_split_all_insns): Add clone method.
1371 2013-08-06 David Malcolm <dmalcolm@redhat.com>
1373 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
1374 (mips_option_override): ...here, porting to new C++ API for passes.
1376 2013-08-06 Jan Hubicka <jh@suse.cz>
1378 * cgraph.c (cgraph_get_body): New function based on lto.c
1380 * cgraph.h (cgraph_get_body): Declare.
1381 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
1383 * cgraphunit.c (expand_function): Get body prior expanding.
1384 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
1385 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
1387 * passes.c (do_per_function_toporder): Get body.
1388 * tree-inline.c (expand_call_inline): Get body prior inlining it.
1389 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
1391 2013-08-06 Martin Jambor <mjambor@suse.cz>
1394 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
1395 is not re-finalized. Rename second parameter to no_collect.
1397 2013-08-06 Martin Jambor <mjambor@suse.cz>
1400 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
1401 MEM_REF has proper alignment information.
1403 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
1406 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
1408 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
1409 rtx (*) (rtx, ...) with insn_gen_fn.
1410 * genoutput.c (output_insn_data): Cast gen_? function pointers to
1411 insn_gen_fn::stored_funcptr. Add initializer braces.
1413 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1415 Rewrite how instances of passes are cloned to remove assumptions
1416 about their sizes (thus allowing pass subclasses to have
1417 additional data fields, albeit non-GC-managed ones at this point).
1419 * passes.c (make_pass_instance): Now that passes have clone
1420 methods, rewrite this function to eliminate XNEW and memcpy
1421 calls that used hardcoded sizes. Since this function no longer
1422 creates pass instances, rename it to...
1423 (add_pass_instance): ...this. Document the old way that passes were
1424 numbered and flagged, and rework this function to continue using it.
1425 (next_pass_1): Add an initial_pass argument for use by
1427 (position_pass): When adding multiple instances of a pass, use
1428 the pass's clone method, rather than relying on the XNEW/memcpy
1429 within the former make_pass_instance (now add_pass_instance).
1430 (pass_manager::pass_manager): When invoking next_pass_1, also supply
1431 the initial instance of the current pass within the pass manager.
1433 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1435 This is the automated part of the conversion of passes from C
1436 structs to C++ classes.
1438 Patch autogenerated by refactor_passes.py from
1439 https://github.com/davidmalcolm/gcc-refactoring-scripts
1440 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
1442 * asan.c (pass_asan): Convert from a global struct to a subclass of
1443 gimple_opt_pass along with...
1444 (pass_data_asan): ...new pass_data instance and...
1445 (make_pass_asan): ...new function.
1446 (pass_asan_O0): Convert from a global struct to a subclass of
1447 gimple_opt_pass along with...
1448 (pass_data_asan_O0): ...new pass_data instance and...
1449 (make_pass_asan_O0): ...new function.
1450 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
1451 subclass of rtl_opt_pass along with...
1452 (pass_data_inc_dec): ...new pass_data instance and...
1453 (make_pass_inc_dec): ...new function.
1454 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
1455 a subclass of rtl_opt_pass along with...
1456 (pass_data_reorder_blocks): ...new pass_data instance and...
1457 (make_pass_reorder_blocks): ...new function.
1458 (pass_duplicate_computed_gotos): Convert from a global struct to a
1459 subclass of rtl_opt_pass along with...
1460 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
1461 (make_pass_duplicate_computed_gotos): ...new function.
1462 (pass_partition_blocks): Convert from a global struct to a subclass of
1463 rtl_opt_pass along with...
1464 (pass_data_partition_blocks): ...new pass_data instance and...
1465 (make_pass_partition_blocks): ...new function.
1466 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
1467 struct to a subclass of rtl_opt_pass along with...
1468 (pass_data_branch_target_load_optimize1): ...new pass_data instance
1470 (make_pass_branch_target_load_optimize1): ...new function.
1471 (pass_branch_target_load_optimize2): Convert from a global struct to a
1472 subclass of rtl_opt_pass along with...
1473 (pass_data_branch_target_load_optimize2): ...new pass_data instance
1475 (make_pass_branch_target_load_optimize2): ...new function.
1476 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
1477 of rtl_opt_pass along with...
1478 (pass_data_jump): ...new pass_data instance and...
1479 (make_pass_jump): ...new function.
1480 (pass_jump2): Convert from a global struct to a subclass of
1481 rtl_opt_pass along with...
1482 (pass_data_jump2): ...new pass_data instance and...
1483 (make_pass_jump2): ...new function.
1484 * cfgexpand.c (pass_expand): Convert from a global struct to a
1485 subclass of rtl_opt_pass along with...
1486 (pass_data_expand): ...new pass_data instance and...
1487 (make_pass_expand): ...new function.
1488 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
1489 of rtl_opt_pass along with...
1490 (pass_data_free_cfg): ...new pass_data instance and...
1491 (make_pass_free_cfg): ...new function.
1492 (pass_into_cfg_layout_mode): Convert from a global struct to a
1493 subclass of rtl_opt_pass along with...
1494 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
1495 (make_pass_into_cfg_layout_mode): ...new function.
1496 (pass_outof_cfg_layout_mode): Convert from a global struct to a
1497 subclass of rtl_opt_pass along with...
1498 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
1499 (make_pass_outof_cfg_layout_mode): ...new function.
1500 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
1501 struct to a subclass of gimple_opt_pass along with...
1502 (pass_data_build_cgraph_edges): ...new pass_data instance and...
1503 (make_pass_build_cgraph_edges): ...new function.
1504 (pass_rebuild_cgraph_edges): Convert from a global struct to a
1505 subclass of gimple_opt_pass along with...
1506 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
1507 (make_pass_rebuild_cgraph_edges): ...new function.
1508 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
1509 subclass of gimple_opt_pass along with...
1510 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
1512 (make_pass_remove_cgraph_callee_edges): ...new function.
1513 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
1514 struct to a subclass of rtl_opt_pass along with...
1515 (pass_data_stack_adjustments): ...new pass_data instance and...
1516 (make_pass_stack_adjustments): ...new function.
1517 * combine.c (pass_combine): Convert from a global struct to a subclass
1518 of rtl_opt_pass along with...
1519 (pass_data_combine): ...new pass_data instance and...
1520 (make_pass_combine): ...new function.
1521 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
1522 global struct to a subclass of rtl_opt_pass along with...
1523 (pass_data_compare_elim_after_reload): ...new pass_data instance
1525 (make_pass_compare_elim_after_reload): ...new function.
1526 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1527 of rtl_opt_pass along with...
1528 (pass_data_rtl_cprop): ...new pass_data instance and...
1529 (make_pass_rtl_cprop): ...new function.
1530 * cse.c (pass_cse): Convert from a global struct to a subclass of
1531 rtl_opt_pass along with...
1532 (pass_data_cse): ...new pass_data instance and...
1533 (make_pass_cse): ...new function.
1534 (pass_cse2): Convert from a global struct to a subclass of
1535 rtl_opt_pass along with...
1536 (pass_data_cse2): ...new pass_data instance and...
1537 (make_pass_cse2): ...new function.
1538 (pass_cse_after_global_opts): Convert from a global struct to a
1539 subclass of rtl_opt_pass along with...
1540 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1541 (make_pass_cse_after_global_opts): ...new function.
1542 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1543 of rtl_opt_pass along with...
1544 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1545 (make_pass_ud_rtl_dce): ...new function.
1546 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1547 rtl_opt_pass along with...
1548 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1549 (make_pass_fast_rtl_dce): ...new function.
1550 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1551 a subclass of rtl_opt_pass along with...
1552 (pass_data_df_initialize_opt): ...new pass_data instance and...
1553 (make_pass_df_initialize_opt): ...new function.
1554 (pass_df_initialize_no_opt): Convert from a global struct to a
1555 subclass of rtl_opt_pass along with...
1556 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1557 (make_pass_df_initialize_no_opt): ...new function.
1558 (pass_df_finish): Convert from a global struct to a subclass of
1559 rtl_opt_pass along with...
1560 (pass_data_df_finish): ...new pass_data instance and...
1561 (make_pass_df_finish): ...new function.
1562 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1563 rtl_opt_pass along with...
1564 (pass_data_rtl_dse1): ...new pass_data instance and...
1565 (make_pass_rtl_dse1): ...new function.
1566 (pass_rtl_dse2): Convert from a global struct to a subclass of
1567 rtl_opt_pass along with...
1568 (pass_data_rtl_dse2): ...new pass_data instance and...
1569 (make_pass_rtl_dse2): ...new function.
1570 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1571 subclass of rtl_opt_pass along with...
1572 (pass_data_dwarf2_frame): ...new pass_data instance and...
1573 (make_pass_dwarf2_frame): ...new function.
1574 * except.c (pass_set_nothrow_function_flags): Convert from a global
1575 struct to a subclass of rtl_opt_pass along with...
1576 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1578 (make_pass_set_nothrow_function_flags): ...new function.
1579 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1580 subclass of rtl_opt_pass along with...
1581 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1583 (make_pass_convert_to_eh_region_ranges): ...new function.
1584 * final.c (pass_compute_alignments): Convert from a global struct to a
1585 subclass of rtl_opt_pass along with...
1586 (pass_data_compute_alignments): ...new pass_data instance and...
1587 (make_pass_compute_alignments): ...new function.
1588 (pass_final): Convert from a global struct to a subclass of
1589 rtl_opt_pass along with...
1590 (pass_data_final): ...new pass_data instance and...
1591 (make_pass_final): ...new function.
1592 (pass_shorten_branches): Convert from a global struct to a subclass of
1593 rtl_opt_pass along with...
1594 (pass_data_shorten_branches): ...new pass_data instance and...
1595 (make_pass_shorten_branches): ...new function.
1596 (pass_clean_state): Convert from a global struct to a subclass of
1597 rtl_opt_pass along with...
1598 (pass_data_clean_state): ...new pass_data instance and...
1599 (make_pass_clean_state): ...new function.
1600 * function.c (pass_instantiate_virtual_regs): Convert from a global
1601 struct to a subclass of rtl_opt_pass along with...
1602 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1603 (make_pass_instantiate_virtual_regs): ...new function.
1604 (pass_leaf_regs): Convert from a global struct to a subclass of
1605 rtl_opt_pass along with...
1606 (pass_data_leaf_regs): ...new pass_data instance and...
1607 (make_pass_leaf_regs): ...new function.
1608 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1609 subclass of rtl_opt_pass along with...
1610 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1612 (make_pass_thread_prologue_and_epilogue): ...new function.
1613 (pass_match_asm_constraints): Convert from a global struct to a
1614 subclass of rtl_opt_pass along with...
1615 (pass_data_match_asm_constraints): ...new pass_data instance and...
1616 (make_pass_match_asm_constraints): ...new function.
1617 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1618 subclass of rtl_opt_pass along with...
1619 (pass_data_rtl_fwprop): ...new pass_data instance and...
1620 (make_pass_rtl_fwprop): ...new function.
1621 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1622 rtl_opt_pass along with...
1623 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1624 (make_pass_rtl_fwprop_addr): ...new function.
1625 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1626 rtl_opt_pass along with...
1627 (pass_data_rtl_pre): ...new pass_data instance and...
1628 (make_pass_rtl_pre): ...new function.
1629 (pass_rtl_hoist): Convert from a global struct to a subclass of
1630 rtl_opt_pass along with...
1631 (pass_data_rtl_hoist): ...new pass_data instance and...
1632 (make_pass_rtl_hoist): ...new function.
1633 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1634 subclass of gimple_opt_pass along with...
1635 (pass_data_lower_cf): ...new pass_data instance and...
1636 (make_pass_lower_cf): ...new function.
1637 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1638 from a global struct to a subclass of gimple_opt_pass along with...
1639 (pass_data_strength_reduction): ...new pass_data instance and...
1640 (make_pass_strength_reduction): ...new function.
1641 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1642 of rtl_opt_pass along with...
1643 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1644 (make_pass_rtl_ifcvt): ...new function.
1645 (pass_if_after_combine): Convert from a global struct to a subclass of
1646 rtl_opt_pass along with...
1647 (pass_data_if_after_combine): ...new pass_data instance and...
1648 (make_pass_if_after_combine): ...new function.
1649 (pass_if_after_reload): Convert from a global struct to a subclass of
1650 rtl_opt_pass along with...
1651 (pass_data_if_after_reload): ...new pass_data instance and...
1652 (make_pass_if_after_reload): ...new function.
1653 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1654 a subclass of rtl_opt_pass along with...
1655 (pass_data_initialize_regs): ...new pass_data instance and...
1656 (make_pass_initialize_regs): ...new function.
1657 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1658 of ipa_opt_pass_d along with...
1659 (pass_data_ipa_cp): ...new pass_data instance and...
1660 (make_pass_ipa_cp): ...new function.
1661 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1662 global struct to a subclass of gimple_opt_pass along with...
1663 (pass_data_inline_parameters): ...new pass_data instance and...
1664 (make_pass_inline_parameters): ...new function.
1665 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1666 subclass of gimple_opt_pass along with...
1667 (pass_data_early_inline): ...new pass_data instance and...
1668 (make_pass_early_inline): ...new function.
1669 (pass_ipa_inline): Convert from a global struct to a subclass of
1670 ipa_opt_pass_d along with...
1671 (pass_data_ipa_inline): ...new pass_data instance and...
1672 (make_pass_ipa_inline): ...new function.
1673 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1674 struct to a subclass of gimple_opt_pass along with...
1675 (pass_data_local_pure_const): ...new pass_data instance and...
1676 (make_pass_local_pure_const): ...new function.
1677 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1678 ipa_opt_pass_d along with...
1679 (pass_data_ipa_pure_const): ...new pass_data instance and...
1680 (make_pass_ipa_pure_const): ...new function.
1681 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1682 to a subclass of ipa_opt_pass_d along with...
1683 (pass_data_ipa_reference): ...new pass_data instance and...
1684 (make_pass_ipa_reference): ...new function.
1685 * ipa-split.c (pass_split_functions): Convert from a global struct to
1686 a subclass of gimple_opt_pass along with...
1687 (pass_data_split_functions): ...new pass_data instance and...
1688 (make_pass_split_functions): ...new function.
1689 (pass_feedback_split_functions): Convert from a global struct to a
1690 subclass of gimple_opt_pass along with...
1691 (pass_data_feedback_split_functions): ...new pass_data instance and...
1692 (make_pass_feedback_split_functions): ...new function.
1693 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1694 global struct to a subclass of simple_ipa_opt_pass along with...
1695 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1697 (make_pass_ipa_function_and_variable_visibility): ...new function.
1698 (pass_ipa_free_inline_summary): Convert from a global struct to a
1699 subclass of simple_ipa_opt_pass along with...
1700 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1701 (make_pass_ipa_free_inline_summary): ...new function.
1702 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1703 subclass of ipa_opt_pass_d along with...
1704 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1706 (make_pass_ipa_whole_program_visibility): ...new function.
1707 (pass_ipa_profile): Convert from a global struct to a subclass of
1708 ipa_opt_pass_d along with...
1709 (pass_data_ipa_profile): ...new pass_data instance and...
1710 (make_pass_ipa_profile): ...new function.
1711 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1712 ipa_opt_pass_d along with...
1713 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1714 (make_pass_ipa_cdtor_merge): ...new function.
1715 * ira.c (pass_ira): Convert from a global struct to a subclass of
1716 rtl_opt_pass along with...
1717 (pass_data_ira): ...new pass_data instance and...
1718 (make_pass_ira): ...new function.
1719 (pass_reload): Convert from a global struct to a subclass of
1720 rtl_opt_pass along with...
1721 (pass_data_reload): ...new pass_data instance and...
1722 (make_pass_reload): ...new function.
1723 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1724 subclass of rtl_opt_pass along with...
1725 (pass_data_cleanup_barriers): ...new pass_data instance and...
1726 (make_pass_cleanup_barriers): ...new function.
1727 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1728 of rtl_opt_pass along with...
1729 (pass_data_loop2): ...new pass_data instance and...
1730 (make_pass_loop2): ...new function.
1731 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1732 rtl_opt_pass along with...
1733 (pass_data_rtl_loop_init): ...new pass_data instance and...
1734 (make_pass_rtl_loop_init): ...new function.
1735 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1736 rtl_opt_pass along with...
1737 (pass_data_rtl_loop_done): ...new pass_data instance and...
1738 (make_pass_rtl_loop_done): ...new function.
1739 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1740 subclass of rtl_opt_pass along with...
1741 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1742 (make_pass_rtl_move_loop_invariants): ...new function.
1743 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1744 rtl_opt_pass along with...
1745 (pass_data_rtl_unswitch): ...new pass_data instance and...
1746 (make_pass_rtl_unswitch): ...new function.
1747 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1748 subclass of rtl_opt_pass along with...
1749 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1751 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1752 (pass_rtl_doloop): Convert from a global struct to a subclass of
1753 rtl_opt_pass along with...
1754 (pass_data_rtl_doloop): ...new pass_data instance and...
1755 (make_pass_rtl_doloop): ...new function.
1756 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1757 a subclass of rtl_opt_pass along with...
1758 (pass_data_lower_subreg): ...new pass_data instance and...
1759 (make_pass_lower_subreg): ...new function.
1760 (pass_lower_subreg2): Convert from a global struct to a subclass of
1761 rtl_opt_pass along with...
1762 (pass_data_lower_subreg2): ...new pass_data instance and...
1763 (make_pass_lower_subreg2): ...new function.
1764 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1765 struct to a subclass of ipa_opt_pass_d along with...
1766 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1767 (make_pass_ipa_lto_gimple_out): ...new function.
1768 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1769 of ipa_opt_pass_d along with...
1770 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1771 (make_pass_ipa_lto_finish_out): ...new function.
1772 * mode-switching.c (pass_mode_switching): Convert from a global struct
1773 to a subclass of rtl_opt_pass along with...
1774 (pass_data_mode_switching): ...new pass_data instance and...
1775 (make_pass_mode_switching): ...new function.
1776 * modulo-sched.c (pass_sms): Convert from a global struct to a
1777 subclass of rtl_opt_pass along with...
1778 (pass_data_sms): ...new pass_data instance and...
1779 (make_pass_sms): ...new function.
1780 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1781 subclass of gimple_opt_pass along with...
1782 (pass_data_expand_omp): ...new pass_data instance and...
1783 (make_pass_expand_omp): ...new function.
1784 (pass_lower_omp): Convert from a global struct to a subclass of
1785 gimple_opt_pass along with...
1786 (pass_data_lower_omp): ...new pass_data instance and...
1787 (make_pass_lower_omp): ...new function.
1788 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1789 of gimple_opt_pass along with...
1790 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1791 (make_pass_diagnose_omp_blocks): ...new function.
1792 * passes.c (pass_early_local_passes): Convert from a global struct to
1793 a subclass of simple_ipa_opt_pass along with...
1794 (pass_data_early_local_passes): ...new pass_data instance and...
1795 (make_pass_early_local_passes): ...new function.
1796 (pass_all_early_optimizations): Convert from a global struct to a
1797 subclass of gimple_opt_pass along with...
1798 (pass_data_all_early_optimizations): ...new pass_data instance and...
1799 (make_pass_all_early_optimizations): ...new function.
1800 (pass_all_optimizations): Convert from a global struct to a subclass
1801 of gimple_opt_pass along with...
1802 (pass_data_all_optimizations): ...new pass_data instance and...
1803 (make_pass_all_optimizations): ...new function.
1804 (pass_all_optimizations_g): Convert from a global struct to a subclass
1805 of gimple_opt_pass along with...
1806 (pass_data_all_optimizations_g): ...new pass_data instance and...
1807 (make_pass_all_optimizations_g): ...new function.
1808 (pass_rest_of_compilation): Convert from a global struct to a subclass
1809 of rtl_opt_pass along with...
1810 (pass_data_rest_of_compilation): ...new pass_data instance and...
1811 (make_pass_rest_of_compilation): ...new function.
1812 (pass_postreload): Convert from a global struct to a subclass of
1813 rtl_opt_pass along with...
1814 (pass_data_postreload): ...new pass_data instance and...
1815 (make_pass_postreload): ...new function.
1816 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1817 subclass of rtl_opt_pass along with...
1818 (pass_data_gcse2): ...new pass_data instance and...
1819 (make_pass_gcse2): ...new function.
1820 * postreload.c (pass_postreload_cse): Convert from a global struct to
1821 a subclass of rtl_opt_pass along with...
1822 (pass_data_postreload_cse): ...new pass_data instance and...
1823 (make_pass_postreload_cse): ...new function.
1824 * predict.c (pass_profile): Convert from a global struct to a subclass
1825 of gimple_opt_pass along with...
1826 (pass_data_profile): ...new pass_data instance and...
1827 (make_pass_profile): ...new function.
1828 (pass_strip_predict_hints): Convert from a global struct to a subclass
1829 of gimple_opt_pass along with...
1830 (pass_data_strip_predict_hints): ...new pass_data instance and...
1831 (make_pass_strip_predict_hints): ...new function.
1832 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1833 of rtl_opt_pass along with...
1834 (pass_data_peephole2): ...new pass_data instance and...
1835 (make_pass_peephole2): ...new function.
1836 (pass_split_all_insns): Convert from a global struct to a subclass of
1837 rtl_opt_pass along with...
1838 (pass_data_split_all_insns): ...new pass_data instance and...
1839 (make_pass_split_all_insns): ...new function.
1840 (pass_split_after_reload): Convert from a global struct to a subclass
1841 of rtl_opt_pass along with...
1842 (pass_data_split_after_reload): ...new pass_data instance and...
1843 (make_pass_split_after_reload): ...new function.
1844 (pass_split_before_regstack): Convert from a global struct to a
1845 subclass of rtl_opt_pass along with...
1846 (pass_data_split_before_regstack): ...new pass_data instance and...
1847 (make_pass_split_before_regstack): ...new function.
1848 (pass_split_before_sched2): Convert from a global struct to a subclass
1849 of rtl_opt_pass along with...
1850 (pass_data_split_before_sched2): ...new pass_data instance and...
1851 (make_pass_split_before_sched2): ...new function.
1852 (pass_split_for_shorten_branches): Convert from a global struct to a
1853 subclass of rtl_opt_pass along with...
1854 (pass_data_split_for_shorten_branches): ...new pass_data instance
1856 (make_pass_split_for_shorten_branches): ...new function.
1857 * ree.c (pass_ree): Convert from a global struct to a subclass of
1858 rtl_opt_pass along with...
1859 (pass_data_ree): ...new pass_data instance and...
1860 (make_pass_ree): ...new function.
1861 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1862 subclass of rtl_opt_pass along with...
1863 (pass_data_stack_regs): ...new pass_data instance and...
1864 (make_pass_stack_regs): ...new function.
1865 (pass_stack_regs_run): Convert from a global struct to a subclass of
1866 rtl_opt_pass along with...
1867 (pass_data_stack_regs_run): ...new pass_data instance and...
1868 (make_pass_stack_regs_run): ...new function.
1869 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1870 subclass of rtl_opt_pass along with...
1871 (pass_data_cprop_hardreg): ...new pass_data instance and...
1872 (make_pass_cprop_hardreg): ...new function.
1873 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1874 subclass of rtl_opt_pass along with...
1875 (pass_data_reginfo_init): ...new pass_data instance and...
1876 (make_pass_reginfo_init): ...new function.
1877 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1878 of rtl_opt_pass along with...
1879 (pass_data_regmove): ...new pass_data instance and...
1880 (make_pass_regmove): ...new function.
1881 * regrename.c (pass_regrename): Convert from a global struct to a
1882 subclass of rtl_opt_pass along with...
1883 (pass_data_regrename): ...new pass_data instance and...
1884 (make_pass_regrename): ...new function.
1885 * reorg.c (pass_delay_slots): Convert from a global struct to a
1886 subclass of rtl_opt_pass along with...
1887 (pass_data_delay_slots): ...new pass_data instance and...
1888 (make_pass_delay_slots): ...new function.
1889 (pass_machine_reorg): Convert from a global struct to a subclass of
1890 rtl_opt_pass along with...
1891 (pass_data_machine_reorg): ...new pass_data instance and...
1892 (make_pass_machine_reorg): ...new function.
1893 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1894 of rtl_opt_pass along with...
1895 (pass_data_sched): ...new pass_data instance and...
1896 (make_pass_sched): ...new function.
1897 (pass_sched2): Convert from a global struct to a subclass of
1898 rtl_opt_pass along with...
1899 (pass_data_sched2): ...new pass_data instance and...
1900 (make_pass_sched2): ...new function.
1901 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1902 to a subclass of rtl_opt_pass along with...
1903 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1904 (make_pass_stack_ptr_mod): ...new function.
1905 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1906 to a subclass of rtl_opt_pass along with...
1907 (pass_data_rtl_store_motion): ...new pass_data instance and...
1908 (make_pass_rtl_store_motion): ...new function.
1909 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1910 of gimple_opt_pass along with...
1911 (pass_data_tracer): ...new pass_data instance and...
1912 (make_pass_tracer): ...new function.
1913 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1914 to a subclass of gimple_opt_pass along with...
1915 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1916 (make_pass_diagnose_tm_blocks): ...new function.
1917 (pass_lower_tm): Convert from a global struct to a subclass of
1918 gimple_opt_pass along with...
1919 (pass_data_lower_tm): ...new pass_data instance and...
1920 (make_pass_lower_tm): ...new function.
1921 (pass_tm_init): Convert from a global struct to a subclass of
1922 gimple_opt_pass along with...
1923 (pass_data_tm_init): ...new pass_data instance and...
1924 (make_pass_tm_init): ...new function.
1925 (pass_tm_mark): Convert from a global struct to a subclass of
1926 gimple_opt_pass along with...
1927 (pass_data_tm_mark): ...new pass_data instance and...
1928 (make_pass_tm_mark): ...new function.
1929 (pass_tm_edges): Convert from a global struct to a subclass of
1930 gimple_opt_pass along with...
1931 (pass_data_tm_edges): ...new pass_data instance and...
1932 (make_pass_tm_edges): ...new function.
1933 (pass_tm_memopt): Convert from a global struct to a subclass of
1934 gimple_opt_pass along with...
1935 (pass_data_tm_memopt): ...new pass_data instance and...
1936 (make_pass_tm_memopt): ...new function.
1937 (pass_ipa_tm): Convert from a global struct to a subclass of
1938 simple_ipa_opt_pass along with...
1939 (pass_data_ipa_tm): ...new pass_data instance and...
1940 (make_pass_ipa_tm): ...new function.
1941 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1942 subclass of gimple_opt_pass along with...
1943 (pass_data_call_cdce): ...new pass_data instance and...
1944 (make_pass_call_cdce): ...new function.
1945 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1946 subclass of gimple_opt_pass along with...
1947 (pass_data_build_cfg): ...new pass_data instance and...
1948 (make_pass_build_cfg): ...new function.
1949 (pass_split_crit_edges): Convert from a global struct to a subclass of
1950 gimple_opt_pass along with...
1951 (pass_data_split_crit_edges): ...new pass_data instance and...
1952 (make_pass_split_crit_edges): ...new function.
1953 (pass_warn_function_return): Convert from a global struct to a
1954 subclass of gimple_opt_pass along with...
1955 (pass_data_warn_function_return): ...new pass_data instance and...
1956 (make_pass_warn_function_return): ...new function.
1957 (pass_warn_function_noreturn): Convert from a global struct to a
1958 subclass of gimple_opt_pass along with...
1959 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1960 (make_pass_warn_function_noreturn): ...new function.
1961 (pass_warn_unused_result): Convert from a global struct to a subclass
1962 of gimple_opt_pass along with...
1963 (pass_data_warn_unused_result): ...new pass_data instance and...
1964 (make_pass_warn_unused_result): ...new function.
1965 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1966 a subclass of gimple_opt_pass along with...
1967 (pass_data_merge_phi): ...new pass_data instance and...
1968 (make_pass_merge_phi): ...new function.
1969 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1970 a subclass of gimple_opt_pass along with...
1971 (pass_data_lower_complex): ...new pass_data instance and...
1972 (make_pass_lower_complex): ...new function.
1973 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1974 gimple_opt_pass along with...
1975 (pass_data_lower_complex_O0): ...new pass_data instance and...
1976 (make_pass_lower_complex_O0): ...new function.
1977 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1978 subclass of gimple_opt_pass along with...
1979 (pass_data_lower_eh): ...new pass_data instance and...
1980 (make_pass_lower_eh): ...new function.
1981 (pass_refactor_eh): Convert from a global struct to a subclass of
1982 gimple_opt_pass along with...
1983 (pass_data_refactor_eh): ...new pass_data instance and...
1984 (make_pass_refactor_eh): ...new function.
1985 (pass_lower_resx): Convert from a global struct to a subclass of
1986 gimple_opt_pass along with...
1987 (pass_data_lower_resx): ...new pass_data instance and...
1988 (make_pass_lower_resx): ...new function.
1989 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1990 of gimple_opt_pass along with...
1991 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1992 (make_pass_lower_eh_dispatch): ...new function.
1993 (pass_cleanup_eh): Convert from a global struct to a subclass of
1994 gimple_opt_pass along with...
1995 (pass_data_cleanup_eh): ...new pass_data instance and...
1996 (make_pass_cleanup_eh): ...new function.
1997 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1998 to a subclass of simple_ipa_opt_pass along with...
1999 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
2000 (make_pass_ipa_lower_emutls): ...new function.
2001 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
2002 a subclass of gimple_opt_pass along with...
2003 (pass_data_if_conversion): ...new pass_data instance and...
2004 (make_pass_if_conversion): ...new function.
2005 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
2006 subclass of gimple_opt_pass along with...
2007 (pass_data_build_ssa): ...new pass_data instance and...
2008 (make_pass_build_ssa): ...new function.
2009 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
2010 global struct to a subclass of gimple_opt_pass along with...
2011 (pass_data_loop_distribution): ...new pass_data instance and...
2012 (make_pass_loop_distribution): ...new function.
2013 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
2014 subclass of gimple_opt_pass along with...
2015 (pass_data_mudflap_1): ...new pass_data instance and...
2016 (make_pass_mudflap_1): ...new function.
2017 (pass_mudflap_2): Convert from a global struct to a subclass of
2018 gimple_opt_pass along with...
2019 (pass_data_mudflap_2): ...new pass_data instance and...
2020 (make_pass_mudflap_2): ...new function.
2021 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
2022 subclass of gimple_opt_pass along with...
2023 (pass_data_mudflap_1): ...new pass_data instance and...
2024 (make_pass_mudflap_1): ...new function.
2025 (pass_mudflap_2): Convert from a global struct to a subclass of
2026 gimple_opt_pass along with...
2027 (pass_data_mudflap_2): ...new pass_data instance and...
2028 (make_pass_mudflap_2): ...new function.
2029 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
2030 gimple_opt_pass along with...
2031 (pass_data_nrv): ...new pass_data instance and...
2032 (make_pass_nrv): ...new function.
2033 (pass_return_slot): Convert from a global struct to a subclass of
2034 gimple_opt_pass along with...
2035 (pass_data_return_slot): ...new pass_data instance and...
2036 (make_pass_return_slot): ...new function.
2037 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2038 to a subclass of gimple_opt_pass along with...
2039 (pass_data_object_sizes): ...new pass_data instance and...
2040 (make_pass_object_sizes): ...new function.
2041 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2042 global struct to a subclass of gimple_opt_pass along with...
2043 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2045 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2046 (pass_fixup_cfg): Convert from a global struct to a subclass of
2047 gimple_opt_pass along with...
2048 (pass_data_fixup_cfg): ...new pass_data instance and...
2049 (make_pass_fixup_cfg): ...new function.
2050 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2051 (make_pass_mudflap_1): ...new function.
2052 (pass_mudflap_2): Replace declaration with that of...
2053 (make_pass_mudflap_2): ...new function.
2054 (pass_asan): Replace declaration with that of...
2055 (make_pass_asan): ...new function.
2056 (pass_asan_O0): Replace declaration with that of...
2057 (make_pass_asan_O0): ...new function.
2058 (pass_tsan): Replace declaration with that of...
2059 (make_pass_tsan): ...new function.
2060 (pass_tsan_O0): Replace declaration with that of...
2061 (make_pass_tsan_O0): ...new function.
2062 (pass_lower_cf): Replace declaration with that of...
2063 (make_pass_lower_cf): ...new function.
2064 (pass_refactor_eh): Replace declaration with that of...
2065 (make_pass_refactor_eh): ...new function.
2066 (pass_lower_eh): Replace declaration with that of...
2067 (make_pass_lower_eh): ...new function.
2068 (pass_lower_eh_dispatch): Replace declaration with that of...
2069 (make_pass_lower_eh_dispatch): ...new function.
2070 (pass_lower_resx): Replace declaration with that of...
2071 (make_pass_lower_resx): ...new function.
2072 (pass_build_cfg): Replace declaration with that of...
2073 (make_pass_build_cfg): ...new function.
2074 (pass_early_tree_profile): Replace declaration with that of...
2075 (make_pass_early_tree_profile): ...new function.
2076 (pass_cleanup_eh): Replace declaration with that of...
2077 (make_pass_cleanup_eh): ...new function.
2078 (pass_sra): Replace declaration with that of...
2079 (make_pass_sra): ...new function.
2080 (pass_sra_early): Replace declaration with that of...
2081 (make_pass_sra_early): ...new function.
2082 (pass_early_ipa_sra): Replace declaration with that of...
2083 (make_pass_early_ipa_sra): ...new function.
2084 (pass_tail_recursion): Replace declaration with that of...
2085 (make_pass_tail_recursion): ...new function.
2086 (pass_tail_calls): Replace declaration with that of...
2087 (make_pass_tail_calls): ...new function.
2088 (pass_tree_loop): Replace declaration with that of...
2089 (make_pass_tree_loop): ...new function.
2090 (pass_tree_loop_init): Replace declaration with that of...
2091 (make_pass_tree_loop_init): ...new function.
2092 (pass_lim): Replace declaration with that of...
2093 (make_pass_lim): ...new function.
2094 (pass_tree_unswitch): Replace declaration with that of...
2095 (make_pass_tree_unswitch): ...new function.
2096 (pass_predcom): Replace declaration with that of...
2097 (make_pass_predcom): ...new function.
2098 (pass_iv_canon): Replace declaration with that of...
2099 (make_pass_iv_canon): ...new function.
2100 (pass_scev_cprop): Replace declaration with that of...
2101 (make_pass_scev_cprop): ...new function.
2102 (pass_empty_loop): Replace declaration with that of...
2103 (make_pass_empty_loop): ...new function.
2104 (pass_record_bounds): Replace declaration with that of...
2105 (make_pass_record_bounds): ...new function.
2106 (pass_graphite): Replace declaration with that of...
2107 (make_pass_graphite): ...new function.
2108 (pass_graphite_transforms): Replace declaration with that of...
2109 (make_pass_graphite_transforms): ...new function.
2110 (pass_if_conversion): Replace declaration with that of...
2111 (make_pass_if_conversion): ...new function.
2112 (pass_loop_distribution): Replace declaration with that of...
2113 (make_pass_loop_distribution): ...new function.
2114 (pass_vectorize): Replace declaration with that of...
2115 (make_pass_vectorize): ...new function.
2116 (pass_slp_vectorize): Replace declaration with that of...
2117 (make_pass_slp_vectorize): ...new function.
2118 (pass_complete_unroll): Replace declaration with that of...
2119 (make_pass_complete_unroll): ...new function.
2120 (pass_complete_unrolli): Replace declaration with that of...
2121 (make_pass_complete_unrolli): ...new function.
2122 (pass_parallelize_loops): Replace declaration with that of...
2123 (make_pass_parallelize_loops): ...new function.
2124 (pass_loop_prefetch): Replace declaration with that of...
2125 (make_pass_loop_prefetch): ...new function.
2126 (pass_iv_optimize): Replace declaration with that of...
2127 (make_pass_iv_optimize): ...new function.
2128 (pass_tree_loop_done): Replace declaration with that of...
2129 (make_pass_tree_loop_done): ...new function.
2130 (pass_ch): Replace declaration with that of...
2131 (make_pass_ch): ...new function.
2132 (pass_ccp): Replace declaration with that of...
2133 (make_pass_ccp): ...new function.
2134 (pass_phi_only_cprop): Replace declaration with that of...
2135 (make_pass_phi_only_cprop): ...new function.
2136 (pass_build_ssa): Replace declaration with that of...
2137 (make_pass_build_ssa): ...new function.
2138 (pass_build_alias): Replace declaration with that of...
2139 (make_pass_build_alias): ...new function.
2140 (pass_build_ealias): Replace declaration with that of...
2141 (make_pass_build_ealias): ...new function.
2142 (pass_dominator): Replace declaration with that of...
2143 (make_pass_dominator): ...new function.
2144 (pass_dce): Replace declaration with that of...
2145 (make_pass_dce): ...new function.
2146 (pass_dce_loop): Replace declaration with that of...
2147 (make_pass_dce_loop): ...new function.
2148 (pass_cd_dce): Replace declaration with that of...
2149 (make_pass_cd_dce): ...new function.
2150 (pass_call_cdce): Replace declaration with that of...
2151 (make_pass_call_cdce): ...new function.
2152 (pass_merge_phi): Replace declaration with that of...
2153 (make_pass_merge_phi): ...new function.
2154 (pass_split_crit_edges): Replace declaration with that of...
2155 (make_pass_split_crit_edges): ...new function.
2156 (pass_pre): Replace declaration with that of...
2157 (make_pass_pre): ...new function.
2158 (pass_profile): Replace declaration with that of...
2159 (make_pass_profile): ...new function.
2160 (pass_strip_predict_hints): Replace declaration with that of...
2161 (make_pass_strip_predict_hints): ...new function.
2162 (pass_lower_complex_O0): Replace declaration with that of...
2163 (make_pass_lower_complex_O0): ...new function.
2164 (pass_lower_complex): Replace declaration with that of...
2165 (make_pass_lower_complex): ...new function.
2166 (pass_lower_vector): Replace declaration with that of...
2167 (make_pass_lower_vector): ...new function.
2168 (pass_lower_vector_ssa): Replace declaration with that of...
2169 (make_pass_lower_vector_ssa): ...new function.
2170 (pass_lower_omp): Replace declaration with that of...
2171 (make_pass_lower_omp): ...new function.
2172 (pass_diagnose_omp_blocks): Replace declaration with that of...
2173 (make_pass_diagnose_omp_blocks): ...new function.
2174 (pass_expand_omp): Replace declaration with that of...
2175 (make_pass_expand_omp): ...new function.
2176 (pass_expand_omp_ssa): Replace declaration with that of...
2177 (make_pass_expand_omp_ssa): ...new function.
2178 (pass_object_sizes): Replace declaration with that of...
2179 (make_pass_object_sizes): ...new function.
2180 (pass_strlen): Replace declaration with that of...
2181 (make_pass_strlen): ...new function.
2182 (pass_fold_builtins): Replace declaration with that of...
2183 (make_pass_fold_builtins): ...new function.
2184 (pass_stdarg): Replace declaration with that of...
2185 (make_pass_stdarg): ...new function.
2186 (pass_early_warn_uninitialized): Replace declaration with that of...
2187 (make_pass_early_warn_uninitialized): ...new function.
2188 (pass_late_warn_uninitialized): Replace declaration with that of...
2189 (make_pass_late_warn_uninitialized): ...new function.
2190 (pass_cse_reciprocals): Replace declaration with that of...
2191 (make_pass_cse_reciprocals): ...new function.
2192 (pass_cse_sincos): Replace declaration with that of...
2193 (make_pass_cse_sincos): ...new function.
2194 (pass_optimize_bswap): Replace declaration with that of...
2195 (make_pass_optimize_bswap): ...new function.
2196 (pass_optimize_widening_mul): Replace declaration with that of...
2197 (make_pass_optimize_widening_mul): ...new function.
2198 (pass_warn_function_return): Replace declaration with that of...
2199 (make_pass_warn_function_return): ...new function.
2200 (pass_warn_function_noreturn): Replace declaration with that of...
2201 (make_pass_warn_function_noreturn): ...new function.
2202 (pass_cselim): Replace declaration with that of...
2203 (make_pass_cselim): ...new function.
2204 (pass_phiopt): Replace declaration with that of...
2205 (make_pass_phiopt): ...new function.
2206 (pass_forwprop): Replace declaration with that of...
2207 (make_pass_forwprop): ...new function.
2208 (pass_phiprop): Replace declaration with that of...
2209 (make_pass_phiprop): ...new function.
2210 (pass_tree_ifcombine): Replace declaration with that of...
2211 (make_pass_tree_ifcombine): ...new function.
2212 (pass_dse): Replace declaration with that of...
2213 (make_pass_dse): ...new function.
2214 (pass_nrv): Replace declaration with that of...
2215 (make_pass_nrv): ...new function.
2216 (pass_rename_ssa_copies): Replace declaration with that of...
2217 (make_pass_rename_ssa_copies): ...new function.
2218 (pass_sink_code): Replace declaration with that of...
2219 (make_pass_sink_code): ...new function.
2220 (pass_fre): Replace declaration with that of...
2221 (make_pass_fre): ...new function.
2222 (pass_check_data_deps): Replace declaration with that of...
2223 (make_pass_check_data_deps): ...new function.
2224 (pass_copy_prop): Replace declaration with that of...
2225 (make_pass_copy_prop): ...new function.
2226 (pass_vrp): Replace declaration with that of...
2227 (make_pass_vrp): ...new function.
2228 (pass_uncprop): Replace declaration with that of...
2229 (make_pass_uncprop): ...new function.
2230 (pass_return_slot): Replace declaration with that of...
2231 (make_pass_return_slot): ...new function.
2232 (pass_reassoc): Replace declaration with that of...
2233 (make_pass_reassoc): ...new function.
2234 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2235 (make_pass_rebuild_cgraph_edges): ...new function.
2236 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2237 (make_pass_remove_cgraph_callee_edges): ...new function.
2238 (pass_build_cgraph_edges): Replace declaration with that of...
2239 (make_pass_build_cgraph_edges): ...new function.
2240 (pass_local_pure_const): Replace declaration with that of...
2241 (make_pass_local_pure_const): ...new function.
2242 (pass_tracer): Replace declaration with that of...
2243 (make_pass_tracer): ...new function.
2244 (pass_warn_unused_result): Replace declaration with that of...
2245 (make_pass_warn_unused_result): ...new function.
2246 (pass_diagnose_tm_blocks): Replace declaration with that of...
2247 (make_pass_diagnose_tm_blocks): ...new function.
2248 (pass_lower_tm): Replace declaration with that of...
2249 (make_pass_lower_tm): ...new function.
2250 (pass_tm_init): Replace declaration with that of...
2251 (make_pass_tm_init): ...new function.
2252 (pass_tm_mark): Replace declaration with that of...
2253 (make_pass_tm_mark): ...new function.
2254 (pass_tm_memopt): Replace declaration with that of...
2255 (make_pass_tm_memopt): ...new function.
2256 (pass_tm_edges): Replace declaration with that of...
2257 (make_pass_tm_edges): ...new function.
2258 (pass_split_functions): Replace declaration with that of...
2259 (make_pass_split_functions): ...new function.
2260 (pass_feedback_split_functions): Replace declaration with that of...
2261 (make_pass_feedback_split_functions): ...new function.
2262 (pass_strength_reduction): Replace declaration with that of...
2263 (make_pass_strength_reduction): ...new function.
2264 (pass_ipa_lower_emutls): Replace declaration with that of...
2265 (make_pass_ipa_lower_emutls): ...new function.
2266 (pass_ipa_function_and_variable_visibility): Replace declaration with
2268 (make_pass_ipa_function_and_variable_visibility): ...new function.
2269 (pass_ipa_tree_profile): Replace declaration with that of...
2270 (make_pass_ipa_tree_profile): ...new function.
2271 (pass_early_local_passes): Replace declaration with that of...
2272 (make_pass_early_local_passes): ...new function.
2273 (pass_ipa_whole_program_visibility): Replace declaration with that
2275 (make_pass_ipa_whole_program_visibility): ...new function.
2276 (pass_ipa_lto_gimple_out): Replace declaration with that of...
2277 (make_pass_ipa_lto_gimple_out): ...new function.
2278 (pass_ipa_increase_alignment): Replace declaration with that of...
2279 (make_pass_ipa_increase_alignment): ...new function.
2280 (pass_ipa_inline): Replace declaration with that of...
2281 (make_pass_ipa_inline): ...new function.
2282 (pass_ipa_free_lang_data): Replace declaration with that of...
2283 (make_pass_ipa_free_lang_data): ...new function.
2284 (pass_ipa_free_inline_summary): Replace declaration with that of...
2285 (make_pass_ipa_free_inline_summary): ...new function.
2286 (pass_ipa_cp): Replace declaration with that of...
2287 (make_pass_ipa_cp): ...new function.
2288 (pass_ipa_reference): Replace declaration with that of...
2289 (make_pass_ipa_reference): ...new function.
2290 (pass_ipa_pure_const): Replace declaration with that of...
2291 (make_pass_ipa_pure_const): ...new function.
2292 (pass_ipa_pta): Replace declaration with that of...
2293 (make_pass_ipa_pta): ...new function.
2294 (pass_ipa_lto_finish_out): Replace declaration with that of...
2295 (make_pass_ipa_lto_finish_out): ...new function.
2296 (pass_ipa_tm): Replace declaration with that of...
2297 (make_pass_ipa_tm): ...new function.
2298 (pass_ipa_profile): Replace declaration with that of...
2299 (make_pass_ipa_profile): ...new function.
2300 (pass_ipa_cdtor_merge): Replace declaration with that of...
2301 (make_pass_ipa_cdtor_merge): ...new function.
2302 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
2304 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2305 (pass_init_datastructures): Replace declaration with that of...
2306 (make_pass_init_datastructures): ...new function.
2307 (pass_fixup_cfg): Replace declaration with that of...
2308 (make_pass_fixup_cfg): ...new function.
2309 (pass_expand): Replace declaration with that of...
2310 (make_pass_expand): ...new function.
2311 (pass_instantiate_virtual_regs): Replace declaration with that of...
2312 (make_pass_instantiate_virtual_regs): ...new function.
2313 (pass_rtl_fwprop): Replace declaration with that of...
2314 (make_pass_rtl_fwprop): ...new function.
2315 (pass_rtl_fwprop_addr): Replace declaration with that of...
2316 (make_pass_rtl_fwprop_addr): ...new function.
2317 (pass_jump): Replace declaration with that of...
2318 (make_pass_jump): ...new function.
2319 (pass_jump2): Replace declaration with that of...
2320 (make_pass_jump2): ...new function.
2321 (pass_lower_subreg): Replace declaration with that of...
2322 (make_pass_lower_subreg): ...new function.
2323 (pass_cse): Replace declaration with that of...
2324 (make_pass_cse): ...new function.
2325 (pass_fast_rtl_dce): Replace declaration with that of...
2326 (make_pass_fast_rtl_dce): ...new function.
2327 (pass_ud_rtl_dce): Replace declaration with that of...
2328 (make_pass_ud_rtl_dce): ...new function.
2329 (pass_rtl_dce): Replace declaration with that of...
2330 (make_pass_rtl_dce): ...new function.
2331 (pass_rtl_dse1): Replace declaration with that of...
2332 (make_pass_rtl_dse1): ...new function.
2333 (pass_rtl_dse2): Replace declaration with that of...
2334 (make_pass_rtl_dse2): ...new function.
2335 (pass_rtl_dse3): Replace declaration with that of...
2336 (make_pass_rtl_dse3): ...new function.
2337 (pass_rtl_cprop): Replace declaration with that of...
2338 (make_pass_rtl_cprop): ...new function.
2339 (pass_rtl_pre): Replace declaration with that of...
2340 (make_pass_rtl_pre): ...new function.
2341 (pass_rtl_hoist): Replace declaration with that of...
2342 (make_pass_rtl_hoist): ...new function.
2343 (pass_rtl_store_motion): Replace declaration with that of...
2344 (make_pass_rtl_store_motion): ...new function.
2345 (pass_cse_after_global_opts): Replace declaration with that of...
2346 (make_pass_cse_after_global_opts): ...new function.
2347 (pass_rtl_ifcvt): Replace declaration with that of...
2348 (make_pass_rtl_ifcvt): ...new function.
2349 (pass_into_cfg_layout_mode): Replace declaration with that of...
2350 (make_pass_into_cfg_layout_mode): ...new function.
2351 (pass_outof_cfg_layout_mode): Replace declaration with that of...
2352 (make_pass_outof_cfg_layout_mode): ...new function.
2353 (pass_loop2): Replace declaration with that of...
2354 (make_pass_loop2): ...new function.
2355 (pass_rtl_loop_init): Replace declaration with that of...
2356 (make_pass_rtl_loop_init): ...new function.
2357 (pass_rtl_move_loop_invariants): Replace declaration with that of...
2358 (make_pass_rtl_move_loop_invariants): ...new function.
2359 (pass_rtl_unswitch): Replace declaration with that of...
2360 (make_pass_rtl_unswitch): ...new function.
2361 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
2362 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2363 (pass_rtl_doloop): Replace declaration with that of...
2364 (make_pass_rtl_doloop): ...new function.
2365 (pass_rtl_loop_done): Replace declaration with that of...
2366 (make_pass_rtl_loop_done): ...new function.
2367 (pass_web): Replace declaration with that of...
2368 (make_pass_web): ...new function.
2369 (pass_cse2): Replace declaration with that of...
2370 (make_pass_cse2): ...new function.
2371 (pass_df_initialize_opt): Replace declaration with that of...
2372 (make_pass_df_initialize_opt): ...new function.
2373 (pass_df_initialize_no_opt): Replace declaration with that of...
2374 (make_pass_df_initialize_no_opt): ...new function.
2375 (pass_reginfo_init): Replace declaration with that of...
2376 (make_pass_reginfo_init): ...new function.
2377 (pass_inc_dec): Replace declaration with that of...
2378 (make_pass_inc_dec): ...new function.
2379 (pass_stack_ptr_mod): Replace declaration with that of...
2380 (make_pass_stack_ptr_mod): ...new function.
2381 (pass_initialize_regs): Replace declaration with that of...
2382 (make_pass_initialize_regs): ...new function.
2383 (pass_combine): Replace declaration with that of...
2384 (make_pass_combine): ...new function.
2385 (pass_if_after_combine): Replace declaration with that of...
2386 (make_pass_if_after_combine): ...new function.
2387 (pass_ree): Replace declaration with that of...
2388 (make_pass_ree): ...new function.
2389 (pass_partition_blocks): Replace declaration with that of...
2390 (make_pass_partition_blocks): ...new function.
2391 (pass_match_asm_constraints): Replace declaration with that of...
2392 (make_pass_match_asm_constraints): ...new function.
2393 (pass_regmove): Replace declaration with that of...
2394 (make_pass_regmove): ...new function.
2395 (pass_split_all_insns): Replace declaration with that of...
2396 (make_pass_split_all_insns): ...new function.
2397 (pass_fast_rtl_byte_dce): Replace declaration with that of...
2398 (make_pass_fast_rtl_byte_dce): ...new function.
2399 (pass_lower_subreg2): Replace declaration with that of...
2400 (make_pass_lower_subreg2): ...new function.
2401 (pass_mode_switching): Replace declaration with that of...
2402 (make_pass_mode_switching): ...new function.
2403 (pass_sms): Replace declaration with that of...
2404 (make_pass_sms): ...new function.
2405 (pass_sched): Replace declaration with that of...
2406 (make_pass_sched): ...new function.
2407 (pass_ira): Replace declaration with that of...
2408 (make_pass_ira): ...new function.
2409 (pass_reload): Replace declaration with that of...
2410 (make_pass_reload): ...new function.
2411 (pass_clean_state): Replace declaration with that of...
2412 (make_pass_clean_state): ...new function.
2413 (pass_branch_prob): Replace declaration with that of...
2414 (make_pass_branch_prob): ...new function.
2415 (pass_value_profile_transformations): Replace declaration with that
2417 (make_pass_value_profile_transformations): ...new function.
2418 (pass_postreload_cse): Replace declaration with that of...
2419 (make_pass_postreload_cse): ...new function.
2420 (pass_gcse2): Replace declaration with that of...
2421 (make_pass_gcse2): ...new function.
2422 (pass_split_after_reload): Replace declaration with that of...
2423 (make_pass_split_after_reload): ...new function.
2424 (pass_branch_target_load_optimize1): Replace declaration with that
2426 (make_pass_branch_target_load_optimize1): ...new function.
2427 (pass_thread_prologue_and_epilogue): Replace declaration with that
2429 (make_pass_thread_prologue_and_epilogue): ...new function.
2430 (pass_stack_adjustments): Replace declaration with that of...
2431 (make_pass_stack_adjustments): ...new function.
2432 (pass_peephole2): Replace declaration with that of...
2433 (make_pass_peephole2): ...new function.
2434 (pass_if_after_reload): Replace declaration with that of...
2435 (make_pass_if_after_reload): ...new function.
2436 (pass_regrename): Replace declaration with that of...
2437 (make_pass_regrename): ...new function.
2438 (pass_cprop_hardreg): Replace declaration with that of...
2439 (make_pass_cprop_hardreg): ...new function.
2440 (pass_reorder_blocks): Replace declaration with that of...
2441 (make_pass_reorder_blocks): ...new function.
2442 (pass_branch_target_load_optimize2): Replace declaration with that
2444 (make_pass_branch_target_load_optimize2): ...new function.
2445 (pass_leaf_regs): Replace declaration with that of...
2446 (make_pass_leaf_regs): ...new function.
2447 (pass_split_before_sched2): Replace declaration with that of...
2448 (make_pass_split_before_sched2): ...new function.
2449 (pass_compare_elim_after_reload): Replace declaration with that of...
2450 (make_pass_compare_elim_after_reload): ...new function.
2451 (pass_sched2): Replace declaration with that of...
2452 (make_pass_sched2): ...new function.
2453 (pass_stack_regs): Replace declaration with that of...
2454 (make_pass_stack_regs): ...new function.
2455 (pass_stack_regs_run): Replace declaration with that of...
2456 (make_pass_stack_regs_run): ...new function.
2457 (pass_df_finish): Replace declaration with that of...
2458 (make_pass_df_finish): ...new function.
2459 (pass_compute_alignments): Replace declaration with that of...
2460 (make_pass_compute_alignments): ...new function.
2461 (pass_duplicate_computed_gotos): Replace declaration with that of...
2462 (make_pass_duplicate_computed_gotos): ...new function.
2463 (pass_variable_tracking): Replace declaration with that of...
2464 (make_pass_variable_tracking): ...new function.
2465 (pass_free_cfg): Replace declaration with that of...
2466 (make_pass_free_cfg): ...new function.
2467 (pass_machine_reorg): Replace declaration with that of...
2468 (make_pass_machine_reorg): ...new function.
2469 (pass_cleanup_barriers): Replace declaration with that of...
2470 (make_pass_cleanup_barriers): ...new function.
2471 (pass_delay_slots): Replace declaration with that of...
2472 (make_pass_delay_slots): ...new function.
2473 (pass_split_for_shorten_branches): Replace declaration with that of...
2474 (make_pass_split_for_shorten_branches): ...new function.
2475 (pass_split_before_regstack): Replace declaration with that of...
2476 (make_pass_split_before_regstack): ...new function.
2477 (pass_convert_to_eh_region_ranges): Replace declaration with that
2479 (make_pass_convert_to_eh_region_ranges): ...new function.
2480 (pass_shorten_branches): Replace declaration with that of...
2481 (make_pass_shorten_branches): ...new function.
2482 (pass_set_nothrow_function_flags): Replace declaration with that of...
2483 (make_pass_set_nothrow_function_flags): ...new function.
2484 (pass_dwarf2_frame): Replace declaration with that of...
2485 (make_pass_dwarf2_frame): ...new function.
2486 (pass_final): Replace declaration with that of...
2487 (make_pass_final): ...new function.
2488 (pass_rtl_seqabstr): Replace declaration with that of...
2489 (make_pass_rtl_seqabstr): ...new function.
2490 (pass_release_ssa_names): Replace declaration with that of...
2491 (make_pass_release_ssa_names): ...new function.
2492 (pass_early_inline): Replace declaration with that of...
2493 (make_pass_early_inline): ...new function.
2494 (pass_inline_parameters): Replace declaration with that of...
2495 (make_pass_inline_parameters): ...new function.
2496 (pass_update_address_taken): Replace declaration with that of...
2497 (make_pass_update_address_taken): ...new function.
2498 (pass_convert_switch): Replace declaration with that of...
2499 (make_pass_convert_switch): ...new function.
2500 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
2501 to a subclass of simple_ipa_opt_pass along with...
2502 (pass_data_ipa_tree_profile): ...new pass_data instance and...
2503 (make_pass_ipa_tree_profile): ...new function.
2504 * tree-sra.c (pass_sra_early): Convert from a global struct to a
2505 subclass of gimple_opt_pass along with...
2506 (pass_data_sra_early): ...new pass_data instance and...
2507 (make_pass_sra_early): ...new function.
2508 (pass_sra): Convert from a global struct to a subclass of
2509 gimple_opt_pass along with...
2510 (pass_data_sra): ...new pass_data instance and...
2511 (make_pass_sra): ...new function.
2512 (pass_early_ipa_sra): Convert from a global struct to a subclass of
2513 gimple_opt_pass along with...
2514 (pass_data_early_ipa_sra): ...new pass_data instance and...
2515 (make_pass_early_ipa_sra): ...new function.
2516 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
2517 subclass of gimple_opt_pass along with...
2518 (pass_data_ccp): ...new pass_data instance and...
2519 (make_pass_ccp): ...new function.
2520 (pass_fold_builtins): Convert from a global struct to a subclass of
2521 gimple_opt_pass along with...
2522 (pass_data_fold_builtins): ...new pass_data instance and...
2523 (make_pass_fold_builtins): ...new function.
2524 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2525 subclass of gimple_opt_pass along with...
2526 (pass_data_copy_prop): ...new pass_data instance and...
2527 (make_pass_copy_prop): ...new function.
2528 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2529 global struct to a subclass of gimple_opt_pass along with...
2530 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2531 (make_pass_rename_ssa_copies): ...new function.
2532 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2533 subclass of gimple_opt_pass along with...
2534 (pass_data_dce): ...new pass_data instance and...
2535 (make_pass_dce): ...new function.
2536 (pass_dce_loop): Convert from a global struct to a subclass of
2537 gimple_opt_pass along with...
2538 (pass_data_dce_loop): ...new pass_data instance and...
2539 (make_pass_dce_loop): ...new function.
2540 (pass_cd_dce): Convert from a global struct to a subclass of
2541 gimple_opt_pass along with...
2542 (pass_data_cd_dce): ...new pass_data instance and...
2543 (make_pass_cd_dce): ...new function.
2544 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2545 subclass of gimple_opt_pass along with...
2546 (pass_data_dominator): ...new pass_data instance and...
2547 (make_pass_dominator): ...new function.
2548 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2549 gimple_opt_pass along with...
2550 (pass_data_phi_only_cprop): ...new pass_data instance and...
2551 (make_pass_phi_only_cprop): ...new function.
2552 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2553 subclass of gimple_opt_pass along with...
2554 (pass_data_dse): ...new pass_data instance and...
2555 (make_pass_dse): ...new function.
2556 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2557 a subclass of gimple_opt_pass along with...
2558 (pass_data_forwprop): ...new pass_data instance and...
2559 (make_pass_forwprop): ...new function.
2560 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2561 struct to a subclass of gimple_opt_pass along with...
2562 (pass_data_tree_ifcombine): ...new pass_data instance and...
2563 (make_pass_tree_ifcombine): ...new function.
2564 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2565 subclass of gimple_opt_pass along with...
2566 (pass_data_ch): ...new pass_data instance and...
2567 (make_pass_ch): ...new function.
2568 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2569 subclass of gimple_opt_pass along with...
2570 (pass_data_tree_loop): ...new pass_data instance and...
2571 (make_pass_tree_loop): ...new function.
2572 (pass_tree_loop_init): Convert from a global struct to a subclass of
2573 gimple_opt_pass along with...
2574 (pass_data_tree_loop_init): ...new pass_data instance and...
2575 (make_pass_tree_loop_init): ...new function.
2576 (pass_lim): Convert from a global struct to a subclass of
2577 gimple_opt_pass along with...
2578 (pass_data_lim): ...new pass_data instance and...
2579 (make_pass_lim): ...new function.
2580 (pass_tree_unswitch): Convert from a global struct to a subclass of
2581 gimple_opt_pass along with...
2582 (pass_data_tree_unswitch): ...new pass_data instance and...
2583 (make_pass_tree_unswitch): ...new function.
2584 (pass_predcom): Convert from a global struct to a subclass of
2585 gimple_opt_pass along with...
2586 (pass_data_predcom): ...new pass_data instance and...
2587 (make_pass_predcom): ...new function.
2588 (pass_vectorize): Convert from a global struct to a subclass of
2589 gimple_opt_pass along with...
2590 (pass_data_vectorize): ...new pass_data instance and...
2591 (make_pass_vectorize): ...new function.
2592 (pass_graphite): Convert from a global struct to a subclass of
2593 gimple_opt_pass along with...
2594 (pass_data_graphite): ...new pass_data instance and...
2595 (make_pass_graphite): ...new function.
2596 (pass_graphite_transforms): Convert from a global struct to a subclass
2597 of gimple_opt_pass along with...
2598 (pass_data_graphite_transforms): ...new pass_data instance and...
2599 (make_pass_graphite_transforms): ...new function.
2600 (pass_check_data_deps): Convert from a global struct to a subclass of
2601 gimple_opt_pass along with...
2602 (pass_data_check_data_deps): ...new pass_data instance and...
2603 (make_pass_check_data_deps): ...new function.
2604 (pass_iv_canon): Convert from a global struct to a subclass of
2605 gimple_opt_pass along with...
2606 (pass_data_iv_canon): ...new pass_data instance and...
2607 (make_pass_iv_canon): ...new function.
2608 (pass_scev_cprop): Convert from a global struct to a subclass of
2609 gimple_opt_pass along with...
2610 (pass_data_scev_cprop): ...new pass_data instance and...
2611 (make_pass_scev_cprop): ...new function.
2612 (pass_record_bounds): Convert from a global struct to a subclass of
2613 gimple_opt_pass along with...
2614 (pass_data_record_bounds): ...new pass_data instance and...
2615 (make_pass_record_bounds): ...new function.
2616 (pass_complete_unroll): Convert from a global struct to a subclass of
2617 gimple_opt_pass along with...
2618 (pass_data_complete_unroll): ...new pass_data instance and...
2619 (make_pass_complete_unroll): ...new function.
2620 (pass_complete_unrolli): Convert from a global struct to a subclass of
2621 gimple_opt_pass along with...
2622 (pass_data_complete_unrolli): ...new pass_data instance and...
2623 (make_pass_complete_unrolli): ...new function.
2624 (pass_parallelize_loops): Convert from a global struct to a subclass
2625 of gimple_opt_pass along with...
2626 (pass_data_parallelize_loops): ...new pass_data instance and...
2627 (make_pass_parallelize_loops): ...new function.
2628 (pass_loop_prefetch): Convert from a global struct to a subclass of
2629 gimple_opt_pass along with...
2630 (pass_data_loop_prefetch): ...new pass_data instance and...
2631 (make_pass_loop_prefetch): ...new function.
2632 (pass_iv_optimize): Convert from a global struct to a subclass of
2633 gimple_opt_pass along with...
2634 (pass_data_iv_optimize): ...new pass_data instance and...
2635 (make_pass_iv_optimize): ...new function.
2636 (pass_tree_loop_done): Convert from a global struct to a subclass of
2637 gimple_opt_pass along with...
2638 (pass_data_tree_loop_done): ...new pass_data instance and...
2639 (make_pass_tree_loop_done): ...new function.
2640 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2641 struct to a subclass of gimple_opt_pass along with...
2642 (pass_data_cse_reciprocals): ...new pass_data instance and...
2643 (make_pass_cse_reciprocals): ...new function.
2644 (pass_cse_sincos): Convert from a global struct to a subclass of
2645 gimple_opt_pass along with...
2646 (pass_data_cse_sincos): ...new pass_data instance and...
2647 (make_pass_cse_sincos): ...new function.
2648 (pass_optimize_bswap): Convert from a global struct to a subclass of
2649 gimple_opt_pass along with...
2650 (pass_data_optimize_bswap): ...new pass_data instance and...
2651 (make_pass_optimize_bswap): ...new function.
2652 (pass_optimize_widening_mul): Convert from a global struct to a
2653 subclass of gimple_opt_pass along with...
2654 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2655 (make_pass_optimize_widening_mul): ...new function.
2656 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2657 subclass of gimple_opt_pass along with...
2658 (pass_data_phiopt): ...new pass_data instance and...
2659 (make_pass_phiopt): ...new function.
2660 (pass_cselim): Convert from a global struct to a subclass of
2661 gimple_opt_pass along with...
2662 (pass_data_cselim): ...new pass_data instance and...
2663 (make_pass_cselim): ...new function.
2664 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2665 subclass of gimple_opt_pass along with...
2666 (pass_data_phiprop): ...new pass_data instance and...
2667 (make_pass_phiprop): ...new function.
2668 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2669 subclass of gimple_opt_pass along with...
2670 (pass_data_pre): ...new pass_data instance and...
2671 (make_pass_pre): ...new function.
2672 (pass_fre): Convert from a global struct to a subclass of
2673 gimple_opt_pass along with...
2674 (pass_data_fre): ...new pass_data instance and...
2675 (make_pass_fre): ...new function.
2676 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2677 subclass of gimple_opt_pass along with...
2678 (pass_data_reassoc): ...new pass_data instance and...
2679 (make_pass_reassoc): ...new function.
2680 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2681 subclass of gimple_opt_pass along with...
2682 (pass_data_sink_code): ...new pass_data instance and...
2683 (make_pass_sink_code): ...new function.
2684 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2685 subclass of gimple_opt_pass along with...
2686 (pass_data_strlen): ...new pass_data instance and...
2687 (make_pass_strlen): ...new function.
2688 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2689 struct to a subclass of gimple_opt_pass along with...
2690 (pass_data_build_alias): ...new pass_data instance and...
2691 (make_pass_build_alias): ...new function.
2692 (pass_build_ealias): Convert from a global struct to a subclass of
2693 gimple_opt_pass along with...
2694 (pass_data_build_ealias): ...new pass_data instance and...
2695 (make_pass_build_ealias): ...new function.
2696 (pass_ipa_pta): Convert from a global struct to a subclass of
2697 simple_ipa_opt_pass along with...
2698 (pass_data_ipa_pta): ...new pass_data instance and...
2699 (make_pass_ipa_pta): ...new function.
2700 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2701 subclass of gimple_opt_pass along with...
2702 (pass_data_uncprop): ...new pass_data instance and...
2703 (make_pass_uncprop): ...new function.
2704 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2705 global struct to a subclass of gimple_opt_pass along with...
2706 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2707 (make_pass_late_warn_uninitialized): ...new function.
2708 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2709 to a subclass of gimple_opt_pass along with...
2710 (pass_data_init_datastructures): ...new pass_data instance and...
2711 (make_pass_init_datastructures): ...new function.
2712 (pass_early_warn_uninitialized): Convert from a global struct to a
2713 subclass of gimple_opt_pass along with...
2714 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2715 (make_pass_early_warn_uninitialized): ...new function.
2716 (pass_update_address_taken): Convert from a global struct to a
2717 subclass of gimple_opt_pass along with...
2718 (pass_data_update_address_taken): ...new pass_data instance and...
2719 (make_pass_update_address_taken): ...new function.
2720 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2721 struct to a subclass of gimple_opt_pass along with...
2722 (pass_data_release_ssa_names): ...new pass_data instance and...
2723 (make_pass_release_ssa_names): ...new function.
2724 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2725 subclass of gimple_opt_pass along with...
2726 (pass_data_stdarg): ...new pass_data instance and...
2727 (make_pass_stdarg): ...new function.
2728 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2729 global struct to a subclass of gimple_opt_pass along with...
2730 (pass_data_convert_switch): ...new pass_data instance and...
2731 (make_pass_convert_switch): ...new function.
2732 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2733 to a subclass of gimple_opt_pass along with...
2734 (pass_data_tail_recursion): ...new pass_data instance and...
2735 (make_pass_tail_recursion): ...new function.
2736 (pass_tail_calls): Convert from a global struct to a subclass of
2737 gimple_opt_pass along with...
2738 (pass_data_tail_calls): ...new pass_data instance and...
2739 (make_pass_tail_calls): ...new function.
2740 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2741 struct to a subclass of gimple_opt_pass along with...
2742 (pass_data_lower_vector): ...new pass_data instance and...
2743 (make_pass_lower_vector): ...new function.
2744 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2745 gimple_opt_pass along with...
2746 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2747 (make_pass_lower_vector_ssa): ...new function.
2748 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2749 to a subclass of gimple_opt_pass along with...
2750 (pass_data_slp_vectorize): ...new pass_data instance and...
2751 (make_pass_slp_vectorize): ...new function.
2752 (pass_ipa_increase_alignment): Convert from a global struct to a
2753 subclass of simple_ipa_opt_pass along with...
2754 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2755 (make_pass_ipa_increase_alignment): ...new function.
2756 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2757 gimple_opt_pass along with...
2758 (pass_data_vrp): ...new pass_data instance and...
2759 (make_pass_vrp): ...new function.
2760 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2761 subclass of simple_ipa_opt_pass along with...
2762 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2763 (make_pass_ipa_free_lang_data): ...new function.
2764 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2765 gimple_opt_pass along with...
2766 (pass_data_tsan): ...new pass_data instance and...
2767 (make_pass_tsan): ...new function.
2768 (pass_tsan_O0): Convert from a global struct to a subclass of
2769 gimple_opt_pass along with...
2770 (pass_data_tsan_O0): ...new pass_data instance and...
2771 (make_pass_tsan_O0): ...new function.
2772 * var-tracking.c (pass_variable_tracking): Convert from a global
2773 struct to a subclass of rtl_opt_pass along with...
2774 (pass_data_variable_tracking): ...new pass_data instance and...
2775 (make_pass_variable_tracking): ...new function.
2776 * web.c (pass_web): Convert from a global struct to a subclass of
2777 rtl_opt_pass along with...
2778 (pass_data_web): ...new pass_data instance and...
2779 (make_pass_web): ...new function.
2780 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2781 declaration with that of...
2782 (make_pass_mode_switch_use): ...new function.
2783 (pass_resolve_sw_modes): Replace declaration with that of...
2784 (make_pass_resolve_sw_modes): ...new function.
2785 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2786 from a global struct to a subclass of rtl_opt_pass along with...
2787 (pass_data_mode_switch_use): ...new pass_data instance and...
2788 (make_pass_mode_switch_use): ...new function.
2789 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2790 from a global struct to a subclass of rtl_opt_pass along with...
2791 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2792 (make_pass_resolve_sw_modes): ...new function.
2793 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2794 struct to a subclass of rtl_opt_pass along with...
2795 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2796 (make_pass_insert_vzeroupper): ...new function.
2797 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2798 global struct to a subclass of rtl_opt_pass along with...
2799 (pass_data_work_around_errata): ...new pass_data instance and...
2800 (make_pass_work_around_errata): ...new function.
2801 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2802 struct to a subclass of rtl_opt_pass along with...
2803 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2804 (make_pass_mips_machine_reorg2): ...new function.
2806 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2808 * passes.c (pass_manager::operator new): New.
2810 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2812 Handwritten part of conversion of passes to C++ classes.
2814 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2815 (toplev.o): Add dep on PASS_MANAGER_H.
2816 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2817 of early local pases to reflect this moving from a global to a
2818 member of gcc::pass_manager.
2819 (cgraph_add_new_function): Likewise.
2820 * lto-cgraph.c (lto_output_node): Update for conversion of
2821 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2822 * passes.c (opt_pass::clone): New.
2823 (opt_pass::gate): New.
2824 (opt_pass::execute): New.
2825 (opt_pass::opt_pass): New.
2826 (pass_manager::execute_early_local_passes): New.
2827 (pass_manager::execute_pass_mode_switching): new.
2828 (finish_optimization_passes): Convert to...
2829 (pass_manager::finish_optimization_passes): ...this.
2830 (finish_optimization_passes): Update for conversion of passes to
2832 (register_dump_files_1): Use has_gate since we cannot portably
2833 check a vtable entry against NULL.
2834 (dump_one_pass): Likewise.
2835 (ipa_write_summaries_2): Likewise.
2836 (ipa_write_optimization_summaries_1): Likewise.
2837 (ipa_read_summaries_1): Likewise.
2838 (ipa_read_optimization_summaries_1): Likewise.
2839 (execute_ipa_stmt_fixups): Likewise.
2840 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2841 pass-creation functions rather than wiring up globals, and
2842 storing the results in fields of pass_manager generated using
2844 (pass_manager::dump_profile_report): Update for conversion of
2845 passes to C++ classes.
2846 (pass_manager::execute_ipa_summary_passes): Likewise.
2847 (execute_one_ipa_transform_pass): Likewise.
2848 (execute_one_pass): Use has_gate and has_execute since we cannot
2849 portably check a vtable entry against NULL.
2850 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2851 (pass_manager): Use pass-instances.def to add fields for the
2852 various pass instances.
2853 * toplev.c (finalize): Update for move of
2854 finish_optimization_passes to a method of gcc::pass_manager.
2855 * toplev.h (finish_optimization_passes): Move to method of class
2857 * tree-pass.h (struct pass_data): New.
2858 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2859 (opt_pass::gate): Convert to virtual function.
2860 (opt_pass::~opt_pass): New.
2861 (opt_pass::clone): New.
2862 (opt_pass::execute): Convert to virtual function.
2863 (opt_pass::opt_pass): New.
2864 (opt_pass::ctxt_): new.
2865 (gimple_opt_pass): Convert to subclass of opt_pass.
2866 (gimple_opt_pass::gimple_opt_pass): New.
2867 (rtl_opt_pass): Convert to subclass of opt_pass.
2868 (rtl_opt_pass::rtl_opt_pass): New.
2869 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2870 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2871 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2872 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2873 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2874 invocation of pass_mode_switching to reflect this moving from a
2875 global to a member of gcc::pass_manager.
2876 (ix86_option_override): Rework how pass_insert_vzeroupper is
2877 added to the pass_manager to reflect autogenerated changes.
2878 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2880 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2882 PR rtl-optimization/57708
2883 * recog.c (peep2_find_free_register): Validate all regs in a
2886 2013-08-05 Jan Hubicka <jh@suse.cz>
2889 * cgraph.c (verify_cgraph_node): Accept local flags from other
2891 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2892 (function_and_variable_visibility): Likewise.
2893 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2895 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2897 * graph.c (init_graph_slim_pretty_print): Remove.
2898 (print_graph_cfg): Do not call it. Use local pretty printer.
2899 (start_graph_dump): Likewise.
2901 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2903 * gimple-pretty-print.c (buffer): Remove.
2904 (initialized): Likewise.
2905 (maybe_init_pretty_print): Likewise.
2906 (print_gimple_stmt): Do not call it. Use non-static local
2907 pretty_printer variable.
2908 (print_gimple_expr): Likewise.
2909 (print_gimple_seq): Likewise.
2910 (gimple_dump_bb): Likewise.
2912 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2914 * asan.c (asan_pp): Remove.
2915 (asan_pp_initialized): Likewise.
2916 (asan_pp_initialize): Likewise.
2917 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2918 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2919 (asan_add_global): Likewise.
2921 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2923 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2924 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2925 * pretty-print.c (pp_formatted_text_data): Likewise.
2926 (pp_indent): Rename from pp_base_indent.
2927 (pp_format): Rename from pp_base_format.
2928 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2929 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2930 (pp_flush): Rename from pp_base_flush.
2931 (pp_set_line_maximum_length): Rename from
2932 pp_base_set_line_maximum_length.
2933 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2934 (pp_set_prefix): Rename from pp_base_set_prefix.
2935 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2936 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2937 (pp_append_text): Rename from pp_base_append_text.
2938 (pp_formatted_text): Rename from pp_base_formatted_text.
2939 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2940 (pp_remaining_character_count_for_line): Rename from
2941 pp_base_remaining_character_count_for_line.
2942 (pp_newline): Rename from pp_base_newline.
2943 (pp_character): Rename from pp_base_character.
2944 (pp_string): Rename from pp_base_string.
2945 (pp_maybe_space): Rename from pp_base_maybe_space.
2946 * asan.c (asan_pp_string): Adjust.
2947 (asan_emit_stack_protection): Likewise.
2948 (asan_add_global): Likewise.
2949 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2950 * tree-mudflap.c (mf_varname_tree): Likewise.
2951 * tree-pretty-print.c (pp_tree_identifier): Rename from
2952 pp_base_tree_identifier.
2953 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2954 Declare as function.
2956 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2958 * pretty-print.h (pp_bar_bar): New.
2959 (pp_ampersand_ampersand): Likewise.
2960 (pp_less_equal): Likewise.
2961 (pp_greater_equal): Likewise.
2962 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2963 printer functions instead of pp_string or operators and punctuators.
2964 (dump_gimple_call): Likewise.
2965 (dump_gimple_omp_for): Likewise.
2966 (dump_gimple_transaction): Likewise.
2967 (dump_gimple_phi): Likewise.
2968 (pp_gimple_stmt_1): Likewise.
2969 * sched-vis.c (print_insn): Likewise.
2970 * tree-mudflap.c (mf_varname_tree): Likewise.
2971 * tree-pretty-print.c (dump_block_node): Likewise.
2972 (dump_generic_node): Likewise.
2974 2013-08-02 Jan Hubicka <jh@suse.cz>
2976 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2978 * lto-streamer-out.c (tree_is_indexable): Results decls and
2979 parm decls are not indexable.
2980 (DFS_write_tree_body): Do not follow args and results.
2981 (hash_tree): Likewise.
2982 (output_functions): Rearrange so struct function is needed
2983 only when real body is output; be able to also ouptut abstract
2984 functions; output DECL_ARGUMENTS and DECL_RESULT.
2985 (lto_output): When not in WPA, ale store abstract functions.
2986 (write_symbol): Do not care about RESULT_DECL.
2987 (output_symbol_p): Handle correctly sbtract decls.
2988 * lto-streamer-in.c (input_function): Rearrange so struct
2989 function can be NULL at entry; allow streaming of
2990 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2991 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2992 sanity check during LTO.
2993 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2994 RESULT_DECl and DECL_ARGUMENTS.
2995 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2998 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3000 * pretty-print.h (pp_underscore): New.
3002 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
3003 printer functions instead of pp_character.
3004 (dump_binary_rhs): Likewise.
3005 (dump_ternary_rhs): Likewise.
3006 (dump_gimple_call_args): Likewise.
3007 (pp_points_to_solution): Likewise.
3008 (dump_gimple_call): Likewise.
3009 (dump_gimple_switch): Likewise.
3010 (dump_gimple_cond): Likewise.
3011 (dump_gimple_bind): Likewise.
3012 (dump_gimple_try): Likewise.
3013 (dump_gimple_omp_for): Likewise.
3014 (dump_gimple_omp_continue): Likewise.
3015 (dump_gimple_omp_single): Likewise.
3016 (dump_gimple_omp_sections): Likewise.
3017 (dump_gimple_omp_block): Likewise.
3018 (dump_gimple_omp_critical): Likewise.
3019 (dump_gimple_transaction): Likewise.
3020 (dump_gimple_asm): Likewise.
3021 (dump_gimple_phi): Likewise.
3022 (dump_gimple_omp_parallel): Likewise.
3023 (dump_gimple_omp_task): Likewise.
3024 (dump_gimple_omp_atomic_load): Likewise.
3025 (dump_gimple_omp_atomic_store): Likewise.
3026 (dump_gimple_mem_ops): Likewise.
3027 (pp_gimple_stmt_1): Likewise.
3028 (pp_cfg_jump): Likewise.
3029 (dump_implicit_edges): Likewise.
3030 (gimple_dump_bb_for_graph): Likewise.
3031 * graph.c (draw_cfg_node): Likewise.
3032 * langhooks.c (lhd_print_error_function): Likewise.
3033 * sched-vis.c (print_exp): Likewise.
3034 (print_value): Likewise.
3035 (print_pattern): Likewise.
3036 (print_insn): Likewise.
3037 (rtl_dump_bb_for_graph): Likewise.
3038 * tree-pretty-print.c (dump_function_declaration): Likewise.
3039 (dump_array_domain): Likewise.
3040 (dump_omp_clause): Likewise.
3041 (dump_location): Likewise.
3042 (dump_generic_node): Likewise.
3043 (print_struct_decl): Likewise.
3044 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3046 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3048 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3049 candidate table when replacing a candidate statement.
3050 (replace_rhs_if_not_dup): Likewise.
3051 (replace_one_candidate): Likewise.
3053 2013-08-02 Jan Hubicka <jh@suse.cz>
3054 Martin Liska <marxin.liska@gmail.com>
3056 * cgraphunit.c (add_new_function): Fix logic when adding from
3058 (assemble_thunk): Rename to ...
3059 (expand_thunk); .. this one; export; get it working with
3060 general functions; make produced gimple valid.
3061 * cgraph.h (expand_thunk): Declare.
3063 2013-08-02 Jan Hubicka <jh@suse.cz>
3065 * ipa-cp.c (gather_context_independent_values): Use
3066 ipa_get_param_move_cost.
3067 (get_replacement_map): Remove PARAM; move parameter folding
3069 (create_specialized_node): Update.
3070 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3071 assert that we have gimple body; update move_cost.
3072 (count_formal_params): Assert that we have gimple body.
3073 (ipa_dump_param): New function.
3074 (ipa_alloc_node_params): Break out from ...
3075 (ipa_initialize_node_params): ... here.
3076 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3077 (ipa_write_node_info): Stream move costs.
3078 (ipa_read_node_info): Read move costs.
3079 (ipa_update_after_lto_read): Do not recompute node params.
3080 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3081 (ipa_get_param): Check we are not in WPA.
3082 (ipa_get_param_move_cost): New.
3083 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3084 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3085 parm numbers to be present.
3087 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3089 PR rtl-optimization/58048
3090 * lra-constraints.c (process_alt_operands): Don't check asm
3091 operand on register.
3093 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3095 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3096 the implied StoreLoad barrier for atomic operations if before.
3098 2013-08-02 Jan Hubicka <jh@suse.cz>
3099 Martin Liska <marxin.liska@gmail.com>
3101 * cgraph.c (cgraph_function_body_availability): Do not check
3103 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3104 symtab_node_availability): Declare.
3105 * ipa.c (can_replace_by_local_alias): New.
3106 (function_and_variable_visibility): Use it.
3107 * symtab.c (symtab_for_node_and_aliases,
3108 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3110 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3112 PR rtl-optimization/57963
3113 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3114 (lra_constraints): Use them.
3116 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3118 * config/arm/types.md (define_attr "type"): Add "load_acq"
3120 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3122 (cortex_a53_store1): Likewise.
3124 2013-08-01 Jan Hubicka <jh@suse.cz>
3126 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3127 partitions are not needed.
3129 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3131 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3132 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3133 MAYBE_NON_Q_CLASS_P where appropriate.
3135 2013-08-01 Jan Hubicka <jh@suse.cz>
3137 * cgraph.h (release_function_body): Declare.
3138 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3139 of unused delcarations.
3141 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3143 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3144 RTL form when subtracting a constant.
3146 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3148 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3149 Generate canonical plus rtx with negated immediate instead of minus
3151 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3153 2013-08-01 Jan Hubicka <jh@suse.cz>
3155 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3156 (cgraph_release_function_body): Likewise.
3157 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3158 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3159 to used_as_abstract_origin.
3160 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3161 symbols used as abstract origins.
3162 * cgraphunit.c (analyze_functions): Update.
3163 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3164 used_as_abstract_origin.
3165 * tree-inline.c (tree_function_versioning): Update
3166 used_as_abstract_origin; be ready for DECL_RESULT and
3167 DECL_ARGUMENTS to be NULL.
3169 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3170 for abstract functions.
3171 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3174 2013-08-01 Jan Hubicka <jh@suse.cz>
3176 * profile.c (compute_value_histograms): Fix thinko.
3178 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3180 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3181 aarch-common-protos.h to extra_headers.
3182 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3183 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3184 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3186 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3188 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3189 Include "../arm/types.md". Define "type" attribute for all patterns.
3190 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3193 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3195 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3196 to support power8 load fusion.
3197 (fusion_gpr_mem_load): Likewise.
3199 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3201 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3202 declarations for power8 load fusion.
3203 (emit_fusion_gpr_load): Likewise.
3205 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3206 tuning for power8, turn on fusion mode by default. Turn on sign
3207 extending fusion mode if normal fusion mode is on, and we are at
3209 (fusion_gpr_load_p): New function, return true if we can fuse an
3210 addis instruction with a dependent load to a GPR.
3211 (emit_fusion_gpr_load): Emit the instructions for power8 load
3214 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3215 (VSX load fusion peepholes): New peepholes to fuse together an
3216 addi instruction with a VSX load instruction.
3218 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3219 peepholes to fuse an addis instruction with a load to a GPR base
3220 register. If we are supporting sign extending fusions, convert
3221 sign extending loads to zero extending loads and add an explicit
3224 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3226 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3227 aarch-common-protos.h to extra_headers.
3228 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3229 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3230 (arm_early_store_addr_dep): Likewise.
3231 (arm_no_early_alu_shift_dep): Likewise.
3232 (arm_no_early_alu_shift_value_dep): Likewise.
3233 (arm_no_early_mul_dep): Likewise.
3234 (arm_no_early_store_addr_dep): Likewise.
3235 (arm_mac_accumulator_is_mul_result): Likewise.
3236 (arm_mac_accumulator_is_result): Likewise.
3237 * config/arm/aarch-common.c: ... here. New file.
3238 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3240 (arm_early_store_addr_dep): Likewise.
3241 (arm_no_early_alu_shift_dep): Likewise.
3242 (arm_no_early_alu_shift_value_dep): Likewise.
3243 (arm_no_early_mul_dep): Likewise.
3244 (arm_no_early_store_addr_dep): Likewise.
3245 (arm_mac_accumulator_is_mul_result): Likewise.
3246 (arm_mac_accumulator_is_result): Likewise.
3247 * config/arm/aarch-common-protos.h: ... here. New file.
3248 * config/arm/t-arm (aarch-common.o): Define.
3250 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3252 * config/arm/arm.md: Include new file "types.md".
3253 (define_attr "type"): Move from here to ...
3254 (define_attr "mul32"): Likewise.
3255 (define_attr "mul64"): Likewise.
3256 * config/arm/types.md: ... here. New file.
3258 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
3260 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
3261 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
3263 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3265 * gen-pass-instances.awk: Fix offset of substr().
3267 2013-07-31 David Malcolm <dmalcolm@redhat.com>
3269 * Makefile.in (pass-instances.def): New.
3270 (passes.o): Replace dependency on passes.def with one on
3273 * gen-pass-instances.awk: New.
3275 * passes.c (pass_manager::pass_manager): Use pass-instances.def
3276 rather than passes.def, updating local definition of NEXT_PASS
3277 macro to add an extra NUM parameter (currently unused).
3279 2013-07-30 David Malcolm <dmalcolm@redhat.com>
3281 * Makefile.in (PASS_MANAGER_H): New.
3282 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
3283 (passes.o): Likewise.
3284 (statistics.o): Likewise.
3285 (cgraphunit.o): Likewise.
3286 (context.o): Depend on PASS_MANAGER_H.
3288 * pass_manager.h: New.
3290 * cgraphunit.c (cgraph_add_new_function): Update for moves
3291 of globals to fields of pass_manager.
3292 (analyze_function): Likewise.
3293 (expand_function): Likewise.
3294 (ipa_passes): Likewise.
3295 (compile): Likewise.
3297 * context.c (context::context): New.
3298 * context.h (context::context): New.
3299 (context::get_passes): New.
3300 (context::passes_): New.
3302 * lto-cgraph.c (input_node): Update for moves of globals to
3303 fields of pass_manager.
3305 * passes.c (all_passes): Remove, in favor of a field of the
3306 same name within the new class pass_manager.
3307 (all_small_ipa_passes): Likewise.
3308 (all_lowering_passes): Likewise.
3309 (all_regular_ipa_passes): Likewise.
3310 (all_late_ipa_passes): Likewise.
3311 (all_lto_gen_passes): Likewise.
3312 (passes_by_id): Likewise.
3313 (passes_by_id_size): Likewise.
3314 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3315 the new class pass_manager.
3316 (set_pass_for_id): Convert to...
3317 (pass_manager::set_pass_for_id): ...method.
3318 (get_pass_for_id): Convert to...
3319 (pass_manager::get_pass_for_id): ...method.
3320 (register_one_dump_file): Move body of implementation into...
3321 (pass_manager::register_one_dump_file): ...here.
3322 (register_dump_files_1): Convert to...
3323 (pass_manager::register_dump_files_1): ...method.
3324 (register_dump_files): Convert to...
3325 (pass_manager::register_dump_files): ...method.
3326 (create_pass_tab): Update for moves of globals to fields of
3328 (dump_passes): Move body of implementation into...
3329 (pass_manager::dump_passes): ...here.
3330 (register_pass): Move body of implementation into...
3331 (pass_manager::register_pass): ...here.
3332 (init_optimization_passes): Convert into...
3333 (pass_manager::pass_manager): ...constructor for new
3334 pass_manager class, and initialize the pass_lists array.
3335 (check_profile_consistency): Update for moves of globals to
3336 fields of pass_manager.
3337 (dump_profile_report): Move body of implementation into...
3338 (pass_manager::dump_profile_report): ...here.
3339 (ipa_write_summaries_1): Update for moves of pass lists from
3340 being globals to fields of pass_manager.
3341 (ipa_write_optimization_summaries): Likewise.
3342 (ipa_read_summaries): Likewise.
3343 (ipa_read_optimization_summaries): Likewise.
3344 (execute_all_ipa_stmt_fixups): Likewise.
3346 * statistics.c (statistics_fini): Update for moves of globals to
3347 fields of pass_manager.
3349 * toplev.c (general_init): Replace call to
3350 init_optimization_passes with construction of the pass_manager
3353 * tree-pass.h (all_passes): Remove, in favor of a field of the
3354 same name within the new class pass_manager.
3355 (all_small_ipa_passes): Likewise.
3356 (all_lowering_passes): Likewise.
3357 (all_regular_ipa_passes): Likewise.
3358 (all_lto_gen_passes): Likewise.
3359 (all_late_ipa_passes): Likewise.
3360 (passes_by_id): Likewise.
3361 (passes_by_id_size): Likewise.
3362 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3363 the new class pass_manager.
3364 (get_pass_for_id): Remove.
3366 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3368 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
3371 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3373 * arm.md (mulhi3): New expand pattern.
3375 2013-07-30 Jan Hubicka <jh@suse.cz>
3376 Martin Liska <marxin.liska@gmail.com>
3378 * profile.c (compute_value_histograms): Do not ICE when
3379 there is mismatch only on some counters.
3381 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3383 PR rtl-optimization/57637
3384 * function.c (move_insn_for_shrink_wrap): Also check the
3385 GEN set of the LIVE problem for the liveness analysis
3386 if it exists, otherwise give up.
3388 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3390 PR tree-optimization/57993
3391 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
3392 replaced statement in the candidate table.
3393 (phi_add_costs): Return infinite cost when the hidden basis does
3394 not dominate all phis on which the candidate is dependent.
3395 (replace_one_candidate): Record replaced statement in the
3398 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
3400 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
3401 (ashlv2si3): New expander.
3402 (*ashlv2si3_i): New define_insn_and_split.
3403 * predicates.md (float_operation): Allow patterns with three
3406 PR rtl-optimization/58021
3407 * mode-switching.c (create_pre_exit): Always split off preceding
3408 insns if we are not at the basic block head.
3410 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
3412 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
3413 (UCLIBC_DYNAMIC_LINKER): New macro.
3414 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
3416 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
3417 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
3419 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
3420 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
3421 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
3422 for SF and DF modes. Use ieee_quad_format for TF mode.
3423 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
3424 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
3425 (mips_option_override): Handle `-mnan=legacy'.
3426 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
3427 `-mabs=2008' and `-mnan=2008'.
3428 (OPTION_DEFAULT_SPECS): Add "nan" default.
3429 (ASM_SPEC): Handle `-mnan='.
3430 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
3431 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
3432 comment accordingly.
3433 (neg<mode>2): Likewise.
3434 * config/mips/mips.opt (mabs, mnan): New options.
3435 * doc/install.texi (Configuration): Document `--with-nan=' option.
3436 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
3438 (MIPS Options): Document them.
3439 * config.gcc <mips*-*-*>: Handle `--with-nan='.
3440 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
3441 * configure: Regenerate.
3442 * config.in: Regenerate.
3444 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3446 * config/i386/i386.md (float post-reload splitters): Do not check
3447 for subregs of SSE registers.
3449 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3450 H.J. Lu <hongjiu.lu@intel.com>
3454 * config/i386/i386.md (post-reload splitter
3455 to avoid partial SSE reg dependency stalls): New pattern.
3457 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
3459 * config/s390/s390.md ("movcc"): Swap load and store instructions.
3461 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
3463 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
3464 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
3466 2013-07-26 Cary Coutant <ccoutant@google.com>
3468 * dwarf2out.c (die_checksum_ordered): Don't include template
3469 instantiations in signature.
3470 (is_template_parameter): New function.
3471 (is_template_instantiation): New function.
3472 (generate_skeleton_bottom_up): Don't include template instantiations
3474 (generate_skeleton): Likewise.
3475 (break_out_comdat_types): Move recursive call to break out nested
3477 (prune_unused_types_mark_generic_parms_dies): Call
3478 is_template_parameter.
3480 2013-07-26 Ian Bolton <ian.bolton@arm.com>
3482 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
3483 uses vector registers.
3484 * config/aarch64/iterators.md: Add attributes rtn and vas.
3486 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3487 Richard Earnshaw <richard.earnshaw@arm.com>
3489 * combine.c (simplify_comparison): Re-canonicalize operands
3491 * config/arm/arm.md (movcond_addsi): New splitter.
3493 2013-07-25 Sterling Augustine <saugustine@google.com>
3495 * dwarf2out.c (size_of_pubnames): Move code to...
3496 (include_pubname_in_output): ...here. New.
3497 (want_pubnames): Rearrange.
3498 (output_pubnames): Call include_pubname_in_output. Move assertion.
3500 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3502 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
3504 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3507 * doc/extend.texi: Remove obsolete builtins. Fix
3508 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
3510 2013-07-25 Jan Hubicka <jh@suse.cz>
3512 * cgraph.c (release_function_body): Break out from ...
3513 (cgraph_release_function_body): ... this one; also release DECL_RESULT
3515 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
3516 old_tree in the map.
3517 (create_specialized_node): Update.
3518 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
3520 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
3521 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
3522 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
3523 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3526 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3528 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3529 addsi3_carryin_alt2_<optab>): Correct output template.
3531 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3533 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3534 Adjust for arm_restrict_it.
3535 Remove trailing whitespace.
3537 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3539 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3540 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3542 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3544 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3546 PR rtl-optimization/57960
3547 * lra-constraints.c (process_alt_operands): Use the right mode
3548 when checking strict_low.
3550 2013-07-25 Jan Hubicka <jh@suse.cz>
3552 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3553 * cgraph.c (cgraph_remove_node): Do not release function body
3554 when in cgraph streaming.
3555 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3556 in other partitions are not considered reachable; fix handling of
3559 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3561 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3563 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3568 * config/arm/arm.md ("*sibcall_insn): Replace use of
3569 Ss with US. Adjust output for v5 and v4t.
3570 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3572 * config/arm/constraints.md ("Ss"): Rename to US.
3574 2013-07-25 Terry Guo <terry.guo@arm.com>
3576 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3577 shift_add/shift_sub0/shift_sub1 RTXs.
3579 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3580 Anton Blanchard <anton@au1.ibm.com>
3582 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3583 (altivec_vpks<VI_char>ss): Likewise.
3584 (altivec_vpks<VI_char>us): Likewise.
3585 (altivec_vpku<VI_char>us): Likewise.
3586 (altivec_vpku<VI_char>um): Likewise.
3588 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3590 Introduce context class.
3592 * Makefile.in (CONTEXT_H): New.
3593 (OBJS): Add context.o.
3594 (toplev.o): Add CONTEXT_H to dependencies.
3597 * toplev.c (general_init): Create the singleton gcc::context instance.
3603 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3605 PR rtl-optimization/57968
3606 * mode-switching.c (create_pre_exit): Allow instructions that
3607 don't set a return register to need a non-exit mode.
3609 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3610 Anton Blanchard <anton@au1.ibm.com>
3612 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3613 operands to vperm for little endian.
3614 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3615 of lvsl to create the control mask for a vperm for little endian.
3617 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3618 Anton Blanchard <anton@au1.ibm.com>
3620 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3621 two operands for little-endian.
3623 2013-07-23 Steve Ellcey <sellcey@mips.com>
3625 * config/mips/mips.c (mips_case_values_threshold): New.
3626 (TARGET_CASE_VALUES_THRESHOLD): Define.
3628 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3629 Anton Blanchard <anton@au1.ibm.com>
3631 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3632 selection of field for vector splat in little endian mode.
3634 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3636 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3637 expanders to rs6000.md.
3638 (ior<mode>3): Likewise.
3639 (and<mode>3): Likewise.
3640 (one_cmpl<mode>2): Likewise.
3641 (nor<mode>3): Likewise.
3642 (andc<mode>3): Likewise.
3643 (eqv<mode>3): Likewise.
3644 (nand<mode>3): Likewise.
3645 (orc<mode>3): Likewise.
3647 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3650 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3651 to split multi-word logical operations.
3652 (rs6000_split_logical_di): Likewise.
3653 (rs6000_split_logical): Likewise.
3655 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3656 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3657 and allow TImode operations in 32-bit.
3658 (vsx_and<mode>3_64bit): Likewise.
3659 (vsx_ior<mode>3_32bit): Likewise.
3660 (vsx_ior<mode>3_64bit): Likewise.
3661 (vsx_xor<mode>3_32bit): Likewise.
3662 (vsx_xor<mode>3_64bit): Likewise.
3663 (vsx_one_cmpl<mode>2_32bit): Likewise.
3664 (vsx_one_cmpl<mode>2_64bit): Likewise.
3665 (vsx_nor<mode>3_32bit): Likewise.
3666 (vsx_nor<mode>3_64bit): Likewise.
3667 (vsx_andc<mode>3_32bit): Likewise.
3668 (vsx_andc<mode>3_64bit): Likewise.
3669 (vsx_eqv<mode>3_32bit): Likewise.
3670 (vsx_eqv<mode>3_64bit): Likewise.
3671 (vsx_nand<mode>3_32bit): Likewise.
3672 (vsx_nand<mode>3_64bit): Likewise.
3673 (vsx_orc<mode>3_32bit): Likewise.
3674 (vsx_orc<mode>3_64bit): Likewise.
3676 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3677 logical types in GPRs.
3679 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3680 logical insns to rs6000.md, and allow TImode operations in
3682 (altivec_ior<mode>3): Likewise.
3683 (altivec_xor<mode>3): Likewise.
3684 (altivec_one_cmpl<mode>2): Likewise.
3685 (altivec_nor<mode>3): Likewise.
3686 (altivec_andc<mode>3): Likewise.
3688 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3689 attributes for moving the 128-bit logical operations into
3691 (BOOL_REGS_OUTPUT): Likewise.
3692 (BOOL_REGS_OP1): Likewise.
3693 (BOOL_REGS_OP2): Likewise.
3694 (BOOL_REGS_UNARY): Likewise.
3695 (BOOL_REGS_AND_CR0): Likewise.
3696 (one_cmpl<mode>2): Add support for DI logical operations on
3697 32-bit, splitting the operations to 32-bit.
3701 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3702 changes to combine the 32/64-bit code, allow logical operations on
3703 TI mode in 32-bit, and to use similar match_operator patterns like
3704 scalar mode uses. Combine the Altivec and VSX code for logical
3705 operations, and move it here.
3706 (ior<mode>3, 128-bit types): Likewise.
3707 (xor<mode>3, 128-bit types): Likewise.
3708 (one_cmpl<mode>3, 128-bit types): Likewise.
3709 (nor<mode>3, 128-bit types): Likewise.
3710 (andc<mode>3, 128-bit types): Likewise.
3711 (eqv<mode>3, 128-bit types): Likewise.
3712 (nand<mode>3, 128-bit types): Likewise.
3713 (orc<mode>3, 128-bit types): Likewise.
3714 (and<mode>3_internal): Likewise.
3715 (bool<mode>3_internal): Likewise.
3716 (boolc<mode>3_internal1): Likewise.
3717 (boolc<mode>3_internal2): Likewise.
3718 (boolcc<mode>3_internal1): Likewise.
3719 (boolcc<mode>3_internal2): Likewise.
3720 (eqv<mode>3_internal1): Likewise.
3721 (eqv<mode>3_internal2): Likewise.
3722 (one_cmpl1<mode>3_internal): Likewise.
3724 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3726 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3727 Rename flag_stack_usage to flag_stack_usage_info.
3729 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3731 * config/microblaze/sync.md: New file.
3732 * config/microblaze/microblaze.md: Include sync.md
3733 * config/microblaze/microblaze.c: Add print_operand 'y'.
3734 * config/microblaze/constraints.md: Add memory_contraint
3735 'Q' which is a single register.
3737 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3739 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3741 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3743 * reload.c (find_reloads): Exit loop once we find this operand
3744 cannot be reloaded somehow for this alternative.
3746 * reload.c (find_reloads): Exit loop once we find a hard register.
3748 * rtlanal.c (computed_jump_p): Exit loop once we find label
3751 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3753 * cfgloopmanip.c (remove_path): Exit loop after setting
3756 * gensupport.c (subst_dup): Avoid loop if code is not
3757 MATCH_DUP nor MATCH_OP_DUP.
3759 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3761 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3763 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3765 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3766 true for SP_REGNUM if mode == ptr_mode.
3767 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3768 with value R0_REGNUM + 31.
3770 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3772 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3773 pad pointer-typed argument downward.
3775 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3777 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3778 and __ILP32__ when the ILP32 model is in use.
3780 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3782 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3783 (aarch64_load_symref_appropriately): In the case of
3784 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3785 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3786 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3787 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3788 if 'mode' doesn't equal to 'ptr_mode'.
3789 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3790 'vcall_offset'; change to call aarch64_emit_move differently depending
3791 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3792 to calculate the upper bound of 'vcall_offset'.
3793 (aarch64_cannot_force_const_mem): Change to also return true if
3795 (aarch64_legitimize_reload_address): In the case of large
3796 displacements, add new local variable 'xmode' and an assertion
3797 based on it; change to use 'xmode' to generate the new rtx and
3799 (aarch64_asm_trampoline_template): Change to generate the template
3800 differently depending on TARGET_ILP32 or not; change to use
3801 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3802 (aarch64_trampoline_size): Removed.
3803 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3804 and replace immediate literals with it. Change to use 'ptr_mode'
3805 instead of 'DImode' and call convert_memory_address if the mode
3806 of 'fnaddr' doesn't equal to 'ptr_mode'.
3807 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3809 (aarch64_elf_asm_destructor): Likewise.
3810 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3811 on TARGET_ILP32 instead of aarch64_trampoline_size.
3812 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3813 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3814 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3815 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3816 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3817 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3818 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3819 depending on the value of 'mode'.
3820 (add_losym_<mode>): New.
3821 (ldr_got_small_<mode>): New, based on ldr_got_small.
3822 (ldr_got_small): Remove.
3823 (ldr_got_small_sidi): New.
3824 * config/aarch64/iterators.md (P): New.
3825 (PTR): Change to 'ptr_mode' in the condition.
3827 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3829 * config.gcc (aarch64*-*-*): Support --with-abi.
3830 (aarch64*-*-elf): Support --with-multilib-list.
3831 (aarch64*-*-linux*): Likewise.
3832 (supported_defaults): Add abi to aarch64*-*-*.
3833 * configure.ac: Mention AArch64 for --with-multilib-list.
3834 * configure: Re-generated.
3835 * config/aarch64/biarchilp32.h: New file.
3836 * config/aarch64/biarchlp64.h: New file.
3837 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3839 (MULTILIB_DEFAULTS): Ditto.
3840 (DRIVER_SELF_SPECS): Ditto.
3841 (ASM_SPEC): Update to also substitute -mabi.
3842 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3843 file whose name depends on -mabi= and -mbig-endian.
3844 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3846 (POINTER_SIZE): New define.
3847 (POINTERS_EXTEND_UNSIGNED): Ditto.
3848 (enum aarch64_abi_type): New enumeration tag.
3849 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3850 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3851 (TARGET_ILP32): New define.
3852 * config/aarch64/aarch64.opt (mabi): New.
3854 (ilp32, lp64): New values for -mabi.
3855 * config/aarch64/t-aarch64 (comma): New define.
3856 (MULTILIB_OPTIONS): Ditto.
3857 (MULTILIB_DIRNAMES): Ditto.
3858 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3859 * doc/invoke.texi: Document -mabi for AArch64.
3861 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3863 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3865 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3866 Anton Blanchard <anton@au1.ibm.com>
3868 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3869 endianness when selecting field to splat.
3871 2013-07-22 Eric Christopher <echristo@gmail.com>
3873 * dwarf2out.c (die_odr_checksum): New function to use
3874 CHECKSUM_ macros and ULEB128 for DIE tag.
3875 (generate_type_signature): Use.
3877 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3879 * config.gcc (sparc*-*-*): Accept leon3 processor.
3880 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3881 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3882 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3883 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3884 (mfix-ut699): Adjust comment.
3885 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3886 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3887 (CPP_CPU_SPEC): Likewise.
3888 (ASM_CPU_SPEC): Likewise.
3889 * config/sparc/sparc.c (leon3_cost): New constant.
3890 (sparc_option_override): Add leon3 support.
3891 (mem_ref): New function.
3892 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3893 (sparc_do_work_around_errata): Look into the instruction in the delay
3894 slot and adjust accordingly. Add fix for the data cache nullify issues
3895 of the UT699. Change insertion position for the NOP.
3896 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3897 (leon3_load): New reservation.
3898 (leon_store): Bump latency to 2.
3899 (grfpu): New automaton.
3900 (grfpu_alu): New unit.
3901 (grfpu_ds): Likewise.
3902 (leon_fp_alu): Adjust.
3903 (leon_fp_mult): Delete.
3904 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3905 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3906 * config/sparc/sparc.md (cpu): Add leon3.
3907 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3909 (atomic_test_and_set): Likewise.
3912 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3914 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3915 default for R5900 targets.
3916 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3917 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3918 * config/mips/mips.c (mips_option_override): Report an error for
3919 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3920 for -march=r5900 -mhard-float.
3922 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3924 * df-problems.c (can_move_insns_across): Exit loop once we
3925 find a non-fixed, non-global register.
3927 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3930 * omega.c (omega_eliminate_red): Break after setting red_found.
3931 (omega_problem_has_red_equations): Similarly after setting found.
3932 (omega_query_variable): Similarly after setting coupled.
3934 2013-07-22 Marek Polacek <polacek@redhat.com>
3936 * gimplify.c: Don't include gimple.h twice.
3938 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3940 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3941 instead of TARGET_THUMB1.
3942 (Pz): New constraint.
3943 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3945 (compare_negsi_si): Likewise.
3946 (compare_addsi2_op0): Likewise.
3947 (compare_addsi2_op1): Likewise.
3948 (addsi3_carryin_<optab>): Likewise.
3949 (addsi3_carryin_alt2_<optab>): Likewise.
3950 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3951 for arm_restrict_it.
3952 (subsi3_carryin): Likewise.
3953 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3954 (minmax_arithsi): Disable for arm_restrict_it.
3955 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3956 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3957 (satsi_<SAT:code>_shift): Likewise.
3958 (arm_shiftsi3): Add alternative for 16-bit encoding.
3959 (arm32_movhf): Disable for arm_restrict_it.
3960 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3961 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3963 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3965 * config/arm/arm.md (attribute "insn"): Delete.
3966 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3967 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3968 (not_shiftsi): Update for attribute change.
3969 (not_shiftsi_compare0): Likewise.
3970 (not_shiftsi_compare0_scratch): Likewise.
3971 (arm_one_cmplsi2): Likewise.
3972 (thumb1_one_cmplsi2): Likewise.
3973 (notsi_compare0): Likewise.
3974 (notsi_compare0_scratch): Likewise.
3975 (thumb1_movdi_insn): Likewise.
3976 (arm_movsi_insn): Likewise.
3977 (movhi_insn_arch4): Likewise.
3978 (movhi_bytes): Likewise.
3979 (arm_movqi_insn): Likewise.
3980 (thumb1_movqi_insn): Likewise.
3981 (arm32_movhf): Likewise.
3982 (thumb1_movhf): Likewise.
3983 (arm_movsf_soft_insn): Likewise.
3984 (thumb1_movsf_insn): Likewise.
3985 (thumb_movdf_insn): Likewise.
3986 (movsicc_insn): Likewise.
3987 (movsfcc_soft_insn): Likewise.
3988 (and_scc): Likewise.
3989 (cond_move): Likewise.
3990 (if_move_not): Likewise.
3991 (if_not_move): Likewise.
3992 (if_shift_move): Likewise.
3993 (if_move_shift): Likewise.
3994 (if_shift_shift): Likewise.
3995 (if_not_arith): Likewise.
3996 (if_arith_not): Likewise.
3997 (cond_move_not): Likewise.
3998 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3999 (neon_mov<mode>): Likewise.
4000 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
4001 (thumb2_movsi_vfp): Likewise.
4002 (movsf_vfp): Likewise.
4003 (thumb2_movsf_vfp): Likewise.
4004 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4006 (cortexa7_older_only): Likewise.
4007 (cortexa7_younger): Likewise.
4008 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
4009 (1020alu_shift_op): Likewise.
4010 (1020alu_shift_reg_op): Likewise.
4011 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
4012 (alu_shift_op): Likewise.
4013 (alu_shift_reg_op): Likewise.
4014 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
4015 (11_alu_shift_op): Likewise.
4016 (11_alu_shift_reg_op): Likewise.
4017 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
4018 (9_alu_shift_reg_op): Likewise.
4019 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4021 (cortex_a15_alu_shift): Likewise.
4022 (cortex_a15_alu_shift_reg): Likewise.
4023 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
4024 (cortex_a5_alu_shift): Likewise.
4025 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
4027 (cortex_a53_alu_shift): Likewise.
4028 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4030 (cortex_a7_alu_reg): Likewise.
4031 (cortex_a7_alu_shift): Likewise.
4032 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4033 (cortex_a8_alu_shift): Likewise.
4034 (cortex_a8_alu_shift_reg): Likewise.
4035 (cortex_a8_mov): Likewise.
4036 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4037 (cortex_a9_dp_shift): Likewise.
4038 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4039 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4040 (cortex_r4_mov): Likewise.
4041 (cortex_r4_alu_shift): Likewise.
4042 (cortex_r4_alu_shift_reg): Likewise.
4043 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4044 (526_alu_shift_op): Likewise.
4045 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4046 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4047 (626te_alu_shift_op): Likewise.
4048 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4049 (726te_alu_op): Likewise.
4050 (726te_alu_shift_op): Likewise.
4051 (726te_alu_shift_reg_op): Likewise.
4052 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4053 (mp626_alu_shift_op): Likewise.
4054 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4055 (pj4_alu_e1_conds): Likewise.
4056 (pj4_alu): Likewise.
4057 (pj4_alu_conds): Likewise.
4058 (pj4_shift): Likewise.
4059 (pj4_shift_conds): Likewise.
4060 (pj4_alu_shift): Likewise.
4061 (pj4_alu_shift_conds): Likewise.
4063 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4065 * config/arm/predicates.md (shiftable_operator_strict_it):
4067 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4068 Disable cond_exec version for arm_restrict_it.
4069 (thumb2_smaxsi3): Convert to generate cond_exec.
4070 (thumb2_sminsi3): Likewise.
4071 (thumb32_umaxsi3): Likewise.
4072 (thumb2_uminsi3): Likewise.
4073 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4074 (thumb2_neg_abssi2): Likewise.
4075 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4076 (thumb2_movsicc_insn): Adjust alternatives.
4077 (thumb2_mov_negscc): Disable for arm_restrict_it.
4078 (thumb2_mov_negscc_strict_it): New pattern.
4079 (thumb2_mov_notscc_strict_it): New pattern.
4080 (thumb2_mov_notscc): Disable for arm_restrict_it.
4081 (thumb2_ior_scc): Likewise.
4082 (thumb2_ior_scc_strict_it): New pattern.
4083 (thumb2_cond_move): Adjust for arm_restrict_it.
4084 (thumb2_cond_arith): Disable for arm_restrict_it.
4085 (thumb2_cond_arith_strict_it): New pattern.
4086 (thumb2_cond_sub): Adjust for arm_restrict_it.
4087 (thumb2_movcond): Likewise.
4088 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4089 (thumb2_zero_extendhisi2_v6): Likewise.
4090 (thumb2_zero_extendqisi2_v6): Likewise.
4091 (orsi_notsi_si): Likewise.
4092 (orsi_not_shiftsi_si): Likewise.
4094 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4096 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4097 instruction sequence is 1 byte shorter.
4099 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4101 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4102 it is not needed after split.
4104 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4107 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4108 second label for nonlocal goto receivers. Don't output pic base labels
4109 unless we're producing PIC; mark that action unreachable().
4110 (ix86_save_reg): If the function contains a nonlocal label, save the
4112 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4113 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4114 (update_pic_label_number_if_needed): New.
4115 (machopic_output_function_base_name): Adjust for nonlocal receiver
4117 (machopic_should_output_picbase_label): New.
4118 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4119 (nonlocal_goto_receiver): New insn and split.
4121 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4123 * config/aarch64/aarch64-builtins.c
4124 (aarch64_fold_builtin): Fold abs in all modes.
4125 * config/aarch64/aarch64-simd-builtins.def
4126 (abs): Enable for all modes.
4127 * config/aarch64/arm_neon.h
4128 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4129 (vabs_f64): Add missing intrinsic.
4131 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4133 * config/aarch64/arm_neon.h (vabs_s64): New function
4135 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4138 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4139 * config/avr/avr.md (adjust_len): Add `round'.
4140 * config/avr/avr-protos.h (avr_out_round): New prototype.
4141 (avr_out_plus): Add `out_label' argument.
4142 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4143 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4144 Handle the case where `insn' is just a pattern.
4145 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4146 (avr_out_round): New function.
4147 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4149 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4151 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4152 Add check for flag_stack_usage to handle -fstack-usage support
4154 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4156 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4157 interaction for new Power8 flags and VSX.
4159 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4162 * tree-inline.c (expand_call_inline): Emit errors during
4163 early_inlining only if optimization is not turned on.
4165 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4169 * passes.c (init_optimization_passes): Move the construction of
4170 the pass hierarchy into a new passes.def file.
4172 * Makefile.in (passes.o): Add dependency on passes.def.
4174 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4176 * passes.c (init_optimization_passes): Introduce macros for
4177 constructing the tree of passes (INSERT_PASSES_AFTER,
4178 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4179 TERMINATE_PASS_LIST).
4181 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4182 Wei Mi <wmi@google.com>
4184 PR rtl-optimization/57878
4185 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4187 (reload_pseudo_compare_func): Check nregs first for reload
4190 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4192 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4194 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4196 * read-rtl.c (validate_const_int): Once an invalid character is
4197 seen, quit the loop.
4199 * gengtype.c (write_roots): Similarly once we find the "deletable"
4200 or "if_marked" option.
4202 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4204 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4205 "xtab" and "sat". Move value "clz" from here to ...
4206 (attriubte "type"): ... here.
4207 (satsi_<SAT:code>): Delete "insn" attribute.
4208 (satsi_<SAT:code>_shift): Likewise.
4209 (arm_zero_extendqisi2addsi): Likewise.
4210 (arm_extendqisi2addsi): Likewise.
4211 (clzsi2): Update for attribute changes.
4212 (rbitsi2): Likewise.
4213 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4215 (arm_usatsihi): Likewise.
4216 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4218 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4220 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4221 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4222 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4223 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4224 in alphabetical order.
4225 (attribute "core_cycles"): Update for attribute changes.
4226 (arm_addsi3): Likewise.
4227 (addsi3_compare0): Likewise.
4228 (addsi3_compare0_scratch): Likewise.
4229 (addsi3_compare_op1): Likewise.
4230 (addsi3_compare_op2): Likewise.
4231 (compare_addsi2_op0): Likewise.
4232 (compare_addsi2_op1): Likewise.
4233 (addsi3_carryin_shift_<optab>): Likewise.
4234 (subsi3_carryin_shift): Likewise.
4235 (rsbsi3_carryin_shift): Likewise.
4236 (arm_subsi3_insn): Likewise.
4237 (subsi3_compare0): Likewise.
4238 (subsi3_compare): Likewise.
4239 (arm_andsi3_insn): Likewise.
4240 (thumb1_andsi3_insn): Likewise.
4241 (andsi3_compare0): Likewise.
4242 (andsi3_compare0_scratch): Likewise.
4243 (zeroextractsi_compare0_scratch
4244 (andsi_not_shiftsi_si): Likewise.
4245 (iorsi3_insn): Likewise.
4246 (iorsi3_compare0): Likewise.
4247 (iorsi3_compare0_scratch): Likewise.
4248 (arm_xorsi3): Likewise.
4249 (thumb1_xorsi3_insn): Likewise.
4250 (xorsi3_compare0): Likewise.
4251 (xorsi3_compare0_scratch): Likewise.
4252 (satsi_<SAT:code>_shift): Likewise.
4254 (arm_shiftsi3): Likewise.
4255 (shiftsi3_compare0): Likewise.
4256 (not_shiftsi): Likewise.
4257 (not_shiftsi_compare0): Likewise.
4258 (not_shiftsi_compare0_scratch): Likewise.
4259 (arm_one_cmplsi2): Likewise.
4260 (thumb_one_complsi2): Likewise.
4261 (notsi_compare0): Likewise.
4262 (notsi_compare0_scratch): Likewise.
4263 (thumb1_zero_extendhisi2): Likewise.
4264 (arm_zero_extendhisi2): Likewise.
4265 (arm_zero_extendhisi2_v6): Likewise.
4266 (arm_zero_extendhisi2addsi): Likewise.
4267 (thumb1_zero_extendqisi2): Likewise.
4268 (thumb1_zero_extendqisi2_v6): Likewise.
4269 (arm_zero_extendqisi2): Likewise.
4270 (arm_zero_extendqisi2_v6): Likewise.
4271 (arm_zero_extendqisi2addsi): Likewise.
4272 (thumb1_extendhisi2): Likewise.
4273 (arm_extendhisi2): Likewise.
4274 (arm_extendhisi2_v6): Likewise.
4275 (arm_extendqisi): Likewise.
4276 (arm_extendqisi_v6): Likewise.
4277 (arm_extendqisi2addsi): Likewise.
4278 (thumb1_extendqisi2): Likewise.
4279 (thumb1_movdi_insn): Likewise.
4280 (arm_movsi_insn): Likewise.
4281 (movsi_compare0): Likewise.
4282 (movhi_insn_arch4): Likewise.
4283 (movhi_bytes): Likewise.
4284 (arm_movqi_insn): Likewise.
4285 (thumb1_movqi_insn): Likewise.
4286 (arm32_movhf): Likewise.
4287 (thumb1_movhf): Likewise.
4288 (arm_movsf_soft_insn): Likewise.
4289 (thumb1_movsf_insn): Likewise.
4290 (movdf_soft_insn): Likewise.
4291 (thumb_movdf_insn): Likewise.
4292 (arm_cmpsi_insn): Likewise.
4293 (cmpsi_shiftsi): Likewise.
4294 (cmpsi_shiftsi_swp): Likewise.
4295 (arm_cmpsi_negshiftsi_si): Likewise.
4296 (movsicc_insn): Likewise.
4297 (movsfcc_soft_insn): Likewise.
4298 (arith_shiftsi): Likewise.
4299 (arith_shiftsi_compare0
4300 (arith_shiftsi_compare0_scratch
4301 (sub_shiftsi): Likewise.
4302 (sub_shiftsi_compare0
4303 (sub_shiftsi_compare0_scratch
4304 (and_scc): Likewise.
4305 (cond_move): Likewise.
4306 (if_plus_move): Likewise.
4307 (if_move_plus): Likewise.
4308 (if_move_not): Likewise.
4309 (if_not_move): Likewise.
4310 (if_shift_move): Likewise.
4311 (if_move_shift): Likewise.
4312 (if_shift_shift): Likewise.
4313 (if_not_arith): Likewise.
4314 (if_arith_not): Likewise.
4315 (cond_move_not): Likewise.
4316 (thumb1_ashlsi3): Set type attribute.
4317 (thumb1_ashrsi3): Likewise.
4318 (thumb1_lshrsi3): Likewise.
4319 (thumb1_rotrsi3): Likewise.
4320 (shiftsi3_compare0_scratch): Likewise.
4321 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
4322 (neon_mov<mode>): Likewise.
4323 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
4325 (thumb2_movsi_insn): Likewise.
4326 (thumb2_cmpsi_neg_shiftsi): Likewise.
4327 (thumb2_extendqisi_v6): Likewise.
4328 (thumb2_zero_extendhisi2_v6): Likewise.
4329 (thumb2_zero_extendqisi2_v6): Likewise.
4330 (thumb2_shiftsi3_short): Likewise.
4331 (thumb2_addsi3_compare0_scratch): Likewise.
4332 (orsi_not_shiftsi_si): Likewise.
4333 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4334 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
4336 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4337 (1020alu_shift_op): Likewise.
4338 (1020alu_shift_reg_op): Likewise.
4339 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4340 (alu_shift_op): Likewise.
4341 (alu_shift_reg_op): Likewise.
4342 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4343 (11_alu_shift_op): Likewise.
4344 (11_alu_shift_reg_op): Likewise.
4345 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4346 (9_alu_shift_reg_op): Likewise.
4347 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4349 (cortex_a15_alu_shift): Likewise.
4350 (cortex_a15_alu_shift_reg): Likewise.
4351 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
4353 (cortex_a5_alu_shift): Likewise.
4354 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
4356 (cortex_a53_alu_shift): Likewise.
4357 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4359 (cortex_a7_alu_reg): Likewise.
4360 (cortex_a7_alu_shift): Likewise.
4361 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
4363 (cortex_a8_alu_shift): Likewise.
4364 (cortex_a8_alu_shift_reg): Likewise.
4365 (cortex_a8_mov): Likewise.
4366 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4367 (cortex_a9_dp_shift): Likewise.
4368 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
4370 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
4372 (cortex_r4_mov): Likewise.
4373 (cortex_r4_alu_shift): Likewise.
4374 (cortex_r4_alu_shift_reg): Likewise.
4375 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4376 (526_alu_shift_op): Likewise.
4377 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4378 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4379 (626te_alu_shift_op): Likewise.
4380 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
4381 (726te_alu_op): Likewise.
4382 (726te_alu_shift_op): Likewise.
4383 (726te_alu_shift_reg_op): Likewise.
4384 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4385 (mp626_alu_shift_op): Likewise.
4386 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
4387 (pj4_alu_e1_conds): Likewise.
4388 (pj4_alu): Likewise.
4389 (pj4_alu_conds): Likewise.
4390 (pj4_shift): Likewise.
4391 (pj4_shift_conds): Likewise.
4392 (pj4_alu_shift): Likewise.
4393 (pj4_alu_shift_conds): Likewise.
4394 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4396 (cortexa7_older_only): Likewise.
4397 (cortexa7_younger): Likewise.
4399 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4401 * ipa-pure-const.c (generate_summary): Rename to...
4402 (pure_const_generate_summary): ... this.
4404 2013-07-17 Iain Sandoe <iain@codesourcery.com>
4406 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
4408 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
4411 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
4414 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4416 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
4417 enabled without -march=zEC12.
4418 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
4421 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4423 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
4424 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
4425 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
4426 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
4427 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
4428 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
4429 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
4431 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
4432 and ISA_HAS_NMADD3_NMSUB3.
4433 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
4434 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
4435 (nmsub4<mode>, nmsub3<mode>): Likewise.
4436 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
4438 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4440 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
4441 TARGET_MIPS5400 checking.
4443 2013-07-16 Jakub Jelinek <jakub@redhat.com>
4444 Peter Bergner <bergner@vnet.ibm.com>
4446 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
4447 registers in the comment.
4448 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
4449 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
4450 rather than FIRST_PSEUDO_REGISTERS.
4452 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
4454 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
4455 enable extra ISA flags with TARGET_HTM.
4457 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4459 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
4462 2013-07-15 Cong Hou <congh@google.com>
4464 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
4465 in compare function for sorting.
4467 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
4469 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
4470 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
4471 * config/rs6000/rs6000.opt: Add -mhtm option.
4472 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
4473 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
4474 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
4475 __HTM__ if the HTM instructions are available.
4476 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
4477 htm_spr_reg_operand): New define_predicates.
4478 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
4479 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
4481 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
4482 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
4483 HTM builtin functions.
4484 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
4485 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
4486 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
4487 (rs6000_builtin_mask_calculate): Likewise.
4488 (rs6000_option_override_internal): Likewise.
4489 (bdesc_htm): Add new HTM builtin support.
4490 (htm_spr_num): New function.
4491 (htm_spr_regno): Likewise.
4492 (rs6000_htm_spr_icode): Likewise.
4493 (htm_expand_builtin): Likewise.
4494 (htm_init_builtins): Likewise.
4495 (rs6000_expand_builtin): Add support for HTM builtin functions.
4496 (rs6000_init_builtins): Likewise.
4497 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
4499 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
4500 (TARGET_HTM, MASK_HTM): Define macros.
4501 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
4502 (FIXED_REGISTERS): Likewise.
4503 (CALL_USED_REGISTERS): Likewise.
4504 (CALL_REALLY_USED_REGISTERS): Likewise.
4505 (REG_ALLOC_ORDER): Likewise.
4506 (enum reg_class): Likewise.
4507 (REG_CLASS_NAMES): Likewise.
4508 (REG_CLASS_CONTENTS): Likewise.
4509 (REGISTER_NAMES): Likewise.
4510 (ADDITIONAL_REGISTER_NAMES): Likewise.
4511 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
4512 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
4513 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
4514 * config/rs6000/htm.md: New file.
4515 * config/rs6000/htmintrin.h: New file.
4516 * config/rs6000/htmxlintrin.h: New file.
4518 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
4520 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4521 Define SYMBOL_TINY_GOT, update comment.
4522 * config/aarch64/aarch64.c
4523 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
4524 (aarch64_expand_mov_immediate): Likewise.
4525 (aarch64_print_operand): Likewise.
4526 (aarch64_classify_symbol): Likewise.
4527 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4528 (ldr_got_tiny): Define.
4530 2013-07-13 Tobias Grosser <tobias@grosser.es>
4532 PR tree-optimization/54094
4533 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4534 scheduling dimension for the parallelism check from the polyhedral
4535 information in the AST.
4536 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4539 2013-07-13 Jason Merrill <jason@redhat.com>
4541 * print-tree.c (debug_vec_tree): Use debug_raw.
4542 (debug_raw (vec<tree, va_gc> &)): New.
4543 (debug_raw (vec<tree, va_gc> *)): New.
4544 * tree.h: Declare them.
4546 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4548 * ifcvt.c (ifcvt_after_combine): New static variable.
4549 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4551 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4552 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4553 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4555 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4557 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4559 2013-07-12 Michael Matz <matz@suse.de>
4562 * convert.c (convert_to_real): Reject non-float inner types.
4564 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4566 * config/aarch64/aarch64-protos.h
4567 (aarch64_simd_immediate_valid_for_move): Remove.
4568 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4569 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4571 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4573 2013-07-11 Steve Ellcey <sellcey@mips.com>
4575 * config/mips/mips.c (mips_conditional_register_usage): Do not
4576 use t[0-7] registers in MIPS16 mode when optimizing for size.
4578 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4580 * config/i386/i386.c (dispatch_function_versions): Fix array
4581 indexing of function_version_info to match actual_versions.
4583 2013-07-11 Teresa Johnson <tejohnson@google.com>
4585 * vec.h (struct va_gc): Move release out-of-line.
4586 (va_gc::release): Call ggc_free on released vec.
4588 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4590 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4591 Require GOT register as additional operand in UNSPEC.
4592 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4593 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4594 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4595 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4596 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4597 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4598 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4600 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4603 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4604 name seen by assembler/linker rather if available.
4606 2013-07-11 Andreas Schwab <schwab@suse.de>
4608 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4610 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4612 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4614 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4616 * doc/tm.texi.in: Move hook documentation to ....
4617 * target.def: ... here.
4619 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4621 * doc/tm.texi: Regenerate.
4623 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4626 * doc/invoke.texi: Document -Wconditionally-supported.
4628 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4631 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4634 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4637 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4638 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4639 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4640 Remove duplicate devices.
4641 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4642 * config/avr/t-multilib: Regenerate.
4643 * config/avr/avr-tables.opt: Regenerate.
4644 * doc/avr-mmcu.texi: Regenerate.
4646 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4649 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4651 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4653 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4654 the cost of MULT when optimizing for size.
4656 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4658 * config/cr16/cr16-protos.h: Don't include target.h.
4660 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4662 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4663 adjust register size for TDmode and TFmode for VSX registers.
4665 2013-07-08 Kai Tietz <ktietz@redhat.com>
4668 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4669 hook_bool_const_tree_true.
4671 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4673 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4674 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4675 * config/s390/s390.md: Define FPR*_REGNUM constants.
4676 Fix FPR2_REGNUM constant (18 -> 17).
4677 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4678 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4679 ("trunc<BFP:mode><DFP_ALL:mode>2")
4680 ("trunc<DFP_ALL:mode><BFP:mode>2")
4681 ("*extend<BFP:mode><DFP_ALL:mode>2")
4682 ("*extend<DFP_ALL:mode><BFP:mode>2")
4683 ("extend<BFP:mode><DFP_ALL:mode>2")
4684 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4687 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4689 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4691 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4693 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4694 and cfun_fpr_bit_p to cfun_fpr_save_p.
4695 (s390_frame_area, s390_register_info, s390_frame_info)
4696 (s390_emit_prologue, s390_emit_epilogue)
4697 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4699 * config/s390/s390.h: Define *_REGNUM macros for floating point
4702 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4704 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4706 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4708 PR rtl-optimization/57786
4709 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4710 and break out of the loop when it is set to false.
4712 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4715 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4716 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4717 (const_int 63)) 0)).
4718 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4719 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4720 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4722 PR rtl-optimization/57829
4723 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4724 mask bits outside of mode are just sign-extension from mode to HWI.
4726 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4728 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4729 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4730 adjust_address instead of change_address to keep info about alignment.
4731 (emit_strmov): Remove.
4732 (emit_memmov): New function.
4733 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4734 (expand_movmem_epilogue): Likewise and return updated rtx for
4736 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4737 destination and source.
4738 (decide_alignment): Refactor, handle vector_loop.
4739 (ix86_expand_movmem): Likewise.
4740 (ix86_expand_setmem): Likewise.
4741 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4743 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4745 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4746 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4748 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4750 * config/i386/sse.md (sse_movlhps): Change alternative 3
4751 of operand 2 to "m".
4753 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4756 * config/i386/sse.md (iptr): New mode attribute.
4757 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4758 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4759 (<sse>_vmmul<mode>3): Ditto.
4760 (<sse>_vmdiv<mode>3): Ditto.
4761 (sse_vmrcpv4sf2): Ditto.
4762 (<sse>_vmsqrt<mode>2): Ditto.
4763 (sse_vmrsqrtv4sf2): Ditto.
4764 (<sse>_vm<code><mode>3): Ditto.
4765 (avx_vmcmp<mode>3): Ditto.
4766 (<sse>_vmmaskcmp<mode>3): Ditto.
4767 (<sse>_comi): Ditto.
4768 (<sse>_ucomi): Ditto.
4769 (*xop_vmfrcz_<mode>): Ditto.
4770 (*fmai_fmadd_<mode>): Ditto.
4771 (*fmai_fmsub_<mode>): Ditto.
4772 (*fmai_fnmadd_<mode>): Ditto.
4773 (*fmai_fnmsub_<mode>): Ditto.
4774 (*fma4i_vmfmadd_<mode>): Ditto.
4775 (*fma4i_vmfmsub_<mode>): Ditto.
4776 (*fma4i_vmfnmadd_<mode>): Ditto.
4777 (*fma4i_vmfnmsub_<mode>): Ditto.
4778 (*xop_vmfrcz_<mode>): Ditto.
4779 (sse_cvtps2pi): Ditto.
4780 (sse_cvttps2pi): Ditto.
4781 (sse_cvtss2si): Ditto.
4782 (sse_cvtss2si_2): Ditto.
4783 (sse_cvtss2siq_2): Ditto.
4784 (sse_cvttss2si): Ditto.
4785 (sse_cvttss2siq): Ditto.
4786 (sse_cvtsd2si): Ditto.
4787 (sse_cvtsd2si_2): Ditto.
4788 (sse_cvtsd2siq_2): Ditto.
4789 (sse_cvttsd2si): Ditto.
4790 (sse_cvttsd2siq): Ditto.
4791 (sse_cvtsd2ss): Ditto.
4792 (sse_cvtss2sd): Ditto.
4793 (avx2_pbroadcast<mode>): Ditto.
4794 (avx2_pbroadcast<mode>_1): Ditto.
4795 (*avx_vperm_broadcast_v4sf): Ditto.
4797 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4798 (sse_movlhps): Ditto.
4799 (sse_storehps): Ditto.
4800 (sse_loadhps): Ditto.
4801 (sse_storelps): Ditto.
4802 (sse_loadlps): Ditto.
4803 (*vec_concatv4sf): Ditto.
4804 (*vec_interleave_highv2df): Ditto.
4805 (*vec_interleave_lowv2df): Ditto.
4806 (*vec_extractv2df_1_sse): Ditto.
4807 (*vec_extractv2df_0_sse): Ditto.
4808 (sse2_storelpd): Ditto.
4809 (sse2_loadlpd): Ditto.
4810 (sse2_movsd): Ditto.
4811 (*vec_concatv4si): Ditto.
4812 (vec_concatv2di): Ditto.
4814 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4815 for Intel asm dialect.
4816 (mmx_punpcklwd): Ditto.
4817 (mmx_punpckldq): Ditto.
4819 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4820 for intel assembler dialect.
4822 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4825 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4826 for BUILT_IN_C{LZ,LRSB}*.
4827 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4828 * tree-vrp.c (extract_range_basic): Handle
4829 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4830 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4831 fall thru to code calling set_value*.
4832 * builtins.c (expand_builtin): Remove *IMAX cases.
4833 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4834 if width is bigger than 2*HWI.
4836 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4838 PR rtl-optimization/55342
4839 * lra-int.h (lra_subreg_reload_pseudos): New.
4840 * lra.c: Add undoing optional reloads to the block diagram.
4841 (lra_subreg_reload_pseudos): New.
4842 (lra_optional_reload_pseudos): Change comments.
4843 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4844 lra_optional_reload_pseudos after undo transformations.
4845 * lra-assigns.c (pseudo_prefix_title): New.
4846 (lra_setup_reg_renumber): Use it.
4847 (spill_for): Ditto. Check subreg reload pseudos too.
4848 (assign_by_spills): Consider subreg reload pseudos too.
4849 * lra-constraints.c (simplify_operand_subreg): Use
4850 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4851 (curr_insn_transform): Recognize and do optional reloads.
4852 (undo_optional_reloads): New.
4853 (lra_undo_inheritance): Call undo_optional_reloads.
4855 2013-07-05 Thomas Quinot <quinot@adacore.com>
4857 * tree-complex.c (expand_complex_operations_1): Fix typo.
4859 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4861 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4862 (tune_params): New member 'const vec_costs'.
4863 * config/aarch64/aarch64.c (generic_vector_cost): New.
4864 (generic_tunings): New member 'generic_vector_cost'.
4865 (aarch64_builtin_vectorization_cost): New.
4866 (aarch64_add_stmt_cost): New.
4867 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4868 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4870 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4873 * config/i386/predicates.md (vsib_address_operand): Disallow
4874 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4876 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4879 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4880 expand_builtin_setjmp_receiver.
4881 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4882 with NULL for the label parameter.
4883 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4884 the frame-pointer. Adjust comments.
4885 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4886 only if LABEL is non-NULL.
4888 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4890 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4891 (ARM_ABI_AAPCS64): Ditto.
4893 (ARM_DEFAULT_ABI): Ditto.
4895 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4897 * config/aarch64/aarch64-builtins.c
4898 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4899 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4901 * config/aarch64/aarch64-simd.md
4902 (aarch64_ld1<VALL:mode>): New.
4903 (aarch64_st1<VALL:mode>): Likewise.
4904 * config/aarch64/arm_neon.h
4905 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4907 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4909 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4911 (ix86_option_override_internal):Turn on all -mavx target flags by
4912 default as they are dependent on AVX anyway.
4914 2013-07-02 Cary Coutant <ccoutant@google.com>
4916 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4918 (loc_checksum_ordered): Likewise.
4919 (hash_loc_operands): Remove inline keyword.
4921 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4923 PR tree-optimization/57741
4924 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4925 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4926 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4927 Allow REAL_CST step_exprs if flag_associative_math.
4928 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4930 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4932 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4934 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4936 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4938 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4940 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4942 (iorsi3_insn): Likewise.
4943 (arm_xorsi3): Likewise.
4945 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4947 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4949 (attribute "type"): ... here, and prefix with "wmmx_".
4950 (attribute "core_cycles"): Update for attribute changes.
4951 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4952 (tbcstv4hi): Likewise.
4953 (tbcstv2si): Likewise.
4954 (iwmmxt_iordi3): Likewise.
4955 (iwmmxt_xordi3): Likewise.
4956 (iwmmxt_anddi3): Likewise.
4957 (iwmmxt_nanddi3): Likewise.
4958 (iwmmxt_arm_movdi): Likewise.
4959 (iwmmxt_movsi_insn): Likewise.
4960 (mov<mode>_internal): Likewise.
4961 (and<mode>3_iwmmxt): Likewise.
4962 (ior<mode>3_iwmmxt): Likewise.
4963 (xor<mode>3_iwmmxt): Likewise.
4964 (add<mode>3_iwmmxt): Likewise.
4965 (ssaddv8qi3): Likewise.
4966 (ssaddv4hi3): Likewise.
4967 (ssaddv2si3): Likewise.
4968 (usaddv8qi3): Likewise.
4969 (usaddv4hi3): Likewise.
4970 (usaddv2si3): Likewise.
4971 (sub<mode>3_iwmmxt): Likewise.
4972 (sssubv8qi3): Likewise.
4973 (sssubv4hi3): Likewise.
4974 (sssubv2si3): Likewise.
4975 (ussubv8qi3): Likewise.
4976 (ussubv4hi3): Likewise.
4977 (ussubv2si3): Likewise.
4978 (mulv4hi3_iwmmxt): Likewise.
4979 (smulv4hi3_highpart): Likewise.
4980 (umulv4hi3_highpart): Likewise.
4981 (iwmmxt_wmacs): Likewise.
4982 (iwmmxt_wmacsz): Likewise.
4983 (iwmmxt_wmacu): Likewise.
4984 (iwmmxt_wmacuz): Likewise.
4985 (iwmmxt_clrdi): Likewise.
4986 (iwmmxt_clrv8qi): Likewise.
4987 (iwmmxt_clr4hi): Likewise.
4988 (iwmmxt_clr2si): Likewise.
4989 (iwmmxt_uavgrndv8qi3): Likewise.
4990 (iwmmxt_uavgrndv4hi3): Likewise.
4991 (iwmmxt_uavgv8qi3): Likewise.
4992 (iwmmxt_uavgv4hi3): Likewise.
4993 (iwmmxt_tinsrb): Likewise.
4994 (iwmmxt_tinsrh): Likewise.
4995 (iwmmxt_tinsrw): Likewise.
4996 (iwmmxt_textrmub): Likewise.
4997 (iwmmxt_textrmsb): Likewise.
4998 (iwmmxt_textrmuh): Likewise.
4999 (iwmmxt_textrmsh): Likewise.
5000 (iwmmxt_textrmw): Likewise.
5001 (iwmxxt_wshufh): Likewise.
5002 (eqv8qi3): Likewise.
5003 (eqv4hi3): Likewise.
5004 (eqv2si3): Likewise.
5005 (gtuv8qi3): Likewise.
5006 (gtuv4hi3): Likewise.
5007 (gtuv2si3): Likewise.
5008 (gtv8qi3): Likewise.
5009 (gtv4hi3): Likewise.
5010 (gtv2si3): Likewise.
5011 (smax<mode>3_iwmmxt): Likewise.
5012 (umax<mode>3_iwmmxt): Likewise.
5013 (smin<mode>3_iwmmxt): Likewise.
5014 (umin<mode>3_iwmmxt): Likewise.
5015 (iwmmxt_wpackhss): Likewise.
5016 (iwmmxt_wpackwss): Likewise.
5017 (iwmmxt_wpackdss): Likewise.
5018 (iwmmxt_wpackhus): Likewise.
5019 (iwmmxt_wpackwus): Likewise.
5020 (iwmmxt_wpackdus): Likewise.
5021 (iwmmxt_wunpckihb): Likewise.
5022 (iwmmxt_wunpckihh): Likewise.
5023 (iwmmxt_wunpckihw): Likewise.
5024 (iwmmxt_wunpckilb): Likewise.
5025 (iwmmxt_wunpckilh): Likewise.
5026 (iwmmxt_wunpckilw): Likewise.
5027 (iwmmxt_wunpckehub): Likewise.
5028 (iwmmxt_wunpckehuh): Likewise.
5029 (iwmmxt_wunpckehuw): Likewise.
5030 (iwmmxt_wunpckehsb): Likewise.
5031 (iwmmxt_wunpckehsh): Likewise.
5032 (iwmmxt_wunpckehsw): Likewise.
5033 (iwmmxt_wunpckelub): Likewise.
5034 (iwmmxt_wunpckeluh): Likewise.
5035 (iwmmxt_wunpckeluw): Likewise.
5036 (iwmmxt_wunpckelsb): Likewise.
5037 (iwmmxt_wunpckelsh): Likewise.
5038 (iwmmxt_wunpckelsw): Likewise.
5039 (ror<mode>3): Likewise.
5040 (ashr<mode>3_iwmmxt): Likewise.
5041 (lshr<mode>3_iwmmxt): Likewise.
5042 (ashl<mode>3_iwmmxt): Likewise.
5043 (ror<mode>3_di): Likewise.
5044 (ashr<mode>3_di): Likewise.
5045 (lshr<mode>3_di): Likewise.
5046 (ashl<mode>3_di): Likewise.
5047 (iwmmxt_wmadds): Likewise.
5048 (iwmmxt_wmaddu): Likewise.
5049 (iwmmxt_tmia): Likewise.
5050 (iwmmxt_tmiaph): Likewise.
5051 (iwmmxt_tmiabb): Likewise.
5052 (iwmmxt_tmiatb): Likewise.
5053 (iwmmxt_tmiabt): Likewise.
5054 (iwmmxt_tmiatt): Likewise.
5055 (iwmmxt_tmovmskb): Likewise.
5056 (iwmmxt_tmovmskh): Likewise.
5057 (iwmmxt_tmovmskw): Likewise.
5058 (iwmmxt_waccb): Likewise.
5059 (iwmmxt_wacch): Likewise.
5060 (iwmmxt_waccw): Likewise.
5061 (iwmmxt_waligni): Likewise.
5062 (iwmmxt_walignr): Likewise.
5063 (iwmmxt_walignr0): Likewise.
5064 (iwmmxt_walignr1): Likewise.
5065 (iwmmxt_walignr2): Likewise.
5066 (iwmmxt_walignr3): Likewise.
5067 (iwmmxt_wsadb): Likewise.
5068 (iwmmxt_wsadh): Likewise.
5069 (iwmmxt_wsadbz): Likewise.
5070 (iwmmxt_wsadhz): Likewise.
5071 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5072 (iwmmxt_wabsdiffb): Likewise.
5073 (iwmmxt_wabsdiffh): Likewise.
5074 (iwmmxt_wabsdiffw): Likewise.
5075 (iwmmxt_waddsubhx): Likewise
5076 (iwmmxt_wsubaddhx): Likewise.
5077 (addc<mode>3): Likewise.
5078 (iwmmxt_avg4): Likewise.
5079 (iwmmxt_avg4r): Likewise.
5080 (iwmmxt_wmaddsx): Likewise.
5081 (iwmmxt_wmaddux): Likewise.
5082 (iwmmxt_wmaddsn): Likewise.
5083 (iwmmxt_wmaddun): Likewise.
5084 (iwmmxt_wmulwsm): Likewise.
5085 (iwmmxt_wmulwum): Likewise.
5086 (iwmmxt_wmulsmr): Likewise.
5087 (iwmmxt_wmulumr): Likewise.
5088 (iwmmxt_wmulwsmr): Likewise.
5089 (iwmmxt_wmulwumr): Likewise.
5090 (iwmmxt_wmulwl): Likewise.
5091 (iwmmxt_wqmulm): Likewise.
5092 (iwmmxt_wqmulwm): Likewise.
5093 (iwmmxt_wqmulmr): Likewise.
5094 (iwmmxt_wqmulwmr): Likewise.
5095 (iwmmxt_waddbhusm): Likewise.
5096 (iwmmxt_waddbhusl): Likewise.
5097 (iwmmxt_wqmiabb): Likewise.
5098 (iwmmxt_wqmiabt): Likewise.
5099 (iwmmxt_wqmiatb): Likewise.
5100 (iwmmxt_wqmiatt): Likewise.
5101 (iwmmxt_wqmiabbn): Likewise.
5102 (iwmmxt_wqmiabtn): Likewise.
5103 (iwmmxt_wqmiatbn): Likewise.
5104 (iwmmxt_wqmiattn): Likewise.
5105 (iwmmxt_wmiabb): Likewise.
5106 (iwmmxt_wmiabt): Likewise.
5107 (iwmmxt_wmiatb): Likewise.
5108 (iwmmxt_wmiatt): Likewise.
5109 (iwmmxt_wmiabbn): Likewise.
5110 (iwmmxt_wmiabtn): Likewise.
5111 (iwmmxt_wmiatbn): Likewise.
5112 (iwmmxt_wmiattn): Likewise.
5113 (iwmmxt_wmiawbb): Likewise.
5114 (iwmmxt_wmiawbt): Likewise.
5115 (iwmmxt_wmiawtb): Likewise.
5116 (iwmmxt_wmiawtt): Likewise.
5117 (iwmmxt_wmiawbbn): Likewise.
5118 (iwmmxt_wmiawbtn): Likewise.
5119 (iwmmxt_wmiawtbn): Likewise.
5120 (iwmmxt_wmiawttn): Likewise.
5121 (iwmmxt_wmerge): Likewise.
5122 (iwmmxt_tandc<mode>3): Likewise.
5123 (iwmmxt_torc<mode>3): Likewise.
5124 (iwmmxt_torvsc<mode>3): Likewise.
5125 (iwmmxt_textrc<mode>3): Likewise.
5126 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5127 (wmmxt_pack): Likewise.
5128 (wmmxt_mult_c1): Likewise.
5129 (wmmxt_mult_c2): Likewise.
5130 (wmmxt_alu_c1): Likewise.
5131 (wmmxt_alu_c2): Likewise.
5132 (wmmxt_alu_c3): Likewise.
5133 (wmmxt_transfer_c1): Likewise.
5134 (wmmxt_transfer_c2): Likewise.
5135 (wmmxt_transfer_c3): Likewise.
5136 (marvell_f_iwmmxt_wstr): Likewise.
5137 (marvell_f_iwmmxt_wldr): Likewise.
5139 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5141 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5143 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5146 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5147 * lra-constraints.c (need_for_split_p): Check call used hard regs
5148 living through calls.
5150 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5151 call used regs for call insn.
5153 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5156 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5157 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5159 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5161 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5163 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5165 * lra-constraints.c (need_for_split_p): Check call used hard regs
5166 living through calls.
5168 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5171 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5172 to tie with any other modes. Eliminate Altivec vector mode tests,
5173 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5174 code, to return 0 if testing MODE2 for a condition, if we've
5175 already tested MODE1 for the same condition.
5177 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5179 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5182 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5184 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5185 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5187 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5189 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5191 (aarch64_symbolic_constant_p): Remove.
5192 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5193 static. Fix line length and white space.
5194 (aarch64_symbolic_constant_p): Remove.
5195 * config/aarch64/predicates.md (aarch64_valid_symref):
5196 Use aarch64_classify_symbol_expression.
5198 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5200 * config/arm/constraints.md (Ts): New constraint.
5201 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5203 (compare_scc): Use "Ts" constraint for operand 0.
5204 (ior_scc_scc): Likewise.
5205 (and_scc_scc): Likewise.
5206 (and_scc_scc_nodom): Likewise.
5207 (ior_scc_scc_cmp): Likewise for operand 7.
5208 (and_scc_scc_cmp): Likewise.
5209 * config/arm/thumb2.md (thumb2_movsi_insn):
5210 Add alternatives for 16-bit encodings.
5211 (thumb2_movhi_insn): Likewise.
5212 (thumb2_movsicc_insn): Likewise.
5213 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5214 (thumb2_negscc): Use "Ts" constraint.
5215 Move mvn instruction outside cond_exec block.
5216 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5217 for 16-bit encodings.
5219 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5221 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5223 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5224 (mulsi3subsi): Likewise.
5225 (mulsidi3adddi): Likewise.
5226 (mulsidi3_v6): Likewise.
5227 (umulsidi3_v6): Likewise.
5228 (umulsidi3adddi_v6): Likewise.
5229 (smulsi3_highpart_v6): Likewise.
5230 (umulsi3_highpart_v6): Likewise.
5231 (mulhisi3tb): Likewise.
5232 (mulhisi3bt): Likewise.
5233 (mulhisi3tt): Likewise.
5234 (maddhisi4): Likewise.
5235 (maddhisi4tb): Likewise.
5236 (maddhisi4tt): Likewise.
5237 (maddhidi4): Likewise.
5238 (maddhidi4tb): Likewise.
5239 (maddhidi4tt): Likewise.
5240 (zeroextractsi_compare0_scratch): Likewise.
5241 (insv_zero): Likewise.
5242 (insv_t2): Likewise.
5243 (anddi_notzesidi_di): Likewise.
5244 (anddi_notsesidi_di): Likewise.
5245 (andsi_notsi_si): Likewise.
5246 (iordi_zesidi_di): Likewise.
5247 (xordi_zesidi_di): Likewise.
5248 (andsi_iorsi3_notsi): Likewise.
5250 (smax_m1): Likewise.
5252 (not_shiftsi): Likewise.
5253 (unaligned_loadsi): Likewise.
5254 (unaligned_loadhis): Likewise.
5255 (unaligned_loadhiu): Likewise.
5256 (unaligned_storesi): Likewise.
5257 (unaligned_storehi): Likewise.
5258 (extv_reg): Likewise.
5259 (extzv_t2): Likewise.
5261 (udivsi3): Likewise.
5262 (arm_zero_extendhisi2addsi): Likewise.
5263 (arm_zero_extendqisi2addsi): Likewise.
5264 (compareqi_eq0): Likewise.
5265 (arm_extendhisi2_v6): Likewise.
5266 (arm_extendqisi2addsi): Likewise.
5267 (arm_movt): Likewise.
5268 (thumb2_ldrd): Likewise.
5269 (thumb2_ldrd_base): Likewise.
5270 (thumb2_ldrd_base_neg): Likewise.
5271 (thumb2_strd): Likewise.
5272 (thumb2_strd_base): Likewise.
5273 (thumb2_strd_base_neg): Likewise.
5274 (arm_negsi2): Add alternative for 16-bit encoding.
5275 (arm_one_cmplsi2): Likewise.
5277 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5279 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
5280 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
5281 (movdfcc): Likewise.
5282 * config/arm/vfp.md (*thumb2_movsf_vfp):
5283 Disable predication for arm_restrict_it.
5284 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
5285 (*thumb2_movdfcc_vfp): Likewise.
5286 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
5287 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
5288 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
5289 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
5290 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
5291 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
5292 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
5293 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
5294 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
5295 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
5296 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
5297 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
5298 Disable predication for arm_restrict_it.
5300 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
5302 * config/i386/bmiintrin.h (_bextr_u32): New.
5303 (_bextr_u64): Ditto.
5305 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
5307 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
5308 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
5309 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
5310 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
5311 * config/mips/n32-elf.h: ...this new file.
5313 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
5316 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
5317 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
5319 2013-06-27 Catherine Moore <clm@codesourcery.com>
5321 * config/mips/mips-tables.opt: Regenerate.
5322 * config/mips/mips-cpus.def: Add m14ke and m14kec.
5323 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
5324 * doc/invoke.texi: Add -m14kc.
5326 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5329 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
5330 constraints of operand 1 and 2.
5333 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
5334 to match RTL canonicalization. Swap predicates and
5335 constraints of operand 1 and 2.
5337 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
5339 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
5340 Process OP_INOUT regs for splitting too.
5342 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5344 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
5345 decl before the loop, initialize to NULL.
5346 (vectorizable_load): Initialize ptr_incr to NULL.
5348 2013-06-27 Martin Jambor <mjambor@suse.cz>
5351 * ipa-ref.h (ipa_maybe_record_reference): Declare.
5352 * ipa-ref.c (ipa_maybe_record_reference): New function.
5353 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
5354 * ipa-cp.c (create_specialized_node): Record potential references from
5356 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
5358 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5360 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
5361 parameter 'mode' of type 'enum machine_mode mode'; change to pass
5362 'mode' to force_reg.
5363 (aarch64_add_offset): Update calls to aarch64_force_temporary.
5364 (aarch64_expand_mov_immediate): Likewise.
5366 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5368 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
5369 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
5371 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5373 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
5374 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5375 (struct machine_function): Add tbegin_p.
5376 (s390_canonicalize_comparison): Fold CC mode compares to
5377 conditional jump if possible.
5378 (s390_emit_jump): Return the emitted jump.
5379 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
5380 Handle CCRAWmode compares.
5381 (s390_option_override): Default to -mhtm if available.
5382 (s390_reg_clobbered_rtx): Handle floating point regs as well.
5383 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
5384 FPRs instead of df_regs_ever_live_p.
5385 (s390_optimize_nonescaping_tx): New function.
5386 (s390_init_frame_layout): Extend clobbered_regs array to cover
5388 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
5389 (s390_expand_tbegin): New function.
5390 (enum s390_builtin): New enum definition.
5391 (code_for_builtin): New array definition.
5392 (s390_init_builtins): New function.
5393 (s390_expand_builtin): New function.
5394 (TARGET_INIT_BUILTINS): Define.
5395 (TARGET_EXPAND_BUILTIN): Define.
5396 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5397 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
5398 (s390_alc_comparison): Likewise.
5399 * config/s390/s390-modes.def: Add CCRAWmode.
5400 * config/s390/s390.h (processor_flags): Add PF_TX.
5401 (TARGET_CPU_HTM): Define macro.
5402 (TARGET_HTM): Define macro.
5403 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
5404 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
5405 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5406 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
5407 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5408 (TBEGIN_MASK, TBEGINC_MASK): New constants.
5409 ("*cc_to_int"): Move up.
5410 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
5411 constants other than 0.
5412 ("*ccraw_to_int"): New insn and splitter definition.
5413 ("tbegin", "tbegin_nofloat", "tbegin_retry")
5414 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
5415 ("tx_assist"): New expander.
5416 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
5417 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
5418 * config/s390/s390.opt: Add -mhtm option.
5419 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
5420 * config/s390/htmxlintrin.h: New file.
5421 * config/s390/htmintrin.h: New file.
5422 * config/s390/s390intrin.h: New file.
5423 * doc/extend.texi: Document htm builtins.
5424 * config.gcc: Add the new header files to extra_headers.
5426 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
5428 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
5429 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
5431 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
5432 Pat Haugen <pthaugen@us.ibm.com>
5433 Peter Bergner <bergner@vnet.ibm.com>
5435 * config/rs6000/power8.md: New.
5436 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
5437 setting for power8 entry.
5438 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
5439 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
5440 test for Power4/Power5 only.
5441 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
5443 (force_new_group): Adjust comment.
5444 * config/rs6000/rs6000.md: Include power8.md.
5446 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
5448 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
5449 * config/arm/arm-protos.h (arm_max_conditional_execute): New
5451 (tune_params): Update comment.
5452 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
5453 (arm_max_conditional_execute): New function.
5454 (thumb2_final_prescan_insn): Use max_insn_skipped and
5455 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
5457 2013-06-25 Jakub Jelinek <jakub@redhat.com>
5459 PR tree-optimization/57705
5460 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
5461 SSA_NAME step, provided that it is not defined inside the loop.
5462 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
5463 (get_initial_def_for_induction): Handle SSA_NAME IV step.
5465 2013-06-25 Martin Jambor <mjambor@suse.cz>
5468 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
5469 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
5471 (ipa_note_param_call): Initialize member_ptr flag.
5472 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
5473 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
5474 (ipa_write_indirect_edge_info): Stream member_ptr flag.
5475 (ipa_read_indirect_edge_info): Likewise.
5477 2013-06-25 Richard Biener <rguenther@suse.de>
5480 * passes.c (init_optimization_passes): Move pass_fold_builtins
5481 and pass_dce earlier with -Og.
5483 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
5485 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
5486 <BIT_FIELD_REF>: Remove trailing TAB.
5487 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
5490 2013-06-24 Martin Jambor <mjambor@suse.cz>
5492 PR tree-optimization/57358
5493 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
5494 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
5495 (ipa_analyze_params_uses): Generate pessimistic info when true.
5497 2013-06-24 Martin Jambor <mjambor@suse.cz>
5499 PR tree-optimization/57539
5500 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
5501 global.inlined_to of the new node to it. All callers changed.
5502 * ipa-inline-transform.c (clone_inlined_nodes): New variable
5503 inlining_into, pass it to cgraph_clone_node.
5504 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
5505 ipa_free_edge_args_substructures.
5506 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
5507 rdesc linked list. Do not assert rdesc edges have inlined caller.
5508 Assert we have found an rdesc in the rdesc list.
5510 2013-06-24 Richard Biener <rguenther@suse.de>
5512 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
5513 (pointer_set_lookup): Declare.
5514 (class pointer_map): New template class implementing a
5515 generic pointer to T map.
5516 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
5517 pointer_map<T>::contains, pointer_map<T>::insert,
5518 pointer_map<T>::traverse): New functions.
5519 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
5520 (pointer_set_lookup): New function.
5521 (pointer_set_contains): Use pointer_set_lookup.
5522 (pointer_set_insert): Likewise.
5523 (insert_aux): Remove.
5524 (struct pointer_map_t): Embed a pointer_set_t.
5525 (pointer_map_create): Adjust.
5526 (pointer_map_destroy): Likewise.
5527 (pointer_map_contains): Likewise.
5528 (pointer_map_insert): Likewise.
5529 (pointer_map_traverse): Likewise.
5530 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5531 pointer_map<unsigned> instead of a pointer_map_t.
5532 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5533 (streamer_tree_cache_lookup): Likewise.
5534 (streamer_tree_cache_create): Likewise.
5535 (streamer_tree_cache_delete): Likewise.
5536 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5537 pointer_map<unsigned> instead of a pointer_map_t.
5538 (lto_init_tree_ref_encoder): Adjust.
5539 (lto_destroy_tree_ref_encoder): Likewise.
5540 * lto-section-out.c (lto_output_decl_index): Likewise.
5541 (lto_record_function_out_decl_state): Likewise.
5542 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5544 2013-06-24 Richard Biener <rguenther@suse.de>
5546 PR tree-optimization/57488
5547 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5549 2013-06-24 Alan Modra <amodra@gmail.com>
5551 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5552 (gen_easy_altivec_constant): Likewise.
5553 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5554 easy_vector_constant_msb): Likewise.
5556 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5559 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5560 add missing return true.
5562 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5565 * config/sh/predicates.md (general_extend_operand): Invoke
5566 general_movsrc_operand for memory operands.
5567 (general_movsrc_operand): Allow reg+reg addressing, do not use
5568 general_operand for memory operands.
5570 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5572 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5573 when current target options does not apply.
5574 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5575 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5576 * config/i386/bmiintrin.h: Pass appropriate target
5577 attributes to header.
5578 * config/i386/mmintrin.h: Ditto.
5579 * config/i386/nmmintrin.h: Ditto.
5580 * config/i386/avx2intrin.h: Ditto.
5581 * config/i386/fxsrintrin.h: Ditto.
5582 * config/i386/tbmintrin.h: Ditto.
5583 * config/i386/xsaveintrin.h: Ditto.
5584 * config/i386/f16cintrin.h: Ditto.
5585 * config/i386/xtestintrin.h: Ditto.
5586 * config/i386/xsaveoptintrin.h: Ditto.
5587 * config/i386/bmi2intrin.h: Ditto.
5588 * config/i386/lzcntintrin.h: Ditto.
5589 * config/i386/smmintrin.h: Ditto.
5590 * config/i386/wmmintrin.h: Ditto.
5591 * config/i386/x86intrin.h: Remove all header include guards.
5592 * config/i386/prfchwintrin.h: Ditto.
5593 * config/i386/pmmintrin.h: Ditto.
5594 * config/i386/tmmintrin.h: Ditto.
5595 * config/i386/xmmintrin.h: Ditto.
5596 * config/i386/popcntintrin.h: Ditto.
5597 * config/i386/rdseedintrin.h: Ditto.
5598 * config/i386/ammintrin.h: Ditto.
5599 * config/i386/emmintrin.h: Ditto.
5600 * config/i386/immintrin.h: Remove all header include guards.
5601 * config/i386/fma4intrin.h: Ditto.
5602 * config/i386/lwpintrin.h: Ditto.
5603 * config/i386/xopintrin.h: Ditto.
5604 * config/i386/ia32intrin.h: Ditto.
5605 * config/i386/avxintrin.h: Ditto.
5606 * config/i386/rtmintrin.h: Ditto.
5607 * config/i386/fmaintrin.h: Ditto.
5608 * config/i386/mm3dnow.h: Ditto.
5610 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5612 * common/config/i386/i386-common.c: Handle LZCNT.
5614 2013-06-22 Andi Kleen <ak@linux.intel.com>
5616 * doc/extend.texi: Use __atomic_store_n instead of
5617 __atomic_store in HLE example.
5619 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5621 * config/sh/sh.c: Remove <cstdlib> workaround.
5623 2013-06-21 Andi Kleen <ak@linux.intel.com>
5625 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5627 2013-06-21 Andi Kleen <ak@linux.intel.com>
5629 * doc/extend.texi: Document that __atomic_clear and
5630 __atomic_test_and_set should only be used with bool.
5632 2013-06-20 Jan Hubicka <jh@suse.cz>
5634 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5636 * tree.c (decls_same_for_odr): New function.
5637 (same_for_edr): New function.
5638 (types_same_for_odr): New function.
5639 (get_binfo_at_offset): Use it.
5640 * tree.h (types_same_for_odr): Declare.
5642 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5643 Jason Merrill <jason@redhat.com>
5645 * system.h: Include <cstdlib> as well as <stdlib.h>.
5647 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5650 * config/i386/i386.c (construct_container): Report error if
5651 long double is used with disabled x87 float returns.
5653 2013-06-20 Jan Hubicka <jh@suse.cz>
5655 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5657 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5659 PR rtl-optimization/57425
5660 PR rtl-optimization/57569
5661 * alias.c (write_dependence_p): Remove parameters mem_mode and
5662 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5663 Changed all callers.
5664 (canon_anti_dependence): Get comments and semantics in sync.
5665 Add parameter mem_canonicalized. Changed all callers.
5666 * rtl.h (canon_anti_dependence): Update prototype.
5668 2013-06-20 Richard Biener <rguenther@suse.de>
5670 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5671 byte case, inline streamer_read_uchar and defer section
5674 2013-06-20 Richard Biener <rguenther@suse.de>
5676 PR tree-optimization/57584
5677 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5678 SSA names into the expanded expression that take part in
5679 abnormal coalescing.
5681 2013-06-19 Sharad Singhai <singhai@google.com>
5683 * gcov.c (print_usage): Handle new option.
5684 (process_args): Ditto.
5685 (get_gcov_intermediate_filename): New function.
5686 (output_intermediate_file): New function.
5687 (output_gcov_file): New function
5688 (generate_results): Handle new option.
5689 (release_function): Relase demangled name.
5690 (read_graph_file): Handle demangled name.
5691 (output_lines): Ditto.
5692 * doc/gcov.texi: Document gcov intermediate format.
5694 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5697 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5698 (lra_emit_add): Use the functions. Add comment about Y as an
5701 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5704 * collect2.c (collect_atexit): New.
5705 (collect_exit): Delete.
5706 (main): Register collect_atexit with atexit.
5707 (collect_wait): Change collect_exit to exit.
5709 * collect2.h (collect_exit): Delete.
5710 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5712 2013-06-19 Wei Mi <wmi@google.com>
5714 PR rtl-optimization/57518
5715 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5716 if regno is used in paradoxical subreg.
5717 (update_equiv_regs): Check pdx_subregs[regno] before
5718 set a reg to be equivalent with a mem.
5720 2013-06-19 Matthias Klose <doko@ubuntu.com>
5723 * file-find.h (find_a_file): Add a mode parameter.
5724 * file-find.c (find_a_file): Likewise.
5725 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5726 with X_OK for the executables.
5727 * collect2.c (main): Call find_a_file with X_OK.
5729 2013-06-19 Steve Ellcey <sellcey@mips.com>
5732 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5733 Use NEXT_INSN instead of next_real_insn.
5735 2013-06-19 Jan Hubicka <jh@suse.cz>
5737 * cgraph.h (const_value_known_p): Replace by ...
5738 (ctor_for_folding): .. this one.
5739 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5740 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5741 * expr.c (expand_expr_real_1): Likewise.
5742 (string_constant): Likewise.
5743 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5744 * ipa.c (process_references): Likewise.
5745 (symtab_remove_unreachable_nodes): Likewise.
5746 * ipa-inline-analysis.c (param_change_prob): Likewise.
5747 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5748 (get_base_constructor): Likwise.
5749 * varpool.c (varpool_remove_node): Likewise.
5750 (varpool_remove_initializer): LIkewise.
5751 (dump_varpool_node): LIkwise.
5752 (const_value_known_p): Rewrite to ...
5753 (ctor_for_folding): ... this one.
5755 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5758 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5759 PERSONALITY in $PATH derived prefixes.
5761 2013-06-19 Jeff Law <law@redhat.com>
5763 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5766 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5767 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5770 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5772 * config/arm/vfp.md: Move VFP instruction classification documentation
5774 * config/arm/arm.md: ... here. Update instruction classification
5777 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5779 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5780 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5781 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5783 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5786 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5787 with NEXT_INSN. Use tablejump_p to check for jump table data
5790 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5793 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5794 that now in C++ the value is correct per the C++ standards.
5796 2013-06-19 Richard Biener <rguenther@suse.de>
5798 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5801 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5804 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5807 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5808 with next_active_insn.
5810 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5812 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5813 functions are inlined during failures to flag an error.
5814 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5815 in early inline pass.
5817 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5819 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5822 2013-06-18 Julian Brown <julian@codesourcery.com>
5824 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5825 Permit virtual register pre-reload if !strict.
5826 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5828 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5830 * config/arm/neon.md (movmisalign<mode>): Use
5831 neon_perm_struct_or_reg_operand instead of
5832 neon_struct_or_register_operand.
5833 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5834 neon_permissive_struct_operand instead of neon_struct_operand.
5835 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5836 neon_vector_mem_operand.
5837 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5838 neon_vector_mem_operand.
5839 (neon_permissive_struct_operand): New.
5840 (neon_struct_or_register_operand): Rename to...
5841 (neon_perm_struct_or_reg_operand): This. Adjust call to
5842 neon_vector_mem_operand.
5844 2013-06-18 Richard Biener <rguenther@suse.de>
5846 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5847 * lto-streamer.h: Include pointer-set.h.
5848 (struct lto_decl_slot): Remove.
5849 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5850 Remove next_index entry.
5851 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5852 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5853 (lto_init_tree_ref_encoder): Adjust.
5854 (lto_destroy_tree_ref_encoder): Likewise.
5855 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5856 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5857 (lto_output_decl_index): Adjust.
5858 (lto_new_out_decl_state): Likewise.
5859 (lto_record_function_out_decl_state): Likewise.
5860 * lto-streamer-out.c (copy_function): Likewise.
5862 2013-06-18 Richard Biener <rguenther@suse.de>
5864 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5865 * cgraphunit.c: Include cfgloop.h.
5866 (init_lowered_empty_function): Initialize the loop tree.
5867 (assemble_thunk): Insert new BBs into loops.
5869 2013-06-18 Richard Biener <rguenther@suse.de>
5871 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5872 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5873 the map from cache entry to cache index optional.
5874 (streamer_tree_cache_replace_tree): Adjust accordingly.
5875 (streamer_tree_cache_append): Likewise.
5876 (streamer_tree_cache_delete): Likewise.
5877 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5878 streamer cache map from cache entry to cache index.
5879 * lto-streamer-out.c (create_output_block): Adjust.
5881 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5883 * config/arm/arm.md (attribute "insn"): Move multiplication and
5884 division attributes to...
5885 (attribute "type"): ... here. Remove mult.
5886 (attribute "mul32"): New attribute.
5887 (attribute "mul64"): Add umaal.
5888 (*arm_mulsi3): Update attributes.
5889 (*arm_mulsi3_v6): Likewise.
5890 (*thumb_mulsi3): Likewise.
5891 (*thumb_mulsi3_v6): Likewise.
5892 (*mulsi3_compare0): Likewise.
5893 (*mulsi3_compare0_v6): Likewise.
5894 (*mulsi_compare0_scratch): Likewise.
5895 (*mulsi_compare0_scratch_v6): Likewise.
5896 (*mulsi3addsi): Likewise.
5897 (*mulsi3addsi_v6): Likewise.
5898 (*mulsi3addsi_compare0): Likewise.
5899 (*mulsi3addsi_compare0_v6): Likewise.
5900 (*mulsi3addsi_compare0_scratch): Likewise.
5901 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5902 (*mulsi3subsi): Likewise.
5903 (*mulsidi3adddi): Likewise.
5904 (*mulsi3addsi_v6): Likewise.
5905 (*mulsidi3adddi_v6): Likewise.
5906 (*mulsidi3_nov6): Likewise.
5907 (*mulsidi3_v6): Likewise.
5908 (*umulsidi3_nov6): Likewise.
5909 (*umulsidi3_v6): Likewise.
5910 (*umulsidi3adddi): Likewise.
5911 (*umulsidi3adddi_v6): Likewise.
5912 (*smulsi3_highpart_nov6): Likewise.
5913 (*smulsi3_highpart_v6): Likewise.
5914 (*umulsi3_highpart_nov6): Likewise.
5915 (*umulsi3_highpart_v6): Likewise.
5916 (mulhisi3): Likewise.
5917 (*mulhisi3tb): Likewise.
5918 (*mulhisi3bt): Likewise.
5919 (*mulhisi3tt): Likewise.
5920 (maddhisi4): Likewise.
5921 (*maddhisi4tb): Likewise.
5922 (*maddhisi4tt): Likewise.
5923 (maddhidi4): Likewise.
5924 (*maddhidi4tb): Likewise.
5925 (*maddhidi4tt): Likewise.
5927 (udivsi3): Likewise.
5928 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5929 (thumb2_mulsi_short_compare0): Likewise.
5930 (thumb2_mulsi_short_compare0_scratch): Likewise.
5931 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5932 (1020mult2): Likewise.
5933 (1020mult3): Likewise.
5934 (1020mult4): Likewise.
5935 (1020mult5): Likewise.
5936 (1020mult6): Likewise.
5937 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5939 (cortex_a15_mult64): Likewise.
5940 (cortex_a15_sdiv): Likewise.
5941 (cortex_a15_udiv): Likewise.
5942 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5948 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5949 (pj4_ir_div): Likewise.
5950 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5951 (11_mult2): Likewise.
5952 (11_mult3): Likewise.
5953 (11_mult4): Likewise.
5954 (11_mult5): Likewise.
5955 (11_mult6): Likewise.
5956 (11_mult7): Likewise.
5957 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5958 (cortex_a8_mla): Likewise.
5959 (cortex_a8_mull): Likewise.
5960 (cortex_a8_smulwy): Likewise.
5961 (cortex_a8_smlald): Likewise.
5962 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5963 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5964 (cortex_r4_mul_3): Likewise.
5965 (cortex_r4_mla_4): Likewise.
5966 (cortex_r4_mla_3): Likewise.
5967 (cortex_r4_smlald): Likewise.
5968 (cortex_r4_mull): Likewise.
5969 (cortex_r4_sdiv): Likewise.
5970 (cortex_r4_udiv): Likewise.
5971 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5972 (cortex_a7_idiv): Likewise.
5973 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5974 (9_mult2): Likewise.
5975 (9_mult3): Likewise.
5976 (9_mult4): Likewise.
5977 (9_mult5): Likewise.
5978 (9_mult6): Likewise.
5979 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5980 (cortex_a53_sdiv): Likewise.
5981 (cortex_a53_udiv): Likewise.
5982 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5983 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5984 (mp626_mult2): Likewise.
5985 (mp626_mult3): Likewise.
5986 (mp626_mult4): Likewise.
5987 * config/arm/fa526.md (526_mult1): Update attribute change.
5988 (526_mult2): Likewise.
5989 * config/arm/arm-generic.md (mult): Update attribute change.
5990 (mult_ldsched_strongarm): Likewise.
5991 (mult_ldsched): Likewise.
5992 (multi_cycle): Likewise.
5993 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5994 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5995 (606te_mult2): Likewise.
5996 (606te_mult3): Likewise.
5997 (606te_mult4): Likewise.
5998 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5999 (cortex_a9_mac16): Likewise.
6000 (cortex_a9_multiply): Likewise.
6001 (cortex_a9_mac): Likewise.
6002 (cortex_a9_multiply_long): Likewise.
6003 * config/arm/fa626te.md (626te_mult1): Update attribute change.
6004 (626te_mult2): Likewise.
6005 (626te_mult3): Likewise.
6006 (626te_mult4): Likewise.
6008 2013-06-18 Richard Biener <rguenther@suse.de>
6011 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
6013 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6016 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6017 with next_active_insn.
6019 2013-06-18 Alan Modra <amodra@gmail.com>
6021 * config/rs6000/rs6000.h (enum data_align): New.
6022 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
6023 (DATA_ABI_ALIGNMENT): Define.
6024 (CONSTANT_ALIGNMENT): Correct comment.
6025 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
6026 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
6028 2013-06-17 David Malcolm <dmalcolm@redhat.com>
6030 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
6031 ATTRIBUTE_UNUSED marking.
6033 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6035 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6036 alternative and update.
6037 (aarch64_dup_lanedi): Delete.
6038 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6039 * config/aarch64/aarch64-simd-builtins.def: Update.
6041 2013-06-17 Richard Biener <rguenther@suse.de>
6043 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6044 (lto_input_scc): Declare.
6045 (lto_input_tree_1): Likewise.
6046 (struct lto_stats_d): Add num_tree_bodies_output and
6047 num_pickle_refs_output.
6048 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6049 (lto_read_tree_1): Split out from ...
6050 (lto_read_tree): ... this.
6051 (lto_input_scc): New function.
6052 (lto_input_tree_1): Split out from ...
6053 (lto_input_tree): ... this. Handle LTO_tree_scc.
6054 (lto_data_in_create): Create the streamer cache without hashes.
6055 * lto-streamer-out.c (create_output_block): Create the streamer
6056 cache with hashes when not doing WPA.
6057 (lto_write_tree_1): Split out from ...
6058 (lto_write_tree): ... this.
6059 (get_symbol_initial_value): New function.
6060 (lto_output_tree_1): Split out from ...
6061 (lto_output_tree): ... this. Write trees as series of SCCs
6062 using a DFS walk via DFS_write_tree.
6063 (struct sccs, struct scc_entry): New types.
6064 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6065 (DFS_write_tree_body): New function.
6066 (DFS_write_tree): Likewise.
6067 (hash_tree): Likewise.
6068 (scc_entry_compare): Likewise.
6069 (hash_scc): Likewise.
6070 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6071 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6072 TREE_CHAIN as regular reference.
6073 (streamer_read_integer_cst): Remove.
6074 (streamer_get_pickled_tree): Adjust.
6075 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6076 of DECL_EXTERNALs in BLOCK_VARS for now.
6077 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6079 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6080 Add hash value argument and record that if hashes are recorded
6082 (streamer_tree_cache_insert_1): Adjust.
6083 (streamer_tree_cache_insert): Likewise.
6084 (streamer_tree_cache_insert_at): Rename to ...
6085 (streamer_tree_cache_replace_tree): ... this and adjust.
6086 (streamer_tree_cache_append): Adjust.
6087 (record_common_node): Likewise.
6088 (streamer_tree_cache_create): Add argument whether to
6089 record hash values together with trees.
6090 (streamer_tree_cache_delete): Adjust.
6091 * tree-streamer.h (struct streamer_tree_cache_d): Add
6093 (streamer_read_integer_cst): Remove.
6094 (streamer_tree_cache_insert): Adjust.
6095 (streamer_tree_cache_append): Likewise.
6096 (streamer_tree_cache_insert_at): Rename to ...
6097 (streamer_tree_cache_replace_tree): ... this and adjust.
6098 (streamer_tree_cache_create): Add argument whether to record hashes.
6099 (streamer_tree_cache_get): Rename to ...
6100 (streamer_tree_cache_get_tree): ... this.
6101 (streamer_tree_cache_get_hash): New function.
6102 * tree.c (cache_integer_cst): New function.
6103 * tree.h (cache_integer_cst): Declare.
6104 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6105 * lto-symtab.c (lto_varpool_replace_node): Only release
6106 DECL_INITIAL of non-prevailing decls.
6107 * varpool.c (varpool_remove_initializer): Do not release
6108 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6110 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6112 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6113 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6114 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6115 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6116 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6117 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6118 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6119 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6120 instead of TARGET_64BIT.
6121 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6122 Require ISA_HAS_<D>DIV.
6124 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6126 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6127 (mips*-*-linux*): Move default with_llsc setting to where other
6129 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6131 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6132 Likewise. Remove default with_tune setting. Move default float
6133 setting to its own block. Handle with_llsc in the same block as above.
6135 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6137 PR rtl-optimization/57425
6138 PR rtl-optimization/57569
6139 * alias.c (write_dependence_p): Add new parameters mem_mode,
6140 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6141 Changed all callers.
6142 (canon_anti_dependence): New function.
6143 * cse.c (check_dependence): Use canon_anti_dependence.
6144 * cselib.c (cselib_invalidate_mem): Likewise.
6145 * rtl.h (canon_anti_dependence): Declare.
6147 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6149 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6150 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6151 ".set mips3" for 64-bit targets.
6153 2013-06-15 Dehao Chen <dehao@google.com>
6155 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6156 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6157 (gimple_check_call_args): Likewise.
6158 * value-prof.c (check_ic_target): Likewise.
6159 * ipa-inline.c (early_inliner): Likewise.
6160 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6161 * cgraph.c (cgraph_create_edge_1): Likewise.
6162 (cgraph_make_edge_direct): Likewise.
6164 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6167 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6168 rs6000_output_move_128bit to handle emitting quad memory
6169 operations. Set attribute length to 8 bytes.
6171 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6173 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6175 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6176 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6177 (aarch64_<su>mlsl<mode>): Likewise.
6179 2013-06-14 Mike Stump <mikestump@comcast.net>
6181 * Makefile.in (TARGET_H): Add insn-codes.h.
6183 2013-06-14 Alan Modra <amodra@gmail.com>
6187 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6188 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6189 bitfield expansion when EXPAND_MEMORY.
6190 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6192 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6194 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6195 test for clearing quad memory on 32-bit later.
6197 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6199 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6200 (fold_negate_expr): Likewise.
6201 (fold_real_zero_addition_p): Handle vectors.
6202 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6204 2013-06-14 Alan Modra <amodra@gmail.com>
6206 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6208 2013-06-13 Jan Hubicka <jh@suse.cz>
6210 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6211 Local comdats are not externally visible.
6212 * symtab.c (dump_symtab_base): Dump externally visible.
6213 (verify_symtab_base): Verify back links in the symtab hash.
6215 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6217 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6218 CONVERT_EXPR as equal nodes.
6220 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6222 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6224 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6226 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6227 Generalize to complex and vector.
6228 * tree.c (build_all_ones_cst): New function.
6229 * tree.h (build_all_ones_cst): Declare it.
6231 2013-06-13 Alan Modra <amodra@gmail.com>
6233 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6234 * config/rs6000/rs6000.md (signbittf2): New insn.
6235 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6236 (abstf2_internal, cmptf_internal2): Likewise.
6237 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6239 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6240 Pat Haugen <pthaugen@us.ibm.com>
6241 Peter Bergner <bergner@vnet.ibm.com>
6243 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6244 power8 byte, half-word, and quad-word atomic instructions.
6245 (emit_store_conditional): Likewise.
6246 (rs6000_expand_atomic_compare_and_swap): Likewise.
6247 (rs6000_expand_atomic_op): Likewise.
6249 * config/rs6000/sync.md (larx): Add new modes for power8.
6251 (AINT): New mode iterator to include TImode as well as normal
6252 integer modes on power8.
6253 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6254 that VSX registers are not considered. Use AINT mode iterator
6255 instead of INT1 to allow inclusion of quad word atomic operations
6257 (load_locked<mode>): Likewise.
6258 (store_conditional<mode>): Likewise.
6259 (atomic_compare_and_swap<mode>): Likewise.
6260 (atomic_exchange<mode>): Likewise.
6261 (atomic_nand<mode>): Likewise.
6262 (atomic_fetch_<fetchop_name><mode>): Likewise.
6263 (atomic_nand_fetch<mode>): Likewise.
6264 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
6266 (ATOMIC): On power8, add QImode, HImode modes.
6267 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
6268 modes that promote to SImode.
6269 (load_lockedti): Convert TImode arguments to PTImode, so that we
6270 get a guaranteed even/odd register pair.
6271 (load_lockedpti): Likewise.
6272 (store_conditionalti): Likewise.
6273 (store_conditionalpti): Likewise.
6275 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
6276 atomic load/store instructions.
6279 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
6281 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
6283 (insn_count): New attribute, with most cases extracted from...
6284 (length): ...here. Redefine most cases in terms of insn_count.
6285 (single_insn): Delete.
6286 (can_delay): Use insn_count to check for single instructions.
6287 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
6288 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
6289 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
6290 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
6291 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
6292 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
6293 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
6294 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
6295 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
6296 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
6297 rather than "length".
6298 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
6299 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
6300 Use "insn_count" rather than "length".
6301 * config/mips/mips-dsp.md
6302 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
6303 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
6306 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
6308 PR tree-optimization/57361
6309 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
6311 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
6313 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
6315 (aarch64_simd_combine<mode>): New instruction expansion.
6316 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
6318 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
6319 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
6321 2013-06-12 Jan Hubicka <jh@suse.cz>
6323 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
6324 decl has when in streaming stage.
6325 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
6326 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
6328 2013-06-12 Roland Stigge <stigge@antcom.de>
6331 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
6333 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6335 PR tree-optimization/57537
6336 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
6337 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
6339 2013-06-12 Richard Biener <rguenther@suse.de>
6341 * data-streamer.h (streamer_write_char_stream): CSE
6342 obs->current_pointer.
6343 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
6344 streamer_write_char_stream manually and optimize the resulting loop.
6345 (streamer_write_hwi_stream): Likewise.
6347 2013-06-12 Jan Hubicka <jh@suse.cz>
6349 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
6350 * cgraph.h (varpool_create_empty_node): Declare.
6351 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
6353 * symtab.c (symtab_unregister_node): Be lax about missin entries
6355 (symtab_get_node): Update comment.
6356 * varpool.c (varpool_create_empty_node): Break out from ...
6357 (varpool_node_for_decl): ... here.
6358 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
6360 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
6362 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
6363 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
6364 part. Use straight-line flow at the end.
6365 <COMPONENT_REF>: Remove superfluous else.
6366 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
6368 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6371 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
6372 target hook even for !TREE_PUBLIC decls. If no resolution info
6373 is available, return false for common and external decls.
6375 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6377 * config/rl78/constraints.md (U): New constraint.
6378 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
6381 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6384 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
6385 to allow returning address to AT_PLATFORM name.
6387 2013-06-11 Jan Hubicka <jh@suse.cz>
6389 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
6390 * cgraph.h (symtab_node_base): Add weakref flag.
6391 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
6392 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
6393 (output_weakrefs): Use weakref flag.
6394 * fold-const.c (simple_operand_p): Handle WEAK.
6395 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
6396 * ipa.c (varpool_externally_visible_p): Drop weakref.
6397 (function_and_variable_visibility): Update comment; fix weakref
6398 sanity checks; do not clear DECL_WEAK on them.
6399 * lto-cgraph.c (lto_output_node): update.
6400 (lto_output_varpool_node): Update.
6401 (input_overwrite_node): Update.
6402 (input_node): Update.
6403 (input_varpool_node): Update.
6404 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
6405 (lto_symtab_merge_symbols): Add sanity check.
6406 (lto_symtab_prevailing_decl): Do not special case weakrefs.
6407 * passes.c (rest_of_decl_compilation): Set static flag, too.
6408 * symtab.c (dump_symtab_base): Dump weakref.
6409 (verify_symtab_base): Sanity check weakrefs.
6410 (symtab_make_decl_local): Remove duplicated code.
6411 (symtab_alias_ultimate_target): Simplify.
6412 * varpool.c (varpool_create_variable_alias): Set weakref flag.
6414 2013-06-11 DJ Delorie <dj@redhat.com>
6416 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
6417 (rl78_unwind_word_mode): New.
6419 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6421 * final.c (debug_prefix_maps): Make static.
6423 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6425 * function.c (initial_trampoline): Remove stray copy.
6427 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
6429 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
6431 2013-06-11 Martin Jambor <mjambor@suse.cz>
6433 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
6434 within bounds at the beginning of the function.
6436 2013-06-11 Alan Modra <amodra@gmail.com>
6438 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
6440 (get_named_section): Don't NULL !DECL_P decl.
6442 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
6444 * doc/invoke.texi (core-avx2): Document.
6446 (atom): Updated with MOVBE.
6448 2013-06-11 Richard Biener <rguenther@suse.de>
6450 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
6452 2013-06-11 Anton Blanchard <anton@samba.org>
6454 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
6455 correct shift value in little-endian mode.
6457 2013-06-11 Jakub Jelinek <jakub@redhat.com>
6460 * varasm.c (get_variable_align): Move #endif to the right place.
6462 2013-06-10 Cary Coutant <ccoutant@google.com>
6464 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
6465 for hash so that hash table traversal order is deterministic.
6467 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
6468 Pat Haugen <pthaugen@us.ibm.com>
6469 Peter Bergner <bergner@vnet.ibm.com>
6471 * config/rs6000/vector.md (GPR move splitter): Do not split moves
6472 of vectors in GPRS if they are direct moves or quad word load or
6475 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
6477 (direct_move_p): Likewise.
6478 (quad_load_store_p): Likewise.
6480 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
6481 classes into bins based on the physical register type.
6482 (reg_class_to_reg_type): Likewise.
6483 (IS_STD_REG_TYPE): Likewise.
6484 (IS_FP_VECT_REG_TYPE): Likewise.
6485 (reload_fpr_gpr): Arrays to determine what insn to use if we can
6486 use direct move instructions.
6487 (reload_gpr_vsx): Likewise.
6488 (reload_vsx_gpr): Likewise.
6489 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
6490 information that is a simplification of register classes. Also
6491 precalculate direct move reload helpers.
6492 (direct_move_p): New function to return true if the operation can
6493 be done as a direct move instruciton.
6494 (quad_load_store_p): New function to return true if the operation
6495 is a quad memory operation.
6496 (rs6000_legitimize_address): If quad memory, only allow register
6497 indirect for TImode addresses.
6498 (rs6000_legitimate_address_p): Likewise.
6499 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
6500 (rs6000_reload_register_type): Likewise.
6501 (register_to_reg_type): Return register type.
6502 (rs6000_secondary_reload_simple_move): New helper function for
6503 secondary reload and secondary memory needed to identify anything
6504 that is a simple move, and does not need reloading.
6505 (rs6000_secondary_reload_direct_move): New helper function for
6506 secondary reload to identify cases that can be done with several
6507 instructions via the direct move instructions.
6508 (rs6000_secondary_reload_move): New helper function for secondary
6509 reload to identify moves between register types that can be done.
6510 (rs6000_secondary_reload): Add support for quad memory operations
6511 and for direct move.
6512 (rs6000_secondary_memory_needed): Likewise.
6513 (rs6000_debug_secondary_memory_needed): Change argument names.
6514 (rs6000_output_move_128bit): New function to return the move to
6515 use for 128-bit moves, including knowing about the various
6516 limitations of quad memory operations.
6518 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
6519 memory operations. call rs6000_output_move_128bit for the actual
6520 instruciton(s) to generate.
6521 (vsx_movti_64bit): Likewise.
6523 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
6524 (UNSPEC_P8V_MTVSRWZ): Likewise.
6525 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6526 (UNSPEC_P8V_MTVSRD): Likewise.
6527 (UNSPEC_P8V_XXPERMDI): Likewise.
6528 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6529 (UNSPEC_FUSION_GPR): Likewise.
6530 (FMOVE128_GPR): New iterator for direct move.
6531 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6534 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6535 loads and direct move instructions.
6536 (zero_extendsidi2_lfiwzx): Likewise.
6537 (extendsidi2_lfiwax): Likewise.
6538 (extendsidi2_nocell): Likewise.
6539 (floatsi<mode>2_lfiwax): Likewise.
6541 (floatunssi<mode>2_lfiwzx): Likewise.
6543 (fix_trunc<mode>_stfiwx): Likewise.
6544 (fixuns_trunc<mode>_stfiwx): Likewise.
6545 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6546 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6547 (parity<mode>2_cmpb): Set length/type attr.
6548 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6549 for 'mr.' to fast_compare.
6550 (bpermd_<mode>): Change type attr to popcnt.
6551 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6552 (p8_mtvsrwz_1): Likewise.
6553 (p8_mtvsrwz_2): Likewise.
6554 (reload_fpr_from_gpr<mode>): Likewise.
6555 (p8_mtvsrd_1): Likewise.
6556 (p8_mtvsrd_2): Likewise.
6557 (p8_xxpermdi_<mode>): Likewise.
6558 (reload_vsx_from_gpr<mode>): Likewise.
6559 (reload_vsx_from_gprsf): Likewise.
6560 (p8_mfvsrd_3_<mode>): LIkewise.
6561 (reload_gpr_from_vsx<mode>): Likewise.
6562 (reload_gpr_from_vsxsf): Likewise.
6563 (p8_mfvsrd_4_disf): Likewise.
6564 (multi-word GPR splits): Do not split direct moves or quad memory
6567 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6569 * tree-into-ssa.c (interesting_blocks): Make static.
6571 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6574 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6575 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6576 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6577 (get_variable_align): New function.
6578 (get_variable_section, emit_bss, emit_common,
6579 assemble_variable_contents, place_block_symbol): Use
6580 get_variable_align instead of DECL_ALIGN.
6581 (assemble_noswitch_variable): Add align argument, use it
6582 instead of DECL_ALIGN.
6583 (assemble_variable): Adjust caller. Use get_variable_align
6584 instead of DECL_ALIGN.
6585 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6587 (DATA_ABI_ALIGNMENT): Define.
6588 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6589 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6590 opt is false, only return the psABI mandated alignment increase.
6591 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6592 (DATA_ABI_ALIGNMENT): ... this.
6593 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6594 (DATA_ABI_ALIGNMENT): ... this.
6595 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6596 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6597 (DATA_ABI_ALIGNMENT): ... this.
6598 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6599 * doc/tm.texi: Regenerated.
6601 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6603 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6604 cmp_code to construct REG_EQUAL note.
6606 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6609 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6610 that operands[2] doesn't overlap with operands[0].
6612 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6613 Jan Hubicka <jh@suse.cz>
6615 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6616 hack to mark symbols as used.
6618 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6620 PR rtl-optimization/57559
6621 * lra-constraints.c (process_alt_operands): Don't discourage
6622 memory with known offset for offsetable memory constraint.
6623 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6625 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6627 * varasm.c (struct oc_local_state): Reorder fields.
6628 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6629 and adjust accordingly.
6630 (output_constructor): Reorder initialization code and adjust call to
6631 output_constructor_bitfield.
6633 2013-06-07 Jan Hubicka <jh@suse.cz>
6635 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6637 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6639 * tree-object-size.c (unknown): Make const.
6641 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6643 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6644 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6645 for last alternative in the cpu_facility attribute.
6647 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6650 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6651 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6652 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6653 * config/arm/constraints.md (Dg): New constraint.
6654 * config/arm/neon.md (xordi3_neon): Remove.
6655 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6656 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6658 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6660 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6661 Clean up alternatives.
6663 2013-06-07 Alan Modra <amodra@gmail.com>
6665 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6668 2013-06-07 Alan Modra <amodra@gmail.com>
6670 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6672 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6674 * config/arm/constraints.md (Df): New constraint.
6675 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6676 Correct length attribute for last two alternatives.
6678 2013-06-07 Alan Modra <amodra@gmail.com>
6680 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6681 override user -mfp-in-toc.
6682 (offsettable_ok_by_alignment): Consider just the current access
6683 rather than the whole object, unless BLKmode. Handle
6684 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6685 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6686 for -mcmodel=medium.
6687 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6688 override user -mfp-in-toc or -msum-in-toc. Default to
6689 -mno-fp-in-toc for -mcmodel=medium.
6691 2013-06-06 DJ Delorie <dj@redhat.com>
6693 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6694 TARGET_VALID_POINTER_MODE.
6696 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6697 Pat Haugen <pthaugen@us.ibm.com>
6698 Peter Bergner <bergner@vnet.ibm.com>
6700 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6701 Document new power8 builtins.
6703 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6704 condition code register, to allow 128-bit logical operations to be
6705 done in the VSX or GPR registers.
6706 (nor<mode>3): Use the canonical form for nor.
6707 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6708 vclz*, and vpopcnt* vector instructions.
6709 (nand<mode>3): Likewise.
6710 (orc<mode>3): Likewise.
6711 (clz<mode>2): LIkewise.
6712 (popcount<mode>2): Likewise.
6714 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6715 that only the GPRs are recognized.
6717 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6718 support for new power8 builtins.
6720 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6722 (xscvdpspn): Likewise.
6728 (vpopcnt): Likewise.
6729 (vpopcntb): Likewise.
6730 (vpopcnth): Likewise.
6731 (vpopcntw): Likewise.
6732 (vpopcntd): Likewise.
6737 (eqv_v16qi3): Likewise.
6738 (eqv_v8hi3): Likewise.
6739 (eqv_v4si3): Likewise.
6740 (eqv_v2di3): Likewise.
6741 (eqv_v4sf3): Likewise.
6742 (eqv_v2df3): Likewise.
6744 (nand_v16qi3): Likewise.
6745 (nand_v8hi3): Likewise.
6746 (nand_v4si3): Likewise.
6747 (nand_v2di3): Likewise.
6748 (nand_v4sf3): Likewise.
6749 (nand_v2df3): Likewise.
6751 (orc_v16qi3): Likewise.
6752 (orc_v8hi3): Likewise.
6753 (orc_v4si3): Likewise.
6754 (orc_v2di3): Likewise.
6755 (orc_v4sf3): Likewise.
6756 (orc_v2df3): Likewise.
6758 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6759 allow power8 quad mode in 64-bit.
6760 (rs6000_builtin_vectorized_function): Add support to vectorize
6761 ISA 2.07 count leading zeros, population count builtins.
6762 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6763 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6764 (builtin_function_type): Add vgbbd builtin function which takes an
6766 (altivec_expand_vec_perm_const): Add support for new power8 merge
6769 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6770 that does not include TImdoe for use with 32-bit.
6771 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6773 (UNSPEC_VSX_CVDPSPN): Likewise.
6774 (vsx_xscvdpspn): Likewise.
6775 (vsx_xscvspdpn): Likewise.
6776 (vsx_xscvdpspn_scalar): Likewise.
6777 (vsx_xscvspdpn_directmove): Likewise.
6778 (vsx_and<mode>3): Split logical operations into 32-bit and
6779 64-bit. Add support to do logical operations on TImode as well as
6780 VSX vector types. Allow logical operations to be done in either
6781 VSX registers or in general purpose registers in 64-bit mode. Add
6782 splitters if GPRs were used. For AND, add clobber of CCmode to
6783 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6785 (vsx_and<mode>3_32bit): Likewise.
6786 (vsx_and<mode>3_64bit): Likewise.
6787 (vsx_ior<mode>3): Likewise.
6788 (vsx_ior<mode>3_32bit): Likewise.
6789 (vsx_ior<mode>3_64bit): Likewise.
6790 (vsx_xor<mode>3): Likewise.
6791 (vsx_xor<mode>3_32bit): Likewise.
6792 (vsx_xor<mode>3_64bit): Likewise.
6793 (vsx_one_cmpl<mode>2): Likewise.
6794 (vsx_one_cmpl<mode>2_32bit): Likewise.
6795 (vsx_one_cmpl<mode>2_64bit): Likewise.
6796 (vsx_nor<mode>3): Likewise.
6797 (vsx_nor<mode>3_32bit): Likewise.
6798 (vsx_nor<mode>3_64bit): Likewise.
6799 (vsx_andc<mode>3): Likewise.
6800 (vsx_andc<mode>3_32bit): Likewise.
6801 (vsx_andc<mode>3_64bit): Likewise.
6802 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6803 and xxlorc instructions.
6804 (vsx_eqv<mode>3_64bit): Likewise.
6805 (vsx_nand<mode>3_32bit): Likewise.
6806 (vsx_nand<mode>3_64bit): Likewise.
6807 (vsx_orc<mode>3_32bit): Likewise.
6808 (vsx_orc<mode>3_64bit): Likewise.
6810 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6812 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6814 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6815 (p8_vmrgow): Likewise.
6816 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6817 GPRs to be split under VSX.
6818 (p8v_clz<mode>2): Add power8 count leading zero support.
6819 (p8v_popcount<mode>2): Add power8 population count support.
6820 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6823 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6826 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6828 (vec_nand): Likewise.
6829 (vec_vclz): Likewise.
6830 (vec_vclzb): Likewise.
6831 (vec_vclzd): Likewise.
6832 (vec_vclzh): Likewise.
6833 (vec_vclzw): Likewise.
6834 (vec_vgbbd): Likewise.
6835 (vec_vmrgew): Likewise.
6836 (vec_vmrgow): Likewise.
6837 (vec_vpopcnt): Likewise.
6838 (vec_vpopcntb): Likewise.
6839 (vec_vpopcntd): Likewise.
6840 (vec_vpopcnth): Likewise.
6841 (vec_vpopcntw): Likewise.
6843 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6845 PR rtl-optimization/57468
6846 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6849 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6851 PR rtl-optimization/57459
6852 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6853 type when setting live regs.
6855 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6857 * config/s390/s390.opt (mlra): New option.
6858 * config/s390/s390.c (s390_decompose_address): Check displacement
6859 for all registers for LRA.
6860 (s390_secondary_reload): Don't used secondary reloads for LRA.
6861 (s390_lra_p): New function.
6862 (TARGET_LRA_P): Define.
6863 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6864 of attribute cpu_facility to zarch for the last alternative.
6865 (*cmpmem_short): Ditto.
6867 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6869 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6870 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6871 (arm_expand_prologue): Likewise.
6873 2013-06-06 Teresa Johnson <tejohnson@google.com>
6876 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6877 as this is now done by redirect_edge_and_branch_force.
6878 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6879 barriers, and fix interaction with splitting.
6880 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6881 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6882 reflect changes made in the routine.
6883 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6884 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6885 since this is called in cfglayout mode, and replace partition fixup
6886 with assert as that is now done by force_nonfallthru_and_redirect.
6887 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6888 already be marked with region crossing note.
6889 (insert_section_boundary_note): Make non-static, gate on flag
6890 has_bb_partition, rewrite to also check for multiple partitions.
6891 (rest_of_handle_reorder_blocks): Remove call to
6892 insert_section_boundary_note, now done later during free_cfg.
6893 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6894 * bb-reorder.h (insert_section_boundary_note): Declare.
6895 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6896 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6897 invoke insert_section_boundary_note.
6898 (try_redirect_by_replacing_jump): Remove unnecessary
6899 check for region crossing note.
6900 (fixup_partition_crossing): New function.
6901 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6902 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6903 in non-cfglayout mode.
6904 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6905 remove old code that tried to do this. Emit barrier correctly
6906 when we are in cfglayout mode.
6907 (last_bb_in_partition): New function.
6908 (rtl_split_edge): Correctly fixup partition boundaries.
6909 (commit_one_edge_insertion): Remove old code that tried to
6910 fixup region crossing edge since this is now handled in
6911 split_block, and set up insertion point correctly since
6912 block may now end in a jump.
6913 (verify_hot_cold_block_grouping): Guard against checking when not in
6914 linearized RTL mode.
6915 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6917 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6918 rtl_verify_flow_info, so not called in cfglayout mode.
6919 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6920 (fixup_reorder_chain): Remove old code that attempted to fixup region
6921 crossing note as this is now handled in force_nonfallthru_and_redirect.
6922 (duplicate_insn_chain): Don't duplicate switch section notes.
6923 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6925 * basic-block.h (emit_barrier_after_bb): Declare.
6927 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6929 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6930 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6931 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6933 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6935 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6937 * config/arm/ldmstm.md: Regenerate.
6939 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6941 * config/arm/sync.md (atomic_loaddi_1):
6942 Disable predication for arm_restrict_it.
6943 (arm_load_exclusive<mode>): Likewise.
6944 (arm_load_exclusivesi): Likewise.
6945 (arm_load_exclusivedi): Likewise.
6946 (arm_load_acquire_exclusive<mode>): Likewise.
6947 (arm_load_acquire_exclusivesi): Likewise.
6948 (arm_load_acquire_exclusivedi): Likewise.
6949 (arm_store_exclusive<mode>): Likewise.
6950 (arm_store_exclusive<mode>): Likewise.
6951 (arm_store_release_exclusivedi): Likewise.
6952 (arm_store_release_exclusive<mode>): Likewise.
6954 2013-06-06 Richard Biener <rguenther@suse.de>
6956 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6958 (lto_tag_is_tree_code_p): Adjust.
6959 (lto_tag_is_gimple_code_p): Likewise.
6960 (lto_gimple_code_to_tag): Likewise.
6961 (lto_tag_to_gimple_code): Likewise.
6962 (lto_tree_code_to_tag): Likewise.
6963 (lto_tag_to_tree_code): Likewise.
6964 * data-streamer.h (streamer_write_hwi_in_range): Use
6965 uhwi streaming to stream the normalized range.
6966 (streamer_read_hwi_in_range): Likewise.
6968 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6970 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6971 (predicable_short_it): Likewise.
6972 (predicated): Likewise.
6973 (enabled): Handle above.
6974 (define_cond_exec): Set predicated attribute to yes.
6976 2013-06-05 Mike Stump <mikestump@comcast.net>
6978 * gdbinit.in (__FUNCTION__): Add.
6980 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6982 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6983 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6985 2013-06-05 Jan Hubicka <jh@suse.cz>
6987 * varasm.c (mark_decl_referenced): Revert the removal until targets
6990 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6992 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6993 instead of mark_decl_referenced.
6995 2013-06-05 Jan Hubicka <jh@suse.cz>
6997 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6998 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6999 and symtab_used_from_object_file_p.
7000 (cgraph_make_node_local_1): Clear forced_by_abi.
7001 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
7002 * cgraph.h (symtab_node_base): Add forced_by_abi.
7003 (decide_is_variable_needed): Remove.
7004 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
7005 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
7006 (decide_is_symbol_needed): ... this one; handle symbols in general;
7007 always analyze virtuals; honnor forced_by_abi.
7008 (cgraph_finalize_function): Update.
7009 (varpool_finalize_decl): Update.
7010 (symbol_defined_and_needed): Remove.
7011 (analyze_functions): Update.
7012 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
7013 output_refs, input_overwrite_node): Handle forced_by_abi.
7014 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
7015 (address_taken_from_non_vtable_p): ... this one.
7016 (comdat_can_be_unshared_p_1): New function.
7017 (cgraph_comdat_can_be_unshared_p): Rename to ...
7018 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
7019 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
7020 (function_and_variable_visibility): Clear forced_by_abi as needed.
7021 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
7022 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
7023 * symtab.c (dump_symtab_base): Dump forced_by_abi.
7024 * varpool.c (decide_is_variable_needed): Remove.
7026 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7028 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
7029 (arm_option_override): Override arm_restrict_it where appropriate.
7030 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
7031 * config/arm/arm.opt (mrestrict-it): New command-line option.
7032 * doc/invoke.texi: Document -mrestrict-it.
7034 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7036 * tsan.c (tsan_atomic_table): Make const.
7038 2013-06-05 Richard Biener <rguenther@suse.de>
7040 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7041 index associated with the tree we are supposed to replace.
7042 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7043 TREE_ASM_WRITTEN as zero for everything but SSA names.
7045 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7047 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7049 2013-06-04 Jan Hubicka <jh@suse.cz>
7051 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7052 (input_node, input_varpool_node): Handle correctly external same
7054 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7055 nodes at ltrans stage.
7057 2013-06-04 Jan Hubicka <jh@suse.cz>
7059 * ipa-inline.c (update_caller_keys): Fix availability test.
7060 (update_callee_keys): Likewise.
7061 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7062 to follow ELF standard.
7064 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7066 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7067 (mips64r5900el-*-elf*): New configurations.
7068 * config/mips/mips-cpus.def (r5900): New processor.
7069 * config/mips/mips-tables.opt: Regenerate.
7070 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7071 (mips_issue_rate): Handle PROCESSOR_R5900.
7072 (mips_reorg_process_insns): Force reorder mode for the R5900.
7073 * config/mips/mips.h (TARGET_MIPS5900): Define.
7074 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7076 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7078 * config/mips/mips.md (processor): Add r5900.
7079 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7081 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7083 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7084 into function to generate MOVI instruction.
7085 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7086 (aarch64_preferred_simd_mode): Turn into wrapper.
7087 (aarch64_output_scalar_simd_mov_immediate): New function.
7088 * config/aarch64/aarch64-protos.h: Add prototype for above.
7090 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7092 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7093 element_char member.
7094 (sizetochar): Return signed char.
7095 (aarch64_simd_valid_immediate): Remove elchar and other
7096 unnecessary variables.
7097 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7098 Calculate element_char as required.
7099 * config/aarch64/aarch64-protos.h: Update and move prototype
7100 for aarch64_output_simd_mov_immediate.
7101 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7104 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7106 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7107 information completed by aarch64_simd_valid_immediate.
7108 (aarch64_legitimate_constant_p): Update arguments.
7109 (aarch64_simd_valid_immediate): Work with struct rather than many
7111 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7112 (aarch64_simd_make_constant): Update arguments.
7113 (aarch64_output_simd_mov_immediate): Work with struct rather than
7114 many pointers. Output immediate directly rather than as operand.
7115 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7117 * config/aarch64/constraints.md (Dn): Update arguments.
7119 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7121 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7123 (aarch64_simd_immediate_valid_for_move): Remove.
7124 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7125 (aarch64_simd_make_constant): Update call.
7126 (aarch64_output_simd_mov_immediate): Update call.
7127 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7129 * config/aarch64/constraints.md (Dn): Update call.
7131 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7133 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7134 return type to bool for prototype.
7135 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7136 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7137 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7139 2013-06-04 Catherine Moore <clm@codesourcery.com>
7141 * config/mips/mips.opt (meva): New.
7142 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7143 (ASM_SPEC): Handle -meva.
7144 * doc/invoke.texi (meva): Document.
7146 2013-06-04 Alan Modra <amodra@gmail.com>
7148 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7151 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7153 * rtl.def: Add extra fourth optional field to define_cond_exec.
7154 * gensupport.c (process_one_cond_exec): Process attributes from
7156 * doc/md.texi: Document fourth field in define_cond_exec.
7158 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7160 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7161 out the processing order as in store_bit_field_1.
7163 2013-06-04 Jan Hubicka <jh@suse.cz>
7166 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7167 non-VAR_DECL node if it does not exist yet.
7169 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7171 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7172 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7173 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7174 target_cpu_default setting.
7176 2013-06-03 Teresa Johnson <tejohnson@google.com>
7178 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7179 default to -fopt-info=optimized instead of all.
7180 * doc/invoke.texi: Ditto.
7181 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7182 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7183 (execute_vect_slp): Emit BB vectorization success under
7184 MSG_OPTIMIZED_LOCATIONS.
7185 * tree-vect-slp.c (vect_slp_transform_bb): Change
7186 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7187 * tree-vect-loop.c (vect_transform_loop): Ditto.
7189 2013-06-03 Jason Merrill <jason@redhat.com>
7192 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7193 Use TARGET_EXPR for C++.
7195 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7197 PR rtl-optimization/57268
7198 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7199 if DEBUG_INSN_P (insn).
7202 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7204 PR rtl-optimization/57268
7205 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7206 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7208 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7210 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7211 (ix86_avoid_lea_for_addr): Likewise.
7212 (exact_dependency_1): Likewise.
7213 (ix86_adjust_cost): Likewise.
7214 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7216 (do_reorder_for_imul): Fix typo, formatting and
7217 !reload_completed check removed.
7218 (ix86_sched_reorder): Fix typo and formatting.
7219 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7222 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7224 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7226 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7228 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7229 <STRING_CST>: Likewise.
7230 <VECTOR_CST>: Likewise.
7232 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7233 Mikael Morin <mikael@gcc.gnu.org>
7235 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7236 * config.in: Regenerated.
7237 * configure: Regenerated.
7239 2013-06-01 Jan Hubicka <jh@suse.cz>
7242 * cgraphunit.c (compile): When weakref is not supported,
7243 set up transparent aliases before final output pass.
7244 * varasm.c (assemble_alias): Do not try to do it here.
7246 2013-06-01 Jan Hubicka <jh@suse.cz>
7249 * passes.c (for_per_function): Skip unanalyzed functions.
7251 2013-06-01 Jan Hubicka <jh@suse.cz>
7253 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7254 (lto_symtab_merge_symbols_1): ... this one.
7255 (lto_symtab_merge_cgraph_nodes): Rename to ...
7256 (lto_symtab_merge_symbols): ... this one; simplify.
7257 * cgraph.c (same_body_aliases_done): Rename to ...
7258 (cpp_implicit_aliases_done): ... this one.
7259 (cgraph_create_function_alias): Update.
7260 (cgraph_same_body_alias): Update.
7261 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
7262 (verify_edge_corresponds_to_fndecl): Simplify.
7263 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
7264 (cgraph_node): Remove same_body_alias.
7265 (varpool_node): Remove alias_of and extra_name_alias.
7266 (same_body_aliases_done): Rename to ..
7267 (cpp_implicit_aliases_done): ... this one.
7268 (symtab_alias_ultimate_target): Add default parameter.
7269 (symtab_resolve_alias): New function.
7270 (fixup_same_cpp_alias_visibility): Declare.
7271 (cgraph_function_node): Add default parameter.
7272 (cgraph_node_asm_name): Likewise.
7273 (cgraph_function_or_thunk_node): Add default parameter; do
7274 not ICE when it is NULL.
7275 (varpool_variable_node): Likewise.
7276 * tree-emutls.c (create_emultls_var): Update.
7277 (ipa_lower_emutls): Update.
7278 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
7279 (cgraph_reset_node): Reset alias info.
7280 (cgraph_finalize_function): Update.
7281 (fixup_same_cpp_alias_visibility): Move to symtab.c.
7282 (analyze_function): Simplify.
7283 (cgraph_process_same_body_aliases): Simplify.
7284 (analyze_functions): Fixup same body aliases.
7285 (handle_alias_pairs): Simplify.
7286 (assemble_thunk): Update.
7287 (assemble_thunks_and_aliases): Update.
7288 (output_weakrefs): Rewrite.
7289 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
7290 (lto_output_varpool_node): Likewise.
7291 (compute_ltrans_boundary): Remve assert.
7292 (get_alias_symbol): New functoin.
7293 (input_node): Rewrite alias handling.
7294 (input_varpool_node): Likewise.
7295 * ipa-pure-const.c (propagate_pure_const): Fix formating.
7296 * ipa.c (process_references): Handle weakrefs correctly.
7297 (symtab_remove_unreachable_nodes): Likewise.
7298 * trans-mem.c (get_cg_data): Update.
7299 (ipa_tm_create_version_alias): Update.
7300 (ipa_tm_execute): Update.
7301 * symtab.c (dump_symtab_base): Dump aliases.
7302 (verify_symtab_base): Verify aliases.
7303 (symtab_node_availability): New function.
7304 (symtab_alias_ultimate_target): Simplify.
7305 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
7306 handle all the fixup cases.
7307 (symtab_resolve_alias): New function.
7308 * passes.c (ipa_write_summaries): Handle weakrefs.
7309 * varpool.c (varpool_analyze_node): Simplify.
7310 (assemble_aliases): Update.
7311 (varpool_create_variable_alias): Simplify.
7312 (varpool_extra_name_alias): Simplify.
7313 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
7314 (lto_symtab_merge_symbols): ... this one.
7316 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
7319 PR rtl-optimization/57268
7320 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7321 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7323 2013-06-01 Tobias Burnus <burnus@net-b.de>
7326 2013-05-31 Tobias Burnus <burnus@net-b.de>
7329 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
7332 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7334 PR rtl-optimization/57268
7335 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7336 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7338 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7340 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
7341 unordered comparison operators when -fno-trapping-math is in effect
7343 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
7344 and implement unordered comparison operators properly on the e500.
7346 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7348 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
7349 for constant scalar integers.
7350 (simplify_relational_operation_1): Likewise.
7352 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
7354 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
7355 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
7358 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7359 Igor Zamyatin <igor.zamyatin@intel.com>
7361 Silvermont (SLM) architecture performance tuning.
7362 * config/i386/i386.h (enum ix86_tune_indices): Add
7363 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
7364 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
7366 * config/i386/i386.c (initial_ix86_tune_features)
7367 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
7368 (ix86_lea_outperforms): Handle Silvermont tuning.
7369 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
7371 (ix86_use_lea_for_mov): Likewise.
7372 (ix86_avoid_lea_for_addr): Likewise.
7373 (ix86_lea_for_add_ok): Likewise.
7374 (exact_dependency_1): New function.
7375 (exact_store_load_dependency): Likewise.
7376 (ix86_adjust_cost): Handle Silvermont tuning.
7377 (do_reoder_for_imul): Likewise.
7378 (swap_top_of_ready_list): New function.
7379 (ix86_sched_reorder): Changed to handle Silvermont tuning.
7381 * config/i386/i386.md (peepholes that split memory operand in fp
7384 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7386 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
7387 Remove un-necessary braces.
7389 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7391 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7392 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
7394 2013-05-31 Tobias Burnus <burnus@net-b.de>
7397 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
7399 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7402 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
7403 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
7404 * config/arm/neon.md (iordi3_neon): Remove.
7405 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
7406 * config/arm/predicates.md (imm_for_neon_logic_operand):
7407 Move to earlier in the file.
7408 (neon_logic_op2): Likewise.
7409 (arm_iordi_operand_neon): New predicate.
7411 2013-05-31 Richard Biener <rguenther@suse.de>
7413 PR tree-optimization/57478
7414 PR tree-optimization/57453
7415 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
7418 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7420 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
7421 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
7423 2013-05-30 Tobias Burnus <burnus@net-b.de>
7424 Thomas Koenig <tkoenig@gcc.gnu.org>
7427 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
7428 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
7430 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
7432 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
7434 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
7436 * target.def (register_usage_leveling_p): New hook.
7437 * targhooks.c (default_register_usage_leveling_p): New.
7438 * targhooks.h (default_register_usage_leveling_p): New prototype.
7439 * lra-assigns.c (register_usage_leveling_p): Use the hook.
7440 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
7441 * doc/tm.texi: Update.
7442 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
7444 2013-05-30 Ian Bolton <ian.bolton@arm.com>
7446 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
7447 (*insv_reg<mode>): New define_insn.
7449 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
7451 PR rtl-optimization/57439
7452 * postreload.c (move2add_valid_value_p): Check that we have
7453 a zero subreg_regno_offset when accessing the register in
7456 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7457 Igor Zamyatin <igor.zamyatin@intel.com>
7459 Silvermont (SLM) architecture pipeline model, tuning and
7461 * config.gcc: Add slm config options and target.
7463 * config/i386/slm.md: New.
7465 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
7467 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
7469 (ix86_target_macros_internal): Likewise.
7471 * gcc/config/i386/i386.c (slm_cost): New cost.
7472 (m_SLM): New macro flag.
7473 (initial_ix86_tune_features): Set m_SLM.
7474 (x86_accumulate_outgoing_args): Likewise.
7475 (x86_arch_always_fancy_math_387): Likewise.
7476 (processor_target_table): Add slm cost.
7477 (cpu_names): Add slm cpu name.
7478 (x86_option_override_internal): Set SLM ISA.
7479 (ix86_issue_rate): New case PROCESSOR_SLM.
7480 (ia32_multipass_dfa_lookahead): Likewise.
7481 (fold_builtin_cpu): Add slm.
7483 * config/i386/i386.h (TARGET_SLM): New target macro.
7484 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
7485 (processor_type): Add PROCESSOR_SLM.
7487 * config/i386/i386.md (cpu): Add new value "slm".
7488 (slm.md): Include slm.md.
7490 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
7491 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7493 * config/arm/arm-protos.h: Add and update function protos.
7494 * config/arm/arm.c (use_simple_return_p): New added.
7495 (thumb2_expand_return): Check simple_return flag.
7496 * config/arm/arm.md: Add simple_return and conditional simple_return.
7497 * config/arm/iterators.md: Add iterator for return and simple_return.
7499 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7501 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
7502 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
7503 (arm_emit_vfp_multi_reg_pop): Likewise.
7504 (thumb2_emit_ldrd_pop): Likewise.
7505 (arm_expand_epilogue): Add misc REG_CFA notes.
7506 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
7508 2013-05-29 Lawrence Crowl <crowl@google.com>
7510 * config/arm/t-arm: Update for below.
7512 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
7513 Change type to hash_table. Update dependent calls and types.
7515 * config/i386/t-cygming: Update for below.
7517 * config/i386/t-interix: Update for below.
7519 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
7520 Change type to hash_table. Update dependent calls and types.
7521 (i386_find_on_wrapper_list::wrappers): Likewise.
7523 * config/ia64/t-ia64: Update for below.
7525 * config/ia64/ia64.c (bundle_state_table):
7526 Change type to hash_table. Update dependent calls and types.
7528 * config/mips/mips.c (mips_reorg_process_insns::htab):
7529 Change type to hash_table. Update dependent calls and types.
7531 * config/sol2.c (solaris_comdat_htab):
7532 Change type to hash_table. Update dependent calls and types.
7534 * config/t-sol2: Update for above.
7536 2013-05-29 Teresa Johnson <tejohnson@google.com>
7538 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7539 functions are not yet marked as defined.
7541 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7542 Pat Haugen <pthaugen@us.ibm.com>
7543 Peter Bergner <bergner@vnet.ibm.com>
7545 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7549 (vrotl<mode>3): Likewise.
7550 (vashl<mode>3): Likewise.
7551 (vlshr<mode>3): Likewise.
7552 (vashr<mode>3): Likewise.
7554 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7555 support for power8 V2DI builtins.
7557 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7558 power8 V2DI builtins.
7559 (vupkhsw): Likewise.
7560 (vupklsw): Likewise.
7561 (vaddudm): Likewise.
7566 (vpkudum): Likewise.
7567 (vpksdss): Likewise.
7568 (vpkudus): Likewise.
7569 (vpksdus): Likewise.
7574 (vsubudm): Likewise.
7575 (vcmpequd): Likewise.
7576 (vcmpgtsd): Likewise.
7577 (vcmpgtud): Likewise.
7578 (vcmpequd_p): Likewise.
7579 (vcmpgtsd_p): Likewise.
7580 (vcmpgtud_p): Likewise.
7581 (vupkhsw): Likewise.
7582 (vupklsw): Likewise.
7583 (vaddudm): Likewise.
7588 (vpksdss): Likewise.
7589 (vpksdus): Likewise.
7590 (vpkudum): Likewise.
7591 (vpkudus): Likewise.
7596 (vsubudm): Likewise.
7598 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7599 support for power8 V2DI instructions.
7601 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7602 power8 V2DI instructions. Combine pack and unpack insns to use an
7603 iterator for each mode. Check whether a particular mode supports
7604 Altivec instructions instead of just checking TARGET_ALTIVEC.
7605 (UNSPEC_VPKUWUM): Likewise.
7606 (UNSPEC_VPKSHSS): Likewise.
7607 (UNSPEC_VPKSWSS): Likewise.
7608 (UNSPEC_VPKUHUS): Likewise.
7609 (UNSPEC_VPKSHUS): Likewise.
7610 (UNSPEC_VPKUWUS): Likewise.
7611 (UNSPEC_VPKSWUS): Likewise.
7612 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7613 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7614 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7615 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7616 (UNSPEC_VUPKHSB): Likewise.
7617 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7618 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7619 (UNSPEC_VUPKHSH): Likewise.
7620 (UNSPEC_VUPKLSB): Likewise.
7621 (UNSPEC_VUPKLSH): Likewise.
7623 (VI_char): Likewise.
7624 (VI_scalar): Likewise.
7625 (VI_unit): Likewise.
7627 (VP_small): Likewise.
7628 (VP_small_lc): Likewise.
7629 (VU_char): Likewise.
7630 (add<mode>3): Likewise.
7631 (altivec_vaddcuw): Likewise.
7632 (altivec_vaddu<VI_char>s): Likewise.
7633 (altivec_vadds<VI_char>s): Likewise.
7634 (sub<mode>3): Likewise.
7635 (altivec_vsubcuw): Likewise.
7636 (altivec_vsubu<VI_char>s): Likewise.
7637 (altivec_vsubs<VI_char>s): Likewise.
7638 (altivec_vavgs<VI_char>): Likewise.
7639 (altivec_vcmpbfp): Likewise.
7640 (altivec_eq<mode>): Likewise.
7641 (altivec_gt<mode>): Likewise.
7642 (altivec_gtu<mode>): Likewise.
7643 (umax<mode>3): Likewise.
7644 (smax<mode>3): Likewise.
7645 (umin<mode>3): Likewise.
7646 (smin<mode>3): Likewise.
7647 (altivec_vpkuhum): Likewise.
7648 (altivec_vpkuwum): Likewise.
7649 (altivec_vpkshss): Likewise.
7650 (altivec_vpkswss): Likewise.
7651 (altivec_vpkuhus): Likewise.
7652 (altivec_vpkshus): Likewise.
7653 (altivec_vpkuwus): Likewise.
7654 (altivec_vpkswus): Likewise.
7655 (altivec_vpks<VI_char>ss): Likewise.
7656 (altivec_vpks<VI_char>us): Likewise.
7657 (altivec_vpku<VI_char>us): Likewise.
7658 (altivec_vpku<VI_char>um): Likewise.
7659 (altivec_vrl<VI_char>): Likewise.
7660 (altivec_vsl<VI_char>): Likewise.
7661 (altivec_vsr<VI_char>): Likewise.
7662 (altivec_vsra<VI_char>): Likewise.
7663 (altivec_vsldoi_<mode>): Likewise.
7664 (altivec_vupkhsb): Likewise.
7665 (altivec_vupkhs<VU_char>): Likewise.
7666 (altivec_vupkls<VU_char>): Likewise.
7667 (altivec_vupkhsh): Likewise.
7668 (altivec_vupklsb): Likewise.
7669 (altivec_vupklsh): Likewise.
7670 (altivec_vcmpequ<VI_char>_p): Likewise.
7671 (altivec_vcmpgts<VI_char>_p): Likewise.
7672 (altivec_vcmpgtu<VI_char>_p): Likewise.
7673 (abs<mode>2): Likewise.
7674 (vec_unpacks_hi_v16qi): Likewise.
7675 (vec_unpacks_hi_v8hi): Likewise.
7676 (vec_unpacks_lo_v16qi): Likewise.
7677 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7678 (vec_unpacks_lo_v8hi): Likewise.
7679 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7680 (vec_pack_trunc_v8h): Likewise.
7681 (vec_pack_trunc_v4si): Likewise.
7682 (vec_pack_trunc_<mode>): Likewise.
7684 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7686 (vec_vmaxsd): Likewise.
7687 (vec_vmaxud): Likewise.
7688 (vec_vminsd): Likewise.
7689 (vec_vminud): Likewise.
7690 (vec_vpksdss): Likewise.
7691 (vec_vpksdus): Likewise.
7692 (vec_vpkudum): Likewise.
7693 (vec_vpkudus): Likewise.
7694 (vec_vrld): Likewise.
7695 (vec_vsld): Likewise.
7696 (vec_vsrad): Likewise.
7697 (vec_vsrd): Likewise.
7698 (vec_vsubudm): Likewise.
7699 (vec_vupkhsw): Likewise.
7700 (vec_vupklsw): Likewise.
7702 2013-05-29 Jan Hubicka <jh@suse.cz>
7704 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7705 flags; reorder rest of fields in more consistent way.
7706 (varpool_node): Remove analyzed, finalized and alias.
7707 (cgraph_ndoe): Likewise.
7708 (symtab_alias_ultimate_target): New function.
7709 (cgraph_function_node): Move offline.
7710 (cgraph_reset_node): Declare.
7711 (cgraph_comdat_can_be_unshared_p): Remove.
7712 (varpool_remove_initializer): Declare.
7713 (varpool_first_defined_variable, varpool_next_defined_variable
7714 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7715 (cgraph_function_with_gimple_body_p): Update.
7716 (varpool_all_refs_explicit_p): Update.
7717 (symtab_alias_target): New function.
7718 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7719 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7720 (cgraph_function_or_thunk_node): Simplify using
7721 symtab_alias_ultimate_target.
7722 (varpool_variable_node): Likewise.
7723 * cgraph.c (cgraph_create_function_alias): Update.
7724 (cgraph_add_thunk): Update.
7725 (cgraph_remove_node): Update.
7726 (dump_cgraph_node): Do not dump removed flags.
7727 (cgraph_function_body_availability): Update.
7728 (cgraph_propagate_frequency): Update.
7729 (verify_cgraph_node): Check sanity of local flag.
7730 (cgraph_function_node): Move here from cgraph.h; revamp for
7731 cgraph_function_or_thunk_node.
7732 * lto-symtab.c (lto_varpool_replace_node): Update.
7733 (lto_symtab_resolve_can_prevail_p): Update.
7734 (lto_symtab_merge_cgraph_nodes): Update.
7735 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7736 propagate_constants_accross_call, devirtualization_time_bonus,
7737 ipcp_propagate_stage): Update.
7738 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7739 * ipa-inline-transform.c (clone_inlined_nodes,
7740 preserve_function_body_p): Update.
7741 * ipa-reference.c (propagate): Update.
7742 (write_node_summary_p): Update.
7743 * toplev.c (wrapup_global_declaration_2): Update.
7744 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7745 (analyze_function) ... this one.
7746 (cgraph_process_new_functions): Update.
7747 (cgraph_reset_node): Export.
7748 (cgraph_finalize_function): Update.
7749 (cgraph_add_new_function): Update.
7750 (process_function_and_variable_attributes): Update.
7751 (varpool_finalize_decl): Update.
7752 (symbol_finalized): Remove.
7753 (symbol_finalized_and_needed): Rename to ...
7754 (symbol_defined_and_needed): ... update.
7755 (cgraph_analyze_functions): Update.
7756 (handle_alias_pairs): Update.
7757 (mark_functions_to_output): Update.
7758 (assemble_thunk): Update.
7759 (output_in_order): Update.
7760 (output_weakrefs): Update.
7761 (finalize_compilation_unit): Update.
7762 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7763 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7764 input_node, input_varpool_node): Update.
7765 * dbxout.c (dbxout_expand_expr): Update.
7766 * cgraphclones.c (cgraph_clone_node): Update.
7767 (cgraph_copy_node_for_versioning): Update.
7768 (cgraph_materialize_clone): Update.
7769 (cgraph_materialize_all_clones): Update.
7770 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7771 propagate_pure_const, propagate_nothrow): Update.
7772 * lto-streamer-out.c (lto_output, write_symbol): Update.
7773 * ipa-utils.c (ipa_reverse_postorder): Update.
7774 * ipa-inline.c (can_inline_edge_p): Update.
7775 (update_caller_keys, ipa_inline): Update.
7776 * dwarf2out.c (reference_to_unused,
7777 premark_types_used_by_global_vars_helper): Update.
7778 * tree-eh.c (tree_could_trap_p): Update.
7779 * ipa-split.c (consider_split, execute_split_functions): Update.
7780 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7781 has_addr_references_p): Update; move ahead in file for better
7783 (process_references): Simplify.
7784 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7786 (cgraph_comdat_can_be_unshared_p): Make static.
7787 (cgraph_externally_visible_p): Update.
7788 (varpool_externally_visible_p): Update.
7789 (function_and_variable_visibility): Update.
7790 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7791 ipa_tm_mark_force_output_node): Update.
7792 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7793 estimate_edge_devirt_benefit, inline_generate_summary,
7794 inline_write_summary): Update.
7795 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7796 * ipa-prop.c (ipa_compute_jump_functions): Update.
7797 (ipa_print_node_params, ipa_prop_read_section,
7798 ipa_update_after_lto_read, read_replacements_section): Update.
7799 * varasm.c (mark_decl_referenced): Update.
7800 (assemble_alias, dump_tm_clone_pairs): Update.
7801 * tree-inline.c (copy_bb): Update.
7802 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7804 * symtab.c (dump_symtab_base): Print new flags.
7805 (verify_symtab_base): Verify new flags.
7806 (symtab_alias_ultimate_target): New function.
7807 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7808 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7810 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7812 * i386.c (ix86_get_function_versions_dispatcher,
7813 ix86_generate_version_dispatcher_body): Update.
7814 (fold_builtin_cpu): Use varpool_add_new_variable.
7815 * varpool.c (varpool_remove_initializer): Break out from ...
7816 (varpool_remove_node): ... this one.
7817 (dump_varpool_node, varpool_node_for_asm,
7818 cgraph_variable_initializer_availability, varpool_analyze_node,
7819 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7820 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7823 2013-05-29 Jan Hubicka <jh@suse.cz>
7825 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7827 2013-05-29 Easwaran Raman <eraman@google.com>
7829 PR tree-optimization/57442
7830 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7831 when control exits the main loop.
7833 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7835 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7837 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7838 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7839 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7842 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7844 PR tree-optimization/57441
7845 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7846 Don't limit size of incr_vec to number of candidates.
7848 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7850 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7851 and mips16 directories.
7852 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7853 (MULTILIB_DIRNAMES): Ditto.
7854 (MULTILIB_EXCEPTIONS): Add new exceptions.
7855 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7856 (MULTILIB_DIRNAMES): Ditto.
7857 (MULTILIB_EXCEPTIONS): Add new exceptions.
7859 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7860 Marcus Shawcroft <marcus.shawcroft@arm.com>
7862 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7863 SYMBOL_TINY_ABSOLUTE.
7864 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7865 SYMBOL_TINY_ABSOLUTE.
7866 (aarch64_expand_mov_immediate): Likewise.
7867 (aarch64_classify_symbol): Likewise.
7868 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7869 Permit SYMBOL_TINY_ABSOLUTE.
7870 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7872 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7873 Marcus Shawcroft <marcus.shawcroft@arm.com>
7875 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7876 Refactor if/switch. Replace gcc_assert with if.
7878 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7880 * config/i386/i386.c (initial_ix86_tune_features): Enable
7881 FP Reassociation for AMD bdver1 and bdver2.
7883 2013-05-29 Martin Jambor <mjambor@suse.cz>
7885 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7886 and IMAGPART_EXPR do not occur within other handled_components.
7888 2013-05-29 Richard Biener <rguenther@suse.de>
7890 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7891 access on whether the use is in the BB we currently try to
7893 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7894 vectorize to vect_bb_slp_scalar_cost.
7896 2013-05-29 Richard Biener <rguenther@suse.de>
7898 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7899 computing scalar cost offsetted by stmts that are kept live
7901 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7902 for computation of scalar cost.
7904 2013-05-28 Steve Ellcey <sellcey@mips.com>
7906 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7908 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7910 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7911 array notation built-in reduction functions.
7912 * doc/passes.texi (Passes): Added documentation about changes done
7914 * doc/invoke.texi (C Dialect Options): Added documentation about
7915 the -fcilkplus flag.
7916 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7917 (BUILTINS_DEF): Depend on cilkplus.def.
7918 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7919 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7920 * cilkplus.def: New file.
7922 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7924 PR rtl-optimization/57439
7925 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7927 2013-05-28 Easwaran Raman <eraman@google.com>
7929 PR tree-optimization/57337
7930 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7931 (find_insert_point): Correctly identify the insertion point
7932 when two statements with the same UID is compared.
7934 2013-05-28 Richard Biener <rguenther@suse.de>
7936 PR tree-optimization/56787
7937 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7938 from the list of data references.
7939 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7941 (vect_analyze_loop_operations): Likewise.
7942 (vect_transform_loop): Remove clobbers.
7944 2013-05-28 Martin Jambor <mjambor@suse.cz>
7946 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7947 and REALPART_EXPRs have scalar type.
7949 2013-05-28 Richard Biener <rguenther@suse.de>
7951 PR tree-optimization/57411
7952 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7954 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7955 virtual operand propagation.
7957 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7959 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7960 destination register for bmasksi_vis.
7961 (vector_init_bshuffle): Likewise.
7962 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7964 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7966 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7967 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7968 mode if the instruction isn't available in the original mode.
7969 * config/sparc/sparc.opt (mfix-ut699): New option.
7970 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7971 (divdf3): Turn into expander.
7972 (divdf3_nofix): New insn.
7973 (divdf3_fix): Likewise.
7974 (divsf3): Disable if -mfix-ut699.
7975 (sqrtdf2): Turn into expander.
7976 (sqrtdf2_nofix): New insn.
7977 (sqrtdf2_fix): Likewise.
7978 (sqrtsf2): Disable if -mfix-ut699.
7980 2013-05-27 Richard Biener <rguenther@suse.de>
7983 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7984 block for the new loop.
7986 2013-05-27 Richard Biener <rguenther@suse.de>
7988 PR tree-optimization/57343
7989 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7990 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7991 (number_of_iterations_cond): Do not build the folded tree.
7993 2013-05-27 Richard Biener <rguenther@suse.de>
7997 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7998 OEP_CONSTANT_ADDRESS_OF retained.
8000 PR tree-optimization/57417
8001 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
8003 (set_ssa_val_to): Compare addresses using
8004 get_addr_base_and_unit_offset.
8006 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
8008 PR rtl-optimization/56833
8009 * postreload.c (move2add_record_mode): New function.
8010 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
8011 (move2add_use_add2_insn): Use move2add_record_sym_value.
8012 (move2add_use_add3_insn): Likewise.
8013 (reload_cse_move2add): Use move2add_valid_value_p and
8014 move2add_record_mode. Invalidate call-clobbered and REG_INC
8015 affected regs by setting reg_mode to VOIDmode.
8016 (move2add_note_store): Don't pretend the inside of a SUBREG is
8017 the actual destination. Invalidate single/leading registers by
8018 setting reg_mode to VOIDmode.
8019 Use move2add_record_sym_value, move2add_valid_value_p and
8020 move2add_record_mode.
8022 2013-05-27 Richard Biener <rguenther@suse.de>
8024 PR tree-optimization/57396
8025 * tree-affine.c (double_int_constant_multiple_p): Properly
8026 return false for val == 0 and div != 0.
8028 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8030 * config/mips/mips.h: Use #elif in preprocessor conditions.
8032 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8035 * config/mips/constraints.md (kl): New constraint.
8036 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8037 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8038 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8040 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8042 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8045 * config/mips/mips.c (mips_can_inline_p): New function.
8046 (TARGET_CAN_INLINE_P): Define.
8048 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8050 * sched-int.h (ds_t, dw_t): Make unsigned int.
8051 Fix documentation that describes how all the ds_t bits are used.
8052 Reserve the last bit for delayed-branch scheduling.
8053 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8054 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8055 (gen_dep_weak_1): Remove prototype.
8056 * sched-deps.c (get_dep_weak_1): Make static.
8057 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8058 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8059 * doc/tm.texi: Regenerate.
8060 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8062 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8065 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8067 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8068 Sandra Loosemore <sandra@codesourcery.com>
8070 * config.gcc (powerpc-*): Allow native for with-cpu.
8072 2013-05-24 Jeff Law <law@redhat.com>
8074 PR tree-optimization/57124
8075 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8076 conversion feeding a condition if the range has an overflow
8077 if -fstrict-overflow. Add warnings for when we do make the
8080 2013-05-24 Dehao Chen <dehao@google.com>
8082 * tree-cfg.c (locus_discrim_map): Fix the typo.
8083 (locus_discrim_hasher): Likewise.
8084 (locus_discrim_hasher::hash): Likewise.
8085 (locus_discrim_hasher::equal): Likewise.
8087 2013-05-24 Martin Jambor <mjambor@suse.cz>
8089 PR tree-optimization/57294
8090 * cgraph.h (ipa_record_stmt_references): Declare.
8091 * cgraphbuild.c (ipa_record_stmt_references): New function.
8092 (build_cgraph_edges): Use ipa_record_stmt_references.
8093 (rebuild_cgraph_edges): Likewise.
8094 (cgraph_rebuild_references): Likewise.
8095 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8096 associated with the old statement and build references from the
8097 newly built statements.
8098 * ipa-ref.c (ipa_remove_stmt_references): New function.
8099 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8101 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8103 * lra-constraints.c (emit_spill_move): Use smaller mode for
8105 (check_and_process_move): Consider mem-reg moves for secondary
8107 (curr_insn_transform): Don't lose insns emitted before for
8108 secondary memory moves.
8109 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8110 reg set up in the current insn.
8112 2013-05-24 Dehao Chen <dehao@google.com>
8114 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8116 (locus_descrim_hasher::equal): Likewise.
8117 (build_gimple_cfg): New discriminator assignment algorithm.
8118 (make_edges): Likewise.
8119 (next_discriminator_for_locus): Likewise.
8120 (same_line_p): Likewise.
8121 (assign_discriminators): Likewise.
8122 (make_cond_expr_edges): Likewise.
8123 (make_gimple_switch_edges): Likewise.
8124 (make_goto_expr_edges): Likewise.
8125 (make_gimple_asm_edges): Likewise.
8127 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8129 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8130 X format specifier to only display bottom 16 bits.
8131 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8132 immediate to match for operand 2, since it will be masked.
8134 2013-05-24 Richard Biener <rguenther@suse.de>
8136 PR tree-optimization/57287
8137 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8138 all SSA names that occur in abnormal PHIs.
8140 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8143 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8144 that index is not negative.
8146 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8148 PR rtl-optimization/55177
8149 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8150 (simplify_byte_swapping_operation): New.
8151 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8152 (simplify_relational_operation_1): Deal with BSWAP.
8154 2013-05-23 Richard Henderson <rth@redhat.com>
8157 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8158 (ix86_reorg): Call it.
8160 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8163 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8164 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8165 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8167 2013-05-23 Christian Bruel <christian.bruel@st.com>
8170 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8172 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8173 Marcus Shawcroft <marcus.shawcroft@arm.com>
8175 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8176 * config/aarch64/constraints.md (Usa): Remove.
8177 * doc/md.texi (AArch64 Usa): Remove.
8179 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8180 Marcus Shawcroft <marcus.shawcroft@arm.com>
8182 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8183 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8184 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8185 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8187 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8189 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8190 instruction (AdvSIMD).
8191 * config/aarch64/aarch64-builtins.c
8192 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8193 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8195 2013-05-23 Martin Jambor <mjambor@suse.cz>
8198 * tree.h (contains_bitfld_component_ref_p): Declare.
8199 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8200 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8202 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8203 not access a bit-field. Assert all final offsets are byte-aligned.
8205 2013-05-23 Richard Biener <rguenther@suse.de>
8207 PR tree-optimization/57380
8208 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8209 least one invariant or re-used load.
8210 * passes.c (init_optimization_passes): Move pass_phiprop before
8213 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8215 * config/aarch64/aarch64-simd.md
8216 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8218 2013-05-23 Richard Biener <rguenther@suse.de>
8221 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8222 OEP_CONSTANT_ADDRESS_OF retained.
8224 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8227 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8228 don't lower unit. Handle unit not being always BITS_PER_WORD.
8230 2013-05-23 Richard Biener <rguenther@suse.de>
8232 PR rtl-optimization/57341
8233 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8234 instead of true_dependence.
8236 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8238 * bb-reorder.c (branch_threshold): Make const.
8239 (exec_threshold): Ditto.
8241 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8242 Pat Haugen <pthaugen@us.ibm.com>
8243 Peter Bergner <bergner@vnet.ibm.com>
8245 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8246 documentation for the power8 crypto builtins.
8248 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8250 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8251 macros for defining power8 builtin functions.
8252 (BU_P8V_AV_2): Likewise.
8253 (BU_P8V_AV_P): Likewise.
8254 (BU_P8V_VSX_1): Likewise.
8255 (BU_P8V_OVERLOAD_1): Likewise.
8256 (BU_P8V_OVERLOAD_2): Likewise.
8257 (BU_CRYPTO_1): Likewise.
8258 (BU_CRYPTO_2): Likewise.
8259 (BU_CRYPTO_3): Likewise.
8260 (BU_CRYPTO_OVERLOAD_1): Likewise.
8261 (BU_CRYPTO_OVERLOAD_2): Likewise.
8262 (XSCVSPDP): Fix typo, point to the correct instruction.
8263 (VCIPHER): Add power8 crypto builtins.
8264 (VCIPHERLAST): Likewise.
8265 (VNCIPHER): Likewise.
8266 (VNCIPHERLAST): Likewise.
8267 (VPMSUMB): Likewise.
8268 (VPMSUMH): Likewise.
8269 (VPMSUMW): Likewise.
8270 (VPERMXOR_V2DI): Likewise.
8271 (VPERMXOR_V4SI: Likewise.
8272 (VPERMXOR_V8HI: Likewise.
8273 (VPERMXOR_V16QI: Likewise.
8274 (VSHASIGMAW): Likewise.
8275 (VSHASIGMAD): Likewise.
8277 (VPERMXOR): Likewise.
8278 (VSHASIGMA): Likewise.
8280 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8281 __CRYPTO__ if the crypto instructions are available.
8282 (altivec_overloaded_builtins): Add support for overloaded power8
8285 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
8286 support for power8 crypto builtins.
8287 (builtin_function_type): Likewise.
8288 (altivec_init_builtins): Add support for builtins that take vector
8289 long long (V2DI) arguments.
8291 * config/rs6000/crypto.md: New file, define power8 crypto
8294 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8295 Pat Haugen <pthaugen@us.ibm.com>
8296 Peter Bergner <bergner@vnet.ibm.com>
8298 * doc/invoke.texi (Option Summary): Add power8 options.
8299 (RS/6000 and PowerPC Options): Likewise.
8301 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
8302 constraints.md instead of rs6000.h. Reorder w* constraints. Add
8303 wm, wn, wr documentation.
8305 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
8306 registers if direct move instructions are enabled.
8307 (wn): New constraint for no registers.
8308 (wq): New constraint for quad word even GPR registers.
8309 (wr): New constraint if 64-bit instructions are enabled.
8310 (wv): New constraint if power8 vector instructions are enabled.
8311 (wQ): New constraint for quad word memory locations.
8313 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
8314 constraint for 0..15 for crypto instructions.
8315 (gpc_reg_operand): If VSX allow registers in VSX registers as well
8316 as GPR and floating point registers.
8317 (int_reg_operand): New predicate to match only GPR registers.
8318 (base_reg_operand): New predicate to match base registers.
8319 (quad_int_reg_operand): New predicate to match even GPR registers
8320 for quad memory operations.
8321 (vsx_reg_or_cint_operand): New predicate to allow vector logical
8322 operations in both GPR and VSX registers.
8323 (quad_memory_operand): New predicate for quad memory operations.
8324 (reg_or_indexed_operand): New predicate for direct move support.
8326 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
8327 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
8328 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
8329 (POWERPC_MASKS): Add power8 options.
8330 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
8333 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8334 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
8336 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
8337 (-mpower8-fusion): New power8 options.
8338 (-mpower8-fusion-sign): Likewise.
8339 (-mpower8-vector): Likewise.
8340 (-mcrypto): Likewise.
8341 (-mdirect-move): Likewise.
8342 (-mquad-memory): Likewise.
8344 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
8346 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
8348 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
8349 (rs6000_debug_vector_unit): Add p8_vector.
8350 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
8351 definitions. Also print fusion state.
8352 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
8353 (rs6000_builtin_mask_calculate): Add power8 builtin support.
8354 (rs6000_option_override_internal): Add support for power8.
8355 (rs6000_common_init_builtins): Add debugging for skipped builtins
8357 (rs6000_adjust_cost): Add power8 support.
8358 (rs6000_issue_rate): Likewise.
8359 (insn_must_be_first_in_group): Likewise.
8360 (insn_must_be_last_in_group): Likewise.
8361 (force_new_group): Likewise.
8362 (rs6000_register_move_cost): Likewise.
8363 (rs6000_opt_masks): Likewise.
8365 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
8366 power8 capable assembler, default to power7 options.
8367 (TARGET_DIRECT_MOVE): Likewise.
8368 (TARGET_CRYPTO): Likewise.
8369 (TARGET_P8_VECTOR): Likewise.
8370 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
8371 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
8372 (VECTOR_MEM_P8_VECTOR_P): Likewise.
8373 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
8374 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
8375 (TARGET_XSCVDPSPN): Likewise.
8376 (TARGET_XSCVSPDPN): Likewsie.
8377 (TARGET_SYNC_HI_QI): Likewise.
8378 (TARGET_SYNC_TI): Likewise.
8379 (MASK_CRYPTO): Likewise.
8380 (MASK_DIRECT_MOVE): Likewise.
8381 (MASK_P8_FUSION): Likewise.
8382 (MASK_P8_VECTOR): Likewise.
8383 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
8384 temporary used by some of the direct move instructions to get two FP
8385 temporary registers does not force creation of a stack frame.
8386 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
8387 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
8388 that any VSX registers are tieable, even if they are also an
8389 Altivec vector mode.
8390 (r6000_reg_class_enum): Add wm, wr, wv constraints.
8391 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
8392 (RS6000_BTM_CRYPTO): Likewise.
8393 (RS6000_BTM_COMMON): Likewise.
8395 * config/rs6000/rs6000.md (cpu attribute): Add power8.
8396 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
8397 (enum rs6000_vector): Add power8 vector support.
8399 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8403 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
8404 (any_sibcall_could_use_r3): this and handle indirect calls.
8405 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
8407 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8409 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
8411 2013-05-22 Richard Biener <rguenther@suse.de>
8414 * profile.c (branch_prob): Do not split blocks that are
8415 abnormally receiving from ECF_RETURNS_TWICE functions.
8417 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8419 * recog.c (offsettable_address_addr_space_p): Fix calculation of
8420 address mode. Move pointer mode initialization to the same place.
8422 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8424 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
8425 while it has any effect.
8427 2013-05-21 Easwaran Raman <eraman@google.com>
8429 PR tree-optimization/57322
8430 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
8431 UID of the statement added to the BB to be 1.
8433 2013-05-21 Jakub Jelinek <jakub@redhat.com>
8435 PR tree-optimization/57331
8436 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
8437 of conversion from pointer type to integral type with integer.
8439 2013-05-21 Martin Jambor <mjambor@suse.cz>
8442 * ipa-prop.c (ipa_read_node_info): Process param_used and
8443 controlled_uses in the same order as when writing.
8445 2013-05-21 Magnus Granberg <baldrick@free.fr>
8448 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
8450 2013-05-21 Richard Biener <rguenther@suse.de>
8452 PR tree-optimization/57318
8453 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
8454 estimate stmts with side-effects as likely eliminated.
8456 2013-05-21 Richard Biener <rguenther@suse.de>
8458 PR tree-optimization/57330
8459 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
8460 preserve the call stmts fntype.
8462 2013-05-21 Richard Biener <rguenther@suse.de>
8464 PR tree-optimization/57303
8465 * tree-ssa-sink.c (statement_sink_location): Improve killing
8466 stmt detection and properly handle self-assignments.
8468 2013-05-21 Christian Bruel <christian.bruel@st.com>
8470 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
8471 spanning registers. LEAF_REG_REMAP is supported only for contiguous
8472 registers. Set register size out of the PARALLEL loop.
8474 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
8477 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
8478 (*fmasf4, *fmasf4_media): New insns.
8480 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8482 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
8483 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
8484 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
8485 (mips_idiv_insns): Update the comments to say that the returned
8486 instruction counts are in units of BASE_INSN_LENGTH.
8487 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
8488 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
8489 using 2 rather than 4 as the length of indirect MIPS16 and
8490 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
8491 length of a NOP. Don't divide MIPS16 lengths by 2.
8492 (mips16_split_long_branches): Assume a branch is long if the
8493 length is greater than 4 rather than 8.
8494 * config/mips/mips.md (length): Give MIPS16 lengths directly,
8495 rather than multiplying them by 2. Multiply instruction counts
8496 by BASE_INSN_LENGTH rather than 4.
8497 (*jump_mips16, tls_get_tp_mips16_<mode>)
8498 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
8500 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8502 * config/mips/mips.md (extended_mips16): Remove branch case.
8503 (length): Remove duplicated extended_mips16 test.
8505 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8507 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
8509 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
8511 * recog.h (Recog_data): Rename to...
8512 (recog_data_d): ...this.
8513 (recog_data): Update accordingly.
8514 * recog.c (recog_data): Likewise.
8515 * reload.c (save_recog_data): Likewise.
8516 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
8517 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
8519 2013-05-17 Julian Brown <julian@codesourcery.com>
8521 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
8522 found in a REG_EQUAL note, invalidate it.
8524 2013-05-17 Easwaran Raman <eraman@google.com>
8526 * tree-ssa-reassoc.c (find_insert_point): New function.
8527 (insert_stmt_after): Likewise.
8528 (get_def_stmt): Likewise.
8529 (ensure_ops_are_available): Likewise.
8530 (not_dominated_by): Likewise.
8531 (rewrite_expr_tree): Do not move statements beyond what is
8532 necessary. Remove call to swap_ops_for_binary_stmt...
8533 (reassociate_bb): ... and move it here.
8534 (build_and_add_sum): Assign UIDs for new statements.
8535 (linearize_expr): Likewise.
8536 (do_reassoc): Renumber gimple statement UIDs.
8538 2013-05-17 Jan Hubicka <jh@suse.cz>
8540 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8542 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8543 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8545 (output_weakrefs): Update.
8547 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8548 Martin Jambor <mjambor@suse.cz>
8551 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8552 value that corresponds to the given aggval is found in values vector.
8554 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8556 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8557 sse, sse2, sse3, ssse3 and sse4a flags to options.
8559 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8561 * gengtype-state.c: (s_expr_writer): New class, to handle
8562 prettifying of output layout of s-expressions.
8563 (state_writer): New class, to write out gtype.state.
8564 (state_written_type_count): Move this variable into member data of
8566 (s_expr_writer::s_expr_writer): New code: constructor for new class
8567 (state_writer::state_writer(): ditto
8568 (s_expr_writer::write_new_line): New function
8569 (s_expr_writer::write_any_indent): ditto
8570 (s_expr_writer::begin_s_expr): ditto
8571 (s_expr_writer::end_s_expr): ditto
8572 (write_state_fileloc): convert to method of state_writer...
8573 (state_writer:: write_state_fileloc): ...and use methods of
8574 s_expr_writer to write indentation into the gtype.state output file
8575 to visually represent the hierarchical structure of the list
8577 (write_state_fields): ditto, renaming to...
8578 (state_writer::write_state_fields)
8579 (write_state_a_string): ditto, renaming to...
8580 (state_writer::write_state_a_string)
8581 (write_state_string_option): ditto, renaming to...
8582 (state_writer::write_state_string_option)
8583 (write_state_type_option): ditto, renaming to...
8584 (state_writer::write_state_type_option)
8585 (write_state_nested_option): ditto, renaming to...
8586 (state_writer::write_state_nested_option)
8587 (write_state_option): ditto, renaming to...
8588 (state_writer::write_state_option)
8589 (write_state_options): ditto, renaming to...
8590 (state_writer::write_state_options)
8591 (write_state_lang_bitmap): ditto, renaming to...
8592 (state_writer::write_state_lang_bitmap)
8593 (write_state_version): ditto, renaming to...
8594 (state_writer::write_state_version)
8595 (write_state_scalar_type): ditto, renaming to...
8596 (state_writer::write_state_scalar_type)
8597 (write_state_string_type): ditto, renaming to...
8598 (state_writer::write_state_string_type)
8599 (write_state_undefined_type): ditto, renaming to...
8600 (state_writer::write_state_undefined_type)
8601 (write_state_struct_union_type): ditto, renaming to...
8602 (state_writer::write_state_struct_union_type)
8603 (write_state_struct_type): ditto, renaming to...
8604 (state_writer::write_state_struct_type)
8605 (write_state_user_struct_type): ditto, renaming to...
8606 (state_writer::write_state_user_struct_type)
8607 (write_state_lang_struct_type): ditto, renaming to...
8608 (state_writer::write_state_lang_struct_type)
8609 (write_state_param_struct_type): ditto, renaming to...
8610 (state_writer::write_state_param_struct_type)
8611 (write_state_pointer_type): ditto, renaming to...
8612 (state_writer::write_state_pointer_type)
8613 (write_state_array_type): ditto, renaming to...
8614 (state_writer::write_state_array_type)
8615 (write_state_gc_used): ditto, renaming to...
8616 (state_writer::write_state_gc_used)
8617 (write_state_common_type_content): ditto, renaming to...
8618 (state_writer::write_state_common_type_content)
8619 (write_state_type): ditto, renaming to...
8620 (state_writer::write_state_type)
8621 (write_state_pair_list): ditto, renaming to...
8622 (state_writer::write_state_pair_list)
8623 (write_state_pair): ditto, renaming to...
8624 (state_writer::write_state_pair)
8625 (write_state_typedefs): ditto, renaming to...
8626 (state_writer::write_state_typedefs)
8627 (write_state_structures): ditto, renaming to...
8628 (state_writer::write_state_structures)
8629 (write_state_param_structs): ditto, renaming to...
8630 (state_writer::write_state_param_structs)
8631 (write_state_variables): ditto, renaming to...
8632 (state_writer::write_state_variables)
8633 (write_state_srcdir): ditto, renaming to...
8634 (state_writer::write_state_srcdir)
8635 (write_state_files_list): ditto, renaming to...
8636 (state_writer::write_state_files_list)
8637 (write_state_languages): ditto, renaming to...
8638 (state_writer::write_state_languages)
8639 (write_state): create a state_writer instance and use it when
8640 writing out the state file
8642 2013-05-17 Mike Stump <mikestump@comcast.net>
8644 PR rtl-optimization/57304
8645 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8646 accessing DF_REF_REAL_LOC.
8648 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8650 PR rtl-optimization/57281
8651 PR rtl-optimization/57300
8652 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8653 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8654 what the other splitter did if the registers are dead.
8656 2013-05-17 Richard Biener <rguenther@suse.de>
8658 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8661 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8663 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8666 2013-05-17 Marek Polacek <polacek@redhat.com>
8668 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8669 length when doing non-zero store of storing '\0' to '\0'.
8671 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8673 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8674 vect_external_def oprnd1 with loop_vinfo, try to emit
8675 optional cast, negation and and stmts on the loop preheader
8676 edge instead of into the pattern def seq.
8678 PR tree-optimization/57051
8679 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8680 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8682 2013-05-16 Nick Clifton <nickc@redhat.com>
8684 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8685 (rl78_is_naked_func): New function.
8686 (rl78_expand_prologue): Skip prologue generation for naked functions.
8687 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8688 * doc/extend.texi (naked): Add RL78 to the list of processors
8689 that supports this attribute.
8691 2013-05-16 Jeff Law <law@redhat.com>
8693 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8695 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8697 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8698 cache parameters using detect_caches_amd also for CYRIX,
8699 NSC and TM2 signatures.
8701 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8702 Dzianis Kahanovich <mahatma@eu.by>
8706 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8707 VIA/Centaur processors and determine their cache parameters
8708 using detect_caches_amd.
8710 2013-05-16 Teresa Johnson <tejohnson@google.com>
8712 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8713 (rtl_verify_edges): New function.
8714 (rtl_verify_bb_insns): Ditto.
8715 (rtl_verify_bb_pointers): Ditto.
8716 (rtl_verify_bb_insn_chain): Ditto.
8717 (rtl_verify_fallthru): Ditto.
8718 (rtl_verify_bb_layout): Ditto.
8719 (rtl_verify_flow_info_1): Outline checks into new functions.
8720 (rtl_verify_flow_info): Ditto.
8722 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8724 * cfghooks.c (copy_bbs): Add update_dominance argument.
8725 * cfghooks.h (copy_bbs): Update prototype.
8726 * tree-cfg.c (gimple_duplicate_sese_region):
8727 Add update_dominance argument.
8728 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8729 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8730 gimple_duplicate_sese_region call.
8731 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8732 Update copy_bbs call.
8733 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8734 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8736 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8738 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8739 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8740 vect_recog_rotate_pattern.
8741 (vect_recog_rotate_pattern): New function.
8743 2013-05-16 Jason Merrill <jason@redhat.com>
8745 * Makefile.in (LLINKER): New variable.
8746 (mostlyclean): Remove link mutex.
8747 * configure.ac: Handle --enable-link-mutex.
8748 * lock-and-run.sh: New script.
8750 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8753 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8756 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8758 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8760 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8762 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8763 * config/arm/arm.c (next_consecutive_mem): New function.
8764 (gen_movmem_ldrd_strd): Likewise.
8765 * config/arm/arm.md (movmemqi): Update condition and code.
8766 (unaligned_loaddi, unaligned_storedi): New patterns.
8768 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8770 * config.gcc: Obsolete *-*-solaris2.9*.
8771 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8773 2013-05-16 Richard Biener <rguenther@suse.de>
8775 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8776 earlier, after GRAPHITE transforms and IV canonicalization.
8778 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8780 * omp-low.c (extract_omp_for_data): For collapsed loops,
8781 if at least one of the loops is known at compile time to
8782 iterate zero times, set count to 0.
8783 (expand_omp_regimplify_p): New function.
8784 (expand_omp_for_generic): For collapsed loops, if at least
8785 one of the loops isn't known to iterate at least once,
8786 add runtime check with setting count to 0.
8787 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8788 For unsigned types if it isn't known at compile time that
8789 the loop will iterate at least once, add runtime check to bypass
8790 the whole loop if initial condition isn't true.
8792 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8794 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8796 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8799 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8800 transformations to avoid an infinite loop.
8802 2013-05-16 Marek Polacek <polacek@redhat.com>
8804 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8806 2013-05-15 Leif Ekblad <leif@rdos.net>
8808 * config/i386/i386.c (ix86_decompose_address): Use
8809 DEFAULT_TLS_SEG_REG to access TLS segment register.
8810 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8811 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8812 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8814 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8817 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8818 sibling calls to functions that would normally be lazily bound,
8819 unless $gp is call-clobbered.
8821 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8823 * config/i386/i386.c (ix86_option_override_internal): Update
8824 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8825 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8826 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8827 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8829 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8831 2013-05-15 Andreas Schwab <schwab@suse.de>
8833 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8834 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8837 2013-05-15 Teresa Johnson <tejohnson@google.com>
8839 * loop-unroll.c (report_unroll_peel): Check decision before
8840 emitting unroll/peel message.
8842 2013-05-15 Teresa Johnson <tejohnson@google.com>
8844 * function.h (has_bb_partition): New rtl_data flag.
8845 (bb_reorder_complete): Ditto.
8846 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8847 instead of flag_reorder_blocks_and_partition.
8848 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8849 with some enhancements.
8850 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8851 * bb-reorder.c (connect_traces): Check for has_bb_partition
8852 instead of flag_reorder_blocks_and_partition.
8853 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8854 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8855 verify_hot_cold_block_grouping.
8856 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8858 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8861 * config/arm/predicates.md (call_insn_operand): New predicate.
8862 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8863 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8864 if insn is not a tail call.
8865 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8867 * config/arm/arm.h (enum reg_class): New caller save register class.
8868 (REG_CLASS_NAMES): Likewise.
8869 (REG_CLASS_CONTENTS): Likewise.
8870 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8873 2013-05-15 Richard Biener <rguenther@suse.de>
8875 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8876 of MSG_OPTIMIZED_LOCATIONS.
8877 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8878 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8880 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8881 of MSG_OPTIMIZED_LOCATIONS.
8882 (execute_vect_slp): Likewise.
8883 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8884 (vect_create_cond_for_alias_checks): Likewise.
8885 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8886 (vect_recog_widen_mult_pattern): Likewise.
8887 (vect_recog_widen_sum_pattern): Likewise.
8888 (vect_recog_over_widening_pattern): Likewise.
8889 (vect_recog_widen_shift_pattern): Likewise.
8890 (vect_recog_vector_vector_shift_pattern): Likewise.
8891 (vect_recog_divmod_pattern): Likewise.
8892 (vect_recog_mixed_size_cond_pattern): Likewise.
8893 (vect_recog_bool_pattern): Likewise.
8894 (vect_pattern_recog_1): Likewise.
8896 2013-05-15 Martin Jambor <mjambor@suse.cz>
8898 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8899 non-functions to builtin_unreachable.
8900 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8901 correct when new direct edges were discovered.
8903 2013-05-15 Martin Jambor <mjambor@suse.cz>
8905 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8906 header, print symbol order instead of node uid, print more information
8907 about indirect edge targets.
8908 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8910 (ipa_make_edge_direct_to_target): Likewise.
8911 (remove_described_reference): Likewise.
8912 (propagate_controlled_uses): Likewise.
8913 (ipa_print_node_params): Also print symbol order.
8914 (ipcp_transform_function): Print symbol order instead of node uids.
8915 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8916 (cgraph_get_create_real_symbol_node): Likewise.
8917 * ipa-cp.c (print_lattice): Likewise.
8918 (print_all_lattices): Likewise.
8919 (determine_versionability): Likewise.
8920 (initialize_node_lattices): Likewise.
8921 (estimate_local_effects): Likewise.
8922 (update_profiling_info): Likewise.
8923 (create_specialized_node): Likewise.
8924 (perhaps_add_new_callers): Likewise.
8925 (decide_about_value): Likewise.
8926 (decide_whether_version_node): Likewise.
8927 (identify_dead_nodes): Likewise.
8928 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8929 (dump_inline_summary): Likewise.
8930 (estimate_node_size_and_time): Likewise.
8931 (inline_analyze_function): Likewise.
8932 * ipa-inline.c (report_inline_failed_reason): Likewise.
8933 (want_early_inline_function_p): Likewise.
8934 (edge_badness): Likewise.
8935 (update_edge_key): Likewise.
8936 (inline_small_functions): Likewise. Add dumping of order to two other
8938 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8939 instead of node uids.
8940 (propagate_pure_const): Likewise.
8941 (propagate_pure_const): Likewise.
8942 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8943 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8944 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8946 * tree-pretty-print.c (dump_function_header): Likewise.
8947 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8948 Print symbol order instead of node uids.
8950 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8952 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8953 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8955 2013-05-15 Richard Biener <rguenther@suse.de>
8957 PR tree-optimization/57275
8958 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8959 return value for fail to do runtime alias checks for gather loads.
8961 2013-05-15 Jan Hubicka <jh@suse.cz>
8965 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8966 weakrefs are not external.
8967 (lto_symtab_merge_decls): Fix thinko when dealing with
8968 non-lto_symtab decls.
8969 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8970 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8971 * varpool.c (dump_varpool_node): Dump more flags.
8973 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8975 * config/i386/i386.c (processor_alias_table): Add instruction
8976 FSGSBASE for AMD bdver3 architecture.
8978 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8980 * tree.c (warn_deprecated_use): Print file:line using locus color.
8981 * diagnostic.c (diagnostic_report_current_module): Print file:line
8982 and file:line:column using locus color.
8984 2013-05-14 Mike Stump <mikestump@comcast.net>
8986 * gdbinit.in: Add __null.
8988 2013-05-14 Mike Stump <mikestump@comcast.net>
8990 * recog.h: Rename struct recog_data to Recog_data.
8991 * recog.c: Likewise.
8992 * reload.c (can_reload_into): Likewise.
8993 * config/picochip/picochip.c: Likewise.
8995 2013-05-14 Mike Stump <mikestump@comcast.net>
8997 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8999 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
9001 * resource.h (struct resources): Remove unch_memory member.
9002 (CLEAR_RESOURCE): Don't clear unch_memory.
9003 * resource.c (mark_referenced_resources): Don't set it.
9004 (mark_set_resources): Likewise.
9005 (mark_target_live_regs): Don't clear it.
9006 (init_resource_info): Likewise.
9007 * reorg.c (resource_conflicts_p): Don't compare it.
9008 (redundant_insn): Don't set it.
9010 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
9012 * emit-rtl.c (next_label): Remove unused function.
9013 (skip_consecutive_labels, link_cc0_insns): Move to ...
9014 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
9015 only place where these functions are used, and make them static.
9017 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9019 * fold-const.c (fold_negate_expr): Handle vectors.
9020 (fold_truth_not_expr): Make it static.
9021 (fold_invert_truthvalue): New static function.
9022 (invert_truthvalue_loc): Handle vectors. Do not call
9023 fold_truth_not_expr directly.
9024 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
9025 <TRUTH_NOT_EXPR>: Do not cast to boolean.
9026 (fold_comparison): Handle vector constants.
9027 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
9028 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
9029 * tree.h (fold_truth_not_expr): Remove declaration.
9031 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
9033 * config/aarch64/aarch64-simd.md
9034 (aarch64_vcond_internal<mode>): Rename to...
9035 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9036 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9037 float modes. Clarify all iterator modes.
9038 (vcond<mode><mode>): Use new name for vcond expanders.
9039 (vcond<v_cmp_result><mode>): Likewise.
9040 (vcondu<mode><mode>: Likewise.
9041 * config/aarch64/iterators.md (VDQF_COND): New.
9043 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9046 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9047 variable for the shift amount. Check that we shift by non-negative
9050 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9053 * config/arm/arm.h (EPILOGUE_USES): Only return true
9054 for LR_REGNUM after epilogue_completed.
9056 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9058 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9061 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9064 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9065 and Solaris 11+/x86 with gld.
9066 * configure: Regenerate.
9068 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9070 * expmed.c (expand_shift_1): Canonicalize rotates by
9071 constant bitsize / 2 to bitsize - 1.
9072 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9073 case ROTATERT>: Likewise.
9076 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9078 * config/i386/i386.md (rotateinv): New code attr.
9079 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9080 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9081 roll $31, %eax, etc.
9083 2013-05-14 Richard Biener <rguenther@suse.de>
9086 * tree-eh.c (sink_clobbers): Give up for successors with
9087 multiple predecessors and no virtual uses.
9089 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9091 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9092 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9094 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9097 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9098 the case when both op0 and op1 have VOIDmode.
9100 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9102 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9103 in multiply-accumulate mode.
9105 2013-05-13 Guozhi Wei <carrot@google.com>
9107 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9109 2013-05-13 Kai Tietz <ktietz@redhat.com>
9112 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9113 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9114 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9115 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9116 (ix86_expand_prologue): Likewise.
9117 (ix86_expand_split_stack_prologue): Likewise.
9118 (legitimate_pic_address_disp_p): Likewise.
9119 (legitimize_pic_address): Likewise.
9120 (legitimize_tls_address): Likewise.
9121 (legitimize_pe_coff_symbol): Likewise.
9122 (output_pic_addr_const): Likewise.
9123 (construct_plt_address): Likewise.
9124 (ix86_expand_call): Likewise.
9125 (x86_output_mi_thunk): Likewise.
9126 (x86_function_profiler): Likewise.
9128 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9130 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9131 similar switch cases.
9132 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9133 (aarch64_simd_mov_to_<mode>low): Delete.
9134 (aarch64_simd_mov_to_<mode>high): Delete.
9135 (move_lo_quad_<mode>): Add w<-r alternative.
9136 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9137 (aarch64_simd_mov_from_*): Update type attribute.
9138 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9141 2013-05-13 Jan Hubicka <jh@suse.cz>
9143 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9144 * config/i386/i386.c (ix86_compute_frame_layout,
9145 ix86_expand_epilogue, emit_i387_cw_initialization,
9146 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9147 ix86_local_alignment): Fix use of size/speed predicates.
9149 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9151 PR tree-optimization/45216
9152 PR tree-optimization/57157
9153 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9154 the (-Y) & (B - 1) variant if OP is |.
9155 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9156 return shifted. Use (-op1) & (prec - 1) as other_amount
9157 instead of prec - op1.
9159 2013-05-13 Martin Jambor <mjambor@suse.cz>
9162 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9163 (ipa_constant_data): New type.
9164 (ipa_jump_func): Use ipa_constant_data to hold information about
9165 constant jump functions.
9166 (ipa_get_jf_constant): Adjust to jump function type changes.
9167 (ipa_get_jf_constant_rdesc): New function.
9168 (ipa_param_descriptor): New field controlled_uses.
9169 (ipa_get_controlled_uses): New function.
9170 (ipa_set_controlled_uses): Likewise.
9171 * ipa-ref.h (ipa_find_reference): Declare.
9172 * ipa-prop.c (ipa_cst_ref_desc): New type.
9173 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9175 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9176 New parameter cs. Adjust all callers.
9177 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9178 (remove_described_reference): New function.
9179 (jfunc_rdesc_usable): Likewise.
9180 (try_make_edge_direct_simple_call): Decrement controlled use count,
9181 attempt to remove reference if it hits zero.
9182 (combine_controlled_uses_counters): New function.
9183 (propagate_controlled_uses): Likewise.
9184 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9185 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9186 (ipa_print_node_params): Print described use counter.
9187 (ipa_write_jump_function): Adjust to jump function type changes.
9188 (ipa_read_jump_function): New parameter CS, pass it to
9189 ipa_set_jf_constant. Adjust caller.
9190 (ipa_write_node_info): Stream controlled use count
9191 (ipa_read_node_info): Likewise.
9192 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9194 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9195 count. Remove cloning-added reference if it reaches zero.
9196 * ipa-ref.c (ipa_find_reference): New function.
9198 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9200 * config/i386/i386.c (processor_target_table): Modified default
9201 alignment values for AMD BD and BT architectures.
9203 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9205 * tree-vect-generic.c (uniform_vector_p): Move ...
9206 * tree.c (uniform_vector_p): ... here.
9207 * tree.h (uniform_vector_p): Declare it.
9208 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9211 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9213 PR tree-optimization/57230
9214 * tree-ssa-strlen.c (handle_char_store): Record length for
9215 array store from STRING_CST.
9217 PR tree-optimization/57230
9218 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9221 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9223 * config/epiphany/epiphany.c (epiphany_init): Check size of
9224 NUM_MODES_FOR_MODE_SWITCHING.
9225 (epiphany_expand_prologue):
9226 Remove CONFIG_REGNUM initial value handling code.
9227 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9228 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9229 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9230 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9231 Don't return 1 for FP_MODE_NONE.
9232 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9233 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9234 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9235 * config/epiphany/epiphany.md (save_config): New pattern.
9237 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9239 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9241 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9243 * config/i386/i386.md (memory): Handle sseishft1.
9244 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9245 (*vec_extractv2di_1): Ditto.
9247 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9249 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9252 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9254 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9255 Add mthumb/march=armv7-a multilib.
9256 Add mthumb/march=armv7-r multilib.
9257 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
9259 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
9261 * config/v850/t-rtems: Add more multilibs.
9263 2013-05-10 Richard Biener <rguenther@suse.de>
9265 PR tree-optimization/57214
9266 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
9267 not propagate from SSA names that occur in abnormal PHI nodes.
9269 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
9271 * stor-layout.c (element_precision): New function.
9272 * machmode.h (element_precision): Declare it.
9273 * tree.c (build_minus_one_cst): New function.
9274 (element_precision): Likewise.
9275 * tree.h (build_minus_one_cst): Declare new function.
9276 (element_precision): Likewise.
9277 * fold-const.c (operand_equal_p): Use element_precision.
9278 (fold_binary_loc): Handle vector types.
9279 * convert.c (convert_to_integer): Use element_precision.
9280 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
9283 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
9285 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
9286 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
9287 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
9288 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9289 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
9290 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
9291 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
9292 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9293 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
9294 (Uuw8): New constraints.
9295 (Usb4): Move into alphabetical order.
9296 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
9297 (sd8_operand, ub8_operand, uw8_operand): New predicates.
9298 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
9299 previously unnamed patterns.
9300 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
9301 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
9302 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
9303 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
9304 of set_attr_alternative/if_then_else. Use extended_mips16 instead
9305 of specific lengths.
9307 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9309 * config/i386/i386.md (rotateinv): New code attr.
9310 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9311 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9312 roll $31, %eax, etc.
9314 PR tree-optimization/45216
9315 PR tree-optimization/57157
9316 * tree-ssa-forwprop.c (simplify_rotate): New function.
9317 (ssa_forward_propagate_and_combine): Call it.
9319 2013-05-10 Richard Biener <rguenther@suse.de>
9321 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
9322 disable peeling when we version for aliasing.
9323 (vector_alignment_reachable_p): Honor explicit user alignment.
9324 (vect_supportable_dr_alignment): Likewise.
9325 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
9326 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
9327 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
9328 then peeling to arrange for the cost-model check to come first.
9330 2013-05-10 Alan Modra <amodra@gmail.com>
9332 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
9333 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
9334 * configure: Regenerate.
9336 2013-05-10 Alan Modra <amodra@gmail.com>
9339 * varasm.c (default_elf_select_section): Move !DECL_P check..
9340 (get_named_section): ..to here before calling get_section_name.
9342 (default_section_type_flags): Add DECL_P check.
9343 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
9344 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
9346 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
9348 * config/epiphany/epiphany.c (epiphany_expand_prologue):
9349 When using gen_stack_adjust_str with a register offset, add a
9350 REG_FRAME_RELATED_EXPR note.
9352 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
9354 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
9355 (*vec_extractv4si_zext_mem): Ditto.
9356 (*vec_extractv2di): Add 0->x and x->x alternatives.
9357 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
9358 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
9360 2013-05-09 Jason Merrill <jason@redhat.com>
9362 N3639 C++1y VLA support
9363 * gimplify.c (gimplify_vla_decl): Don't touch an existing
9366 * tree.c (build_constructor_va): New.
9367 * tree.h: Declare it.
9369 2013-05-09 Martin Jambor <mjambor@suse.cz>
9372 * gimple-fold.c (canonicalize_constructor_val): Call
9373 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
9375 2013-05-09 Jan Hubicka <jh@suse.cz>
9376 Richard Biener <rguenther@suse.de>
9379 * symtab.c (symtab_make_decl_local): Do not add private names.
9381 2013-05-09 Jan Hubicka <jh@suse.cz>
9384 * symtab.c (insert_to_assembler_name_hash): Handle clones.
9385 (unlink_from_assembler_name_hash): Likewise.
9386 (symtab_prevail_in_asm_name_hash, symtab_register_node,
9387 symtab_unregister_node, symtab_initialize_asm_name_hash,
9388 change_decl_assembler_name): Update.
9390 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
9392 * config/aarch64/aarch64.md: New movtf split.
9393 (*movtf_aarch64): Update.
9394 (aarch64_movdi_tilow): Handle TF modes and rename to
9395 aarch64_movdi_<mode>low.
9396 (aarch64_movdi_tihigh): Handle TF modes and rename to
9397 aarch64_movdi_<mode>high
9398 (aarch64_movtihigh_di): Handle TF modes and rename to
9399 aarch64_mov<mode>high_di
9400 (aarch64_movtilow_di): Handle TF modes and rename to
9401 aarch64_mov<mode>low_di
9402 (aarch64_movtilow_tilow): Remove spurious whitespace.
9403 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
9405 (aarch64_print_operand): Update.
9407 2013-05-09 Alan Modra <amodra@gmail.com>
9409 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
9411 * configure: Regenerate.
9413 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9415 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
9416 splitter preparation statements.
9417 * config/i386/sse.md (*vec_extract* splitters): Ditto.
9418 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
9421 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9423 * gimple-ssa-strength-reduction.c (count_candidates): Change
9424 return value to int.
9425 (analyze_candidates_and_replace): Change type of length to int.
9427 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9429 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
9430 (*vec_extract<mode>): Use VI12_128 mode iterator.
9431 (*vec_extract<mode>_mem): Ditto.
9432 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
9435 2013-05-08 Diego Novillo <dnovillo@google.com>
9440 2012-08-17 Diego Novillo <dnovillo@google.com>
9443 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
9444 * config.in: Regenerate.
9445 * configure: Regenerate.
9446 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
9448 2013-05-08 Jan Hubicka <jh@suse.cz>
9451 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
9452 * cgraph.h (symtab_node_base): Add unique_name.
9453 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9454 input_overwrite_node, input_varpool_node): Stream unique_name.
9455 * cgraphclones.c (cgraph_create_virtual_clone,
9456 cgraph_function_versioning): Set unique_name.
9457 * ipa.c (function_and_variable_visibility): Set unique_name.
9459 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9461 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
9462 (alloc_cand_and_find_basis): Restrict conditional candidate
9463 processing to CAND_MULTs.
9465 2013-05-08 Jan Hubicka <jh@suse.cz>
9468 lto-symtab.c (lto_symtab_symbol_p): New function.
9469 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
9470 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
9471 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
9472 Skip static symbols.
9474 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
9476 PR tree-optimization/57200
9477 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
9478 Only call inform if the preceding warning_at returns true.
9480 2013-05-07 Han Shen <shenhan@google.com>
9482 * cfgexpand.c (record_or_union_type_has_array_p): New function.
9483 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
9484 * common.opt (fstack-protector-strong): New option.
9485 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
9486 * doc/invoke.texi (Optimization Options): Document
9487 "-fstack-protector-strong".
9488 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
9490 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9492 * config/mips/mips.c (mips_machine_reorg2): Return 0.
9494 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
9496 * ira.c (update_equiv_regs): Add insn having equiv memory even if
9497 it is not lhs of the insn.
9498 (setup_reg_equiv): Remove insn having equiv memory which it is not
9500 * lra-constraints.c (process_address): Try to improve generation
9501 code for address base + disp.
9502 (lra_constraints): Make correct the code for checking insn setting
9503 up backward equivalence. Remove insn only if it is in the init
9505 * lra-eliminations.c (update_reg_eliminate): Change return value.
9506 (lra_eliminate): Use the result.
9508 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
9510 * config/i386/sse.md (ssescalarnummask): New mode attribute.
9511 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
9512 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
9513 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
9514 register target operands.
9515 (*vec_extractv8hi_sse2): New pattern.
9516 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
9517 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
9518 (*vec_extract<mode>_mem): New insn and split pattern.
9520 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
9522 * config/arm/arm.c (arm_asan_shadow_offset): New function.
9523 (TARGET_ASAN_SHADOW_OFFSET): Define.
9524 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9525 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9527 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9529 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9530 (incr_vec_index): Return -1 if increment not found.
9531 (create_add_on_incoming_edge): Assert if increment not found.
9532 (record_increment): Limit number of increments recorded.
9533 (all_phi_incrs_profitable): Return false if an increment not found.
9534 (replace_profitable_candidates): Don't process increments that were
9536 (analyze_candidates_and_replace): Limit size of incr_vec.
9538 2013-05-07 Richard Biener <rguenther@suse.de>
9540 * calls.c (special_function_p): setjmp-like functions are leaf.
9541 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9542 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9544 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9546 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9547 (aarch64_simd_mov<mode>): New expander.
9548 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9549 (aarch64_simd_mov_to_<mode>high): Likewise.
9550 (aarch64_simd_mov_from_<mode>low): Likewise.
9551 (aarch64_simd_mov_from_<mode>high): Likewise.
9552 (aarch64_dup_lane<mode>): Update.
9553 (aarch64_dup_lanedi): New instruction pattern.
9554 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9555 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9557 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9559 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9560 (replace_mult_candidate): Remove unnecessary argument; remove
9561 unnecessary parameter from call to introduce_cast_before_cand.
9562 (replace_unconditional_candidate): Remove unnecessary parameter
9563 from call to replace_mult_candidate.
9564 (replace_conditional_candidate): Likewise.
9565 (insert_initializers): Use make_temp_ssa_name.
9566 (introduce_cast_before_cand): Remove unnecessary argument; use
9568 (replace_one_candidate): Remove unnecessary argument; remove
9569 unnecessary parameter from calls to introduce_cast_before_cand.
9570 (replace_profitable_candidates): Remove unnecessary parameters
9571 from calls to replace_one_candidate.
9573 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9575 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9576 phi def as possibly hiding a basis for a CAND_ADD whose operands
9577 have been commuted in the analysis.
9578 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9580 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9582 * config/aarch64/aarch64.md
9583 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9584 shift value between 0-4.
9586 2013-05-07 Richard Biener <rguenther@suse.de>
9588 * double-int.h (rshift): New overload.
9589 * double-int.c (rshift): New function.
9590 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9591 (create_reference_ops_from_ref): Remove.
9592 (vn_reference_insert): Use shared ops for constructing the
9593 reference and copy it.
9595 2013-05-07 Richard Biener <rguenther@suse.de>
9598 * tree-eh.c (sink_clobbers): Properly propagate
9599 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9601 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9603 PR tree-optimization/57149
9604 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9605 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9606 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9607 uninit_undefined_value_p instead of ssa_undefined_value_p.
9610 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9611 for modifier == EXPAND_INITIALIZER.
9613 2013-05-07 Anton Blanchard <anton@samba.org>
9615 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9616 for powerpc64 little endian.
9617 * configure: Regenerate.
9619 2013-05-06 Graham Stott <grahams@btinternet.com>
9621 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9622 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9623 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9624 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9626 2013-05-06 Graham Stott <grahams@btinternet.com>
9628 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9629 codes which allow non-lvalues.
9631 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9633 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9634 components are all 1s.
9635 (integer_minus_onep): New function.
9636 * tree.h (integer_minus_onep): Declare it.
9637 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9638 integer_minus_onep instead of integer_all_onesp.
9640 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9643 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9644 variations of these patterns.
9646 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9648 * config/i386/i386.md (isa): Add x64_sse4 member.
9649 (enabled): Handle x64_sse4.
9650 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9651 instruction for 64bit SSE4_1 targets. Update insn attributes.
9652 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9653 instruction for SSE4_1 targets. Update insn attributes.
9654 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9655 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9656 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9658 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9659 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9660 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9662 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9665 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9667 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9669 * final.c (do_assembler_dialects): Don't handle curly braces and
9670 vertical bar escaped by % as dialect delimiters.
9671 (output_asm_insn): Print curly braces and vertical bar if escaped
9672 by % and ASSEMBLER_DIALECT defined.
9673 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9674 * doc/tm.texi: Regenerated.
9676 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9678 * config/mips/mips.c: Include tree-pass.h.
9679 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9680 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9681 (pass_mips_machine_reorg2): New machine specific pass.
9682 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9683 (mips_option_override): Register the new pass.
9684 * rtl.h (cleanup_barriers): Remove prototype.
9685 (dbr_schedule): Likewise.
9686 * jump.c (cleanup_barriers): Make static.
9687 * reorg.c (dbr_schedule): Likewise.
9689 2013-05-06 Richard Biener <rguenther@suse.de>
9691 PR tree-optimization/57185
9692 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9695 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9698 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9699 between "first_arg" and "insn", not "last" and "insn".
9701 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9703 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9704 (find_candidates_in_block): Re-enable slsr_process_phi.
9705 (create_phi_basis): Fix double counting of candidate adjustment.
9707 2013-05-06 Richard Biener <rguenther@suse.de>
9710 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9711 the edge is also fallthru, preserve it and just clear the
9713 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9714 also complex, preserve that and just clear the fallthru flag.
9715 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9716 update virtual operands.
9718 2013-05-06 Alan Modra <amodra@gmail.com>
9720 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9721 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9722 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9723 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9724 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9725 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9727 2013-05-06 Alan Modra <amodra@gmail.com>
9729 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9730 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9731 (DEFAULT_ASM_ENDIAN): Define.
9732 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9733 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9734 Update -K PIC clause from sysv4.h.
9735 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9736 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9738 2013-05-06 Alan Modra <amodra@gmail.com>
9740 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9741 twice for little-endian.
9742 (ashrdi3_no_power, ashrdi3): Support little-endian.
9744 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9747 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9748 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9750 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9751 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9752 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9755 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9756 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9758 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9759 * config/pa/pa.opt: Make mbig-switch a no-op.
9760 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9761 (CASE_VECTOR_MODE): Always return SImode.
9762 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9763 for the !TARGET_BIG_SWITCH case.
9764 * config/pa/pa-linux.h: Likewise.
9765 * config/pa/pa-openbsd.h: Likewise.
9766 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9767 * config/pa/pa.md (short_jump): Remove define_insn.
9768 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9769 (casesi0): Remove define_insn.
9770 (type): Remove btable_branch.
9771 (pa_combine_type): Likewise.
9772 (in_nullified_branch_delay): Likewise.
9773 (in_call_delay): Likewise.
9774 (define_delay): Likewise.
9775 (define_insn_reservation "Z3"): Likewise.
9776 (define_insn_reservation "Z4"): Likewise.
9777 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9778 (pa_adjust_insn_length): Remove adjustment for btable branches.
9779 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9782 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9784 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9785 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9786 Add m->r,x alternatives.
9787 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9788 splitters using SWI48x mode iterator.
9789 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9790 TARGET_64BIT. Add m->x alternative.
9791 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9792 Add o->x alternative. Enable for TARGET_SSE.
9793 (sse_storeq): Remove expander.
9794 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9795 with memory input operand.
9796 (*vec_extractv2di_1 splitter): New.
9797 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9798 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9800 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9802 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9803 (INT_LOWPART): Delete.
9804 (extract_MB): Adjust.
9805 (extract_ME): Adjust.
9806 (print_operand): Adjust.
9808 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9810 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9811 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9812 (reg_or_logical_cint_operand, easy_fp_constant,
9813 logical_const_operand): Delete "CONST_DOUBLE" case.
9814 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9815 "HOST_BITS_PER_WIDE_INT == 64" test.
9816 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9817 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9818 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9819 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9821 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9822 CONST_DOUBLE DImode/VOIDmode case.
9823 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9824 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9825 CONST_DOUBLE VOIDmode case.
9826 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9827 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9828 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9829 Delete CONST_DOUBLE case.
9830 (splitters for mov FMOVE64 const_double): Delete
9831 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9832 "HOST_BITS_PER_WIDE_INT >= 64" test.
9833 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9835 (mov DI const_double): Delete.
9837 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9839 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9840 on op shows all bits zero in mode of a lowpart subreg, return zero.
9842 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9845 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9846 to save TFmode registers and DImode to save TImode registers for
9847 caller save operations.
9848 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9849 mark being partially clobbered since they only use the first
9852 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9853 and TDmode only use the upper 64-bits of each VSX register.
9855 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9857 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9858 (find_candidates_in_block): Disable slsr_process_phi.
9860 2013-05-03 Guozhi Wei <carrot@google.com>
9862 * coverage.c (coverage_obj_init): Move the construction of gcov
9864 (build_init_ctor): ... here.
9866 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9868 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9869 (slsr_cand_d): Redefine def_phi.
9870 (stride_status, phi_adjust_status, count_phis_status): New enums.
9871 (find_phi_def): New.
9872 (find_basis_for_base_expr): New.
9873 (find_basis_for_candidate): Handle hidden bases.
9874 (alloc_cand_and_find_basis): Handle phi candidates.
9875 (slsr_process_phi): New.
9876 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9877 (create_mul_imm_cand): Likewise.
9878 (create_add_ssa_cand): Exclude phi base candidates.
9879 (create_add_imm_cand): Likewise.
9880 (slsr_process_cast): Likewise.
9881 (slsr_process_copy): Likewise.
9882 (find_candidates_in_block): Handle phi candidates.
9883 (dump_candidate): Likewise.
9884 (unconditional_cands): Delete.
9885 (unconditional_cands_with_known_stride_p): Delete.
9886 (phi_dependent_cand_p): New.
9887 (cand_increment): Handle phi-dependent candidates.
9888 (replace_dependent): Delete.
9889 (replace_mult_candidate): New.
9890 (replace_unconditional_candidate): New.
9891 (incr_vec_index): Move to avoid forward reference.
9892 (create_add_on_incoming_edge): New.
9893 (create_phi_basis): New.
9894 (replace_dependents): Delete.
9895 (replace_conditional_candidate): New.
9896 (phi_add_costs): New.
9897 (replace_uncond_cands_and_profitable_phis): New.
9898 (record_increment): Handle phi adjustments.
9899 (record_phi_increments): New.
9900 (record_increments): Handle phi adjustments.
9901 (phi_incr_cost): New.
9902 (lowest_cost_path): Handle phis.
9903 (total_savings): Likewise.
9904 (analyze_increments): Likewise.
9905 (ncd_with_phi): New.
9906 (ncd_of_cand_and_phis): New.
9907 (nearest_common_dominator_for_cands): Handle phi increments.
9908 (all_phi_incrs_profitable): New.
9909 (replace_profitable_candidates): Handle phi-dependent candidates.
9910 (analyze_candidates_and_replace): Likewise.
9912 2013-05-03 Teresa Johnson <tejohnson@google.com>
9915 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9916 do not exceed REG_BR_PROB_BASE.
9918 2013-05-03 Jeff Law <law@redhat.com>
9920 PR tree-optimization/57144
9921 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9922 operand of the condition will bit into the new type when eliminating
9923 a cast feeding a condition.
9925 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9927 PR rtl-optimization/57130
9928 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9929 of COMPARE as in_code to the recursive call if needed.
9931 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9933 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9934 (enabled): Handle new members.
9935 * config/i386/sse.md (*vec_concatv2si): Merge from
9936 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9937 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9939 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9941 PR tree-optimization/57027
9942 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9943 for fnms opportunity, check we got the prerequisite kind
9944 of tree / gimple before using accessor functions.
9946 2013-05-03 Richard Biener <rguenther@suse.de>
9948 * double-int.h (lshift): New overload without precision
9950 (operator *=, operator +=, operator -=): Move ...
9951 * double-int.c (operator *=, operator +=, operator -=): ... here
9952 and implement more efficiently.
9953 (mul_double_with_sign): Remove.
9954 (lshift_double): Adjust to take unsinged shift argument, push
9955 dispatching code to callers.
9956 (mul_double_wide_with_sign): Add early out for callers that
9957 are not interested in high parts or overflow.
9958 (lshift): New function.
9959 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9961 (lrotate, rrotate): Use logical shifts.
9962 * expr.c (get_inner_reference): Use lshift.
9963 * fixed-value.c (do_fixed_divide): Likewise.
9964 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9965 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9966 (indirect_refs_may_alias_p): Likewise.
9967 (stmt_kills_ref_p_1): Likewise.
9969 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9971 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9973 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9975 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9976 scalar form of FABD instruction.
9978 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9980 * lra-constraints.c (process_alt_operands): Add checking alt
9981 number to choose the best alternative.
9983 2013-05-02 Richard Biener <rguenther@suse.de>
9985 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9986 bitmap and its handling.
9987 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9989 2013-05-02 Richard Biener <rguenther@suse.de>
9992 * tree-inline.c (copy_loops): Properly handle removed loops.
9993 (copy_cfg_body): Mark destination loops for fixup if source
9996 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9999 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
10000 generating simple_return for naked functions.
10002 2013-05-02 Martin Jambor <mjambor@suse.cz>
10004 PR middle-end/56988
10005 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
10006 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
10008 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
10009 ipa_agg_replacement_value structures.
10010 (known_aggs_to_agg_replacement_list): Likewise.
10011 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
10012 (read_agg_replacement_chain): Likewise.
10013 (ipcp_transform_function): Also check that by_ref flags match.
10015 2013-05-02 Richard Biener <rguenther@suse.de>
10017 * graphds.h (struct graph): Add obstack member.
10018 * graphds.c (new_graph): Initialize obstack and allocate
10020 (add_edge): Allocate edge from the obstack.
10021 (free_graph): Free the obstack instead of all edges and vertices.
10023 2013-05-02 Teresa Johnson <tejohnson@google.com>
10025 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
10027 * cfg.c (update_bb_profile_for_threading): Ditto.
10028 * tree-inline.c (copy_bb): Ditto.
10029 (copy_edges_for_bb): Ditto.
10030 (initialize_cfun): Ditto.
10031 (copy_cfg_body): Ditto.
10032 (expand_call_inline): Ditto.
10033 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10034 (estimate_node_size_and_time): Ditto.
10035 (inline_merge_summary): Ditto.
10036 * cgraphclones.c (cgraph_clone_edge): Ditto.
10037 (cgraph_clone_node): Ditto.
10038 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10039 (compute_trg_info): Ditto.
10041 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10043 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10044 S reg when fp attribute set.
10045 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10047 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10049 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10051 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10052 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10053 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10055 2013-05-02 Richard Biener <rguenther@suse.de>
10057 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10058 (struct instantiate_cache_entry): New type.
10059 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10060 (struct instantiate_cache_type): New type.
10061 (set_instantiated_value, get_instantiated_value): Remove.
10062 (get_instantiated_value_entry): New function.
10063 (instantiate_scev_name): Use the new cache and adjust.
10064 (instantiate_scev_poly): Adjust.
10065 (instantiate_scev_binary): Likewise.
10066 (instantiate_array_ref): Likewise.
10067 (instantiate_scev_convert): Likewise.
10068 (instantiate_scev_not): Likewise.
10069 (instantiate_scev_3): Likewise.
10070 (instantiate_scev_2): Likewise.
10071 (instantiate_scev_r): Likewise.
10072 (instantiate_scev): Likewise.
10073 (resolve_mixers): Likewise.
10075 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10078 * lra-constraints.c (best_small_class_operands_num): Remove.
10079 (process_alt_operands): Remove small_class_operands_num. Take
10080 small classes operands into losers and only if the operand is not
10081 matched. Modify debugging output.
10082 (curr_insn_transform): Remove best_small_class_operands_num.
10085 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10087 * config/aarch64/aarch64-builtins.c
10088 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10089 * config/aarch64/aarch64-simd-builtins.def
10090 (reduc_splus_): Add new modes.
10091 (reduc_uplus_): New.
10092 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10093 (reduc_uplus_v4sf): Likewise.
10094 (reduc_splus_v4sf): Likewise.
10095 (aarch64_addv<mode>): Likewise.
10096 (reduc_uplus_<mode>): Likewise.
10097 (reduc_splus_<mode>): Likewise.
10098 (aarch64_addvv2di): Likewise.
10099 (reduc_uplus_v2di): Likewise.
10100 (reduc_splus_v2di): Likewise.
10101 (aarch64_addvv2si): Likewise.
10102 (reduc_uplus_v2si): Likewise.
10103 (reduc_splus_v2si): Likewise.
10104 (reduc_<sur>plus_<mode>): New.
10105 (reduc_<sur>plus_v2di): Likewise.
10106 (reduc_<sur>plus_v2si): Likewise.
10107 (reduc_<sur>plus_v4sf): Likewise.
10108 (aarch64_addpv4sf): Likewise.
10109 * config/aarch64/arm_neon.h
10110 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10111 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10112 add UNSPEC_SADDV, UNSPEC_UADDV.
10114 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10116 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10118 * config/aarch64/arm_neon.h
10119 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10121 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10123 * config/aarch64/aarch64-builtins
10124 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10126 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10128 * config/aarch64/aarch64-simd-builtins.def
10129 (reduc_smax_): New.
10130 (reduc_smin_): Likewise.
10131 (reduc_umax_): Likewise.
10132 (reduc_umin_): Likewise.
10133 (reduc_smax_nan_): Likewise.
10134 (reduc_smin_nan_): Likewise.
10137 (smax): Update for V2SF, V4SF and V2DF modes.
10140 (smin_nan): Likewise.
10141 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10142 (<su><maxmin><mode>3): ...This, refactor.
10143 (s<maxmin><mode>3): New.
10144 (<maxmin_uns><mode>3): Likewise.
10145 (reduc_<maxmin_uns>_<mode>): Refactor.
10146 (reduc_<maxmin_uns>_v4sf): Likewise.
10147 (reduc_<maxmin_uns>_v2si): Likewise.
10148 (aarch64_<fmaxmin><mode>: Remove.
10149 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10151 (vmin<q>_f<32,64>): Likewise.
10152 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10154 (su): Add mappings for smax, smin, umax, umin.
10156 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10157 (FMAXMIN): Rename as...
10158 (FMAXMIN_UNS): ...This.
10160 (fmaxminv): Likewise.
10161 (fmaxmin): Likewise.
10163 (maxmin_uns_op): Likewise.
10165 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10167 * config/aarch64/arm_neon.h
10168 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10169 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10170 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10172 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10174 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10175 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10177 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10179 * config/aarch64/aarch64-simd.md
10180 (vcond<mode>_internal): Handle special cases for constant masks.
10181 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10182 (vcondu<mode><mode>): Likewise.
10183 (vcond<v_cmp_result><mode>): New.
10185 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10187 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10188 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10189 * config/aarch64/aarch64-simd-builtins.def
10190 (cmeq): Update to BUILTIN_VALLDI.
10195 * config/aarch64/arm_neon.h
10196 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10197 to builtins or C as appropriate.
10199 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10201 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10203 (cmhi): Rename to...
10205 * config/aarch64/aarch64-simd.md
10206 (simd_mode): Add SF.
10207 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10208 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10209 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10210 (cstore<mode>_neg): ...This.
10211 * config/aarch64/iterators.md
10213 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10214 (COMPARISONS): New.
10215 (UCOMPARISONS): Likewise.
10216 (optab): Add missing comparisons.
10222 (VCMP_S): Likewise.
10223 (VCMP_U): Likewise.
10224 (V_cmp_result): Add DF, SF modes.
10225 (v_cmp_result): Likewise.
10227 (vmtype): Likewise.
10228 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10230 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10232 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10233 define_insn to define_insn_and_split.
10234 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10235 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10236 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10237 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10238 (thumb2_negscc): Likewise.
10240 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10242 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10244 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10246 * config/arm/thumb2.md: Remove trailing whitespaces.
10248 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10250 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10251 Use gen_int_mode rather than GEN_INT.
10253 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10255 * value-prof.c (stream_in_histogram_value): Remove the strayed
10258 2013-04-30 Richard Biener <rguenther@suse.de>
10260 PR middle-end/57122
10261 * cfghooks.c (split_edge): Properly check for the loop latch edge.
10263 2013-04-30 Richard Biener <rguenther@suse.de>
10265 PR middle-end/57107
10266 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
10268 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
10270 PR rtl-optimization/56957
10271 PR rtl-optimization/57105
10272 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
10273 variable. Use just INSN_UID for determining whether an insn
10274 should be only disconnected from the insn stream.
10275 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
10277 2013-04-30 Jakub Jelinek <jakub@redhat.com>
10279 PR tree-optimization/57104
10280 * tsan.c (instrument_expr): Don't instrument accesses to
10281 DECL_HARD_REGISTER VAR_DECLs.
10283 2013-04-30 Richard Biener <rguenther@suse.de>
10285 * function.h (loops_for_fn): New inline function.
10286 (set_loops_for_fn): Likewise.
10287 * cfgloop.h (place_new_loop): Add struct function parameter.
10288 (get_loop): Likewise.
10289 (get_loops): Likewise.
10290 (number_of_loops): Likewise.
10291 (fel_next): Adjust.
10292 (fel_init): Likewise.
10293 * cfg.c (get_loop_copy): Adjust.
10294 * cfgloop.c (flow_loops_dump): Likewise.
10295 (record_loop_exits): Likewise.
10296 (verify_loop_structure): Likewise.
10297 * cfgloopanal.c (mark_irreducible_loops): Likewise.
10298 (estimate_reg_pressure_cost): Likewise.
10299 (mark_loop_exit_edges): Likewise.
10300 * cfgloopmanip.c (place_new_loop): Likewise.
10301 (add_loop): Likewise.
10302 (duplicate_loop): Likewise.
10303 * graph.c (draw_cfg_nodes): Likewise.
10304 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
10305 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
10306 (extract_affine_chrec): Likewise.
10307 (build_scop_iteration_domain): Likewise.
10308 * graphite.c (graphite_initialize): Likewise.
10309 * ira-build.c (create_loop_tree_nodes): Likewise.
10310 (more_one_region_p): Likewise.
10311 (rebuild_regno_allocno_maps): Likewise.
10312 (mark_loops_for_removal): Likewise.
10313 (mark_all_loops_for_removal): Likewise.
10314 (remove_unnecessary_regions): Likewise.
10315 (ira_build): Likewise.
10316 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
10317 * loop-init.c (fix_loop_structure): Likewise.
10318 (gate_rtl_move_loop_invariants): Likewise.
10319 (gate_rtl_unswitch): Likewise.
10320 (gate_rtl_unroll_and_peel_loops): Likewise.
10321 (rtl_doloop): Likewise.
10322 * lto-streamer-in.c (input_cfg): Likewise.
10323 * lto-streamer-out.c (output_cfg): Likewise.
10324 * modulo-sched.c (sms_schedule): Likewise.
10325 * predict.c (tree_estimate_probability): Likewise.
10326 (tree_estimate_probability_driver): Likewise.
10327 (estimate_loops): Likewise.
10328 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
10329 (move_sese_region_to_fn): Likewise.
10330 (debug_loop_num): Likewise.
10331 * tree-chrec.c (chrec_evaluate): Likewise.
10332 (hide_evolution_in_other_loops_than_loop): Likewise.
10333 (chrec_component_in_loop_num): Likewise.
10334 (reset_evolution_in_loop): Likewise.
10335 (evolution_function_is_invariant_rec_p): Likewise.
10336 * tree-if-conv.c (main_tree_if_conversion): Likewise.
10337 * tree-inline.c (copy_loops): Likewise.
10338 (copy_cfg_body): Likewise.
10339 (tree_function_versioning): Likewise.
10340 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
10341 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
10343 (add_to_evolution_1): Likewise.
10344 (scev_const_prop): Likewise.
10345 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
10346 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
10347 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
10348 (tree_ssa_lim_initialize): Likewise.
10349 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
10350 (verify_loop_closed_ssa): Likewise.
10351 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
10352 (tree_ssa_loop_im): Likewise.
10353 (tree_ssa_loop_unswitch): Likewise.
10354 (tree_vectorize): Likewise.
10355 (check_data_deps): Likewise.
10356 (tree_ssa_loop_ivcanon): Likewise.
10357 (tree_ssa_loop_bounds): Likewise.
10358 (tree_complete_unroll): Likewise.
10359 (tree_complete_unroll_inner): Likewise.
10360 (tree_parallelize_loops): Likewise.
10361 (tree_ssa_loop_prefetch): Likewise.
10362 (tree_ssa_loop_ivopts): Likewise.
10363 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
10364 * tree-vectorizer.c (vectorize_loops): Likewise.
10366 2013-04-29 Mike Frysinger <vapier@gentoo.org>
10368 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
10369 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
10370 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
10371 with EABI_LINK_SPEC.
10373 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10376 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
10379 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
10382 * lra-constraints.c (process_alt_operands): Discourage a bit more
10383 using memory for pseudos. Print cost dump for alternatives.
10384 Modify cost values for conflicts with early clobbers.
10385 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
10387 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10390 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
10392 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10394 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
10395 from/to S register.
10396 (movdi_aarch64): Support LDR/STR from/to D register.
10398 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10400 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
10401 or higher by default.
10403 2013-04-29 Richard Biener <rguenther@suse.de>
10405 PR middle-end/57075
10406 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
10407 even if not adding abnormal edges for calls that can make
10410 2013-04-29 Richard Biener <rguenther@suse.de>
10412 PR middle-end/57103
10413 * tree-cfg.c (move_stmt_op): Fix condition under which to update
10415 (move_stmt_r): Remove redundant checking.
10417 2013-04-29 Teresa Johnson <tejohnson@google.com>
10420 * basic-block.h (apply_scale): New function.
10421 (apply_probability): Use apply_scale.
10422 * gimple-streamer-in.c (input_bb): Ditto.
10423 * lto-streamer-in.c (input_cfg): Ditto.
10424 * lto-cgraph.c (merge_profile_summaries): Ditto.
10425 * tree-optimize.c (execute_fixup_cfg): Ditto.
10426 * tree-inline.c (copy_bb): Update comment to use apply_scale.
10427 (copy_edges_for_bb): Ditto.
10428 (copy_cfg_body): Ditto.
10430 2013-04-29 Tom de Vries <tom@codesourcery.com>
10432 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
10433 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
10434 (tail_merge_optimize): Handle current_loops == NULL.
10436 2013-04-26 Jeff Law <law@redhat.com>
10438 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
10439 (simplify_cond_using_ranges): Generalize code to simplify
10440 COND_EXPRs where one argument is a constant and the other
10441 is an SSA_NAME created by an integral type conversion.
10443 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10445 * config/arm/arm.md (store_minmaxsi): Use only when
10446 optimize_insn_for_size_p.
10448 2013-04-29 Christian Bruel <christian.bruel@st.com>
10451 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
10453 2013-04-29 Richard Biener <rguenther@suse.de>
10455 PR middle-end/57089
10456 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
10457 loop tree make sure to schedule a fixup for the child as well.
10458 (expand_omp_for_generic): Properly add loops.
10459 (expand_omp_for_static_nochunk): Likewise.
10460 (expand_omp_for_static_chunk): Likewise.
10461 (expand_omp_for): For the degenerate case fixup loops.
10462 (expand_omp_sections): Fix default bb placement in loops.
10463 (expand_omp_atomic_pipeline): Properly add loops.
10465 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10467 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
10469 2013-04-29 Tom de Vries <tom@codesourcery.com>
10471 * tree-ssa-tail-merge.c: Update header comment.
10473 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10475 * config/aarch64/arm_neon.h
10476 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
10477 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
10478 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
10479 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
10480 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
10481 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
10482 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
10483 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
10485 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10487 * config/aarch64/aarch64-simd.md
10488 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
10489 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
10490 fix_trunc, fixuns_trunc.
10491 (ftrunc<VDQF:mode>2): New.
10492 * config/aarch64/iterators.md (optab): Add fix, fixuns.
10493 (fix_trunc_optab): New.
10495 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10497 * config/aarch64/aarch64-builtins.c
10498 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
10499 iceilf, lround, iroundf.
10501 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10504 * config/i386/i386.h (enum ix86_tune_indices)
10505 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
10506 New, split from X86_TUNE_INTER_UNIT_MOVES.
10507 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
10508 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
10509 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
10510 (TARGET_INTER_UNIT_MOVES): Remove.
10511 * config/i386/i386.c (initial_ix86_tune_features): Update.
10512 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
10513 (ix86_expand_convert_uns_didf_sse): Use
10514 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10515 (ix86_expand_vector_init_one_nonzero): Ditto.
10516 (ix86_expand_vector_init_interleave): Ditto.
10517 (inline_secondary_memory_needed): Return true for moves from SSE class
10518 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
10519 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
10520 * config/i386/constraints.md (Yi, Ym): Depend on
10521 TARGET_INTER_UNIT_MOVES_TO_VEC.
10522 (Yj, Yn): New constraints.
10523 * config/i386/i386.md (*movdi_internal): Change constraints of
10524 operand 1 from Yi to Yj and from Ym to Yn.
10525 (*movsi_internal): Ditto.
10526 (*movdf_internal): Ditto.
10527 (*movsf_internal): Ditto.
10528 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10529 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10530 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10531 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10532 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10533 * config/i386/sse.md (movdi_to_sse): Ditto.
10534 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10535 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10536 TARGET_INTER_UNIT_MOVES.
10537 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10538 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10539 instead of TARGET_INTER_UNIT_MOVES.
10540 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10541 operand 1 from Yi to Yj and from Ym to Yn.
10543 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10545 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10546 (float_truncate_hi_): Likewise.
10547 (float_extend_lo_): Likewise.
10548 (float_truncate_lo_): Likewise.
10549 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10550 (aarch64_float_extend_lo_v2df): Likewise.
10551 (vec_unpacks_hi_v4sf): Likewise.
10552 (aarch64_float_truncate_lo_v2sf): Likewise.
10553 (aarch64_float_truncate_hi_v4sf): Likewise.
10554 (vec_pack_trunc_v2df): Likewise.
10555 (vec_pack_trunc_df): Likewise.
10557 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10559 * config/aarch64/aarch64-builtins.c
10560 (aarch64_fold_builtin): Fold float conversions.
10561 * config/aarch64/aarch64-simd-builtins.def
10562 (floatv2si, floatv4si, floatv2di): New.
10563 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10564 * config/aarch64/aarch64-simd.md
10565 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10566 * config/aarch64/iterators.md (FLOATUORS): New.
10567 (optab): Add float, floatuns.
10568 (su_optab): Likewise.
10570 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10572 * config/aarch64/aarch64-builtins.c
10573 (aarch64_builtin_vectorized_function): Use new names for
10575 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10576 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10577 (fcvtzu): Split as...
10578 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10579 (fcvtas): Split as...
10580 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10581 (fcvtau): Split as...
10582 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10583 (fcvtps): Split as...
10584 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10585 (fcvtpu): Split as...
10586 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10587 (fcvtms): Split as...
10588 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10589 (fcvtmu): Split as...
10590 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10591 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10592 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10593 (lfrintnusf, lfrintnudf): Likewise.
10594 * config/aarch64/aarch64-simd.md
10595 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10597 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10598 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10599 (fcvt_pattern): Likewise.
10601 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10603 * config/aarch64/aarch64-simd.md
10604 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10605 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10607 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10609 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10610 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10611 (vrnd<a,m,n,p>_f32): Implement using builtins.
10612 (vrnd<i,x><q>_f<32, 64>): New.
10614 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10616 * config/aarch64/aarch64-builtins.c
10617 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10618 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10619 (frintz): Rename to...
10621 (frintp): Rename to...
10623 (frintm): Rename to...
10625 (frinti): Rename to...
10626 (nearbyint): ...this.
10627 (frintx): Rename to...
10629 (frinta): Rename to...
10631 * config/aarch64/aarch64-simd.md
10632 (aarch64_frint<frint_suffix><mode>): Delete.
10633 (<frint_pattern><mode>2): Convert to insn.
10634 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10635 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10636 (frint_pattern): Likewise.
10637 (frint_suffix): Likewise.
10639 2013-04-29 Richard Biener <rguenther@suse.de>
10641 PR tree-optimization/57081
10642 * loop-init.c: Include tree-flow.h.
10643 (loop_optimizer_finalize): Free number of iteration estimates.
10644 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10646 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10648 PR tree-optimization/57083
10649 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10650 non-singleton shift count range, zero extend low_bound for uns case.
10652 * config/i386/predicates.md (general_vector_operand): New predicate.
10653 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10654 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10655 if they aren't nonimmediate operands. If their original values
10656 satisfy const_vector_equal_evenodd_p, don't shift them.
10657 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10658 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10659 if not nonimmediate_operand.
10660 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10661 instead of register_operand.
10662 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10664 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10666 * stor-layout.c (finalize_size_functions): Allocate a structure and
10667 reset cfun before dumping the functions.
10669 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10671 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10674 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10675 use xop_pmacsdqh if uns_p.
10676 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10677 the immediate rotate count.
10679 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10681 * rtl.h (struct rtx_def): Add comment for field jump.
10682 (LRA_SUBREG_P): New macro.
10683 * recog.c (register_operand): Check LRA_SUBREG_P.
10684 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10686 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10687 (lra_final_code_change): Skip subreg change for operators.
10688 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10689 if there are no operand changes.
10690 * lra-constraints.c (curr_insn_set): New.
10691 (match_reload): Set LRA_SUBREG_P.
10692 (emit_spill_move): Ditto.
10693 (check_and_process_move): Use curr_insn_set. Process only single
10694 set insns. Don't initialize sec_mem_p and change_p.
10695 (simplify_operand_subreg): Use LRA_SUBREG_P.
10696 (reg_in_class_p): New function.
10697 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10698 of #ifdef. Add code to remove cycling.
10699 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10700 non-null disp. Reload inner instead of disp when base and index
10701 are null. Try to put lo_sum into register.
10702 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10703 (check_and_process_move): Move code for move cost check to
10704 simple_move_p. Remove equiv_substitution.
10705 (simple_move_p): New function.
10706 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10707 curr_insn_set. Call check_and_process_move only for single set
10708 insns. Use the new function. Move call of check_and_process_move
10709 after operand equiv substitution and address process.
10711 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10714 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10715 with nonlocal goto receivers or returns twice calls, ignore
10716 unininitialized values from abnormal edges to nl goto receiver
10717 or returns twice call.
10719 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10721 PR tree-optimization/57051
10722 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10723 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10726 2013-04-26 Richard Biener <rguenther@suse.de>
10728 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10729 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10730 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10731 (fixup_loop_arrays_after_move): New function.
10732 (move_sese_region_to_fn): Properly outline the loop tree parts
10733 of the SESE region.
10735 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10737 * config/i386/i386.md (type, unit): Fix long lines.
10739 2013-04-26 Richard Biener <rguenther@suse.de>
10741 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10742 (lto-streamer-out.o): Likewise.
10743 * cfgloop.c (init_loops_structure): Export, add struct function
10744 argument and adjust.
10745 (flow_loops_find): Adjust.
10746 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10747 (init_loops_structure): Declare.
10748 * lto-streamer-in.c: Include cfgloop.h.
10749 (input_cfg): Input the loop tree.
10750 * lto-streamer-out.c: Include cfgloop.h.
10751 (output_cfg): Output the loop tree.
10752 (output_struct_function_base): Do not drop PROP_loops.
10754 2013-03-26 Richard Biener <rguenther@suse.de>
10756 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10757 (pass_build_cfg): Provide PROP_loops.
10758 (move_sese_region_to_fn): Remove loops that are outlined into fn
10760 * tree-inline.c: Include cfgloop.h.
10761 (initialize_cfun): Do not drop PROP_loops.
10762 (copy_loops): New function.
10763 (copy_cfg_body): Copy loop structure.
10764 (tree_function_versioning): Initialize destination loop tree.
10765 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10766 (pass_parallelize_loops): Do IL verification.
10767 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10768 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10769 the CFG make sure we fixup loops as well.
10770 * tree-ssa-tail-merge.c: Include cfgloop.h.
10771 (replace_block_by): When merging loop latches mark loops for fixup.
10772 * lto-streamer-out.c (output_struct_function_base): Drop
10773 PROP_loops for now.
10774 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10775 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10776 * ipa-split.c: Include cfgloop.h.
10777 (split_function): Add the new return block to the loop tree root.
10778 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10779 whether we have removed the forwarder block.
10780 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10781 * cfgloop.h (place_new_loop): Declare.
10782 * cfgloopmanip.c (place_new_loop): Export.
10783 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10784 (tree-switch-conversion.o): Likewise.
10785 (tree-complex.o): Likewise.
10786 (tree-inline.o): Likewise.
10787 (tree-ssa-tailmerge.o): Likewise.
10788 (ipa-split.o): Likewise.
10789 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10790 (tree-ssa-copy.o): Likewise.
10791 * tree-switch-conversion.c: Include cfgloop.h
10792 (process_switch): If we emit a bit-test cascade, schedule loops
10794 * tree-complex.c: Include cfgloop.h.
10795 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10796 * asan.c: Include cfgloop.h.
10797 (create_cond_insert_point): Properly add new basic-blocks to
10798 loops, schedule loop fixup.
10799 * cfgloop.c (verify_loop_structure): Check that looks are not
10801 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10803 (expand_omp_for_generic): Likewise.
10804 (expand_omp_sections): Likewise.
10805 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10806 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10807 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10808 is initialized, not when loops are present.
10809 * tree-parloops.c (parallelize_loops): Remove checking here.
10810 * passes.c (init_optimization_passes): Schedule a copy-propagation
10811 pass before complete unrolling of inner loops.
10813 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10815 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10816 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10817 (colorize_init): Add argument to _WIN32 version.
10818 * toplev.c: Include diagnostic-color.h.
10819 (process_options): Default to -fdiagnostics-color=auto if
10820 GCC_COLORS env var is in the environment.
10821 * common.opt (fdiagnostics-color=): Add Var and Init.
10822 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10823 env var is in the environment, the default is auto rather than never.
10825 * diagnostic.h (file_name_as_prefix): Add context argument.
10826 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10827 the string as locus.
10828 * langhooks.c (lhd_print_error_function): Adjust caller.
10830 2013-04-25 Lawrence Crowl <crowl@google.com>
10832 * var-tracking.c (shared_hash_def::htab):
10833 Change type to hash_table. Update dependent calls and types.
10835 2013-04-25 Lawrence Crowl <crowl@google.com>
10837 * Makefile.in: Update as needed below.
10839 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10840 Move declaration to after the type's method definitons.
10842 * attribs.c (htab_t scoped_attributes::attribute_hash):
10843 Change type to hash_table. Update dependent calls and types.
10845 * bitmap.c (htab_t bitmap_desc_hash):
10846 Change type to hash_table. Update dependent calls and types.
10848 * cselib.c (htab_t cselib_hash_table):
10849 Change type to hash_table. Update dependent calls and types.
10851 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10852 (hash_string_slot_node): Move implementation into lto-streamer.h
10853 struct string_slot_hasher.
10854 (eq_string_slot_node): Likewise.
10856 * data-streamer-out.c: Update output_block::string_hash_table
10857 dependent calls and types.
10859 * dwarf2cfi.c (htab_t trace_index):
10860 Change type to hash_table. Update dependent calls and types.
10862 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10863 Change type to hash_table. Update dependent calls and types.
10864 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10865 (htab_t optimize_external_refs::map): Likewise.
10866 (htab_t output_comp_unit::extern_map): Likewise.
10867 (htab_t output_comdat_type_unit::extern_map): Likewise.
10868 (htab_t output_macinfo::macinfo_htab): Likewise.
10869 (htab_t optimize_location_lists::htab): Likewise.
10870 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10872 * except.c (htab_t ehspec_hash_type):
10873 Change type to hash_table. Update dependent calls and types.
10874 (assign_filter_values::ttypes): Likewise.
10875 (assign_filter_values::ehspec): Likewise.
10876 (sjlj_assign_call_site_values::ar_hash): Likewise.
10877 (convert_to_eh_region_ranges::ar_hash): Likewise.
10879 * gcse.c (htab_t pre_ldst_table):
10880 Change type to hash_table. Update dependent calls and types.
10882 * ggc-common.c (htab_t saving_htab):
10883 Change type to hash_table. Update dependent calls and types.
10884 (htab_t loc_hash): Likewise.
10885 (htab_t ptr_hash): Likewise.
10886 (call_count): Rename ggc_call_count.
10887 (call_alloc): Rename ggc_call_alloc.
10888 (loc_descriptor): Rename make_loc_descriptor.
10889 (add_statistics): Rename ggc_add_statistics.
10891 * ggc-common.c (saving_htab):
10892 Change type to hash_table. Update dependent calls and types.
10894 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10895 (push_gimplify_context): Likewise.
10896 (pop_gimplify_context): Likewise.
10897 (struct gimple_temp_hash_elt): Added.
10898 (struct gimplify_hasher): Likewise.
10899 (struct gimplify_ctx.temp_htab):
10900 Change type to hash_table. Update dependent calls and types.
10902 * gimple-fold.c: Include gimplify-ctx.h.
10904 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10905 Change type to hash_table. Update dependent calls and types.
10906 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10907 avoid potential global name collision.
10909 * gimplify.c: Include gimplify-ctx.h.
10910 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10911 (htab_t gimplify_ctx::temp_htab):
10912 Update dependent calls and types for new type hash_table.
10913 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10914 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10916 * gimplify-ctx.h: New.
10917 (struct gimple_temp_hash_elt): Move from gimplify.c.
10918 (class gimplify_hasher): New.
10919 (struct gimplify_ctx): Move from gimple.h.
10920 (htab_t gimplify_ctx::temp_htab):
10921 Change type to hash_table. Update dependent calls and types.
10923 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10924 (htab_t ivs_params::newivs_index):
10925 Change type to hash_table. Update dependent calls and types.
10926 (htab_t ivs_params::params_index): Likewise.
10927 (htab_t print_generated_program::params_index): Likewise.
10928 (htab_t gloog::newivs_index): Likewise.
10929 (htab_t gloog::params_index): Likewise.
10931 * graphite.c: Include graphite-htab.h.
10932 4htab_t graphite_transform_loops::bb_pbb_mapping):
10933 Change type to hash_table. Update dependent calls and types.
10935 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10936 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10937 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10939 * graphite-dependences.c: Include graphite-htab.h.
10940 (loop_is_parallel_p): Change hash table type of parameter.
10942 * graphite-htab.h: New.
10943 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10944 (extern find_pbb_via_hash): Move from graphite-poly.h.
10945 (extern loop_is_parallel_p): Move from graphite-poly.h.
10946 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10948 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10949 (extern loop_is_parallel_p): Move to graphite-htab.h.
10950 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10952 * haifa-sched.c (htab_t delay_htab):
10953 Change type to hash_table. Update dependent calls and types.
10954 (htab_t delay_htab_i2): Likewise.
10956 * ira-color.c (htab_t allocno_hard_regs_htab):
10957 Change type to hash_table. Update dependent calls and types.
10959 * ira-costs.c (htab_t cost_classes_htab):
10960 Change type to hash_table. Update dependent calls and types.
10962 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10963 Change type to hash_table. Update dependent calls and types.
10965 * loop-iv.c (htab_t bivs):
10966 Change type to hash_table. Update dependent calls and types.
10968 * loop-unroll.c (htab_t opt_info::insns_to_split):
10969 Change type to hash_table. Update dependent calls and types.
10970 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10972 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10973 (struct string_slot_hasher): New.
10974 (htab_t output_block::string_hash_table):
10975 Change type to hash_table. Update dependent calls and types.
10977 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10978 (htab_t file_name_hash_table):
10979 Change type to hash_table. Update dependent calls and types.
10981 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10984 * lto-streamer.c (htab_t tree_htab):
10985 Change type to hash_table. Update dependent calls and types.
10987 * omp-low.c: Include gimplify-ctx.h.
10989 * passes.c (htab_t name_to_pass_map):
10990 Change type to hash_table. Update dependent calls and types.
10991 (pass_traverse): Rename to passes_pass_traverse.
10993 * plugin.c (htab_t event_tab):
10994 Change type to hash_table. Update dependent calls and types.
10996 * postreload-gcse.c (htab_t expr_table):
10997 Change type to hash_table. Update dependent calls and types.
10998 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
11000 * sese.c (debug_rename_map_1): Make extern.
11001 (htab_t copy_bb_and_scalar_dependences::rename_map):
11002 Change type to hash_table. Update dependent calls and types.
11004 * sese.h (extern debug_rename_map): Move to .c file.
11006 * store-motion.c (htab_t store_motion_mems_table):
11007 Change type to hash_table. Update dependent calls and types.
11009 * trans-mem.c (htab_t tm_new_mem_hash):
11010 Change type to hash_table. Update dependent calls and types.
11012 * tree-browser.c (htab_t TB_up_ht):
11013 Change type to hash_table. Update dependent calls and types.
11015 * tree-cfg.c (htab_t discriminator_per_locus):
11016 Change type to hash_table. Update dependent calls and types.
11018 * tree-complex.c: Include tree-hasher.h
11019 (htab_t complex_variable_components):
11020 Change type to hash_table. Update dependent calls and types.
11022 * tree-eh.c (htab_t finally_tree):
11023 Change type to hash_table. Update dependent calls and types.
11025 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
11026 struct int_tree_hasher.
11027 (extern int_tree_map_eq): Likewise.
11028 (uid_decl_map_hash): Removed.
11029 (extern decl_tree_map_eq): Likewise.
11031 * tree-hasher.h: New.
11032 (struct int_tree_hasher): New.
11033 (typedef int_tree_htab_type): New.
11035 * tree-inline.c: Include gimplify-ctx.h.
11037 * tree-mudflap.c: Include gimplify-ctx.h.
11039 * tree-parloops.c: Include tree-hasher.h.
11040 (htab_t eliminate_local_variables_stmt::decl_address):
11041 Change type to hash_table. Update dependent calls and types.
11042 (htab_t separate_decls_in_region::decl_copies): Likewise.
11044 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11045 Change type to hash_table. Update dependent calls and types.
11047 * tree-sra.c (candidates):
11048 Change type to hash_table. Update dependent calls and types.
11050 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11052 (int_tree_map_hash): Likewise.
11054 * tree-ssa-dom.c (htab_t avail_exprs):
11055 Change type to hash_table. Update dependent calls and types.
11057 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11058 Change type to hash_table. Update dependent calls and types.
11060 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11061 Change type to hash_table. Update dependent calls and types.
11063 * tree-ssa-phiopt.c (seen_ssa_names):
11064 Change type to hash_table. Update dependent calls and types.
11066 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11067 Change type to hash_table. Update dependent calls and types.
11069 * tree-ssa-uncprop.c (equiv):
11070 Change type to hash_table. Update dependent calls and types.
11072 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11074 PR rtl-optimization/57003
11075 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11076 call note_stores with kill_clobbered_value callback again after
11077 killing regs_invalidated_by_call.
11079 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11081 * config/aarch64/aarch64-simd.md
11082 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11083 (aarch64_simd_bsl<mode>): Likewise.
11084 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11086 2013-04-25 Marek Polacek <polacek@redhat.com>
11088 PR tree-optimization/57066
11089 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11091 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11093 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11095 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11097 * config/aarch64/aarch64-builtins.c
11098 (aarch64_fold_builtin): New.
11099 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11100 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11101 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11102 * config/aarch64/arm_neon.h
11103 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11105 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11106 Tejas Belagod <tejas.belagod@arm.com>
11108 * config/aarch64/aarch64-builtins.c
11109 (aarch64_gimple_fold_builtin): New.
11110 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11111 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11112 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11113 (addvv4sf): Update.
11114 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11116 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11118 * config/aarch64/aarch64.md
11119 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11121 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11123 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11124 (*ngcsi_uxtw): New pattern.
11126 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11127 Julian Brown <julian@codesourcery.com>
11129 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11130 (TB_DREG): Add T_V4HF.
11131 (v4hf_UP): New macro.
11132 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11133 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11134 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11136 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11137 (arm_vector_mode_supported_p): Handle V4HF.
11138 (arm_mangle_map): Handle V4HFmode.
11139 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11140 * config/arm/arm_neon_builtins.def: Add entries for
11141 vcvtv4hfv4sf, vcvtv4sfv4hf.
11142 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11143 (neon_vcvtv4hfv4sf): Likewise.
11144 * config/arm/neon-gen.ml: Handle half-precision floating point
11146 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11147 * config/arm/arm_neon.h: Regenerate.
11148 * config/arm/neon.ml (type elts): Add F16.
11149 (type vectype): Add T_float16x4, T_floatHF.
11150 (type vecmode): Add V4HF.
11151 (type features): Add Requires_FP_bit feature.
11152 (elt_width): Handle F16.
11153 (elt_class): Likewise.
11154 (elt_of_class_width): Likewise.
11155 (mode_of_elt): Refactor.
11156 (type_for_elt): Handle F16, fix error messages.
11157 (vectype_size): Handle T_float16x4.
11158 (vcvt_sh): New function.
11159 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11160 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11161 (string_of_mode): Handle V4HF.
11162 * doc/arm-neon-intrinsics.texi: Regenerate.
11164 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11166 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11167 format specifier in 'X' case.
11169 2013-04-25 Alan Modra <amodra@gmail.com>
11172 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11173 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11174 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11175 Repeat for many other rotate/shift and mask patterns using subregs.
11176 Name lshiftrt insns.
11177 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11178 on WORDS_BIG_ENDIAN.
11180 2013-04-25 Alan Modra <amodra@gmail.com>
11182 * config.gcc: Support little-endian powerpc-linux targets.
11183 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11184 (LINK_OS_LINUX_SPEC): Define.
11185 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11186 Preserve MASK_LITTLE_ENDIAN.
11187 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11188 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11189 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11190 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11191 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11192 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11193 Correct fp word order for little-endian. Don't shift toc entries
11194 smaller than a word for little-endian.
11195 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11196 (bswapdi2 splits): Correct low-part subreg for little-endian.
11197 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11198 low/high where such is correct only for be.
11199 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11200 little-endian for -mcall-aixdesc.
11202 2013-04-25 Alan Modra <amodra@gmail.com>
11204 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11205 replace_equiv_address_nv.
11207 2013-04-25 Alan Modra <amodra@gmail.com>
11209 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11211 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11214 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11215 * rtl.h (struct rtx_def): ...
11217 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11219 PR rtl-optimizations/57046
11220 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11221 for multi-reg splits.
11223 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11225 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11227 2013-04-24 Sterling Augustine <saugustine@google.com>
11229 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11230 (comp_dir_string, debug_str_dwo_section): New.
11231 (DEBUG_STR_DWO_SECTION): Rename to ...
11232 (DEBUG_DWO_STR_SECTION): ... this.
11233 (DEBUG_NORM_STR_SECTION): Delete.
11234 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11235 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11236 (find_AT_string): Move most logic to ...
11237 (find_AT_string_in_table): ... here. New.
11238 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11239 add_skeleton_AT_string. Delete logic.
11240 (output_skeleton_debug_sections): Remove call to
11241 add_top_level_skeleton_die_attrs.
11242 (add_comp_dir_attribute): Move logic to comp_dir_string.
11243 (dwarf2out_init): Initialize debug_str_dwo_section.
11244 (output_indirect_string): Call find_string_form.
11245 (output_indirect_strings): Rewrite.
11246 (prune_unused_types): Empty skeleton_debug_str_hash.
11247 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11248 (dwarf2out_finish): Call output_indirect_strings.
11250 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11252 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11254 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11256 * rtl.h (struct rtx_def): Add comment for field jump.
11257 (LRA_SUBREG_P): New macro.
11258 * recog.c (register_operand): Check LRA_SUBREG_P.
11259 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11261 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11262 (lra_final_code_change): Skip subreg change for operators.
11263 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11264 if there are no operand changes.
11265 * lra-constraints.c (curr_insn_set): New.
11266 (match_reload): Set LRA_SUBREG_P.
11267 (emit_spill_move): Ditto.
11268 (check_and_process_move): Use curr_insn_set. Process only single
11269 set insns. Don't initialize sec_mem_p and change_p.
11270 (simplify_operand_subreg): Use LRA_SUBREG_P.
11271 (reg_in_class_p): New function.
11272 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11273 of #ifdef. Add code to remove cycling.
11274 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11275 non-null disp. Reload inner instead of disp when base and index
11276 are null. Try to put lo_sum into register.
11277 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11278 (check_and_process_move): Move code for move cost check to
11279 simple_move_p. Remove equiv_substitution.
11280 (simple_move_p): New function.
11281 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11282 curr_insn_set. Call check_and_process_move only for single set
11283 insns. Use the new function. Move call of check_and_process_move
11284 after operand equiv substitution and address process.
11286 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
11288 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
11289 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
11290 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
11292 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11294 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
11296 2013-04-24 Marek Polacek <polacek@redhat.com>
11298 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
11299 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
11300 (select_loops_exit_conditions): Likewise.
11301 (number_of_iterations_for_all_loops): Likewise.
11302 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
11303 (scev_analysis): Likewise.
11305 2013-04-02 Catherine Moore <clm@codesourcery.com>
11306 Chao-ying Fu <fu@mips.com>
11308 * config/mips/micromips.md (jraddiusp): New pattern.
11309 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
11310 instruction if possible.
11312 2013-04-24 Alan Modra <amodra@gmail.com>
11314 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
11316 2013-04-24 Julian Brown <julian@codesourcery.com>
11317 Chung-Lin Tang <cltang@codesourcery.com>
11319 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
11320 dependency behavior in enumeration type DIE generation. Add TODO note
11321 to comments about future DW_FORM_sdata/udata re-work of related code.
11323 2013-04-23 Lawrence Crowl <crowl@google.com>
11325 * Makefile.in: Update as needed below.
11327 * hash-table.h (class hash_table):
11328 Correct many methods with parameter types compare_type to the correct
11329 value_type. (Correct code was unlikely to notice the change.)
11330 (hash_table::elements_with_deleted) New.
11331 (class hashtable::iterator): New.
11332 (hashtable::begin()): New.
11333 (hashtable::end()): New.
11334 (FOR_EACH_HASH_TABLE_ELEMENT): New.
11336 * statistics.c (statistics_hashes):
11337 Change type to hash_table. Update dependent calls and types.
11339 * tree-into-ssa.c (var_infos):
11340 Change type to hash_table. Update dependent calls and types.
11342 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
11343 Change type to hash_table. Update dependent calls and types.
11345 * tree-ssa-loop-im.c (struct mem_ref.refs):
11346 Change type to hash_table. Update dependent calls and types.
11348 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
11349 Change type to hash_table. Update dependent calls and types.
11351 * tree-ssa-sccvn.c (vn_tables_s::nary):
11352 Change type to hash_table. Update dependent calls and types.
11353 (vn_tables_s::phis): Likewise.
11354 (vn_tables_s::references): Likewise.
11356 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
11357 (vn_reference_eq): Update parameter and return types.
11359 * tree-ssa-structalias.c (pointer_equiv_class_table):
11360 Change type to hash_table. Update dependent calls and types.
11361 (location_equiv_class_table): Likewise.
11363 * tree-vect-data-refs.c: Consequential changes for making
11364 peeling a hash_table.
11366 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
11367 (destroy_loop_vec_info): Dependent hash_table update.
11369 * tree-vectorizer.h (peeling_htab):
11370 Change type to hash_table. Update dependent calls and types.
11372 2013-04-23 Shiva Chen <shiva0217@gmail.com>
11374 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
11375 to check the register content is equal or not.
11376 * lra-constraints.c (match_reload): Use lra_assign_reg_val
11377 to assign register content record.
11378 * lra-eliminations.c (update_reg_eliminate): Use
11379 lra_update_reg_val_offset to update register content offset.
11380 * lra-int.h (struct lra_reg): Add offset member.
11381 (lra_reg_val_equal_p): New static inline function.
11382 (lra_update_reg_val_offset): New static inline function.
11383 (lra_assign_reg_val): New static inline function.
11384 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
11385 to assign register content record.
11386 (initialize_lra_reg_info_element): Initial offset to zero.
11388 2013-04-23 Catherine Moore <clm@codesourcery.com>
11390 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
11391 operands. Record compression.
11393 2013-04-23 Xinliang David Li <davidxl@google.com>
11395 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
11397 2013-04-23 Richard Biener <rguenther@suse.de>
11399 PR middle-end/57036
11400 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
11401 parameter, only add abnormal goto edges from the copied body
11402 if the call could perform abnormal gotos.
11403 (copy_cfg_body): Adjust.
11405 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
11407 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
11409 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
11411 * coretypes.h (gimple_stmt_iterator): Add struct to make
11414 2013-04-23 Richard Biener <rguenther@suse.de>
11416 PR tree-optimization/57026
11417 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
11418 from SSA names occuring in abnormal PHI nodes.
11420 2013-04-22 Andi Kleen <ak@linux.intel.com>
11422 * lto/lto.c (print_lto_report_1): Fix LTO report names.
11424 2013-04-22 Andi Kleen <ak@linux.intel.com>
11426 * lto/lto.c (print_lto_report_1): Declare early.
11427 (read_cgraph_and_symbols): Call print_lto_report_1 early.
11429 2013-04-22 Andi Kleen <ak@linux.intel.com>
11431 * common.opt (-flto-report-wpa): Add.
11432 * doc/invoke.texi (-flto-report-wpa): Add.
11433 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
11436 2013-04-22 Xinliang David Li <davidxl@google.com>
11438 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
11439 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
11440 * Makefile.in: New dependency
11442 David Daney <ddaney.cavm@gmail.com>
11444 * configure.ac (gcc_cv_as_micromips_support): Use the
11445 --fatal-warnings option.
11446 * configure: Regenerate.
11448 2013-04-22 Marek Polacek <polacek@redhat.com>
11451 * tsan.c (instrument_expr): Don't instrument expression
11452 in case its size is zero.
11454 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
11458 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
11460 * config/alpha/alpha.c (TARGET_LRA_P): New define.
11462 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11464 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
11465 (gimple_stmt_iterator): New typedef.
11466 * gimple.h (gimple_stmt_iterator): Rename to...
11467 (gimple_stmt_iterator_d): ... This.
11468 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
11469 trees be valid for GIMPLE and GENERIC.
11470 (TARGET_GIMPLE_FOLD_BUILTIN): New.
11471 * gimple-fold.c (gimple_fold_call): Call target hook
11472 gimple_fold_builtin.
11473 * hooks.c (hook_bool_gsiptr_false): New.
11474 * hooks.h (hook_bool_gsiptr_false): New.
11475 * target.def (fold_stmt): New.
11476 * doc/tm.texi: Regenerate.
11478 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
11481 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
11482 a set sp if no stack realignment.
11484 2013-04-22 Nick Clifton <nickc@redhat.com>
11486 * config.gcc (tilegx-linux): Extend extra_objs rather than
11488 (tilepro-linux): Likewise.
11490 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11492 * config/aarch64/aarch64-builtins.c
11494 (CF0, CF1, CF2, CF3, CF4, CF10): New.
11495 (VAR<1-12>): Add MAP parameter.
11496 (BUILTIN_*): Likewise.
11497 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
11498 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
11499 (aarch64_ushl_n<mode>): Likewise.
11500 (aarch64_sshr_n<mode>): Likewise.
11501 (aarch64_ushr_n<mode>): Likewise.
11502 (aarch64_<maxmin><mode>): Likewise.
11503 (aarch64_sqrt<mode>): Likewise.
11504 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
11505 (vshr<q>_n_*): Likewise.
11507 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11509 * config/aarch64/aarch64-builtins.c
11510 (aarch64_simd_builtin_type_mode): Handle SF types.
11512 (BUILTIN_GPF): Define.
11513 (aarch64_init_simd_builtins): Handle SF types.
11514 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
11515 (frecps): Likewise.
11516 (frecpx): Likewise.
11517 * config/aarch64/aarch64-simd.md
11518 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
11519 (aarch64_frecpe<mode>): New.
11520 (aarch64_frecps<mode>): Likewise.
11521 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
11522 (v8type): Add frecp<esx>.
11523 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
11524 (aarch64_frecps<mode>): Likewise.
11525 * config/aarch64/iterators.md (FRECP): New.
11526 (frecp_suffix): Likewise.
11527 * config/aarch64/arm_neon.h
11528 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11530 2013-04-22 Christian Bruel <christian.bruel@st.com>
11533 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11534 (REG_CLASS_NAMES): Idem.
11535 (REG_CLASS_CONTENTS): Idem.
11536 (REGCLASS_HAS_FP_REG): Idem.
11537 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11538 (sh_conditional_register_usage): Idem.
11540 2013-04-21 Jeff Law <law@redhat.com>
11542 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11543 (ssa_forward_propagate_and_combine): Use it.
11545 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11547 * lra.c: Update the flow chart diagram.
11549 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11551 PR rtl-optimization/56847
11552 * lra-constraints.c (process_alt_operands): Discourage alternative
11553 with non-matche doffsettable memory constraint fro memory with
11556 2013-04-19 Richard Biener <rguenther@suse.de>
11558 PR tree-optimization/56982
11559 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11561 * gimplify.c (gimplify_call_expr): Notice special calls.
11562 (gimplify_modify_expr): Likewise.
11563 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11564 abnormal control flow receivers.
11565 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11566 in the same way as cfun->has_nonlocal_labels.
11567 (gimple_purge_dead_abnormal_call_edges): Likewise.
11568 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11569 receivers start a basic-block.
11571 2013-04-19 Richard Biener <rguenther@suse.de>
11573 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11575 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11576 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11577 (SLP_TREE_LOAD_PERMUTATION): Add.
11578 (vect_transform_slp_perm_load): Adjust prototype.
11579 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11580 (vect_free_slp_instance): Likewise.
11581 (vect_create_new_slp_node): Likewise.
11582 (vect_supported_slp_permutation_p): Remove.
11583 (vect_slp_rearrange_stmts): Adjust.
11584 (vect_supported_load_permutation_p): Likewise. Inline
11585 vect_supported_slp_permutation_p here.
11586 (vect_analyze_slp_instance): Compute load permutations per
11587 slp node instead of per instance.
11588 (vect_get_slp_defs): Adjust.
11589 (vect_transform_slp_perm_load): Likewise.
11590 (vect_schedule_slp_instance): Remove redundant code.
11591 (vect_schedule_slp): Remove hack for PR56270, add it ...
11592 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11593 CSE loads for SLP. Adjust.
11595 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11597 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11598 spelling in two comments.
11600 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11603 * config/arm/arm.c (load_multiple_sequence): Require SP
11604 as base register for loads if SP is in the register list.
11606 2013-04-19 Martin Jambor <mjambor@suse.cz>
11608 PR tree-optimization/56718
11609 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11610 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11611 and made public. Adjusted all callers.
11612 (ipa_intraprocedural_devirtualization): New function.
11613 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11614 (ipa_intraprocedural_devirtualization): Likewise.
11615 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11617 2013-04-19 Richard Biener <rguenther@suse.de>
11619 PR tree-optimization/57000
11620 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11622 2013-04-19 Terry Guo <terry.guo@arm.com>
11624 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11626 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11627 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11628 (cortex_m4_fmacs): Use new reservations.
11629 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11631 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11633 PR rtl-optimization/56999
11634 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11636 (lra_coalesce): Remove split_origin_bitmap and related code.
11637 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11638 ranges if necessary.
11640 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11642 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11644 (ix86_expand_call): Remove clobbered_registers array and use
11645 x86_64_ms_sysv_extra_clobbered_registers instead.
11646 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11648 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11650 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11651 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11652 (*call_value_rex64_ms_sysv): Ditto.
11654 2013-04-18 Cary Coutant <ccoutant@google.com>
11656 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11657 parent instead of die_mark.
11659 2013-04-18 Diego Novillo <dnovillo@google.com>
11661 * gimple.c (create_gimple_tmp): New.
11662 (get_expr_type): New.
11663 (build_assign): New.
11664 (build_type_cast): New.
11665 * gimple.h (enum ssa_mode): Define.
11666 (gimple_seq_set_location): New.
11667 * asan.c (build_check_stmt): Change some gimple_build_* calls
11668 to use build_assign and build_type_cast.
11670 2013-04-18 Richard Biener <rguenther@suse.de>
11672 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11673 handle negative step. Remove redundant checks.
11674 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11675 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11676 for negative step and grouped loads fail to vectorize.
11678 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11680 * emit-rtl.c (reset_insn_used_flags): New function.
11681 (reset_all_used_flags): Use it.
11682 (verify_insn_sharing): New function.
11683 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11685 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11687 PR tree-optimization/56984
11688 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11689 and (x >> M) >= N don't register any assertion if N << M is the
11692 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11694 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11695 do not delete_insn insns that have not yet been emitted, only
11696 unlink them with remove_insn.
11697 * df-scan.c (df_insn_delete): Revert r197492.
11699 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11701 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11702 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11704 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11706 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11707 define_insn_and_split.
11708 (and_scc,ior_scc,negscc): Likewise.
11709 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11711 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11713 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11714 can benefit from using a sequence of LDRD instructions in epilogue
11715 instead of a single LDM instruction.
11717 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11720 * doc/extend.texi: Fix typo.
11722 2013-04-17 Richard Biener <rguenther@suse.de>
11724 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11725 (vect_build_slp_tree): ... here.
11726 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11727 match. Remove special-casing of mismatched complex loads.
11728 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11729 re-try the match with swapped commutative operands.
11730 (vect_supported_load_permutation_p): Remove special-casing of
11731 mismatched complex loads.
11732 (vect_analyze_slp_instance): Adjust.
11734 2013-04-17 Richard Biener <rguenther@suse.de>
11736 PR rtl-optimization/56921
11737 * cfgloop.h (struct loop): Add simple_loop_desc member.
11738 (struct niter_desc): Mark with GTY(()).
11739 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11740 * loop-iv.c (get_simple_loop_desc): Likewise.
11741 (free_simple_loop_desc): Likewise.
11744 2013-04-16 Richard Biener <rguenther@suse.de>
11746 PR rtl-optimization/56921
11747 * loop-init.c (pass_rtl_move_loop_invariants): Add
11748 TODO_do_not_ggc_collect to todo_flags_finish.
11749 (pass_rtl_unswitch): Same.
11750 (pass_rtl_unroll_and_peel_loops): Same.
11751 (pass_rtl_doloop): Same.
11753 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11755 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11756 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11757 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11759 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11760 * tree-streamer.c (record_common_node): Adjust reference in comment.
11762 2013-04-17 Terry Guo <terry.guo@arm.com>
11764 * config/arm/cortex-m4.md: Add a new bypass.
11766 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11768 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11770 (*subs_<optab><mode>_multp2): New pattern.
11771 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11772 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11774 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11776 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11777 (*subs_mul_imm_<mode>): New pattern.
11779 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11782 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11783 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11784 (vsx_movti_32bit): Change j->wa to O->wa.
11786 2013-04-16 Richard Biener <rguenther@suse.de>
11788 PR rtl-optimization/56921
11789 * loop-init.c (pass_rtl_move_loop_invariants): Add
11790 TODO_do_not_ggc_collect to todo_flags_finish.
11791 (pass_rtl_unswitch): Same.
11792 (pass_rtl_unroll_and_peel_loops): Same.
11793 (pass_rtl_doloop): Same.
11795 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11797 * config/arm/arm.c (emit_multi_reg_push): New declaration
11798 for an existing function.
11799 (arm_emit_strd_push): New function.
11800 (arm_expand_prologue): Used here.
11801 (arm_emit_ldrd_pop): New function.
11802 (arm_expand_epilogue): Used here.
11803 (arm_get_frame_offsets): Update condition.
11804 (arm_emit_multi_reg_pop): Add a special case for load of a single
11805 register with writeback.
11807 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11809 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11812 2013-04-16 Richard Biener <rguenther@suse.de>
11814 PR tree-optimization/56756
11815 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11816 (first_mem_ref_loc): New.
11817 (execute_sm): Place the load temporarily before a previous
11818 access instead of in the latch edge to ensure its SSA dependencies
11819 are defined at points dominating the load.
11821 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11823 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11824 correct fix by moving header and footer insn to the footer of
11825 the merged basic block. Clear BB_END of the merged-away block.
11827 PR middle-end/43631
11828 * emit-rtl.c (make_note_raw): New function.
11829 (link_insn_into_chain): New static inline function.
11830 (add_insn): Use it.
11831 (add_insn_before, add_insn_after): Factor insn chain linking code...
11832 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11833 using link_insn_into_chain.
11834 (note_outside_basic_block_p): New helper function for emit_note_after
11835 and emit_note_before.
11836 (emit_note_after): Use nobb variant of add_insn_after if the note
11837 should not be contained in a basic block.
11838 (emit_note_before): Use nobb variant of add_insn_before if the note
11839 should not be contained in a basic block.
11840 (emit_note_copy): Use make_note_raw.
11841 (emit_note): Likewise.
11842 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11843 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11844 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11845 the moved barrier the tail of the basic block it follows.
11846 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11848 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11850 PR tree-optimization/56962
11851 * gimple-ssa-strength-reduction.c (record_increment): Only set
11852 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11853 either rhs1 or rhs2 is equal to c->base_expr.
11855 2013-04-15 Richard Biener <rguenther@suse.de>
11857 PR tree-optimization/56933
11858 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11860 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11861 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11862 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11863 dependence check ...
11864 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11866 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11867 GROUP_READ_WRITE_DEPENDENCE.
11869 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11871 * emit-rtl.c (reset_all_used_flags): New function.
11872 (verify_rtl_sharing): Call reset_all_used_flags before and after
11873 performing the checks.
11875 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11877 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11878 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11879 * config/arm/constraints.md (De): New constraint.
11880 * config/arm/neon.md (anddi3_neon): Delete.
11881 (neon_vand<mode>): Expand to standard anddi3 pattern.
11882 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11883 Move earlier in the file.
11884 (neon_inv_logic_op2): Likewise.
11885 (arm_anddi_operand_neon): New predicate.
11887 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11889 * configure.ac (gcc_cv_ld_as_needed): Set
11890 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11891 Use -z ignore, -z record on *-*-solaris2*.
11892 (HAVE_LD_AS_NEEDED): Update comment.
11893 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11894 * configure: Regenerate.
11895 * config.in: Regenerate.
11896 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11897 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11898 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11899 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11900 equivalents. Fix markup.
11901 * doc/tm.texi: Regenerate.
11903 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11905 * config/i386/i386.opt: New option mstack-protector-guard=.
11906 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11907 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11908 TARGET_SSP_TLS_GUARD.
11909 * config/i386/i386.c (ix86_option_override_internal): Set
11910 ix86_stack_protector_guard.
11911 * config/i386/i386.md (stack_protect_set): Enable for
11912 TARGET_SSP_TLS_GUARD only.
11913 (stack_protect_set_<mode>): Ditto.
11914 (stack_protect_test): Ditto.
11915 (stack_protect_test_<mode>): Ditto.
11916 * doc/invoke.texi (i386 Option): Document.
11918 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11921 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11922 (S_MODES): Set H_MODE bit.
11923 (SF_MODES): Set only S_MODE and SF_MODE bits.
11924 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11925 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11926 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11927 <MODE_FLOAT>: Likewise.
11929 2013-04-15 Joey Ye <joey.ye@arm.com>
11931 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11933 2013-04-15 Joey Ye <joey.ye@arm.com>
11935 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11937 (thumb_far_jump_used_p): Count instruction size and set
11940 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11942 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11943 * resource.c (mark_target_live_regs): Reformat conditional block.
11945 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11947 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11948 notes, they are emitted only just before final.
11949 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11951 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11953 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11954 * cfgrtl.c (delete_insn): Call it here instead.
11955 * lra-spills.c (lra_final_code_change): Use delete_insn.
11956 * haifa-sched.c (sched_remove_insn): Likewise.
11957 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11958 returning to the nop pool.
11959 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11960 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11962 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11964 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11965 * doc/tm.texi: Regenerated.
11967 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11969 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11972 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11974 * df-core.c (df_find_def): Compare register numbers.
11975 (df_find_use): Likewise.
11977 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11980 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11981 lra_in_progress for return.
11983 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11985 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11986 define_insn into define_insn_and_split and emit movsicc patterns.
11988 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11990 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11992 2013-04-12 Richard Biener <rguenther@suse.de>
11994 * tree-pass.h (TODO_do_not_ggc_collect): New.
11995 * passes.c (execute_one_ipa_transform_pass): Honor
11996 TODO_do_not_ggc_collect.
11997 (execute_one_pass): Likewise.
12000 2013-04-10 Richard Biener <rguenther@suse.de>
12002 * passes.c (init_optimization_passes): Remove reload pass.
12003 * ira.c (do_reload): Merge into ...
12005 (rest_of_handle_reload): Remove.
12006 (pass_reload): Likewise.
12007 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12008 of reload for vzeroupper pass placement.
12010 2013-04-12 Jakub Jelinek <jakub@redhat.com>
12012 PR tree-optimization/56918
12013 PR tree-optimization/56920
12014 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
12015 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
12016 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
12017 use wide_mul_with_sign method.
12019 2013-04-12 Richard Biener <rguenther@suse.de>
12021 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
12022 not be considered a gimple constant.
12024 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
12026 * fold-const.c (const_binop): Handle vector shifts by a scalar.
12027 (fold_binary_loc): Call const_binop also for mixed vector-scalar
12030 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
12031 Jakub Jelinek <jakub@redhat.com>
12033 * opts.c: Include diagnostic-color.h.
12034 (common_handle_option): Handle OPT_fdiagnostics_color_.
12035 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12036 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12037 (diagnostic-color.o): New.
12038 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12039 (diagnostic_color_rule): New enum.
12040 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12041 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12042 the location string.
12043 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12044 either NULL, or color kind.
12045 * diagnostic-color.c: New file.
12046 * diagnostic-color.h: New file.
12047 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12049 * doc/invoke.texi (-fdiagnostics-color): Document.
12050 * pretty-print.h (pp_show_color): Define.
12051 (struct pretty_print_info): Add show_color field.
12052 * diagnostic.c: Include diagnostic-color.h.
12053 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12054 macros. Colorize error:, warning: etc. strings and also the location
12056 (diagnostic_show_locus): Colorize the caret line.
12057 * pretty-print.c: Include diagnostic-color.h.
12058 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12059 inside of %< %> quotes or quoted through q format modifier.
12061 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12063 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12065 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12067 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12068 code in CC_NZ mode.
12069 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12072 2013-04-11 Marek Polacek <polacek@redhat.com>
12074 PR tree-optimization/48184
12075 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12077 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12079 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12080 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12081 (skip_simple_arithmetic): Tidy up.
12082 * tree.h (skip_simple_constant_arithmetic): Declare.
12084 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12086 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12088 2013-04-11 Richard Biener <rguenther@suse.de>
12090 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12091 generate vector constants.
12093 2013-04-11 Richard Biener <rguenther@suse.de>
12095 PR tree-optimization/56878
12096 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12097 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12099 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12100 Prefer to align the DR with the most invariant base address.
12102 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12104 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12107 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12109 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12110 floating-point vector comparisons against 0.
12112 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12114 PR tree-optimization/56899
12115 * fold-const.c (extract_muldiv_1): Apply distributive law
12116 only if TYPE_OVERFLOW_WRAPS (ctype).
12118 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12121 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12122 loading has side effect.
12124 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12126 * config/sparc/sparc.c: Include tree-pass.h.
12127 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12128 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12129 head of file. Change return type. Split off gate function.
12130 (sparc_gate_work_around_errata): New function.
12131 (pass_work_around_errata): New pass definition.
12132 (insert_pass_work_around_errata) New pass insert definition to
12133 insert pass_work_around_errata just after delayed-branch scheduling.
12134 (sparc_option_override): Insert the pass.
12135 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12137 2013-04-10 David S. Miller <davem@davemloft.net>
12139 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12140 or -mcpu=hypersparc.
12142 * target.def (cstore_mode): New hook.
12143 * target.h: Include insn-codes.h
12144 * targhooks.c: Likewise.
12145 (default_cstore_mode): New function.
12146 * targhooks.h: Declare it.
12147 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12148 * doc/tm.texi: Rebuild.
12149 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12150 target hook, rather than inspecting the insn_data.
12151 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12152 (TARGET_CSTORE_MODE): Redefine.
12153 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12155 * config/sparc/predicates.md (cstore_result_operand): New special
12157 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12158 Use it for operand 0.
12159 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12160 (*snesi_special): Likewise.
12161 (*snesi_zero): Likewise.
12162 (*seqsi_zero): Likewise.
12163 (*sltu_insn): Likewise.
12164 (*sgeu_insn): Likewise.
12165 (*seqdi_special): Make operand 0 and comparison operation be of
12167 (*snedi_special): Likewise.
12168 (*snedi_special_vis3): Likewise.
12169 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12170 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12171 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12172 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12173 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12174 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12175 (*sltu_extend_sp64): Likewise.
12176 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12177 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12178 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12179 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12180 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12182 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12184 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12185 (aarch64_start_file): Use the new function.
12187 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12188 Jason Merrill <jason@redhat.com>
12190 * common.opt: Add -gdwarf.
12191 * opts.c (common_handle_option): Handle it.
12192 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12194 2013-04-10 Richard Biener <rguenther@suse.de>
12196 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12197 (execute_one_ipa_transform_pass): But unconditionally here.
12198 (execute_one_pass): And here.
12199 (init_optimization_passes): Remove reload pass.
12200 * tree-pass.h (TODO_ggc_collect): Remove.
12201 (pass_reload): Likewise.
12202 * ira.c (do_reload): Merge into ...
12204 (rest_of_handle_reload): Remove.
12205 (pass_reload): Likewise.
12206 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12207 of reload for vzeroupper pass placement.
12208 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12209 and todo_flags_finish of all passes.
12211 2013-04-10 Richard Biener <rguenther@suse.de>
12213 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12214 first_const_oprnd field, rename first_def_type to first_op_type.
12215 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12216 (vect_get_and_check_slp_defs): Always use the type of the
12217 operand. Allow mixed vect_external_def, vect_constant_def types.
12218 (vect_get_constant_vectors): Handle mixed vect_external_def,
12219 vect_constant_def types.
12221 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12223 PR tree-optimization/55524
12224 * tree-ssa-math-opts.c
12225 (convert_mult_to_fma): Don't use an fms construct
12226 when we don't have an fms operation, but fnma, and it looks
12227 likely that we'll be able to use the latter.
12229 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12231 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12233 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12234 inline fail caused by overwritable functions.
12236 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12238 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12239 unnecessary bits in the constant power of two case.
12241 2013-04-10 Richard Biener <rguenther@suse.de>
12243 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12244 broken code swapping operands.
12245 (vect_build_slp_tree): Do not compute load permutations here.
12246 (vect_analyze_slp_instance): Compute load permutations here,
12247 after building the SLP tree.
12249 2013-04-09 Christian Bruel <christian.bruel@st.com>
12251 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12252 of next/prev_real_insn.
12254 2013-04-09 Jan Hubicka <jh@suse.cz>
12256 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12257 Drop aliased parameter.
12258 (function_and_variable_visibility): Do not handle alias pairs.
12259 * cgraph.c (varpool_externally_visible_p): Update prototype.
12260 * varpool.c (varpool_add_new_variable): Update.
12262 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12264 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
12266 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
12268 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
12270 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
12272 2013-04-09 Marek Polacek <polacek@redhat.com>
12274 PR tree-optimization/48762
12275 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
12277 2013-04-09 Richard Biener <rguenther@suse.de>
12279 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
12281 (vect_build_slp_tree): Likewise.
12282 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
12283 calculating the cost of a SLP instance.
12284 (vect_analyze_slp_instance): Use it from here, after building
12287 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12289 PR middle-end/56883
12290 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
12291 expand_omp_for_static_chunk): Use simple_p = true in
12292 force_gimple_operand_gsi calls when assigning to addressable decls.
12294 2013-04-09 Jeff Law <law@redhat.com>
12296 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
12297 when the boolean was created by converting a wider object which
12298 had a boolean range.
12300 2013-04-09 Richard Biener <rguenther@suse.de>
12302 * tree-vectorizer.h (slp_void_p): Remove.
12303 (slp_tree): Typedef before _slp_tree declaration.
12304 (struct _slp_tree): Use a vector of slp_tree as children.
12305 (vect_get_place_in_interleaving_chain): Remove.
12306 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
12308 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
12310 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
12311 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
12312 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
12313 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
12314 Use slp_node instead of slp_void_p and adjust.
12316 2013-04-09 Richard Biener <rguenther@suse.de>
12318 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
12319 work that is not necessary.
12321 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12323 PR tree-optimization/56854
12324 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
12325 forward into clobber stmts if it would change MEM_REF lhs into
12328 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
12330 * tree.c (type_hash_lookup, type_hash_add): Make static.
12331 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
12333 2013-04-09 Richard Biener <rguenther@suse.de>
12335 * tree.h (unsave_expr_now): Remove.
12336 * tree-inline.c (mark_local_for_remap_r): Remove.
12337 (unsave_expr_1): Likewise.
12338 (unsave_r): Likewise.
12339 (unsave_expr_now): Likewise.
12340 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
12341 (propagate_tree_value): Likewise.
12343 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
12345 * doc/rtl.texi (sequence): Rewrite documentation to match the
12346 current use of SEQUENCE rtl objects.
12347 * rtl.def (SEQUENCE): Likewise.
12349 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
12350 Update documentation.
12351 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
12352 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
12354 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
12356 2013-04-08 Teresa Johnson <tejohnson@google.com>
12358 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
12359 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
12361 (estimate_edge_size_and_time): Add comment to suggest using rounding
12363 (estimate_node_size_and_time): Ditto.
12364 (remap_edge_change_prob): Use helper rounding divide methods.
12365 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
12366 (gimple_mod_pow2_value_transform): Ditto.
12367 (gimple_mod_subtract_transform): Ditto.
12368 (gimple_ic_transform): Ditto.
12369 (gimple_stringops_transform): Ditto.
12370 * stmt.c (conditional_probability): Ditto.
12371 (emit_case_dispatch_table): Ditto.
12372 * lto-cgraph.c (merge_profile_summaries): Ditto.
12373 * tree-optimize.c (execute_fixup_cfg): Ditto.
12374 * cfgcleanup.c (try_forward_edges): Ditto.
12375 * cfgloopmanip.c (scale_loop_profile): Ditto.
12377 (duplicate_loop_to_header_edge): Ditto.
12378 (lv_adjust_loop_entry_edge): Ditto.
12379 * tree-vect-loop.c (vect_transform_loop): Ditto.
12380 * profile.c (compute_branch_probabilities): Ditto.
12381 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
12382 * lto-streamer-in.c (input_cfg): Ditto.
12383 * gimple-streamer-in.c (input_bb): Ditto.
12384 * ipa-cp.c (update_profiling_info): Ditto.
12385 (update_specialized_profile): Ditto.
12386 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
12387 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
12389 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12390 (compute_trg_info): Ditto.
12391 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
12392 (purge_dead_edges): Ditto.
12393 * loop-unswitch.c (unswitch_loop): Ditto.
12394 * cgraphclones.c (cgraph_clone_edge): Ditto.
12395 (cgraph_clone_node): Ditto.
12396 * tree-inline.c (copy_bb): Ditto.
12397 (copy_edges_for_bb): Ditto.
12398 (initialize_cfun): Ditto.
12399 (copy_cfg_body): Ditto.
12400 (expand_call_inline): Ditto.
12402 2013-04-08 Kai Tietz <ktietz@redhat.com>
12404 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
12405 TARGET_CYGWIN64 by TARGET_64BIT.
12407 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12409 * config/epiphany/epiphany.md (GPR_1): New constant.
12410 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
12411 * config/epiphany/epiphany.c (gen_compare_reg):
12412 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
12413 is already in place.
12414 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
12415 Don't require being called during rtl expansion; If y operlaps r0,
12417 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
12418 (epiphany_expand_epilogue): Likewise.
12420 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
12421 Don't use CC_FPmode for ORDERED / UNORDERED.
12422 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
12424 * config/epiphany/constraints.md (CnL): New constraint.
12425 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
12426 * config/epiphany/predicates.md (add_operand): Allow 1024.
12428 * config/epiphany/epiphany.md (logical_op): New code iterator.
12429 (op_mnc): New code attribute.
12430 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
12431 (mov_f+1, mov_f+2): New peephole2 patterns.
12433 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
12434 (cstoresi4): Also allow re-use of zero result when doing a NE
12435 comparison to a non-zero operand.
12436 Use (clobber (scratch)) for first insn if the gpr output is not needed.
12438 * config/epiphany/epiphany.md (<insn_opname>v2si3):
12439 Use gen_addsi3_i / gen_subsi3_i.
12441 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12445 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
12446 contain anything but clobbers, at most one __builtin_stack_restore,
12447 optionally debug stmts and final resx, and if it has at least one
12448 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
12449 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
12450 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
12451 which isn't defaut definition, remove them.
12452 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
12453 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
12454 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
12455 with MEM_REF LHS with SSA_NAME address.
12457 2013-04-08 Jeff Law <law@redhat.com>
12459 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
12461 2013-04-08 Richard Biener <rguenther@suse.de>
12463 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
12465 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
12466 determined vector type.
12467 (vect_analyze_data_refs): Likewise.
12468 (vect_get_new_vect_var): Adjust.
12469 (vect_create_destination_var): Preserve SSA name versions.
12470 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
12471 not dump anything here.
12473 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12475 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
12476 Add member lr_slot_known.
12477 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
12479 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12480 Remove code that sets lr_slot_offset according to what a previous
12481 version of epiphany_emit_save_restore used to do.
12482 (epiphany_emit_save_restore): When doing an lr save or restore,
12483 set/verify lr_slot_known and lr_slot_offset.
12485 2013-04-08 Xinyu Qi <xyqi@marvell.com>
12488 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
12491 2013-04-08 Richard Biener <rguenther@suse.de>
12493 * alias.c (find_base_term): Fix thinko in previous change.
12495 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12497 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
12498 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
12499 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
12500 if possible to compute val.
12501 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
12502 For QImode integers don't require anything about precision. Use
12503 const_with_all_bytes_same to find out if the constant doesn't have
12504 repeated bytes in it.
12506 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12508 * config/s390/s390.c (s390_expand_insv): Only accept insertions
12511 2013-04-08 Marek Polacek <polacek@redhat.com>
12513 PR rtl-optimization/48182
12514 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
12517 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12520 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
12521 nuses, make sure we have a label.
12523 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12526 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12527 (rs6000_emit_swdiv_low_precision): Remove.
12528 (rs6000_emit_swdiv): Rewrite to handle between one and four
12529 iterations of Newton-Raphson generally; modify required number of
12530 iterations for some cases.
12531 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12533 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12535 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12536 set-but-unused variable.
12538 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12539 basic blocks of released function bodies garbage-collectable.
12541 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12542 (struct rtl_opt_pass): Add TODO_df_finish.
12544 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12546 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12548 * config/arm/constraints.md (q): New constraint.
12549 * config/arm/ldrdstrd.md: New file.
12550 * config/arm/arm.md (ldrdstrd.md) New include.
12551 (arm_movdi): Use "q" instead of "r" constraint
12552 for double-word memory access.
12553 (movdf_soft_insn): Likewise.
12554 * config/arm/vfp.md (movdi_vfp): Likewise.
12555 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12556 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12557 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12558 (mem_ok_for_ldrd_strd): Likewise.
12559 (output_move_double): Update assertion.
12561 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12563 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12565 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12567 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12568 define_insn_and_split.
12569 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12571 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12573 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12574 define_insn_and_split.
12575 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12576 (shiftsi3_compare): New pattern.
12577 (rrx): New pattern.
12578 * config/arm/unspecs.md (UNSPEC_RRX): New.
12580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12582 * config/arm/arm.md (negdi_extendsidi): New pattern.
12583 (negdi_zero_extendsidi): Likewise.
12585 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12587 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12588 define_insn_and_split.
12589 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12590 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12592 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12594 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12595 define_insn_and_split.
12596 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12597 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12599 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12601 * config/arm/arm.md (subsi3_carryin): New pattern.
12602 (subsi3_carryin_const): Likewise.
12603 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12604 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12606 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12608 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12610 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12612 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12613 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12615 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12617 * config/arm/arm.c (arm_expand_builtin): Change fcode
12618 type to unsigned int.
12620 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12622 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12624 2013-04-04 Ian Lance Taylor <iant@google.com>
12626 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12629 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12631 PR middle-end/56729
12632 * df-scan.c (df_insn_delete): Disable failing assert.
12634 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12636 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12637 New function prototype.
12638 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12639 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12640 (arm_builtin_vectorized_function): New function.
12642 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12644 * config/arm/arm_neon_builtins.def: New file.
12645 * config/arm/arm.c (neon_builtin_data): Move contents to
12646 arm_neon_builtins.def.
12647 (enum arm_builtins): Include neon builtin definitions.
12648 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12649 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12651 2013-04-04 Marek Polacek <polacek@redhat.com>
12653 PR tree-optimization/48186
12654 * predict.c (maybe_hot_frequency_p): Return false if
12655 HOT_BB_FREQUENCY_FRACTION is 0.
12656 (cgraph_maybe_hot_edge_p): Likewise.
12658 2013-04-04 Richard Biener <rguenther@suse.de>
12660 PR tree-optimization/56826
12661 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12664 2013-04-04 Richard Biener <rguenther@suse.de>
12666 PR tree-optimization/56213
12667 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12668 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12670 2013-04-04 Richard Biener <rguenther@suse.de>
12672 PR tree-optimization/56837
12673 * tree-loop-distribution.c (classify_partition): For non-zero
12674 values require that the value has the same precision as its
12675 mode to be useful as memset value.
12677 2013-04-03 Nick Clifton <nickc@redhat.com>
12679 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12680 (fmssf4): Use fmsf.s on E3V5 architectures.
12681 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12682 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12684 2013-04-03 Jeff Law <law@redhat.com>
12686 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12687 (lra-eliminations.o): Likewise.
12689 2013-04-03 Teresa Johnson <tejohnson@google.com>
12691 * gcov-io.c (compute_working_sets): Moved most of body of old
12692 compute_working_sets here from profile.c.
12693 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12694 (gcov_working_set_t): Moved typedef here from basic-block.h
12695 (compute_working_set): Declare.
12696 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12697 (get_working_sets): Renamed from compute_working_set,
12698 replace most of body with call to new compute_working_sets.
12699 (get_exec_counts): Replace call to compute_working_sets
12700 to get_working_sets.
12701 * profile.h (get_working_sets): Renamed from compute_working_set.
12702 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12703 to get_working_sets.
12704 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12705 * gcov-dump.c (dump_working_sets): New function.
12707 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12709 * hwint.c (sext_hwi, zext_hwi): New functions.
12710 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12711 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12712 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12713 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12714 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12715 (sext_hwi, zext_hwi): New functions.
12717 2013-04-03 Jeff Law <law@redhat.com>
12719 PR tree-optimization/56799
12720 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12721 back test for widening conversion erroneously dropped in prior change.
12723 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12726 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12727 instead of next_real_insn.
12729 2013-04-03 Marek Polacek <polacek@redhat.com>
12732 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12734 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12737 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12739 (thumb1_output_casesi): Likewise.
12740 (thumb2_output_casesi): Likewise.
12742 2013-04-03 Richard Biener <rguenther@suse.de>
12744 PR tree-optimization/56817
12745 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12747 (tree_unroll_loops_completely_1): ... new function to manually
12748 walk the loop tree, properly defering outer loops of unrolled
12749 loops to later iterations.
12751 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12753 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12754 (vectorizable_load): Likewise.
12755 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12756 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12758 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12760 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12763 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12765 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12767 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12769 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12771 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12773 PR tree-optimization/56790
12774 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12777 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12779 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12781 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12782 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12785 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12788 * tree.h (force_folding_builtin_constant_p): New decl.
12789 * builtins.c (force_folding_builtin_constant_p): New variable.
12790 (fold_builtin_constant_p): Fold immediately also if
12791 force_folding_builtin_constant_p.
12793 2013-04-03 Richard Biener <rguenther@suse.de>
12795 PR tree-optimization/56812
12796 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12797 DRs of the same interleaving chain are independent.
12799 2013-04-02 Jason Merrill <jason@redhat.com>
12801 * gdbinit.in (pbb): Use debug fn.
12803 2013-04-02 Lawrence Crowl <crowl@google.com>
12805 * sese.h (struct ivtype_map_elt_s): Remove unused.
12806 (extern debug_ivtype_map): Remove unused.
12807 (extern eq_ivtype_map_elts): Remove unused.
12808 * sese.c (debug_ivtype_map): Removed unused.
12809 (debug_ivtype_map_1): Removed unused.
12810 (debug_ivtype_elt): Remove unused.
12811 (eq_ivtype_map_elts): Remove unused.
12814 2013-04-02 Kai Tietz <ktietz@redhat.com>
12817 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12818 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12819 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12821 (legitimize_pe_coff_symbol): Likewise.
12822 (is_imported_p): New helper-function.
12823 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12824 for Windows x64 targets.
12825 (ix86_expand_prologue): Optimize for pe-coff targets.
12826 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12827 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12828 medium/large code-model.
12829 (legitimize_pic_address): Likewise.
12830 (legitimize_tls_address): Likewise.
12831 (ix86_expand_call): Likewise.
12832 (x86_output_mi_thunk): Likewise.
12833 (get_dllimport_decl): Add new beimport argument.
12834 (construct_plt_address): Don't assert for x64 pe-coff targets.
12835 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12837 (SYMBOL_FLAG_STUBVAR): New macro.
12838 (SYMBOL_REF_STUBVAR_P): Likewise.
12839 * config/i386/winnt.c (stub_list): New structure.
12840 (stub_head): New local variable.
12841 (i386_pe_record_stub): New function.
12842 (i386_pe_file_end): Emit refptr-stubs.
12844 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12846 PR rtl-optimization/56745
12847 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12848 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12851 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12852 and both of them are MEM_REFs, just compare first argument for
12853 equality and attempt to deal even with differing offsets.
12856 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12857 of gimple_clobber_p to be MEM_REF.
12858 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12859 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12860 after gimplification.
12861 * asan.c (get_mem_ref_of_assignment): Don't instrument
12862 gimple_clobber_p stmts.
12863 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12864 gimple_clobber_p stmt if they have MEM_REF lhs and
12865 are dead because of another gimple_clobber_p stmt.
12866 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12867 gimple_clobber_p stmts like debug stmts.
12868 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12869 that refer to unused VAR_DECLs or uninitialized values.
12870 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12871 gimple_clobber_p stmts if they refer to removed parameters.
12872 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12875 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12877 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12878 using SWI48 mode attribute.
12880 2013-04-02 Wei Mi <wmi@google.com>
12882 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12883 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12884 *<rotate_insn><mode>3_mask in i386.md.
12886 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12888 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12890 2013-04-02 Richard Biener <rguenther@suse.de>
12892 PR tree-optimization/56778
12893 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12894 Runtime alias tests are not supported for gather loads.
12895 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12896 stmts referenced from SSA operands before updating SSA form.
12898 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12899 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12901 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12902 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12903 * config/arm/cortex-a53.md: New file.
12904 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12905 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12906 * config/arm/arm.c (arm_issue_rate): Likewise.
12907 * config/arm/arm-tune.md: Regenerate
12908 * config/arm/arm-tables.opt: Regenerate.
12909 * config/arm/arm-cores.def: Add cortex-a53.
12911 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12913 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12916 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12918 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12919 scalar load/store operations using B/H registers.
12920 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12922 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12924 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12926 * config/aarch64/aarch64.c
12927 (aarch64_simd_scalar_immediate_valid_for_move): New.
12928 * config/aarch64/aarch64-protos.h
12929 (aarch64_simd_scalar_immediate_valid_for_move): New.
12930 * config/aarch64/constraints.md (Dh, Dq): New.
12931 * config/aarch64/iterators.md (hq): New.
12933 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12935 * reorg.c (get_branch_condition): Deal with conditional returns.
12936 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12938 2013-04-01 Wei Mi <wmi@google.com>
12940 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12941 Truncate operand 2 using %b asm operand modifier.
12942 (*<shift_insn><mode>3_mask): Ditto.
12943 (*<rotate_insn><mode>3_mask): Ditto.
12945 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12947 PR middle-end/56798
12948 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12950 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12952 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12954 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12956 2013-03-30 Lawrence Crowl <crowl@google.com>
12958 * dse.c (clear_alias_sets): Remove never set.
12959 (disqualified_clear_alias_sets): Remove never set.
12960 (clear_alias_mode_pool): Remove never set.
12961 (dse_step0): Remove condition that is never true.
12962 (canon_address): Remove condition that is never true.
12963 (dse_step7): Remove condition that is never true.
12964 (rest_of_handle_dse): Remove condition that is never true.
12965 (rest_of_handle_dse::did_global): Remove never read from above.
12966 (dse_step2_spill): Remove never called from above.
12967 (dse_step5_spill): Remove never called from above.
12969 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12971 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12972 JUMP_TABLE_DATA changes.
12973 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12974 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12975 (Insns) <jump_table_data>: New entry.
12976 * doc/tm.texi: Regenerate.
12978 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12980 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12981 for table jump at the end of a basic block using tablejump_p.
12982 * targhooks.c (default_invalid_within_doloop): Likewise.
12983 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12984 target hook implementation that is identical to the default hook.
12985 (rs6000_invalid_within_doloop): Remove.
12987 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12988 unused variable from tablejump_p call.
12990 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12991 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12992 (INSN_DELETED_P): Likewise.
12993 (emit_jump_table_data): New prototype.
12994 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12995 after 4th as unused.
12996 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12997 * sched-vis.c (print_insn): Likewise.
12998 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12999 insn for compatibility with back ends that use next_active_insn to
13000 identify jump table data.
13001 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
13002 (remove_insn): Likewise.
13003 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
13005 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
13006 (emit_jump_table_data): New function.
13008 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
13009 basic block, a JUMP_TABLE_DATA never is.
13010 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
13011 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
13012 off from code handling real insns.
13013 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
13014 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
13015 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
13016 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
13017 is not a NONDEBUG_INSN_P.
13018 * ira-costs.c (scan_one_insn): Likewise.
13019 * jump.c (mark_all_labels): Likewise.
13020 (mark_jump_label_1): Likewise.
13021 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13022 * lra.c (get_insn_freq): Expect all insns reaching here to be in
13024 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
13025 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
13026 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
13027 JUMP_TABLE_DATA_P insns.
13028 (calculate_elim_costs_all_insns): Likewise.
13029 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
13030 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
13031 (delete_output_reload): Code style fixups.
13032 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
13033 insn flags on this non-insn.
13034 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13035 as scheduling barriers, for pre-change compatibility.
13036 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13037 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13039 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13040 redundant JUMP_TABLE_DATA_P test.
13041 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13042 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13043 (frv_for_each_packet): Likewise.
13044 * config/i386/i386.c (min_insn_size): Likewise.
13045 (ix86_avoid_jump_mispredicts): Likewise.
13046 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13047 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13048 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13049 (mips16_insn_length): Robustify.
13050 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13051 (mips16_split_long_branches): Likewise.
13052 * config/pa/pa.c (pa_combine_instructions): Likewise.
13053 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13054 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13055 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13056 as contributing to pool range lengths.
13057 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13058 Remove redundant JUMP_TABLE_DATA_P test.
13059 (sh_loop_align): Likewise.
13060 (split_branches): Likewise.
13061 (sh_insn_length_adjustment): Likewise.
13062 * config/spu/spu.c (get_branch_target): Likewise.
13064 2013-03-29 Jan Hubicka <jh@suse.cz>
13066 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13067 gcov streaming; stream hot bb threshold to ltrans.
13068 * predict.c (get_hot_bb_threshold): Break out from ....
13069 (maybe_hot_count_p): ... here.
13070 (set_hot_bb_threshold): New function.
13071 * lto-section-in.c (lto_section_name): Add profile.
13072 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13073 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13074 and data-streamer.h
13075 (histogram_entry): New structure.
13076 (histogram, histogram_pool): New global vars.
13077 (histogram_hash): New structure.
13078 (histogram_hash::hash): New method.
13079 (histogram_hash::equal): Likewise.
13080 (account_time_size): New function.
13081 (cmp_counts): New function.
13082 (dump_histogram): New function.
13083 (ipa_profile_generate_summary): New function.
13084 (ipa_profile_write_summary): New function.
13085 (ipa_profile_read_summary): New function.
13086 (ipa_profile): Decide on threshold.
13087 (pass_ipa_profile): Add ipa_profile_write_summary and
13088 ipa_profile_read_summary.
13089 * Makefile.in (ipa.o): Update dependencies.
13090 * lto-streamer.h (LTO_section_ipa_profile): New section.
13092 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13094 * tree.h (VAR_P): New.
13096 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13099 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13101 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13103 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13104 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13105 (control_flow_insn_p): Likewise.
13106 * cfgrtl.c (duplicate_insn_chain): Likewise.
13107 * final.c (get_attr_length_1): Likewise.
13108 (shorten_branches): Likewise.
13109 (final_scan_insn): Likewise.
13110 * function.c (instantiate_virtual_regs): Likewise.
13111 * gcse.c (insert_insn_end_basic_block): Likewise.
13112 * ira-costs.c (scan_one_insn): Likewise.
13113 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13114 * lra.c (check_rtl): Likewise.
13115 * reload1.c (elimination_costs_in_insn): Likewise.
13116 * reorg.c (follow_jumps): Likewise.
13118 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13119 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13120 (thumb_far_jump_used_p): Likewise.
13121 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13122 (workaround_speculation): Likewise.
13123 (add_sched_insns_for_speculation): Likewise.
13124 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13125 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13126 (frv_for_each_packet): Likewise.
13127 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13128 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13129 (final_emit_insn_group_barriers): Likewise.
13130 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13131 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13132 (mips16_insn_length): Likewise.
13133 * config/pa/pa.c (pa_reorg): Likewise.
13134 (pa_combine_instructions): Likewise.
13135 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13136 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13137 (sh_reorg): Likewise.
13138 (split_branches): Likewise.
13139 * config/spu/spu.c (get_branch_target): Likewise.
13141 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13144 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13146 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13147 Fix declaration name.
13149 2013-03-28 Lawrence Crowl <crowl@google.com>
13151 * graphds.h (struct graph.indicies): Remove unused.
13152 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13153 (SCOP_ORIGINAL_PDDRS): Remove unused.
13154 * sese.h (extern insert_loop_close_phis): Removed unused.
13155 (extern insert_guard_phis): Removed unused.
13156 (extern ivtype_map_elt_info): Removed unused.
13157 (new_ivtype_map_elt): Removed unused.
13158 * sese.c (ivtype_map_elt_info): Removed unused.
13160 2013-03-28 Lawrence Crowl <crowl@google.com>
13162 * Makefile.in: Add several missing include dependences.
13164 (test-dump.o): New. This object is not added to any executable,
13165 but is present for ad-hoc testing.
13167 (debug (const bitmap_head_def &)): New.
13168 (debug (const bitmap_head_def *)): New.
13170 (extern debug (const bitmap_head_def &)): New.
13171 (extern debug (const bitmap_head_def *)): New.
13173 (debug (edge_def &)): New.
13174 (debug (edge_def *)): New.
13176 (debug (basic_block_def &)): New.
13177 (debug (basic_block_def *)): New.
13179 (dump_node (const_tree, int, FILE *)): Correct source file.
13181 (debug (die_struct &)): New.
13182 (debug (die_struct *)): New.
13184 (extern debug (die_struct &)): New.
13185 (extern debug (die_struct *)): New.
13186 * gimple-pretty-print.c
13187 (debug (gimple_statement_d &)): New.
13188 (debug (gimple_statement_d *)): New.
13189 * gimple-pretty-print.h
13190 (extern debug (gimple_statement_d &)): New.
13191 (extern debug (gimple_statement_d *)): New.
13193 (debug (ira_allocno_copy &)): New.
13194 (debug (ira_allocno_copy *)): New.
13195 (debug (ira_allocno &)): New.
13196 (debug (ira_allocno *)): New.
13198 (extern debug (ira_allocno_copy &)): New.
13199 (extern debug (ira_allocno_copy *)): New.
13200 (extern debug (ira_allocno &)): New.
13201 (extern debug (ira_allocno *)): New.
13203 (debug (live_range &)): New.
13204 (debug (live_range *)): New.
13206 (debug (lra_live_range &)): New.
13207 (debug (lra_live_range *)): New.
13209 (debug (lra_live_range &)): New.
13210 (debug (lra_live_range *)): New.
13212 (debug (omega_pb_d &)): New.
13213 (debug (omega_pb_d *)): New.
13215 (extern debug (omega_pb_d &)): New.
13216 (extern debug (omega_pb_d *)): New.
13218 (debug (const rtx_def &)): New.
13219 (debug (const rtx_def *)): New.
13221 (debug_tree (tree): Move within file.
13222 (debug_raw (const tree_node &)): New.
13223 (debug_raw (const tree_node *)): New.
13224 (dump_tree_via_hooks (const tree_node *, int)): New.
13225 (debug (const tree_node &)): New.
13226 (debug (const tree_node *)): New.
13227 (debug_verbose (const tree_node &)): New.
13228 (debug_verbose (const tree_node *)): New.
13229 (debug_head (const tree_node &)): New.
13230 (debug_head (const tree_node *)): New.
13231 (debug_body (const tree_node &)): New.
13232 (debug_body (const tree_node *)): New.
13233 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13234 (debug (vec<tree, va_gc> &)): New.
13235 (debug (vec<tree, va_gc> *)): New.
13237 (extern debug (const rtx_def &)): New.
13238 (extern debug (const rtx_def *)): New.
13240 (debug_raw (simple_bitmap_def &)): New.
13241 (debug_raw (simple_bitmap_def *)): New.
13242 (debug (simple_bitmap_def &)): New.
13243 (debug (simple_bitmap_def *)): New.
13245 (extern debug (simple_bitmap_def &)): New.
13246 (extern debug (simple_bitmap_def *)): New.
13247 (extern debug_raw (simple_bitmap_def &)): New.
13248 (extern debug_raw (simple_bitmap_def *)): New.
13250 (debug (vinsn_def &)): New.
13251 (debug (vinsn_def *)): New.
13252 (debug_verbose (vinsn_def &)): New.
13253 (debug_verbose (vinsn_def *)): New.
13254 (debug (expr_def &)): New.
13255 (debug (expr_def *)): New.
13256 (debug_verbose (expr_def &)): New.
13257 (debug_verbose (expr_def *)): New.
13258 (debug (vec<rtx> &)): New.
13259 (debug (vec<rtx> *)): New.
13261 (extern debug (vinsn_def &)): New.
13262 (extern debug (vinsn_def *)): New.
13263 (extern debug_verbose (vinsn_def &)): New.
13264 (extern debug_verbose (vinsn_def *)): New.
13265 (extern debug (expr_def &)): New.
13266 (extern debug (expr_def *)): New.
13267 (extern debug_verbose (expr_def &)): New.
13268 (extern debug_verbose (expr_def *)): New.
13269 (extern debug (vec<rtx> &)): New.
13270 (extern debug (vec<rtx> *)): New.
13272 (_list_iter_cond_expr): Make inline instead of static.
13274 (debug (sreal &)): New.
13275 (debug (sreal *)): New.
13277 (extern debug (sreal &)): New.
13278 (extern debug (sreal *)): New.
13280 (extern debug_raw (const tree_node &)): New.
13281 (extern debug_raw (const tree_node *)): New.
13282 (extern debug (const tree_node &)): New.
13283 (extern debug (const tree_node *)): New.
13284 (extern debug_verbose (const tree_node &)): New.
13285 (extern debug_verbose (const tree_node *)): New.
13286 (extern debug_head (const tree_node &)): New.
13287 (extern debug_head (const tree_node *)): New.
13288 (extern debug_body (const tree_node &)): New.
13289 (extern debug_body (const tree_node *)): New.
13290 (extern debug (vec<tree, va_gc> &)): New.
13291 (extern debug (vec<tree, va_gc> *)): New.
13293 (debug (struct loop &)): New.
13294 (debug (struct loop *)): New.
13295 (debug_verbose (struct loop &)): New.
13296 (debug_verbose (struct loop *)): New.
13297 * tree-dump.c: Add header dependence.
13299 (extern debug (struct loop &)): New.
13300 (extern debug (struct loop *)): New.
13301 (extern debug_verbose (struct loop &)): New.
13302 (extern debug_verbose (struct loop *)): New.
13304 (debug (data_reference &)): New.
13305 (debug (data_reference *)): New.
13306 (debug (vec<data_reference_p> &)): New.
13307 (debug (vec<data_reference_p> *)): New.
13308 (debug (vec<ddr_p> &)): New.
13309 (debug (vec<ddr_p> *)): New.
13311 (extern debug (data_reference &)): New.
13312 (extern debug (data_reference *)): New.
13313 (extern debug (vec<data_reference_p> &)): New.
13314 (extern debug (vec<data_reference_p> *)): New.
13315 (extern debug (vec<ddr_p> &)): New.
13316 (extern debug (vec<ddr_p> *)): New.
13318 (debug (pt_solution &)): New.
13319 (debug (pt_solution *)): New.
13321 (extern debug (pt_solution &)): New.
13322 (extern debug (pt_solution *)): New.
13324 (debug (_var_map &)): New.
13325 (debug (_var_map *)): New.
13326 (debug (tree_live_info_d &)): New.
13327 (debug (tree_live_info_d *)): New.
13329 (extern debug (_var_map &)): New.
13330 (extern debug (_var_map *)): New.
13331 (extern debug (tree_live_info_d &)): New.
13332 (extern debug (tree_live_info_d *)): New.
13334 2013-03-28 Jan Hubicka <jh@suse.cz>
13336 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
13338 2013-03-28 Ian Bolton <ian.bolton@arm.com>
13340 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
13341 record only when desired or required.
13343 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
13345 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
13346 *vec_extractv2di_1_rex64. Use x64 isa attribute.
13348 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13350 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
13351 (*andsi3_compare0_uxtw): New pattern.
13352 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
13353 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
13355 2013-03-28 Jan Hubicka <jh@suse.cz>
13357 * data-streamer-in.c (streamer_read_gcov_count): New function.
13358 * gimple-streamer-out.c: Include value-prof.h.
13359 (output_gimple_stmt): Output histogram.
13360 (output_bb): Use streamer_write_gcov_count.
13361 * value-prof.c: Include data-streamer.h
13362 (dump_histogram_value): Add HIST_TYPE_MAX.
13363 (stream_out_histogram_value): New function.
13364 (stream_in_histogram_value): New function.
13365 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
13366 (stream_out_histogram_value, stream_in_histogram_value): Declare.
13367 * data-streamer-out.c (streamer_write_gcov_count): New function.
13368 (streamer_write_gcov_count_stream): New function.
13369 * lto-cgraph.c (lto_output_edge): Update counter streaming.
13370 (lto_output_node): Likewise.
13371 (input_node, input_edge): Likewise.
13372 * lto-streamer-out.c (output_cfg): Update streaming.
13373 * lto-streamer-in.c (input_cfg): Likewise.
13374 * data-streamer.h (streamer_write_gcov_count,
13375 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
13376 * gimple-streamer-in.c: Include value-prof.h
13377 (input_gimple_stmt): Input histograms.
13378 (input_bb): Update profile streaming.
13380 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13382 * genmodes.c (emit_max_int): New function.
13383 (emit_insn_modes_h): Added call to emit_max_function.
13384 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
13386 * machmode.def: Fixed comment.
13388 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13390 * combine.c (try_combine): Removed useless assert.
13391 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
13393 2013-03-28 Marek Polacek <polacek@redhat.com>
13394 Richard Biener <rguenther@suse.de>
13396 PR tree-optimization/56695
13397 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
13398 build signed result of a vector comparison.
13399 * tree-cfg.c (verify_gimple_comparison): Check that a result
13400 of a vector comparison has signed type.
13402 2013-03-28 Richard Biener <rguenther@suse.de>
13404 PR tree-optimization/37021
13405 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
13406 do not restrict gaps between groups.
13407 * tree-vect-stmts.c (vectorizable_load): Properly account for
13408 a gap between groups.
13410 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
13412 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
13413 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
13416 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
13418 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
13419 * doc/extend.texi (Named Address Spaces): Ditto.
13420 (Variable Attributes): Ditto.
13422 2013-03-27 Kai Tietz <ktietz@redhat.com>
13424 * config.build: Add support for cygwin x64 target.
13425 * config.gcc: Likewise.
13426 * config.host: Likewise.
13427 * configure.ac: Likewise
13428 * configure: Regenerated.
13430 2013-03-27 Kai Tietz <ktietz@redhat.com>
13432 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
13433 * config/i386/t-cygwin-w64: New file.
13434 * config/i386/cygwin-w64.h: New file.
13435 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
13436 and add support for x64-cygwin target.
13437 (CPP_SPEC): Likewise.
13438 (CXX_WRAP_SPEC_LIST): Undefine before define.
13439 (LIBGCJ_SONAME): Use 15 as version.
13441 2013-03-27 Richard Biener <rguenther@suse.de>
13443 PR tree-optimization/56716
13444 * tree-ssa-structalias.c (perform_var_substitution): Adjust
13445 dumping for ref nodes.
13447 2013-03-27 Martin Jambor <mjambor@suse.cz>
13449 PR tree-optimization/55334
13450 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
13451 restricted pointers to arrays.
13453 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
13455 * Makefile.in (.SUFFIXES): Add .cc.
13456 (.c.o): Apply same recipe for implicit rule .cc.o.
13458 2013-03-27 Richard Biener <rguenther@suse.de>
13460 PR tree-optimization/37021
13461 * tree-vect-data-refs.c (vect_check_strided_load): Allow
13462 REALPART/IMAGPART_EXPRs around the supported refs.
13463 * tree-ssa-structalias.c (find_func_aliases): Assume that
13464 floating-point values are not used to transfer pointers.
13466 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
13468 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
13469 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
13470 * gcc/doc/tm.texi: Regenerate.
13471 * gcc/targhooks.h (default_has_ifunc_p): New.
13472 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
13473 * gcc/config/linux-protos.h: New file.
13474 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
13475 this hook for linux which disables support of indirect functions in
13477 * gcc/config/linux-android.c: New file.
13478 * gcc/config/t-linux-android.c: Ditto.
13479 * gcc/config.gcc: Added new object file linux-android.o.
13480 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
13481 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
13482 * gcc/varasm.c (do_assemble_alias): Likewise.
13483 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
13484 doesn't support indirect functions.
13485 * configure: Regenerate.
13487 2013-03-27 Bin Cheng <bin.cheng@arm.com>
13490 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
13491 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
13494 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13496 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
13498 2013-03-27 Terry Guo <terry.guo@arm.com>
13500 * config/arm/arm-cores.def: Added core cortex-r7.
13501 * config/arm/arm-tune.md: Regenerated.
13502 * config/arm/arm-tables.opt: Regenerated.
13503 * doc/invoke.texi: Added entry for core cortex-r7.
13505 2013-03-27 Walter Lee <walt@tilera.com>
13507 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
13508 double-decrement of next_scratch_regno.
13510 2013-03-27 Walter Lee <walt@tilera.com>
13512 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
13514 (insn_v1mulus): Ditto.
13515 (insn_v2muls): Ditto.
13517 2013-03-27 Walter Lee <walt@tilera.com>
13519 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
13520 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
13522 2013-03-27 Walter Lee <walt@tilera.com>
13524 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13525 (*sibcall_value): Ditto.
13527 2013-03-27 Walter Lee <walt@tilera.com>
13529 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13530 (insn_mnz_v8qi): ... this ...
13531 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13533 (insn_v<n>mnz): Replaced by ...
13534 (insn_v1mnz): ... this ...
13535 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13537 (insn_mz_<mode>): Replaced by ...
13538 (insn_mz_v8qi): ... this ...
13539 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13541 (insn_v<n>mz): Replaced by ...
13542 (insn_v1mz): ... this ...
13543 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13546 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13548 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13550 2013-03-26 Roland McGrath <mcgrathr@google.com>
13552 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13553 than fprintf with a non-constant, non-format string.
13555 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13557 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13558 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13559 operand 0 predicate.
13560 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13561 attribute. Use general_x64nomem_operand as operand 1 predicate.
13562 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13563 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13564 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13565 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13566 (mov<mode>_insv_1): Remove expander. Merge insn with
13567 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13568 Use general_x64nomem_operand as operand 1 predicate.
13569 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13570 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13571 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13572 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13573 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13574 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13575 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13576 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13577 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13578 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13579 (general_x64nomem_operand): Ditto.
13581 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13583 * config/rtems.opt: Add -pthread option.
13585 2013-03-26 Richard Biener <rguenther@suse.de>
13587 * alias.c (find_base_term): Avoid redundant and not used recursion.
13588 (base_alias_check): Get the initial base term from the caller.
13589 (true_dependence_1): Compute and pass base terms to base_alias_check.
13590 (write_dependence_p): Likewise.
13591 (may_alias_p): Likewise.
13593 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13595 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13596 PC-relative load in SI modes and above only.
13598 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13600 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13601 * config/arm/iwmmxt.md (WCGR0): Update.
13602 (WCGR1, WCGR2, WCGR3): Likewise.
13604 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13606 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13607 Use x64 and nox64 isa attributes.
13609 2013-03-26 Richard Biener <rguenther@suse.de>
13611 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13612 alignment computations and rely on get_object_alignment_1
13613 for the !TYPE_P case.
13614 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13616 2013-03-26 Walter Lee <walt@tilera.com>
13618 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13619 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13621 2013-03-25 Jeff Law <law@redhat.com>
13623 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13624 check for INTEGRAL_TYPE_P that was missing due to checking in
13625 wrong version of prior patch.
13627 2013-03-25 Walter Lee <walt@tilera.com>
13629 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13630 TILEGX_INSN_SHUFFLEBYTES1.
13631 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13633 (tilegx_builtins): Ditto.
13634 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13636 2013-03-25 Walter Lee <walt@tilera.com>
13638 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13639 (floatunssisf2): New pattern.
13640 (floatsidf2): New pattern.
13641 (floatunssidf2): New pattern.
13643 2013-03-25 Walter Lee <walt@tilera.com>
13645 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13646 tests for constraint J, K, N, P.
13648 2013-03-25 Walter Lee <walt@tilera.com>
13650 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13651 Use indirect/pcrel encoding.
13652 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13655 2013-03-25 Steve Ellcey <sellcey@mips.com>
13657 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13658 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13659 * config/mips/mips.c (mips_option_override): Set IMADD default.
13660 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13661 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13662 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13663 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13664 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13666 2013-03-25 Jeff Law <law@redhat.com>
13668 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13669 slightly to avoid creating and folding useless trees. Simplify
13670 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13672 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13674 * config/i386/i386.md (*zero_extendsidi2): Merge with
13675 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13676 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13677 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13678 targets. Clarify comment.
13680 2013-03-25 Martin Jambor <mjambor@suse.cz>
13682 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13683 pass-through jump functions differently.
13684 (ipa_read_jump_function): Likewise. Also use setter functions to set
13687 2013-03-25 Martin Jambor <mjambor@suse.cz>
13689 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13690 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13692 (ipa_get_indirect_edge_target): New function.
13693 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13694 ipa_get_indirect_edge_target. Update all callers.
13695 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13696 ipa_get_indirect_edge_target_1 instead of calling
13697 ipa_get_indirect_edge_target.
13698 (create_specialized_node): Pass aggvlas to
13699 ipcp_discover_new_direct_edges.
13701 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13703 * config/arm/arm.md (f_sels, f_seld): New types.
13704 (*cmov<mode>): New pattern.
13705 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13708 2013-03-25 Kai Tietz <ktietz@redhat.com>
13710 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13711 POSIX-printf for mingw-hosted builds.
13713 2013-03-25 Richard Biener <rguenther@suse.de>
13715 PR middle-end/56694
13716 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13717 must-not-throw stmt location.
13719 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13721 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13722 Emit load-acquire versions when acq is true.
13723 (arm_emit_store_exclusive): Add rel parameter.
13724 Emit store-release versions when rel is true.
13725 (arm_split_compare_and_swap): Use acquire-release instructions
13727 of barriers when appropriate.
13728 (arm_split_atomic_op): Likewise.
13729 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13730 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13731 (VUNSPEC_SLX): Likewise.
13732 (VUNSPEC_LDA): Likewise.
13733 (VUNSPEC_STL): Likewise.
13734 * config/arm/sync.md (atomic_load<mode>): New pattern.
13735 (atomic_store<mode>): Likewise.
13736 (arm_load_acquire_exclusive<mode>): Likewise.
13737 (arm_load_acquire_exclusivesi): Likewise.
13738 (arm_load_acquire_exclusivedi): Likewise.
13739 (arm_store_release_exclusive<mode>): Likewise.
13741 2013-03-25 Catherine Moore <clm@codesourcery.com>
13743 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13744 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13745 * config/mip/predicates.md (lwsp_swsp_operand,
13746 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13747 sb16_operand, db4_operand, db7_operand, ib3_operand,
13748 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13749 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13750 andi16_operand): New predicates.
13751 * config/mips/mips.md (compression): New attribute.
13752 (enabled): New attribute.
13753 (length): Consider compression in computing length.
13754 (shift_compression): New code attribute.
13755 (*add<mode>3): New operands. Record compression.
13756 (sub<mode>3): Likewise.
13757 (one_cmpl<mode>2): Likewise.
13758 (*and<mode>3): Likewise.
13759 (*ior<mode>3): Likewise.
13760 (unnamed pattern for xor): Likewise.
13761 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13762 (*<optab><mode>3): Likewise.
13763 (*mov<mode>_internal: Likewise.
13764 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13765 (mips_unsigned_immediate_p): New.
13766 (umips_lwsp_swsp_address_p): New.
13767 (m16_based_address_p): New.
13768 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13769 (mips_unsigned_immediate_p): New prototype.
13770 (lwsp_swsp_address_p): New prototype.
13771 (m16_based_address_p): New prototype.
13772 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13773 (mips_signed_immediate_p): New function.
13774 (m16_based_address_p): New function.
13775 (lwsp_swsp_address_p): New function.
13776 (mips_print_operand_punctuation): Recognize short delay slot insns
13777 for microMIPS.add<mode>3"
13779 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13782 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13783 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13785 2013-03-25 Richard Biener <rguenther@suse.de>
13787 PR tree-optimization/56689
13788 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13791 2013-03-25 Richard Biener <rguenther@suse.de>
13793 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13795 (memory_references): Likewise.
13796 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13797 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13798 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13799 (gather_mem_refs_in_loops): Fold into ...
13800 (analyze_memory_references): ... this. Move initialization
13801 to tree_ssa_lim_initialize.
13802 (fill_always_executed_in): Rename to ...
13803 (fill_always_executed_in_1): ... this.
13804 (fill_always_executed_in): Move contains_call computation to
13805 this new function from ...
13806 (tree_ssa_lim_initialize): ... here.
13807 (tree_ssa_lim): Call fill_always_executed_in.
13809 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13811 * postreload.c (reload_combine): Fix code detecting returns.
13813 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13815 * function.c (emit_use_return_register_into_block): On cc0 targets,
13816 do not emit the sequence between cc0 setter and user.
13818 2013-03-25 Kai Tietz <ktietz@redhat.com>
13820 * config/i386/predicates.md (local_symbolic_operand): Interpret
13821 dll-imported symbols as none-local.
13823 2013-03-25 Richard Biener <rguenther@suse.de>
13825 * tree-ssa-loop-im.c (struct depend): Remove.
13826 (struct lim_aux_data): Make depends a vec of gimples.
13827 (free_lim_aux_data): Adjust.
13828 (add_dependency): Likewise.
13829 (set_level): Likewise.
13831 2013-03-25 Richard Biener <rguenther@suse.de>
13833 PR middle-end/56434
13834 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13835 the pointer returned by calls with ECF_MALLOC set.
13837 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13839 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13841 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13843 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13844 using MMXMODE mode iterator.
13845 (*move<mode>_internal): Merge with *movv2sf_internal and
13846 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13848 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13850 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13851 (record_last_mem_set_info): Likewise.
13853 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13854 of XNEWVEC followed by memset.
13855 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13857 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13859 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13860 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13861 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13862 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13863 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13864 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13865 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13866 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13867 BARRIER_P instead of GET_CODE.
13869 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13871 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13872 inaccuracy in the probing code.
13874 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13875 (ctrapdi4): Likewise.
13877 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13879 * calls.c (expand_call): Add missing guard to code handling return
13880 of non-BLKmode structures in MSB.
13881 * function.c (expand_function_end): Likewise.
13883 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13885 * combine.c (try_combine): Adjust comment. Do not add the set of
13886 insn #0 if the destination indirectly is set or dies in insn #2.
13887 Tidy up code to distribute a new note.
13889 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13891 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13892 also for alternatives 16 and 17.
13894 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13896 * config/i386/sse.md (*mov<mode>_internal): Merge with
13897 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13898 Emit insn template depending on type attribute. Use
13899 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13900 movd instead of movq mnemonic for interunit moves. Rewrite mode
13901 attribute calculation. Remove unit attribute calculation.
13902 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13903 Set prefix_data16 attribute for DImode ssemov types.
13904 Use Ym instead of y for SSE-MMX conversion alternatives.
13905 Reorder operand constraints.
13907 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13909 * df.h (df_insn_delete): Adjust prototype.
13910 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13911 and let it decide whether mark the basic block dirty.
13912 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13913 * df-scan.c (df_insn_info_delete): New helper function, split
13914 off from df_insn_delete.
13915 (df_scan_free_bb_info): Use it.
13916 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13918 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13919 that the insn is actually an insn and it has a non-NULL basic block.
13920 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13922 2013-03-22 Richard Biener <rguenther@suse.de>
13924 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13926 (mem_ref_alloc): Do not allocate them.
13927 (refs_independent_p): Do not query or maintain a cache.
13929 2013-03-22 Richard Biener <rguenther@suse.de>
13931 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13932 (gather_mem_refs_in_loops): Do not compute it.
13933 (analyze_memory_references): Do not allocate it.
13934 (tree_ssa_lim_finalize): Do not free it.
13935 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13937 2013-03-22 Richard Biener <rguenther@suse.de>
13939 * is-a.h (as_a): Use gcc_checking_assert.
13941 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13943 * config/aarch64/aarch64.c (aarch64_print_operand): New
13944 format specifier for printing a constant in hex.
13945 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13946 format specifier for printing second operand.
13948 2013-03-22 Richard Biener <rguenther@suse.de>
13950 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13952 (gather_mem_refs_in_loops): Perform store accumulation here.
13953 (create_vop_ref_mapping_loop): Remove.
13954 (create_vop_ref_mapping): Likewise.
13955 (analyze_memory_references): Initialize refs_stored_in_loop.
13956 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13957 (record_indep_loop): Remove.
13958 (record_dep_loop): New function.
13959 (ref_indep_loop_p_1): Adjust to only walk over references
13960 in the loop, not its subloops.
13961 (ref_indep_loop_p): Rename to ...
13962 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13963 maintaining a more fine-grained cache.
13964 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13965 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13967 2013-03-22 Richard Biener <rguenther@suse.de>
13969 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13970 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13971 aggregate mem_ref_loc.
13972 (free_mem_ref_locs): Inline into ...
13973 (memref_free): ... this and adjust.
13974 (mem_ref_alloc): Adjust.
13975 (mem_ref_locs_alloc): Remove.
13976 (record_mem_ref_loc): Adjust.
13977 (get_all_locs_in_loop): Rewrite into ...
13978 (for_all_locs_in_loop): ... this iterator.
13979 (rewrite_mem_ref_loc): New functor.
13980 (rewrite_mem_refs): Use for_all_locs_in_loop.
13981 (sm_set_flag_if_changed): New functor.
13982 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13983 (ref_always_accessed): New functor.
13984 (ref_always_accessed_p): Use for_all_locs_in_loop.
13986 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13988 * tree-pass.h (PROP_gimple_lvec): New.
13989 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13990 (init_optimization_passes): Move pass_lower_vector.
13991 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13993 (pass_lower_vector): Provide PROP_gimple_lvec.
13994 (pass_lower_vector_ssa): Likewise.
13995 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13997 2013-03-21 Mark Wielaard <mjw@redhat.com>
13999 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
14001 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14003 * config/i386/i386.md (*movdi_internal): Disparage slightly
14004 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
14005 conversion alternatives.
14007 2013-03-21 Jakub Jelinek <jakub@redhat.com>
14009 PR middle-end/48087
14010 * diagnostic.def (DK_WERROR): New kind.
14011 * diagnostic.h (werrorcount): Define.
14012 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
14013 promoted to DK_ERROR, increment DK_WERROR counter instead of
14015 * toplev.c (toplev_main): Call print_ignored_options even if
14016 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
14017 even if just werrorcount is non-zero.
14020 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
14022 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
14023 (string_cst_pool_decl): New function.
14024 (optimize_one_addr_into_implicit_ptr): New function.
14025 (resolve_addr_in_expr): Optimize DWARF location expression
14026 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
14027 which doesn't live in memory, but has DW_AT_location or
14028 DW_AT_const_value, or refers to a string literal, into
14029 DW_OP_GNU_implicit_pointer.
14030 (optimize_location_into_implicit_ptr): New function.
14031 (resolve_addr): If removing DW_AT_location of a variable because
14032 it was DW_OP_addr of address of the variable, but the variable doesn't
14033 live in memory, try to emit const value attribute for the initializer.
14035 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14037 * tree.h (VECTOR_TYPE_P): New macro.
14038 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14039 TYPE_MODE): Use it.
14040 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14041 VEC_COND_EXPR cannot be lvalues.
14042 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14044 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14046 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14047 Restrict the transformation to equal modes.
14049 2013-03-21 Richard Biener <rguenther@suse.de>
14051 PR tree-optimization/39326
14052 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14053 (MEM_ANALYZABLE): Adjust.
14054 (record_mem_ref_loc): Move bitmap ops ...
14055 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14056 unanalyzable refs, do not record locations for it.
14057 (analyze_memory_references): Allocate ref zero as shared
14059 (refs_independent_p): Do not test for unanalyzed mems here.
14060 (ref_indep_loop_p_1): Special-case disambiguation against
14061 the unanalyzed ref.
14062 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14064 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14066 * config/arm/arm-protos.h (tune_params): Add
14067 prefer_neon_for_64bits field.
14068 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14069 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14070 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14071 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14072 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14073 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14074 (arm_option_override): Handle -mneon-for-64bits new option.
14075 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14076 (prefer_neon_for_64bits): Declare new variable.
14077 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14078 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14079 (arch_enabled): Handle new arch types. Remove support for onlya8
14081 (one_cmpldi2): Use new arch names.
14082 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14083 * config/arm/arm.opt (mneon-for-64bits): Add option.
14084 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14085 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14086 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14088 * doc/invoke.texi (-mneon-for-64bits): Document.
14090 2013-03-21 Richard Biener <rguenther@suse.de>
14092 PR tree-optimization/39326
14093 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14094 (sort_bbs_in_loop_postorder_cmp): New function.
14095 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14097 2013-03-21 Richard Biener <rguenther@suse.de>
14099 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14100 (vect_insert_into_interleaving_chain): Likewise.
14101 (vect_drs_dependent_in_basic_block): Inline ...
14102 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14104 (vect_analyze_data_ref_dependence): ... here. Simplify.
14105 (vect_check_interleaving): Simplify.
14106 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14107 (vect_slp_analyze_data_ref_dependences): ... this new function.
14108 (dr_group_sort_cmp): New function.
14109 (vect_analyze_data_ref_accesses): Compute data-reference groups
14110 here instead of in vect_analyze_data_ref_dependence. Use
14111 a more efficient algorithm.
14112 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14113 vect_slp_analyze_data_ref_dependences. Call
14114 vect_analyze_data_ref_accesses earlier.
14115 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14116 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14117 (vect_slp_analyze_data_ref_dependences): New prototype.
14119 2013-03-21 Richard Biener <rguenther@suse.de>
14121 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14122 ref is stored in the loop.
14123 (find_refs_for_sm): Walk only over all stores.
14124 (store_motion_loop): Allocate from lim_bitmap_obstack.
14125 (store_motion): Likewise.
14127 2013-03-21 Richard Biener <rguenther@suse.de>
14129 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14130 Update virtual SSA form.
14132 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14134 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14135 * configure: Regenerate.
14136 * config.in: Regenerate.
14137 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14138 if !HAVE_LD_EH_FRAME_CIEV3.
14140 2013-03-21 Richard Biener <rguenther@suse.de>
14142 * tree-cfg.c (verify_expr_no_block): New function.
14143 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14144 nor DECL_VALUE_EXPR have locations with associated blocks.
14145 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14146 (clear_unused_block_pointer): Remove code dealing with
14147 blocks in DECL_DEBUG_EXPR locations.
14149 2013-03-21 Richard Biener <rguenther@suse.de>
14151 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14152 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14153 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14154 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14155 instead of DECL_DEBUG_EXPR_IS_FROM.
14156 * gimplify.c (gimplify_modify_expr): Likewise.
14157 * tree-cfg.c (verify_expr_location_1): Likewise.
14158 * tree-complex.c (create_one_component_var): Likewise.
14159 * tree-sra.c (create_access_replacement): Likewise.
14160 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14161 (clear_unused_block_pointer): Likewise.
14162 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14163 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14164 * var-tracking.c (var_debug_decl): Likewise.
14165 (track_expr_p): Likewise.
14166 * tree-inline.c (add_local_variables): Likewise. Set
14167 DECL_HAS_DEBUG_EXPR_P after copying it.
14168 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14169 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14171 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14174 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14175 * configure: Regenerate.
14176 * config.in: Regenerate.
14177 * config/i386/i386.md (*movdf_internal): Use
14178 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14179 movd instead of movq mnemonic for interunit moves.
14180 (*movdi_internal): Ditto.
14182 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14184 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14185 (abd<mode>_3): New pattern.
14186 (aba<mode>_3): New pattern.
14187 (fabd<mode>_3): New pattern.
14189 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14191 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14192 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14193 occurrence of REGISTER_PREFIX as its empty string.
14195 2013-03-20 Jeff Law <law@redhat.com>
14197 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14198 addititional equivalences for equality comparisons between an SSA_NAME
14199 and a constant where the SSA_NAME was set from a widening conversion.
14201 2013-03-20 Walter Lee <walt@tilera.com>
14203 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14205 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14207 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14208 depending on type attribute.
14209 (*movti_internal): Ditto.
14210 (*movtf_internal): Ditto.
14211 (*movxf_internal): Ditto.
14212 (*movdf_internal): Ditto.
14213 (*movsf_internal): Ditto.
14215 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14217 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14218 maybe_vex for sselog1 and ssemov types.
14219 (*movdi_internal): Reorder operand constraints.
14220 (*movsi_internal): Ditto. Set prefix attribute to
14221 maybe_vex for sselog1 and ssemov types.
14222 (*movtf_internal): Set prefix attribute to maybe_vex
14223 for sselog1 and ssemov types.
14224 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14225 DImode ssemov types. Reorder operand constraints.
14226 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14227 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14228 attribute for SImode ssemov types. Reorder operand constraints.
14230 2013-03-20 Martin Jambor <mjambor@suse.cz>
14232 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14233 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14235 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14237 * config/rs6000/predicates.md (indexed_address, update_address_mem
14238 update_indexed_address_mem): New predicates.
14239 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14240 attribute for load/store instructions.
14241 * config/rs6000/dfp.md (movsd_store): Likewise.
14242 (movsd_load): Likewise.
14243 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14244 (unnamed HI->DI extend define_insn): Likewise.
14245 (unnamed SI->DI extend define_insn): Likewise.
14246 (unnamed QI->SI extend define_insn): Likewise.
14247 (unnamed QI->HI extend define_insn): Likewise.
14248 (unnamed HI->SI extend define_insn): Likewise.
14249 (unnamed HI->SI extend define_insn): Likewise.
14250 (extendsfdf2_fpr): Likewise.
14251 (movsi_internal1): Likewise.
14252 (movsi_internal1_single): Likewise.
14253 (movhi_internal): Likewise.
14254 (movqi_internal): Likewise.
14255 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14256 attribute for load/store instructions.
14257 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
14259 (mov<mode>_softfloat): Likewise.
14260 (mov<mode>_hardfloat32): Likewise.
14261 (mov<mode>_hardfloat64): Likewise.
14262 (mov<mode>_softfloat64): Likewise.
14263 (movdi_internal32): Likewise.
14264 (movdi_internal64): Likewise.
14265 (probe_stack_<mode>): Likewise.
14267 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
14269 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
14270 floating point, and decimal floating point to reload iterator.
14272 * config/rs6000/constraints.md (wl constraint): New constraints to
14273 return FLOAT_REGS if certain options are used to reduce the number
14274 of separate patterns that exist in the file.
14275 (wx constraint): Likewise.
14276 (wz constraint): Likewise.
14278 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14279 -mdebug=reg, print wg, wl, wx, and wz constraints.
14280 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14281 Initialize the reload functions for 64-bit binary/decimal floating
14283 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14284 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14285 create the buffer on the stack to overcome not having a 32-bit
14287 (rs6000_emit_move): Likewise.
14288 (rs6000_secondary_memory_needed_rtx): Likewise.
14289 (rs6000_alloc_sdmode_stack_slot): Likewise.
14290 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14291 via xxlxor, just like DFmode 0.0.
14293 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
14294 define as 1 if we are running on a power7 or newer.
14295 (enum r6000_reg_class_enum): Add new constraints.
14297 * config/rs6000/dfp.md (movsd): Delete, combine with binary
14298 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14299 with other moves by using conditional constraits (wg). Use LFIWZX
14300 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
14301 (movsd splitter): Likewise.
14302 (movsd_hardfloat): Likewise.
14303 (movsd_softfloat): Likewise.
14305 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
14306 binary and decimal floating point moves.
14307 (fmove_ok): New attributes to combine binary and decimal floating
14308 point moves, and to combine power6x (mfpgpr) moves along normal
14310 (real_value_to_target): Likewise.
14311 (f32_lr): Likewise.
14312 (f32_lm): Likewise.
14313 (f32_li): Likewise.
14314 (f32_sr): Likewise.
14315 (f32_sm): Likewise.
14316 (f32_si): Likewise.
14317 (movsf): Combine binary and decimal floating point moves. Combine
14318 power6x (mfpgpr) moves with other moves by using conditional
14319 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
14320 (mov<mode> for SFmode/SDmode); Likewise.
14321 (SFmode/SDmode splitters): Likewise.
14322 (movsf_hardfloat): Likewise.
14323 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
14324 (movsf_softfloat): Likewise.
14325 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
14327 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
14328 wx and wz constraints.
14330 * config/rs6000/constraints.md (wg constraint): New constraint to
14331 return FLOAT_REGS if -mmfpgpr (power6x) was used.
14333 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
14336 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14337 -mdebug=reg, print wg, wl, wx, and wz constraints.
14338 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14339 Initialize the reload functions for 64-bit binary/decimal floating
14341 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14342 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14343 create the buffer on the stack to overcome not having a 32-bit
14345 (rs6000_emit_move): Likewise.
14346 (rs6000_secondary_memory_needed_rtx): Likewise.
14347 (rs6000_alloc_sdmode_stack_slot): Likewise.
14348 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14349 via xxlxor, just like DFmode 0.0.
14351 * config/rs6000/dfp.md (movdd): Delete, combine with binary
14352 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14353 with other moves by using conditional constraits (wg). Use LFIWZX
14354 and STFIWX for loading SDmode on power7.
14355 (movdd splitters): Likewise.
14356 (movdd_hardfloat32): Likewise.
14357 (movdd_softfloat32): Likewise.
14358 (movdd_hardfloat64_mfpgpr): Likewise.
14359 (movdd_hardfloat64): Likewise.
14360 (movdd_softfloat64): Likewise.
14362 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
14363 64-bit binary and decimal floating point moves.
14364 (FMOVE64X): Likewise.
14365 (movdf): Combine 64-bit binary and decimal floating point moves.
14366 Combine power6x (mfpgpr) moves with other moves by using
14367 conditional constraits (wg).
14368 (mov<mode> for DFmode/DDmode): Likewise.
14369 (DFmode/DDmode splitters): Likewise.
14370 (movdf_hardfloat32): Likewise.
14371 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
14372 (movdf_softfloat32): Likewise.
14373 (movdf_hardfloat64_mfpgpr): Likewise.
14374 (movdf_hardfloat64): Likewise.
14375 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
14376 (movdf_softfloat64): Likewise.
14377 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
14378 (reload_<mode>_load): Move to later in the file so they aren't in
14379 the middle of the floating point move insns.
14380 (reload_<mode>_store): Likewise.
14382 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
14385 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
14386 constraint if -mdebug=reg.
14387 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
14388 Enable using dd reload support if needed.
14390 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
14391 binary and decimal floating point moves in rs6000.md.
14392 (movtd_internal): Likewise.
14394 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
14395 decimal floating point moves.
14397 (movtf_internal): Likewise.
14398 (mov<mode>_internal, TDmode/TFmode): Likewise.
14399 (movtf_softfloat): Likewise.
14400 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
14402 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
14403 movdi_internal64, using wg constraint for move direct operations.
14404 (movdi_internal64): Likewise.
14406 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
14407 MODES_TIEABLE_P for selected modes. Print the numerical value of
14408 the various virtual registers. Use GPR/FPR first/last values,
14409 instead of hard coding the register numbers. Print which modes
14410 have reload functions registered.
14411 (rs6000_option_override_internal): If -mdebug=reg, trace the options
14412 settings before/after setting cpu, target and subtarget settings.
14413 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
14414 and for secondary reload failures in rs6000_secondary_reload_inner.
14415 (rs6000_secondary_reload_fail): Likewise.
14416 (rs6000_secondary_reload_inner): Likewise.
14418 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
14419 macros for first/last GPR and FPR registers.
14420 (LAST_GPR_REGNO): Likewise.
14421 (FIRST_FPR_REGNO): Likewise.
14422 (LAST_FPR_REGNO): Likewise.
14424 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
14425 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
14426 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
14427 (vcond<mode><mode>): Likewise.
14428 (vcondu<mode><mode>): Likewise.
14429 (vector_gtu<mode>): Likewise.
14430 (vector_gte<mode>): Likewise.
14431 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
14432 to prevent the compiler from converting DImode operations to TImode.
14433 (ior<mode>3): Likewise.
14434 (and<mode>3): Likewise.
14435 (one_cmpl<mode>2): Likewise.
14436 (nor<mode>3): Likewise.
14437 (andc<mode>3): Likewise.
14439 * config/rs6000/constraints.md (wt constraint): New constraint
14440 that returns VSX_REGS if TImode is allowed in VSX registers.
14442 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
14443 constant under VSX.
14445 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
14446 similar to TImode, but it is restricted to being in the GPRs.
14448 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
14449 TImode to occupy a single VSX register.
14451 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
14452 -mvsx-timode for power7/power8.
14453 (power7 cpu): Likewise.
14454 (power8 cpu): Likewise.
14456 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
14457 sure that TFmode/TDmode take up two registers if they are ever
14458 allowed in the upper VSX registers.
14459 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
14461 (rs6000_init_hard_regno_mode_ok): Likewise.
14462 (rs6000_debug_reg_global): Add debugging for PTImode and wt
14463 constraint. Print if LRA is turned on.
14464 (rs6000_option_override_internal): Give an error if -mvsx-timode
14465 and VSX is not enabled.
14466 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
14467 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
14468 to reg+offset addressing. Use PTImode when checking offset
14469 addresses for validity.
14470 (reg_offset_addressing_ok_p): Likewise.
14471 (rs6000_legitimate_offset_address_p): Likewise.
14472 (rs6000_legitimize_address): Likewise.
14473 (rs6000_legitimize_reload_address): Likewise.
14474 (rs6000_legitimate_address_p): Likewise.
14475 (rs6000_eliminate_indexed_memrefs): Likewise.
14476 (rs6000_emit_move): Likewise.
14477 (rs6000_secondary_reload): Likewise.
14478 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
14479 reloads to fpr registers to continue to use reg+offset addressing,
14480 but 64-bit reloads to altivec registers need reg+reg addressing.
14481 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
14482 it. Treat LO_SUM like a PLUS operation.
14483 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
14484 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
14485 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
14486 registers to share a register with a smaller sized type, since VSX
14487 puts scalars in the upper 64-bits.
14488 (print_operand): Add support for PTImode.
14489 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
14490 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
14491 registers, but don't have arithmetic support.
14492 (rs6000_memory_move_cost): Add test for VSX.
14493 (rs6000_opt_masks): Add -mvsx-timode.
14495 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
14498 (VSr): Use wt constraint for TImode.
14499 (VSv): Drop TImode support.
14500 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
14501 (vsx_movti_64bit): Likewise.
14502 (vsx_movti_32bit): Likewise.
14503 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
14504 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
14505 one '?' on the appropriate output constraint. Do not allow TImode
14506 logical operations on 32-bit systems.
14507 (vsx_ior<mode>3): Likewise.
14508 (vsx_xor<mode>3): Likewise.
14509 (vsx_one_cmpl<mode>2): Likewise.
14510 (vsx_nor<mode>3): Likewise.
14511 (vsx_andc<mode>3): Likewise.
14512 (vsx_concat_<mode>): Likewise.
14513 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
14515 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
14516 OPTION_MASK_VSX_TIMODE.
14517 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
14518 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
14520 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
14521 (TI2 iterator): New iterator for TImode, PTImode.
14522 (wd mode attribute): Add values for vector types.
14523 (movti_string): Replace TI move operations with operations for TImode
14524 and PTImode. Add support for TImode being allowed in VSX registers.
14525 (mov<mode>_string, TImode/PTImode): Likewise.
14526 (movti_ppc64): Likewise.
14527 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14528 (TI mode splitters): Likewise.
14530 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14533 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14535 PR tree-optimization/56355
14536 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14537 Also handle integers with undefined overflow.
14539 2013-03-20 Catherine Moore <clm@codesourcery.com>
14540 Maciej W. Rozycki <macro@codesourcery.com>
14541 Tom de Vries <tom@codesourcery.com>
14542 Nathan Sidwell <nathan@codesourcery.com>
14543 Iain Sandoe <iain@codesourcery.com>
14544 Nathan Froyd <froydnj@codesourcery.com>
14545 Chao-ying Fu <fu@mips.com>
14547 * doc/extend.texi: (micromips, nomicromips, nocompression):
14548 Document new function attributes.
14549 * doc/invoke.texi (minterlink-compressed, mmicromips,
14550 m14k, m14ke, m14kec): Document new options.
14551 (minterlink-mips16): Update documentation.
14552 * doc/md.texi (ZC, ZD): Document new constraints.
14553 * configure.ac (gcc_cv_as_micromips): Check if linker
14554 supports the .set micromips directive.
14555 * configure: Regenerate.
14556 * config.in: Regenerate.
14557 * config/mips/mips-tables.opt: Regenerate.
14558 * config/mips/micromips.md: New file.
14559 * constraints.md (ZC, ZD): New constraints.
14560 * config/mips/predicates.md (movep_src_register): New predicate.
14561 (movep_src_operand): New predicate.
14562 (non_volatile_mem_operand): New predicate.
14563 * config/mips/mips.md (multimem): New type.
14564 (length): Differentiate between 17-bit and 18-bit branch offsets.
14565 (MOVEP1, MOVEP2): New mode iterator.
14566 (mov_<load>l): Use ZC constraint.
14567 (mov_<load>r): Likewise.
14568 (mov_<store>l): Likewise.
14569 (mov_<store>r): Likewise.
14570 (*branch_equality<mode>_inverted): Add microMIPS support.
14571 (*branch_equality<mode>): Likewise.
14572 (*jump_absolute): Likewise.
14573 (indirect_jump_<mode>): Likewise.
14574 (tablejump_<mode>): Likewise.
14575 (<optab>_internal): Likewise.
14576 (sibcall_internal): Likewise.
14577 (sibcall_value_internal): Likewise.
14578 (prefetch): Use constraint ZD.
14579 * config/mips/mips.opt (minterlink-compressed): New option.
14580 (minterlink-mips16): Now an alias for minterlink-compressed.
14581 (mmicromips): New option.
14582 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14583 (compare_and_swap_12): Likewise.
14584 (sync_add<mode>): Likewise.
14585 (sync_<optab>_12): Likewise.
14586 (sync_old_<optab>_12): Likewise.
14587 (sync_new_<optab>_12): Likewise.
14588 (sync_nand_12): Likewise.
14589 (sync_old_nand_12): Likewise.
14590 (sync_new_nand_12): Likewise.
14591 (sync_sub<mode>): Likewise.
14592 (sync_old_add<mode>): Likewise.
14593 (sync_old_sub<mode>): Likewise.
14594 (sync_new_add<mode>): Likewise.
14595 (sync_new_sub<mode>): Likewise.
14596 (sync_<optab><mode>): Likewise.
14597 (sync_old_<optab><mode>): Likewise.
14598 (sync_new_<optab><mode>): Likewise.
14599 (sync_nand<mode>): Likewise.
14600 (sync_old_nand<mode>): Likewise.
14601 (sync_new_nand<mode>): Likewise.
14602 (sync_lock_test_and_set<mode>): Likewise.
14603 (test_and_set_12): Likewise.
14604 (atomic_compare_and_swap<mode>): Likewise.
14605 (atomic_exchange<mode>_llsc): Likewise.
14606 (atomic_fetch_add<mode>_llsc): Likewise.
14607 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14608 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14609 (umips_save_restore_pattern_p): Likewise.
14610 (umips_load_store_pair_p): Likewise.
14611 (umips_output_load_store_pair): Likewise.
14612 (umips_movep_target_p): Likewise.
14613 (umips_12bit_offset_address_p): Likewise.
14614 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14615 (mips_base_mips16): Rename this...
14616 (mips_base_compression_flags): ...to this. Update all uses.
14617 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14618 (mips_mips16_decl_p): Delete.
14619 (mips_nomips16_decl_p): Delete.
14620 (mips_get_compress_on_flags): New function.
14621 (mips_get_compress_off_flags): New function.
14622 (mips_get_compress_mode): New function.
14623 (mips_get_compress_on_name): New function.
14624 (mips_get_compress_off_name): New function.
14625 (mips_insert_attributes): Support multiple compression types.
14626 (mips_merge_decl_attributes): Likewise.
14627 (umips_12bit_offset_address_p): New function.
14628 (mips_start_function_definition): Emit .set micromips directive.
14629 (mips_call_may_need_jalx_p): New function.
14630 (mips_function_ok_for_sibcall): Add microMIPS support.
14631 (mips_print_operand_punctuation): Support short delay slots and
14633 (umips_swm_mask, umips_swm_encoding): New.
14634 (umips_build_save_restore): New function.
14635 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14636 (was_mips16_p): Remove.
14637 (old_compression_mode): New.
14638 (mips_set_compression_mode): New function.
14639 (mips_set_current_function): Add microMIPS support.
14640 (mips_option_override): Likewise.
14641 (umips_save_restore_pattern_p): New function.
14642 (umips_output_save_restore): New function.
14643 (umips_load_store_pair_p_1): New function.
14644 (umips_load_store_pair_p): New function.
14645 (umips_output_load_store_pair_1): New function.
14646 (umips_output_load_store_pair): New function.
14647 (umips_movep_target_p) New function.
14648 (mips_prepare_pch_save): Add microMIPS support.
14649 * config/mips/mips.h (TARGET_COMPRESSION): New.
14650 (TARGET_CPU_CPP_BUILTINS): Update macro
14651 to use new compression flags and to support microMIPS.
14652 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14653 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14654 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14655 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14656 (ASM_SPEC): Support mmicromips and mno-micromips.
14657 (M16STORE_REG_P): New macro.
14658 (MIPS_CALL): Support TARGET_MICROMIPS.
14659 (MICROMIPS_J): New macro.
14660 (mips_base_mips16): Rename this...
14661 (mips_base_compression_flags): ...to this.
14662 (UMIPS_12BIT_OFFSET_P): New macro.
14663 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14664 (MULTILIB_DIRNAMES): Likewise.
14665 2013-03-20 Richard Biener <rguenther@suse.de>
14667 PR tree-optimization/56661
14668 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14669 the result does not have to be distinct.
14671 2013-03-20 Richard Biener <rguenther@suse.de>
14673 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14676 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14677 Steven Bosscher <steven@gcc.gnu.org>
14679 PR rtl-optimization/56605
14680 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14682 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14685 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14686 that require movd instead of movq.
14688 2013-03-20 Richard Biener <rguenther@suse.de>
14690 * tree-ssa-structalias.c (struct variable_info): Add pointer
14691 to the first field of an aggregate with sub-vars. Make
14692 this and the pointer to the next subfield its ID.
14693 (vi_next): New function.
14694 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14695 storedanything_id, integer_id): Increment by one.
14696 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14697 get_call_clobber_vi): Adjust.
14698 (solution_set_expand): Simplify and speedup.
14699 (solution_set_add): Inline into ...
14700 (set_union_with_increment): ... this. Adjust accordingly.
14701 (do_sd_constraint): Likewise.
14702 (do_ds_constraint): Likewise.
14703 (do_complex_constraint): Simplify.
14704 (build_pred_graph): Adjust.
14705 (solve_graph): Likewise. Simplify and speedup.
14706 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14707 get_constraint_for_component_ref, get_constraint_for_1,
14708 first_vi_for_offset, first_or_preceding_vi_for_offset,
14709 create_function_info_for, create_variable_info_for_1,
14710 create_variable_info_for, intra_create_variable_infos): Adjust.
14711 (init_base_vars): Push NULL for ID zero.
14712 (compute_points_to_sets): Adjust.
14714 2013-03-20 Richard Biener <rguenther@suse.de>
14716 * cfgloop.c (verify_loop_structure): Streamline and avoid
14717 ICEing on corrupt loop tree.
14718 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14721 2013-03-20 Richard Biener <rguenther@suse.de>
14723 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14724 check whether an SSA update is needed.
14726 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14728 * config/mips/constraints.md (T): Rename to...
14732 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14733 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14735 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14737 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14738 (*subsi3_carryin_uxtw): Likewise.
14740 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14742 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14743 (*rorsi3_insn_uxtw): Likewise.
14745 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14747 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14748 (*extrsi5_insn_uxtw): Likewise.
14750 2013-03-19 Richard Biener <rguenther@suse.de>
14752 PR tree-optimization/56273
14753 * passes.c (init_optimization_passes): Move second VRP after DOM.
14755 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14757 * config/i386/i386.md (*movti_internal): Merge from
14758 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14759 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14760 nox64 isa attributes.
14762 2013-03-18 Richard Biener <rguenther@suse.de>
14764 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14766 (merge_node_constraints): Likewise.
14767 (build_succ_graph): Likewise.
14768 (valid_graph_edge): Inline into single caller.
14769 (unify_nodes): Likewise. Use bitmap_set_bit return value
14771 (scc_visit): Fix formatting and variable use.
14772 (do_sd_constraint): Use gcc_checking_assert.
14773 (do_ds_constraint): Likewise.
14774 (do_complex_constraint): Likewise.
14775 (condense_visit): Likewise. Cleanup.
14776 (dump_pred_graph): New function.
14777 (perform_var_substitution): Dump the pred-graph before
14778 variable substitution.
14779 (find_equivalent_node): Use gcc_checking_assert.
14780 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14782 2013-03-18 Richard Biener <rguenther@suse.de>
14784 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14785 Remove cond_expr_stmt_list argument and do not gimplify the
14787 (vect_loop_versioning): Adjust.
14788 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14789 Cleanup to use less temporaries.
14790 (vect_create_data_ref_ptr): Cleanup.
14792 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14794 PR tree-optimization/56635
14795 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14796 require types_compatible_p types.
14798 2013-03-18 Nick Clifton <nickc@redhat.com>
14800 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14801 spurious backslash.
14803 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14804 Add missing line to comment describing function.
14806 2013-03-18 Richard Biener <rguenther@suse.de>
14808 PR tree-optimization/56210
14809 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14810 Handle string / character search functions.
14811 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14813 2013-03-18 Richard Biener <rguenther@suse.de>
14815 PR middle-end/56483
14816 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14817 and implement properly.
14818 * gimple.h (gimple_cond_single_var_p): Remove.
14820 2013-03-18 Richard Biener <rguenther@suse.de>
14822 * tree-data-ref.h (find_data_references_in_loop): Declare.
14823 * tree-data-ref.c (get_references_in_stmt): Use a stack
14824 vector pre-allocated in the callers.
14825 (find_data_references_in_stmt): Adjust.
14826 (graphite_find_data_references_in_stmt): Likewise.
14827 (create_rdg_vertices): Likewise.
14828 (find_data_references_in_loop): Export.
14829 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14830 Compute dependences here...
14831 (vect_analyze_data_refs): ...not here. When we encounter
14832 a non-vectorizable data reference in basic-block vectorization
14833 truncate the data reference vector. Do not bother to
14834 fixup data-dependence information for gather loads.
14835 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14836 of data references, as reported.
14838 2013-03-18 Richard Biener <rguenther@suse.de>
14840 PR tree-optimization/3713
14841 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14842 has_constants and expr.
14843 (stmt_has_constants): Properly valueize SSA names when deciding
14844 whether the stmt has constants.
14846 2013-03-18 Richard Biener <rguenther@suse.de>
14848 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14849 whole function when there is nothing to do.
14850 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14851 * tree-vectorizer.c (vectorize_loops): Update virtual and
14852 loop-closed SSA once.
14853 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14855 2013-03-18 Richard Biener <rguenther@suse.de>
14857 PR middle-end/56113
14858 * domwalk.c (bb_postorder): New global static.
14859 (cmp_bb_postorder): New function.
14860 (walk_dominator_tree): Replace scheme imposing an order for
14861 visiting dominator sons by one sorting them at the time they
14862 are pushed on the stack.
14864 2013-03-18 Richard Biener <rguenther@suse.de>
14866 PR tree-optimization/39326
14867 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14868 (struct mem_ref): Replace mem member with ao_ref typed member.
14869 (MEM_ANALYZABLE): Adjust.
14870 (memref_eq): Likewise.
14871 (mem_ref_alloc): Likewise.
14872 (gather_mem_refs_stmt): Likewise.
14873 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14874 (execute_sm_if_changed_flag_set): Adjust.
14875 (execute_sm): Likewise.
14876 (ref_always_accessed_p): Likewise.
14877 (refs_independent_p): Likewise.
14878 (can_sm_ref_p): Likewise.
14880 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14883 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14884 return 1 even for !unsignedp.
14886 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14888 * config/i386/i386.md (isa): Add x64 and nox64.
14889 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14890 (*pushtf): Enable *roF alternative for x64 isa only.
14891 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14892 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14893 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14894 (*movtf_internal): Merge from *movtf_internal_rex64 and
14895 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14896 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14897 nox64 isa attributes.
14898 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14899 nox64 isa attributes.
14900 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14902 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14904 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14906 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14909 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14910 class names. Remove trailing comma after "ALL_REGS".
14912 2013-03-16 Jan Hubicka <jh@suse.cz>
14914 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14915 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14916 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14917 of cgraph_get_create_node.
14918 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14920 2013-03-16 Jason Merrill <jason@redhat.com>
14923 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14924 with DW_AT_default_value.
14926 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14928 * BASE-VER: Set to 4.9.0.
14930 2013-03-14 Andi Kleen <ak@linux.intel.com>
14933 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14934 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14935 Document _x* TSX intrinsics.
14937 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14938 David Holsgrove <david.holsgrove@xilinx.com>
14940 * configure.ac: Add MicroBlaze TLS support detection.
14941 * configure: Regenerate.
14942 * config/microblaze/microblaze-protos.h
14943 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14944 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14945 * config/microblaze/microblaze.c (microblaze_address_type): Add
14946 ADDRESS_TLS and tls_reloc address types.
14947 (microblaze_address_info): Add tls_reloc.
14948 (TARGET_HAVE_TLS): Define.
14949 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14950 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14951 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14952 load_tls_operand, microblaze_call_tls_get_addr,
14953 microblaze_legitimize_tls_address): New functions.
14954 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14955 (get_base_reg): Use microblaze_tls_symbol_p.
14956 (microblaze_classify_address): Handle TLS.
14957 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14958 label_mentioned_p and microblaze_tls_referenced_p.
14959 (microblaze_legitimize_address): Handle TLS.
14960 (microblaze_address_insns): Handle ADDRESS_TLS.
14961 (pic_address_needs_scratch): Handle TLS.
14962 (print_operand_address): Handle TLS.
14963 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14964 (microblaze_expand_move): Handle TLS.
14965 (microblaze_legitimate_constant_p): Check
14966 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14967 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14968 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14969 (PIC_OFFSET_TABLE_REGNUM): Set.
14970 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14971 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14972 (addsi3, movsi_internal2, movdf_internal): Update constraints
14973 * config/microblaze/predicates.md (arith_plus_operand): Define
14974 (move_operand): Redefine as move_src_operand,
14975 check microblaze_tls_referenced_p.
14977 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14979 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14980 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14982 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14984 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14987 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14989 PR tree-optimization/53265
14990 * common.opt (Waggressive-loop-optimizations): New option.
14991 * tree-ssa-loop-niter.c: Include tree-pass.h.
14992 (do_warn_aggressive_loop_optimizations): New function.
14993 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14994 if number_of_latch_executions returned constant.
14995 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14996 early. If number_of_latch_executions returned constant, set
14997 nb_iterations_upper_bound back to it.
14998 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
15000 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
15001 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
15003 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
15004 (MULTILIB_OSDIRNAMES): Set.
15005 * genmultilib: If defaultosdirname doesn't start with :: , set
15006 defaultosdirname2 instead, clear it and emit two . multilib_raw
15007 entries instead of just one.
15009 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
15011 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15012 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
15013 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15014 (SUBTARGET_OVERRIDE_OPTIONS): New.
15016 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
15019 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
15020 (musermode): Convert to Var(TARGET_USERMODE).
15021 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
15022 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
15023 * config/sh/sh.c (sh_option_override): Use
15024 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
15025 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
15027 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
15029 (udivsi3_i4_single, divsi3_i4_single): Use
15030 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
15032 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15034 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15037 2013-03-13 Richard Biener <rguenther@suse.de>
15039 PR tree-optimization/56608
15040 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15041 calls when vectorizing basic-blocks.
15043 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15046 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15049 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15051 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15053 2013-03-11 Jan Hubicka <jh@suse.cz>
15056 * lto-streamer-out.c (output_symbol_p): Skip references from
15057 constructors of external variables.
15059 2013-03-11 Jan Hubicka <jh@suse.cz>
15061 PR middle-end/56571
15062 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15064 * emit-rtl.c (verify_rtx_sharing): Likewise.
15065 (copy_insn_1): Likewise.
15066 * rtl.c (copy_rtx): Likewise.
15068 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15071 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15072 output_operand_lossage message.
15074 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15077 * arm.c (shift_op): Validate RTL pattern on the fly.
15078 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15081 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15084 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15085 new call patterns that clobber %fr12.
15086 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15087 split and postreload patterns.
15088 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15089 registers %fr12 and %fr12R as call used.
15091 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15093 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15094 (canon_address, record_store, replace_read, check_mem_read_rtx,
15095 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15096 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15097 rest_of_handle_dse): Likewise.
15099 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15101 PR middle-end/56524
15102 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15104 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15105 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15106 (save_optabs_if_changed): Replace with...
15107 (init_tree_optimization_optabs): ...this.
15108 * optabs.c (save_optabs_if_changed): Rename to...
15109 (init_tree_optimization_optabs): ...this. Take the optimization node
15110 as argument. Do nothing if the base optabs are already correct.
15111 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15112 to recompute optabs.
15113 * function.h (function): Remove optabs field.
15114 * function.c (invoke_set_current_function_hook): Call
15115 init_tree_optimization_optabs. Use the result to initialize
15118 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15120 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15121 if GTMA_HAS_NO_INSTRUMENTATION.
15122 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15123 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15124 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15125 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15126 GTMA_HAS_NO_INSTRUMENTATION.
15128 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15130 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15133 2013-03-08 Marek Polacek <polacek@redhat.com>
15134 Jakub Jelinek <jakub@redhat.com>
15136 PR tree-optimization/56478
15137 * predict.c (is_comparison_with_loop_invariant_p): Change the
15138 type of loop_step to tree.
15139 (predict_loops): Adjust.
15140 (predict_iv_comparison): Perform the computations on double_ints.
15142 2013-03-08 Richard Biener <rguenther@suse.de>
15144 PR tree-optimization/56570
15145 * tree-cfg.c (verify_expr_location_1): Verify locations for
15147 * tree-sra.c (create_access_replacement): Strip locations
15148 from DECL_DEBUG_EXPRs.
15150 2013-03-08 Richard Biener <rguenther@suse.de>
15152 * tree-inline.c (expand_call_inline): Do not associate
15153 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15154 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15156 2013-03-08 Richard Biener <rguenther@suse.de>
15158 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15159 or block changes with -Og. Fix for location / block encoding
15160 changes and PHI arguments with locations.
15162 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15164 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15166 (struct output_info): Likewise.
15167 (register_overhead): Remove bad gcc_assert.
15168 (bitmap_find_bit): If there is only a single bitmap element, do not
15169 count a miss as a search.
15170 (print_statistics): Update for counter type changes.
15171 (dump_bitmap_statistics): Likewise. Print headers such that they
15172 are properly lined up with the printed counters.
15174 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15176 PR tree-optimization/56559
15177 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15178 check that it has only a single use.
15180 2013-03-07 Richard Biener <rguenther@suse.de>
15182 * doc/invoke.texi (fwhole-program): Discourage use in combination
15185 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15187 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15189 PR tree-optimization/56539
15190 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15191 instead of GSI_CONTINUE_LINKING as last argument to
15192 force_gimple_operand_gsi. Adjust function comment.
15194 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15197 PR middle-end/56548
15198 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15199 promoted mode, convert the result back to the original mode.
15201 2013-03-06 Richard Biener <rguenther@suse.de>
15203 PR middle-end/56294
15204 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15205 (insert_updated_phi_nodes_compare_uids): New function.
15206 (update_ssa): Sort symbols_to_rename after UID before
15207 traversing it to insert PHI nodes.
15209 2013-03-06 Richard Biener <rguenther@suse.de>
15211 PR middle-end/50494
15212 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15213 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15216 2013-02-13 Richard Biener <rguenther@suse.de>
15219 * varasm.c (output_constant_def_1): Get the decl representing
15220 the constant as argument.
15221 (output_constant_def): Wrap output_constant_def_1.
15222 (make_decl_rtl): Use output_constant_def_1 with the decl
15223 representing the constant.
15224 (build_constant_desc): Optionally re-use a decl already
15225 representing the constant.
15226 (tree_output_constant_def): Adjust.
15228 2013-03-06 Joey Ye <joey.ye@arm.com>
15231 * gcc.c (convert_white_space): New function.
15232 (main): Handles white space in function name.
15234 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15237 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15238 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15239 to SH_DIV_CALL_TABLE for TARGET_SH2.
15240 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15242 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15243 call-table options.
15245 2013-03-05 Sterling Augustine <saugustine@google.com>
15246 Cary Coutant <ccoutant@google.com>
15249 * dwarf2out.c (resolve_addr): Don't call
15250 remove_loc_list_addr_table_entries a second time for the same
15253 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15256 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15257 (avoid_complex_debug_insns): New function.
15258 (expand_debug_locations): Call it.
15260 PR rtl-optimization/56484
15261 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
15262 lifetimes of hard registers on small register class machines.
15264 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15266 * config/microblaze/microblaze-protos.h: Rename
15267 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
15268 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
15270 (microblaze_fast_interrupt_function_p): New function.
15271 (microblaze_is_interrupt_handler): Rename to
15272 microblaze_is_interrupt_variant and add fast_interrupt check.
15273 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
15274 (save_restore_insns): Likewise.
15275 (compute_frame_size): Likewise.
15276 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
15277 (microblaze_globalize_label): Likewise.
15278 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
15279 * config/microblaze/microblaze.md: Use wrapper
15280 microblaze_is_interrupt_variant.
15282 2013-03-05 Kai Tietz <ktietz@redhat.com>
15284 * sdbout.c (sdbout_one_type): Switch to current function's section
15285 supporting cold/hot.
15287 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15289 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
15292 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15294 PR middle-end/56461
15295 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
15296 if VALGRIND_GET_VBITS is defined, temporarily make object
15297 memory all defined, and restore previous valgrind addressability
15298 and definability afterwards. Free this_object at the end.
15300 PR middle-end/56461
15301 * lra.c (lra): Call lra_clear_live_ranges if live_p,
15302 right before calling lra_create_live_ranges, also call it
15303 when clearing live_p. Only call lra_clear_live_ranges
15304 at the end if live_p.
15306 PR middle-end/56461
15307 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
15309 2013-03-05 Richard Biener <rguenther@suse.de>
15311 PR tree-optimization/56521
15312 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
15315 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
15318 * except.h (remove_unreachable_eh_regions): New prototype.
15319 * except.c (remove_eh_handler_splicer): New function, split out
15320 of remove_eh_handler.
15321 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
15322 warning about running it on many EH regions one at a time.
15323 (remove_unreachable_eh_regions_worker): New function, walk the
15324 EH tree in depth-first order and remove non-marked regions.
15325 (remove_unreachable_eh_regions): New function.
15326 * tree-eh.c (mark_reachable_handlers): New function, split out
15327 from remove_unreachable_handlers.
15328 (remove_unreachable_handlers): Use mark_reachable_handlers and
15329 remove_unreachable_eh_regions.
15330 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
15331 and remove_unreachable_eh_regions.
15333 2013-03-05 Richard Biener <rguenther@suse.de>
15335 PR middle-end/56525
15336 * loop-init.c (fix_loop_structure): Remove loops in two stages,
15337 not freeing them until the end.
15339 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15341 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
15343 2013-03-05 Richard Biener <rguenther@suse.de>
15345 PR tree-optimization/56270
15346 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
15347 of loads after scheduling an SLP instance.
15349 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15351 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
15353 (check_gcc_parallelize): Run guality.exp as a separate job from
15354 vect.exp with unsorted.exp and $(dg_target_exps) separately from
15355 struct-layout-1.exp with stackalign.exp.
15357 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
15359 PR middle-end/56461
15360 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
15361 load_index sbitmap even if some bit in it isn't set.
15363 PR middle-end/56461
15364 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
15365 (discover_iteration_bound_by_body_walk): Change queues to
15366 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
15367 spelling in comment. Call safe_push on queues[bound_index] directly.
15368 Release queues[queue_index] in every iteration unconditionally.
15369 Release bounds vector.
15371 PR middle-end/56461
15372 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
15373 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
15374 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
15377 2013-03-05 Richard Biener <rguenther@suse.de>
15380 * tree-inline.c (remap_blocks_to_null): New function.
15381 (expand_call_inline): When expanding a call stmt without
15382 an associated block inline remap all callee blocks to NULL.
15384 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15386 PR rtl-optimization/56494
15387 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
15388 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
15389 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
15391 PR middle-end/56461
15392 * sel-sched-ir.c (free_sched_pools): Release
15393 succs_info_pool.stack[succs_info_pool.max_top] vectors too
15394 if succs_info_pool.max_top isn't -1.
15397 * opts.c (opts_obstack, opts_concat): Moved to...
15398 * opts-common.c (opts_obstack, opts_concat): ... here.
15400 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15402 PR middle-end/56461
15403 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
15405 2013-03-04 Martin Jambor <mjambor@suse.cz>
15407 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
15408 all appropriate places.
15410 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
15412 PR tree-optimization/56424
15413 * ipa-split.c (split_function): Do not set the RSO flag if result is
15414 not by reference and its type is a register type.
15416 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
15418 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
15419 (microblaze_legitimate_pic_operand): Likewise
15420 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
15421 new function microblaze_legitimate_pic_operand
15422 * config/microblaze/microblaze-protos.h
15423 (microblaze_legitimate_pic_operand): Declare.
15425 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15427 * config/microblaze/predicates.md (call_insn_simple_operand):
15428 New predicate for supported rtx code types.
15429 * config/microblaze/microblaze.md (call_internal1): Use
15430 call_insn_simple_operand predicate.
15432 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15434 PR middle-end/56461
15435 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
15436 partitions.ordered_remove.
15438 PR middle-end/56461
15439 * tree-vect-stmts.c (vectorizable_conversion): Don't call
15440 vec_oprnds0.create (1) for modifier == NONE.
15442 PR middle-end/56461
15443 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
15444 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
15445 vec_oprnds1 right before pushing anything to it for
15448 PR middle-end/56461
15449 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
15450 set nbbs to 0 instead of having separate code path.
15451 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
15452 instead of false as last argument if returning NULL.
15454 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
15456 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
15457 the attribute is now called "target" instead of "option".
15458 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
15459 * doc/tm.texi.in (Target Attributes): Likewise document the correct
15460 attribute/pragma name for TARGET_OPTION_VALID_P and
15461 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
15462 * doc/tm.texi: Regenerated.
15464 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
15466 * config/microblaze/microblaze.c:
15467 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
15468 * config/microblaze/microblaze.h: Add -mxl-reorder to
15470 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
15471 instructions emitted if TARGET_REORDER.
15472 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
15473 or 0 for -m/-mno case, but initialises as 2 to detect default use case
15476 2013-03-01 Xinliang David Li <davidxl@google.com>
15478 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
15481 2013-03-01 Jakub Jelinek <jakub@redhat.com>
15483 PR middle-end/56461
15484 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
15485 vector even when returning true. Fix up function comment formatting.
15487 PR middle-end/56461
15488 * ira-build.c (ira_loop_nodes_count): New variable.
15489 (create_loop_tree_nodes): Initialize it.
15490 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
15492 PR middle-end/56461
15493 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
15494 method on dr_chain and result_chain.
15495 * tree-vect-stmts.c (vectorizable_store): Only call
15496 result_chain.create if j == 0.
15498 PR middle-end/56461
15499 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
15500 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
15501 before overwriting it.
15503 2013-03-01 Tobias Burnus <burnus@net-b.de>
15505 * doc/extended.texi (C Extensions): Change order in @menu
15507 (Other MIPS Built-in Functions): Move last MIPS entry before
15508 "picoChip Built-in Functions".
15509 (SH Built-in Functions): Move after RX Built-in Functions.
15510 * doc/gcc.texi (Introduction): Change order in @menu
15512 * doc/md.texi (Constraints): Ditto.
15513 * gty.texi (Type Information): Ditto.
15514 (User-provided marking routines for template types): Make
15516 * doc/invoke.texi (AArch64 Options): Move before
15517 "Adapteva Epiphany Options".
15519 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
15520 Jakub Jelinek <jakub@redhat.com>
15523 * asan.c (gate_asan): Lookup no_sanitize_address instead of
15524 no_address_safety_analysis attribute.
15525 * doc/extend.texi (no_address_safety_attribute): Rename to
15526 no_sanitize_address attribute, mention no_address_safety_analysis
15527 attribute as deprecated alias.
15529 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15531 PR middle-end/56461
15532 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15533 type to vec<vec<tree> > *.
15534 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15535 to be vec<tree> instead of vec<tree> *, set vec_defs
15536 to vNULL and call vec_defs.create (number_of_vects), adjust other
15538 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15539 vectorizable_condition): Adjust vect_get_slp_defs callers.
15541 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15543 * config/aarch64/aarch64.c
15544 (aarch64_float_const_representable): Remove unused variable.
15546 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15548 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15550 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15552 * config/aarch64/aarch64-builtins.c
15553 (aarch64_init_simd_builtins): Make static.
15555 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15557 * config/aarch64/aarch64.c
15558 (aarch64_simd_make_constant): Make static.
15560 2013-02-28 Martin Jambor <mjambor@suse.cz>
15562 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15563 with no initialization to the RHS of debug statements.
15565 2013-02-28 Martin Jambor <mjambor@suse.cz>
15567 PR tree-optimization/56294
15568 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15570 (get_access_replacement): Do not call create_access_replacement.
15571 Assert a replacement exists.
15572 (get_repl_default_def_ssa_name): Create the replacement declaration
15575 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15577 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15578 final_end_function.
15580 2013-02-28 Marek Polacek <polacek@redhat.com>
15582 PR rtl-optimization/56466
15583 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15584 if we're changing a loop.
15585 (peel_loops_completely): Likewise.
15587 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15590 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15592 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15595 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15596 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15597 INTX_FTYPE_FX, FX_FTYPE_INTX.
15598 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15600 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15602 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15603 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15604 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15605 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15606 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15607 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15608 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15609 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15610 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15611 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15612 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15613 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15614 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15615 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15616 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15617 (avrxmega6): Increase max flash segments from 5 to 6.
15618 * config/avr/t-multilib: Regenerate.
15619 * config/avr/avr-tables.opt: Regenerate.
15620 * doc/avr-mmcu.texi: Regenerate.
15622 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15624 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15625 (avr_device_to_arch): Rename to avr_device_to_ld.
15626 (avr_device_to_as): New prototype.
15627 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15628 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15629 * config/avr/driver-avr.c (avr_device_to_as): New.
15630 (avr_device_to_arch): Rename to avr_device_to_ld.
15632 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15634 PR middle-end/56461
15635 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15636 method on dr_chain and result_chain.
15638 PR middle-end/56461
15639 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15640 pointer_set_destroy on not_executed_last_iteration.
15642 PR middle-end/56461
15643 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15645 PR middle-end/56461
15646 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15647 FOR_EACH_DEFINED_FUNCTION when freeing state.
15649 PR middle-end/56461
15650 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15652 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15655 PR middle-end/56461
15656 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15657 known_aggs[i].items and release known_aggs vector.
15659 PR middle-end/56461
15660 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15662 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15664 * config/microblaze/microblaze.c (microblaze_emit_compare):
15665 Use xor for EQ/NE comparisions.
15666 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15667 (cbranchsf4): Adjust operator to comparison_operator.
15669 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15671 PR middle-end/56461
15672 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15674 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15675 vec_safe_push, always update *slot.
15676 (redirect_edge_var_map_clear): Use vec_free.
15677 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15678 (free_var_map_entry): Use vec_free.
15679 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15680 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15682 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15684 PR middle-end/45472
15685 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15686 when the may_trap_p bit of the exprs being merged differs.
15687 Reorder tests for speculativeness in the logical and operator.
15689 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15691 * incpath.c (add_standard_paths): Use reconcat instead of concat
15692 where appropriate and avoid leaking memory.
15694 * opts.h: Include obstack.h.
15695 (opts_concat): New prototype.
15696 (opts_obstack): New declaration.
15697 * opts.c (opts_concat): New function.
15698 (opts_obstack): New variable.
15699 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15700 (finish_options): Use opts_concat instead of concat
15701 and XOBNEWVEC instead of XNEWVEC.
15702 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15703 generate_option): Likewise.
15704 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15705 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15708 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15709 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15711 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15713 PR middle-end/56461
15714 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15716 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15718 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15719 (arm_block_move_unaligned_straight): Likewise.
15720 (arm_adjust_block_mem): Likewise.
15722 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15725 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15726 temp, cond and label.
15727 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15730 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15731 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15734 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15735 of prologue/epilogue functions.
15738 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15739 Remove unused variable cfa_offset.
15740 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15743 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15744 type promotion to unsigned.
15747 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15748 for HOST_WIDE_INT of 32 bit / same size as int.
15749 (arm_block_move_unaligned_straight): Likewise.
15750 (arm_adjust_block_mem): Likewise.
15753 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15756 2013-02-26 Marek Polacek <polacek@redhat.com>
15758 PR tree-optimization/56426
15759 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15761 2013-02-26 Richard Biener <rguenther@suse.de>
15764 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15765 unused variable loops.
15767 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15769 PR tree-optimization/56448
15770 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15771 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15772 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15773 later operands of the references, or even first operand for
15774 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15776 PR tree-optimization/56443
15777 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15778 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15779 to type_for_mode langhook.
15781 2013-02-25 Matt Turner <mattst88@gmail.com>
15783 * doc/invoke.texi: Document r4700.
15785 2013-02-25 Richard Biener <rguenther@suse.de>
15787 PR tree-optimization/56175
15788 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15790 (simplify_bitwise_binary): ... here. Also guard the conversion
15791 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15793 2013-02-25 Catherine Moore <clm@codesourcery.com>
15796 2013-02-24 Catherine Moore <clm@codesourcery.com>
15797 Maciej W. Rozycki <macro@codesourcery.com>
15798 Tom de Vries <tom@codesourcery.com>
15799 Nathan Sidwell <nathan@codesourcery.com>
15800 Iain Sandoe <iain@codesourcery.com>
15801 Nathan Froyd <froydnj@codesourcery.com>
15802 Chao-ying Fu <fu@mips.com>
15804 * doc/extend.texi: (micromips, nomicromips, nocompression):
15805 Document new function attributes.
15806 * doc/invoke.texi (minterlink-compressed, mmicromips,
15807 m14k, m14ke, m14kec): Document new options.
15808 (minterlink-mips16): Update documentation.
15809 * doc/md.texi (ZC, ZD): Document new constraints.
15810 * configure.ac (gcc_cv_as_micromips): Check if linker
15811 supports the .set micromips directive.
15812 * configure: Regenerate.
15813 * config.in: Regenerate.
15814 * config/mips/mips-tables.opt: Regenerate.
15815 * config/mips/micromips.md: New file.
15816 * constraints.md (ZC, AD): New constraints.
15817 * config/mips/predicates.md (movep_src_register): New predicate.
15818 (movep_src_operand): New predicate.
15819 (non_volatile_mem_operand): New predicate.
15820 * config/mips/mips.md (multimem): New type.
15821 (length): Differentiate between 17-bit and 18-bit branch offsets.
15822 (MOVEP1, MOVEP2): New mode iterator.
15823 (mov_<load>l): Use ZC constraint.
15824 (mov_<load>r): Likewise.
15825 (mov_<store>l): Likewise.
15826 (mov_<store>r): Likewise.
15827 (*branch_equality<mode>_inverted): Add microMIPS support.
15828 (*branch_equality<mode>): Likewise.
15829 (*jump_absolute): Likewise.
15830 (indirect_jump_<mode>): Likewise.
15831 (tablejump_<mode>): Likewise.
15832 (<optab>_internal): Likewise.
15833 (sibcall_internal): Likewise.
15834 (sibcall_value_internal): Likewise.
15835 (prefetch): Use constraint ZD.
15836 * config/mips/mips.opt (minterlink-compressed): New option.
15837 (minterlink-mips16): Now an alias for minterlink-compressed.
15838 (mmicromips): New option.
15839 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15840 (compare_and_swap_12): Likewise.
15841 (sync_add<mode>): Likewise.
15842 (sync_<optab>_12): Likewise.
15843 (sync_old_<optab>_12): Likewise.
15844 (sync_new_<optab>_12): Likewise.
15845 (sync_nand_12): Likewise.
15846 (sync_old_nand_12): Likewise.
15847 (sync_new_nand_12): Likewise.
15848 (sync_sub<mode>): Likewise.
15849 (sync_old_add<mode>): Likewise.
15850 (sync_old_sub<mode>): Likewise.
15851 (sync_new_add<mode>): Likewise.
15852 (sync_new_sub<mode>): Likewise.
15853 (sync_<optab><mode>): Likewise.
15854 (sync_old_<optab><mode>): Likewise.
15855 (sync_new_<optab><mode>): Likewise.
15856 (sync_nand<mode>): Likewise.
15857 (sync_old_nand<mode>): Likewise.
15858 (sync_new_nand<mode>): Likewise.
15859 (sync_lock_test_and_set<mode>): Likewise.
15860 (test_and_set_12): Likewise.
15861 (atomic_compare_and_swap<mode>): Likewise.
15862 (atomic_exchange<mode>_llsc): Likewise.
15863 (atomic_fetch_add<mode>_llsc): Likewise.
15864 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15865 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15866 (umips_save_restore_pattern_p): Likewise.
15867 (umips_load_store_pair_p): Likewise.
15868 (umips_output_load_store_pair): Likewise.
15869 (umips_movep_target_p): Likewise.
15870 (umips_12bit_offset_address_p): Likewise.
15871 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15872 (mips_base_mips16): Rename this...
15873 (mips_base_compression_flags): ...to this. Update all uses.
15874 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15875 (mips_mips16_decl_p): Delete.
15876 (mips_nomips16_decl_p): Delete.
15877 (mips_get_compress_on_flags): New function.
15878 (mips_get_compress_off_flags): New function.
15879 (mips_get_compress_mode): New function.
15880 (mips_get_compress_on_name): New function.
15881 (mips_get_compress_off_name): New function.
15882 (mips_insert_attributes): Support multiple compression types.
15883 (mips_merge_decl_attributes): Likewise.
15884 (umips_12bit_offset_address_p): New function.
15885 (mips_start_function_definition): Emit .set micromips directive.
15886 (mips_call_may_need_jalx_p): New function.
15887 (mips_function_ok_for_sibcall): Add microMIPS support.
15888 (mips_print_operand_punctuation): Support short delay slots and
15890 (umips_swm_mask, umips_swm_encoding): New.
15891 (umips_build_save_restore): New function.
15892 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15893 (was_mips16_p): Remove.
15894 (old_compression_mode): New.
15895 (mips_set_compression_mode): New function.
15896 (mips_set_current_function): Add microMIPS support.
15897 (mips_option_override): Likewise.
15898 (umips_save_restore_pattern_p): New function.
15899 (umips_output_save_restore): New function.
15900 (umips_load_store_pair_p_1): New function.
15901 (umips_load_store_pair_p): New function.
15902 (umips_output_load_store_pair_1): New function.
15903 (umips_output_load_store_pair): New function.
15904 (umips_movep_target_p) New function.
15905 (mips_prepare_pch_save): Add microMIPS support.
15906 * config/mips/mips.h (TARGET_COMPRESSION): New.
15907 (TARGET_CPU_CPP_BUILTINS): Update macro
15908 to use new compression flags and to support microMIPS.
15909 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15910 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15911 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15912 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15913 (ASM_SPEC): Support mmicromips and mno-micromips.
15914 (M16STORE_REG_P): New macro.
15915 (MIPS_CALL): Support TARGET_MICROMIPS.
15916 (MICROMIPS_J): New macro.
15917 (mips_base_mips16): Rename this...
15918 (mips_base_compression_flags): ...to this.
15919 (UMIPS_12BIT_OFFSET_P): New macro.
15920 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15921 (MULTILIB_DIRNAMES): Likewise.
15923 2013-02-25 Tom de Vries <tom@codesourcery.com>
15925 PR rtl-optimization/56131
15926 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15927 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15928 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15930 2013-02-25 Tobias Burnus <burnus@net-b.de>
15932 * doc/invoke.texi (-fsanitize=): Move from optimization
15933 to debugging options.
15935 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15937 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15939 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15940 Alexander Monakov <amonakov@ispras.ru>
15942 PR middle-end/56077
15943 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15944 flush pending lists also on non-jumps. Adjust comment.
15946 2013-02-24 Catherine Moore <clm@codesourcery.com>
15947 Maciej W. Rozycki <macro@codesourcery.com>
15948 Tom de Vries <tom@codesourcery.com>
15949 Nathan Sidwell <nathan@codesourcery.com>
15950 Iain Sandoe <iain@codesourcery.com>
15951 Nathan Froyd <froydnj@codesourcery.com>
15952 Chao-ying Fu <fu@mips.com>
15954 * doc/extend.texi: (micromips, nomicromips, nocompression):
15955 Document new function attributes.
15956 * doc/invoke.texi (minterlink-compressed, mmicromips,
15957 m14k, m14ke, m14kec): Document new options.
15958 (minterlink-mips16): Update documentation.
15959 * doc/md.texi (ZC, ZD): Document new constraints.
15960 * configure.ac (gcc_cv_as_micromips): Check if linker
15961 supports the .set micromips directive.
15962 * configure: Regenerate.
15963 * config.in: Regenerate.
15964 * config/mips/mips-tables.opt: Regenerate.
15965 * config/mips/micromips.md: New file.
15966 * constraints.md (ZC, AD): New constraints.
15967 * config/mips/predicates.md (movep_src_register): New predicate.
15968 (movep_src_operand): New predicate.
15969 (non_volatile_mem_operand): New predicate.
15970 * config/mips/mips.md (multimem): New type.
15971 (length): Differentiate between 17-bit and 18-bit branch offsets.
15972 (MOVEP1, MOVEP2): New mode iterator.
15973 (mov_<load>l): Use ZC constraint.
15974 (mov_<load>r): Likewise.
15975 (mov_<store>l): Likewise.
15976 (mov_<store>r): Likewise.
15977 (*branch_equality<mode>_inverted): Add microMIPS support.
15978 (*branch_equality<mode>): Likewise.
15979 (*jump_absolute): Likewise.
15980 (indirect_jump_<mode>): Likewise.
15981 (tablejump_<mode>): Likewise.
15982 (<optab>_internal): Likewise.
15983 (sibcall_internal): Likewise.
15984 (sibcall_value_internal): Likewise.
15985 (prefetch): Use constraint ZD.
15986 * config/mips/mips.opt (minterlink-compressed): New option.
15987 (minterlink-mips16): Now an alias for minterlink-compressed.
15988 (mmicromips): New option.
15989 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15990 (compare_and_swap_12): Likewise.
15991 (sync_add<mode>): Likewise.
15992 (sync_<optab>_12): Likewise.
15993 (sync_old_<optab>_12): Likewise.
15994 (sync_new_<optab>_12): Likewise.
15995 (sync_nand_12): Likewise.
15996 (sync_old_nand_12): Likewise.
15997 (sync_new_nand_12): Likewise.
15998 (sync_sub<mode>): Likewise.
15999 (sync_old_add<mode>): Likewise.
16000 (sync_old_sub<mode>): Likewise.
16001 (sync_new_add<mode>): Likewise.
16002 (sync_new_sub<mode>): Likewise.
16003 (sync_<optab><mode>): Likewise.
16004 (sync_old_<optab><mode>): Likewise.
16005 (sync_new_<optab><mode>): Likewise.
16006 (sync_nand<mode>): Likewise.
16007 (sync_old_nand<mode>): Likewise.
16008 (sync_new_nand<mode>): Likewise.
16009 (sync_lock_test_and_set<mode>): Likewise.
16010 (test_and_set_12): Likewise.
16011 (atomic_compare_and_swap<mode>): Likewise.
16012 (atomic_exchange<mode>_llsc): Likewise.
16013 (atomic_fetch_add<mode>_llsc): Likewise.
16014 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16015 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16016 (umips_save_restore_pattern_p): Likewise.
16017 (umips_load_store_pair_p): Likewise.
16018 (umips_output_load_store_pair): Likewise.
16019 (umips_movep_target_p): Likewise.
16020 (umips_12bit_offset_address_p): Likewise.
16021 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16022 (mips_base_mips16): Rename this...
16023 (mips_base_compression_flags): ...to this. Update all uses.
16024 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16025 (mips_mips16_decl_p): Delete.
16026 (mips_nomips16_decl_p): Delete.
16027 (mips_get_compress_on_flags): New function.
16028 (mips_get_compress_off_flags): New function.
16029 (mips_get_compress_mode): New function.
16030 (mips_get_compress_on_name): New function.
16031 (mips_get_compress_off_name): New function.
16032 (mips_insert_attributes): Support multiple compression types.
16033 (mips_merge_decl_attributes): Likewise.
16034 (umips_12bit_offset_address_p): New function.
16035 (mips_start_function_definition): Emit .set micromips directive.
16036 (mips_call_may_need_jalx_p): New function.
16037 (mips_function_ok_for_sibcall): Add microMIPS support.
16038 (mips_print_operand_punctuation): Support short delay slots and
16040 (umips_swm_mask, umips_swm_encoding): New.
16041 (umips_build_save_restore): New function.
16042 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16043 (was_mips16_p): Remove.
16044 (old_compression_mode): New.
16045 (mips_set_compression_mode): New function.
16046 (mips_set_current_function): Add microMIPS support.
16047 (mips_option_override): Likewise.
16048 (umips_save_restore_pattern_p): New function.
16049 (umips_output_save_restore): New function.
16050 (umips_load_store_pair_p_1): New function.
16051 (umips_load_store_pair_p): New function.
16052 (umips_output_load_store_pair_1): New function.
16053 (umips_output_load_store_pair): New function.
16054 (umips_movep_target_p) New function.
16055 (mips_prepare_pch_save): Add microMIPS support.
16056 * config/mips/mips.h (TARGET_COMPRESSION): New.
16057 (TARGET_CPU_CPP_BUILTINS): Update macro
16058 to use new compression flags and to support microMIPS.
16059 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16060 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16061 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16062 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16063 (ASM_SPEC): Support mmicromips and mno-micromips.
16064 (M16STORE_REG_P): New macro.
16065 (MIPS_CALL): Support TARGET_MICROMIPS.
16066 (MICROMIPS_J): New macro.
16067 (mips_base_mips16): Rename this...
16068 (mips_base_compression_flags): ...to this.
16069 (UMIPS_12BIT_OFFSET_P): New macro.
16070 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16071 (MULTILIB_DIRNAMES): Likewise.
16073 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16076 * target-globals.c (save_target_globals): For init_reg_sets and
16077 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16079 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16081 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16082 * config/aarch64/t-aarch64
16083 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16085 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16087 PR inline-asm/56148
16088 * lra-constraints.c (process_alt_operands): Reload operand
16089 conflicting with earlier clobber only if no more other conflicting
16092 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16095 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16096 if not linking a shared library.
16098 2013-02-22 Seth LaForge <sethml@google.com>
16100 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16102 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16104 * config/arm/arm.md (split for extendsidi): Update condition.
16105 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16106 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16107 (qhs_zextenddi_cstr): Likewise.
16109 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16111 PR middle-end/56420
16112 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16113 avoid signed wrapping.
16114 (expand_mult): Handle properly multiplication by
16115 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16116 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16117 in the compiler if coeff is HOST_WIDE_INT_MIN.
16118 (expand_divmod): Don't make ext_op1 static, change it's type to
16119 uhwi. Avoid undefined behavior in -INTVAL (op1).
16121 PR rtl-optimization/50339
16122 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16124 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16125 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16126 into splitting_ashiftrt field.
16127 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16129 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16132 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16134 PR middle-end/56108
16135 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16136 are sure to go irrevocable.
16138 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16140 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16141 scalars are valid operands.
16143 2013-02-21 Martin Jambor <mjambor@suse.cz>
16145 PR tree-optimization/56310
16146 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16147 only matching indices and non-negative final offsets.
16148 (intersect_aggregates_with_edge): Pass src_idx to
16149 agg_replacements_to_vector. Pass src_idx insstead of index to
16150 intersect_with_agg_replacements.
16152 2013-02-21 Martin Jambor <mjambor@suse.cz>
16154 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16155 instead of hard-wired defaults.
16157 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16159 * doc/invoke.texi (MIPS Options): Update documentation of the
16160 floating-point multiply-accumulate instruction restrictions.
16162 2013-02-21 Kostya Serebryany <kcc@google.com>
16164 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16165 asan_shadow_offset on x86_64 linux.
16167 2013-02-21 Richard Biener <rguenther@suse.de>
16169 PR tree-optimization/56415
16171 2013-02-11 Richard Biener <rguenther@suse.de>
16173 PR tree-optimization/56273
16174 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16177 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16180 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16183 PR inline-asm/56405
16184 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16185 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16187 2013-02-20 Jan Hubicka <jh@suse.cz>
16189 PR tree-optimization/56265
16190 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16191 when target is referenced for first time.
16193 2013-02-20 Richard Biener <rguenther@suse.de>
16195 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16196 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16197 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16198 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16199 not return anything.
16200 (rename_ssa_copies): Do not remove unused locals.
16201 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16202 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16203 * passes.c (execute_function_todo): Do not schedule unused locals
16204 removal if cleanup_tree_cfg did something.
16205 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16206 about the number of removed locals.
16208 2013-02-20 Richard Biener <rguenther@suse.de>
16210 PR tree-optimization/56398
16211 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16213 2013-02-20 Martin Jambor <mjambor@suse.cz>
16215 PR tree-optimization/55334
16216 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16217 restricted pointers to arrays.
16219 2013-02-20 Richard Biener <rguenther@suse.de>
16220 Jakub Jelinek <jakub@redhat.com>
16222 PR tree-optimization/56396
16223 * tree-ssa-ccp.c (n_const_val): New static variable.
16224 (get_value): Return NULL for SSA names we don't have a lattice
16226 (ccp_initialize): Initialize n_const_val.
16227 * tree-ssa-copy.c (n_copy_of): New static variable.
16228 (init_copy_prop): Initialize n_copy_of.
16229 (get_value): Return NULL_TREE for SSA names we don't have a
16232 2013-02-20 Martin Jambor <mjambor@suse.cz>
16234 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16236 2013-02-20 Richard Biener <rguenther@suse.de>
16238 * genpreds.c (write_lookup_constraint): Do not compare first
16239 letter of the constraint again.
16241 2013-02-20 Richard Biener <rguenther@suse.de>
16243 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16245 (get_use_iv_cost): Terminate hashtable walk when coming across
16248 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16250 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16251 reassociation for avx2 targets.
16253 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16255 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16256 Add version check for v8.10.a to enable microblaze_has_clz
16257 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16258 version and TARGET_PATTERN_COMPARE check
16259 * config/microblaze/microblaze.md: New clzsi2 instruction
16261 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16263 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
16264 function before branching.
16266 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16268 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
16270 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
16272 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16274 PR middle-end/55889
16275 * sel-sched.c: Include ira.h.
16276 (implicit_clobber_conflict_p): New function.
16277 (moveup_expr): Use it.
16278 * Makefile.in (sel-sched.o): Depend on ira.h.
16280 2013-02-19 Richard Biener <rguenther@suse.de>
16282 PR tree-optimization/56384
16283 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
16284 (vn_hash_type): Split out from ...
16285 (vn_hash_constant_with_type): ... here.
16286 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
16287 (vn_phi_eq): Compare types from vn_phi_s structure.
16288 (vn_phi_lookup): Populate vn_phi_s type.
16289 (vn_phi_insert): Likewise.
16291 2013-02-19 Jakub Jelinek <jakub@redhat.com>
16293 PR tree-optimization/56350
16294 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
16295 if haven't found reduction or nested cycle operand, rather than
16296 asserting we must find it.
16298 PR tree-optimization/56381
16299 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
16302 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
16303 Jakub Jelinek <jakub@redhat.com>
16306 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
16307 (swap_optab_enable): Same.
16308 (init_all_optabs): Use argument instead of global.
16309 * tree.h (struct tree_optimization_option): New field target_optabs.
16310 * expr.h (init_all_optabs): Add argument to prototype.
16311 (TREE_OPTIMIZATION_OPTABS): New.
16312 (save_optabs_if_changed): Protoize.
16313 * optabs.h: Declare this_fn_optabs.
16314 * optabs.c (save_optabs_if_changed): New.
16315 Declare this_fn_optabs.
16316 (init_optabs): Add argument to init_all_optabs() call.
16317 * function.c (invoke_set_current_function_hook): Handle per
16319 * function.h (struct function): New field optabs.
16320 * config/mips/mips.c (mips_set_mips16_mode): Handle when
16321 optimization_current_node has changed.
16322 * target-globals.h (save_target_globals_default_opts): Protoize.
16323 * target-globals.c (save_target_globals_default_opts): New.
16325 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16328 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
16329 registers %fr12 and %fr12R as call used.
16332 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
16333 and HImode, require all displacements to be an integer multiple of
16335 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
16336 only allow QImode and HImode when reload is in progress and strict is
16337 true. Likewise for symbolic addresses. Use base14_operand to check
16338 displacements in REG+BASE addresses.
16340 2013-02-18 Richard Biener <rguenther@suse.de>
16342 PR tree-optimization/56366
16343 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16344 handle sign-conversion of outer-loop initial induction value.
16346 2013-02-18 Richard Biener <rguenther@suse.de>
16348 PR middle-end/56349
16349 * cfghooks.c (merge_blocks): If we merge a latch into another
16350 block adjust references to it.
16351 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
16352 (verify_loop_structure): Verify that a recorded latch is in fact
16355 2013-02-18 Richard Biener <rguenther@suse.de>
16357 PR tree-optimization/56321
16358 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
16359 order SSA name release and virtual operand unlinking.
16361 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16363 * config/microblaze/microblaze.md (save_stack_block): Define.
16364 (restore_stack_block): Likewise.
16366 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16368 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
16369 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
16370 * config/microblaze/microblaze.c (microblaze_option_override):
16371 Bail out early for PIC modes when target does not support PIC.
16373 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16375 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
16376 Replace with a microblaze version.
16377 (microblaze_trampoline_init): Adapt for microblaze.
16378 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
16381 2013-02-16 Jakub Jelinek <jakub@redhat.com>
16382 Dodji Seketeli <dodji@redhat.com>
16385 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
16386 (instrument_mem_region_access): Do not forget to always put
16387 instrumentation of the of 'base' and 'base + len' in a "if (len !=
16388 0) statement, even for cases where either 'base' or 'base + len'
16389 are not instrumented -- because they have been previously
16390 instrumented. Simplify the logic by putting all the statements
16391 instrument 'base + len' inside a sequence, and then insert that
16392 sequence right before the current insertion point. Then, to
16393 instrument 'base + len', just get an iterator on that statement.
16394 And do not forget to update the pointer to iterator the function
16395 received as argument.
16397 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
16399 PR rtl-optimization/56348
16400 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
16402 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
16404 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
16405 (clean_graph_dump_file): Pass base to start_graph_dump.
16407 2013-02-14 Richard Henderson <rth@redhat.com>
16410 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
16412 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
16414 * collect2-aix.h: Define F_LOADONLY.
16416 2013-02-14 Richard Biener <rguenther@suse.de>
16419 * varasm.c (output_constant_def_1): Get the decl representing
16420 the constant as argument.
16421 (output_constant_def): Wrap output_constant_def_1.
16422 (make_decl_rtl): Use output_constant_def_1 with the decl
16423 representing the constant.
16424 (build_constant_desc): Optionally re-use a decl already
16425 representing the constant.
16426 (tree_output_constant_def): Adjust.
16428 2013-02-14 Dodji Seketeli <dodji@redhat.com>
16431 * asan.c (instrument_builtin_call): Really put the length of the
16432 second source argument into src1_len.
16434 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16436 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
16437 argument. If it is false, don't create edge from then_bb to
16439 (insert_if_then_before_iter): Pass true to it.
16440 (build_check_stmt): Pass false to it.
16441 (transform_statements): Flush hash table only on extended basic
16442 block boundaries, rather than at the beginning of every bb.
16443 Don't flush hash table on nonfreeing_call_p calls.
16444 * tree-flow.h (nonfreeing_call_p): New prototype.
16445 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
16447 2013-02-13 David S. Miller <davem@davemloft.net>
16449 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
16451 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
16454 * ira.c (max_regno_before_ira): Move from ...
16456 (fix_reg_equiv_init): Use max_regno_before_ira instead of
16459 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16461 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
16463 2013-02-13 Richard Biener <rguenther@suse.de>
16466 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
16467 globals in MEM_REFs.
16469 2013-02-13 Richard Biener <rguenther@suse.de>
16471 * loop-init.c (loop_optimizer_init): Clear loop state when
16472 re-initializing preserved loops.
16473 * loop-unswitch.c (unswitch_single_loop): Return whether
16474 we unswitched the loop. Do not verify loop state here.
16475 (unswitch_loops): When we unswitched a loop discover new loops.
16477 2013-02-13 Kostya Serebryany <kcc@google.com>
16479 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
16481 * sanitizer.def: Rename __asan_init to __asan_init_v1.
16483 2013-02-12 Dodji Seketeli <dodji@redhat.com>
16485 Avoid instrumenting duplicated memory access in the same basic block
16486 * Makefile.in (asan.o): Add new dependency on hash-table.h
16487 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
16488 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
16489 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
16490 (free_mem_ref_resources, has_mem_ref_been_instrumented)
16491 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
16492 (get_mem_ref_of_assignment): New functions.
16493 (get_mem_refs_of_builtin_call): Extract from
16494 instrument_builtin_call and tweak a little bit to make it fit with
16496 (instrument_builtin_call): Use the new
16497 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
16498 of is_gimple_builtin_call.
16499 (instrument_derefs, instrument_mem_region_access): Insert the
16500 instrumented memory reference into the hash table.
16501 (maybe_instrument_assignment): Renamed instrument_assignment into
16502 this, and change it to advance the iterator when instrumentation
16503 actually happened and return true in that case. This makes it
16504 homogeneous with maybe_instrument_assignment, and thus give a
16505 chance to callers to be more 'regular'.
16506 (transform_statements): Clear the memory reference hash table
16507 whenever we enter a new BB, when we cross a function call, or when
16508 we are done transforming statements. Use
16509 maybe_instrument_assignment instead of instrumentation. No more
16510 need to special case maybe_instrument_assignment and advance the
16511 iterator after calling it; it's now handled just like
16512 maybe_instrument_call. Update comment.
16514 2013-02-13 Richard Biener <rguenther@suse.de>
16516 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
16517 Fix loop discovery code.
16519 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
16521 PR inline-asm/56148
16522 * lra-constraints.c (process_alt_operands): Match early clobber
16523 operand with itself. Check conflicts with earlyclobber only if
16524 the operand is not reloaded. Prefer to reload conflicting operand
16525 if earlyclobber and matching operands are the same.
16527 2013-02-12 Richard Biener <rguenther@suse.de>
16530 * lto-streamer-out.c (write_symbol): Do not output symbols
16531 for hard register variables.
16533 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16536 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16537 (umulsidi3_insn, mulsidi3_insn): New insns.
16539 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16541 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16542 (struct tune_params): Add vec_costs field.
16543 * config/arm/arm.c (arm_builtin_vectorization_cost)
16544 (arm_add_stmt_cost): New functions.
16545 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16546 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16547 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16548 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16549 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16550 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16551 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16553 2013-02-12 Richard Biener <rguenther@suse.de>
16556 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16557 decls again if possible.
16559 2013-02-12 Richard Biener <rguenther@suse.de>
16561 PR middle-end/56288
16562 * tree-ssa.c (verify_ssa_name): Fix check, move
16563 SSA_NAME_IN_FREE_LIST check up.
16565 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16566 Steven Bosscher <steven@gcc.gnu.org>
16568 PR rtl-optimization/56151
16569 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16570 equal to op0 or op1, and last_insn pattern is CODE operation
16571 with MEM dest and one of the operands matches that MEM.
16573 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16575 * doc/extend.texi: Document Function Multiversioning and "default"
16576 parameter string to target attribute.
16577 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16578 target attribute parameter is "default".
16579 (ix86_compare_version_priority): Remove checks for target attribute.
16580 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16581 Remove check for target attribute equal to NULL. Add assert.
16582 (ix86_generate_version_dispatcher_body): Change error to sorry.
16584 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16585 Jack Howarth <howarth@bromo.med.uc.edu>
16586 Patrick Marlier <patrick.marlier@gmail.com>
16589 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16590 define ENDFILE_SPEC as TM_DESTRUCTOR.
16591 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16593 2013-02-11 Alexander Potapenko <glider@google.com>
16594 Jack Howarth <howarth@bromo.med.uc.edu>
16595 Jakub Jelinek <jakub@redhat.com>
16598 * config/darwin.c (cdtor_record): Rename ctor_record.
16599 (sort_cdtor_records): Rename sort_ctor_records.
16600 (finalize_dtors): New routine to sort destructors by
16601 priority before use in assemble_integer.
16602 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16604 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16606 PR rtl-optimization/56275
16607 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16608 offset is non-negative and less than cmode size before
16609 calling simplify_subreg.
16611 2013-02-11 Richard Biener <rguenther@suse.de>
16613 PR tree-optimization/56264
16614 * cfgloop.h (fix_loop_structure): Adjust prototype.
16615 * loop-init.c (fix_loop_structure): Return the number of
16616 newly discovered loops.
16617 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16618 are discovered, do a full loop-closed SSA rewrite.
16620 2013-02-11 Richard Biener <rguenther@suse.de>
16622 PR tree-optimization/56273
16623 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16625 (check_array_ref): Fix missing newline in dumps.
16626 (search_for_addr_array): Likewise.
16628 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16630 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16632 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16635 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16637 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16639 PR rtl-optimization/56246
16640 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16642 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16643 constraints are satisfied.
16645 2013-02-08 Jeff Law <law@redhat.com>
16648 * emit-rtl.c (reg_is_parm_p): New function.
16649 * regs.h (reg_is_parm_p): New prototype.
16650 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16651 callee-clobbered registers.
16653 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16656 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16657 If there is no implicit builtin declaration, just return NULL.
16659 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16661 * config/i386/sse.md (FMAMODEM): New mode iterator.
16662 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16663 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16665 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16667 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16668 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16669 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16671 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16673 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16674 (microblaze*-*-elf): Likewise.
16675 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16677 * config/microblaze/microblaze-c.c: Add builtin defines for
16678 _LITTLE_ENDIAN and _BIG_ENDIAN.
16679 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16680 add to TARGET_DEFAULT flags.
16681 Expand ASM_SPEC and LINK_SPEC.
16682 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16683 * config/microblaze/microblaze.md: Update extendsidi2 and
16684 movdi_internal instructions to use low-order / high-order reg
16686 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16687 options and inversemask / mask of LITTLE_ENDIAN.
16688 * config/microblaze/t-microblaze: Expand multilib options to
16689 include mlittle-endian (le) and update exceptions patterns.
16691 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16693 PR rtl-optimization/56195
16694 * lra-constraints.c (get_reload_reg): Don't reuse regs
16695 if they have smaller mode than requested, if they have
16696 wider mode than requested, try to return a SUBREG.
16698 PR tree-optimization/56250
16699 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16700 if type is unsigned and code isn't MULT_EXPR.
16702 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16704 PR tree-optimization/56064
16705 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16706 bits according to mode.
16707 * fixed-value.h (fixed_from_double_int)
16708 (const_fixed_from_double_int): Adjust comments.
16710 2013-02-08 Richard Biener <rguenther@suse.de>
16713 * lto-streamer.h (struct data_in): Remove current_file, current_line
16714 and current_col members.
16715 * lto-streamer-out.c (lto_output_location): Stream changed bits
16716 en-block for efficiency.
16717 * lto-streamer-in.c (clear_line_info): Remove.
16718 (lto_input_location): Cache current file, line and column
16719 globally via local statics. Read changed bits en-block.
16720 (input_function): Do not call clear_line_info.
16721 (lto_read_body): Likewise.
16722 (lto_input_toplevel_asms): Likewise.
16724 2013-02-08 Michael Matz <matz@suse.de>
16726 PR tree-optimization/52448
16727 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16728 (nt_call_phase): New static.
16729 (add_or_mark_expr): Only mark accesses with newer phase than any
16731 (nonfreeing_call_p): New.
16732 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16733 (nt_fini_block): Keep blocks marked as visited.
16734 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16736 2013-02-08 Richard Biener <rguenther@suse.de>
16738 * ira.c (ira): Free broken dominator information.
16740 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16742 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16744 2013-02-08 Marek Polacek <polacek@redhat.com>
16746 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16748 2013-02-08 Richard Biener <rguenther@suse.de>
16750 PR middle-end/56181
16751 * cfgloop.h (flow_loops_find): Adjust.
16752 (bb_loop_header_p): Declare.
16753 * cfgloop.c (bb_loop_header_p): New function split out from ...
16754 (flow_loops_find): ... here. Adjust function signature,
16755 support incremental loop structure update.
16756 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16757 * cfgloopmanip.c (fix_loop_structure): Move ...
16758 * loop-init.c (fix_loop_structure): ... here.
16759 (apply_loop_flags): Split out from ...
16760 (loop_optimizer_init): ... here.
16761 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16762 in incremental mode, only remove dead loops here.
16764 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16767 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16768 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16769 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16770 (*round<mode>3.libgcc): New insns for fixed-modes.
16771 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16772 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16773 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16774 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16775 implementations. Define to __builtin_avr_absFX,
16776 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16777 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16778 __builtin_avr_countlsFX, respectively.
16779 * config/avr/avr-c.c (target.h): Include it.
16780 (enum avr_builtin_id): New enum.
16781 (avr_resolve_overloaded_builtin): New static function.
16782 (avr_register_target_pragmas): Use it to set
16783 targetm.resolve_overloaded_builtin.
16784 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16785 tree nodes used by DEF_BUILTIN.
16786 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16787 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16788 <AVR_BUILTIN_xxBITS>: Same.
16790 2013-02-08 Richard Biener <rguenther@suse.de>
16792 * cfgloop.c (verify_loop_structure): Properly handle
16793 a loop exiting to another loop header.
16794 * ira-int.h (ira_loops): Remove.
16795 * ira.c (ira_loops): Remove.
16796 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16797 (do_reload): Use loop_optimizer_finalize.
16798 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16799 number_of_loops to access the loop tree.
16800 (more_one_region_p): Likewise.
16801 (finish_loop_tree_nodes): Likewise.
16802 (rebuild_regno_allocno_maps): Likewise.
16803 (mark_loops_for_removal): Likewise.
16804 (mark_all_loops_for_removal): Likewise.
16805 (remove_unnecessary_regions): Likewise.
16806 (ira_build): Likewise.
16807 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16809 2013-02-08 Richard Biener <rguenther@suse.de>
16811 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16812 * ipa-pure-const.c (analyze_function): Avoid calling
16813 mark_irreducible_loops twice.
16814 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16816 2013-02-07 David S. Miller <davem@davemloft.net>
16818 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16820 * var-tracking.c (vt_add_function_parameter): Test the presence of
16821 HAVE_window_save properly and do not remap argument registers when
16822 we have a leaf function.
16824 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16827 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16829 * config/i386/i386.c (ix86_print_operand): Ditto.
16831 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16833 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16835 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16837 PR rtl-optimization/56225
16838 * lra-constraints.c (process_alt_operands): Check that reload hard
16839 reg can hold value for strict_low_part.
16841 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16844 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16845 dwarf2out_end_function.
16846 (in_first_function_p, maybe_at_text_label_p,
16847 first_loclabel_num_not_at_text_label): New variables.
16848 (dwarf2out_var_location): In the first function find out
16849 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16850 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16853 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16855 PR rtl-optimization/56178
16856 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16857 SUBREG of a register. Tidy up related block of code.
16858 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16859 note if the source is a register or a SUBREG of a register.
16861 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16864 * config/rs6000/rs6000.md (ptrm): New mode attr.
16865 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16866 call_value_indirect_aix<pttrsize>,
16867 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16870 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16872 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16873 if -bnortl. Convert to strcmp and strncmp.
16875 2013-02-07 Alan Modra <amodra@gmail.com>
16878 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16879 addresses won't wrap when offsetting.
16880 (rs6000_secondary_reload): Provide secondary reloads needed for
16881 wrapping LO_SUM addresses.
16883 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16885 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16886 MACH, just __MACH__.
16888 2013-02-06 Richard Biener <rguenther@suse.de>
16890 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16891 instead of calling fix_loop_structure.
16893 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16895 PR middle-end/56217
16896 * omp-low.c (use_pointer_for_field): Return false if
16897 lower_send_shared_vars doesn't generate any copy-out code.
16899 2013-02-06 Tom de Vries <tom@codesourcery.com>
16901 PR rtl-optimization/56131
16902 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16903 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16904 of the label is NULL. Add comment.
16906 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16908 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16911 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16912 (STATIC_LIBTSAN_LIBS): Likewise.
16913 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16914 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16915 is defined, don't add anything else beyond that.
16916 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16917 (LINK_COMMAND_SPEC): Use them.
16919 PR tree-optimization/56205
16920 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16921 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16922 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16924 2013-02-05 Richard Biener <rguenther@suse.de>
16926 PR tree-optimization/53342
16927 PR tree-optimization/53185
16928 * tree-vectorizer.h (vect_check_strided_load): Remove.
16929 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16930 not disallow peeling for vectorized strided loads.
16931 (vect_check_strided_load): Make static and simplify.
16932 (vect_analyze_data_refs): Adjust.
16933 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16934 correctly when vectorizing strided loads.
16936 2013-02-05 Richard Biener <rguenther@suse.de>
16938 * doc/install.texi: Refer to ISL, not PPL.
16940 2013-02-05 Jan Hubicka <jh@suse.cz>
16942 PR tree-optimization/55789
16943 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16945 2013-02-05 Jan Hubicka <jh@suse.cz>
16947 PR tree-optimization/55789
16948 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16949 the dead call anyway.
16951 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16954 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16956 2013-02-04 Alexander Potapenko <glider@google.com>
16957 Jack Howarth <howarth@bromo.med.uc.edu>
16958 Jakub Jelinek <jakub@redhat.com>
16961 * config/darwin.c (sort_ctor_records): Stabilized qsort
16962 on constructor priority by using original position.
16963 (finalize_ctors): New routine to sort constructors by
16964 priority before use in assemble_integer.
16965 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16967 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16970 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16971 about visibility on artificial decls.
16972 * config/sol2.c (solaris_assemble_visibility): Likewise.
16974 2013-02-04 Kai Tietz <ktietz@redhat.com>
16977 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16978 argument and improve checking of return-argument types for 16-byte
16980 (ix86_function_value_1): Add additional valtype argument on call
16981 of function_value_64.
16982 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16983 handling infunction_value_64 function.
16985 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16987 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16989 2013-02-04 Richard Biener <rguenther@suse.de>
16991 PR tree-optimization/56188
16992 * tree-ssa-structalias.c (label_visit): Consider case with
16993 initially non-empty points-to set.
16994 (perform_var_substitution): Dump node mapping and clean up.
16996 2013-02-04 Richard Guenther <rguenther@suse.de>
16999 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
17000 node prevail as last resort.
17001 (lto_symtab_merge_decls): Remove guard on LTRANS here.
17002 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
17004 2013-02-04 Richard Biener <rguenther@suse.de>
17006 PR tree-optimization/56113
17007 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
17009 (equiv_class_lookup_or_add): ... this.
17010 (label_visit): Adjust and fix error in previous patch.
17011 (perform_var_substitution): Adjust.
17013 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
17015 * config/sh/divtab.c: Fix formatting and comments throughout the file.
17016 * config/sh/sh4-300.md: Likewise.
17017 * config/sh/sh4a.md: Likewise.
17018 * config/sh/constraints.md: Likewise.
17019 * config/sh/sh.md: Likewise.
17020 * config/sh/netbsd-elf.h: Likewise.
17021 * config/sh/predicates.md: Likewise.
17022 * config/sh/sh-protos.h: Likewise.
17023 * config/sh/ushmedia.h: Likewise.
17024 * config/sh/linux.h: Likewise.
17025 * config/sh/sh.c: Likewise.
17026 * config/sh/superh.h: Likewise.
17027 * config/sh/elf.h: Likewise.
17028 * config/sh/sh4.md: Likewise.
17029 * config/sh/sh.h: Likewise.
17031 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17033 * config/pa/constraints.md: Adjust unused letters. Change "T"
17034 constraint to match_test floating_point_store_memory_operand().
17035 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17036 (base14_operand): New.
17037 (floating_point_store_memory_operand): New.
17038 (integer_store_memory_operand): Revise to use base14_operand and
17039 reg_plus_base_memory_operand.
17040 (move_dest_operand): Allow symbolic_memory_operands.
17041 (symbolic_memory_operand): Check for LO_SOM.
17042 (symbolic_operand): Change default case to break.
17043 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17044 CONST_DOUBLE values to be reloaded by putting them into memory when
17045 the destination is a floating point register.
17046 (movdf): Remove code to handle CONST_DOUBLE.
17048 (reload_indf_r1): New.
17049 (reload_insf_r1): New.
17050 Consistently use "Q" and "T" constraints with integer and floating
17051 point move instructions, respectively.
17052 (movdi): Remove FAIL.
17053 Change predicate for source operand unamed DImode move from
17054 general_operand to move_src_operand.
17055 (umulsidi3): Change predicate for destination operand to
17057 Likewise for similar unamed patterns.
17058 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17059 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17060 (hppa_legitimize_address): Simplify mask calculation.
17061 (pa_emit_move_sequence): Revised handling of secondary reloads from
17062 REG+D addresses for floating point loads and stores. Directly handle
17063 loading CONST0_RTX (mode) to a floating point register.
17064 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17065 to floating point registers. Don't restrict secondary reloads to
17066 floating point registers to integer modes. Revise some comments and
17068 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17069 (pa_legitimate_address_p): New.
17070 (pa_legitimize_reload_address): New.
17071 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17072 (STRICT_REG_OK_FOR_BASE_P): New.
17073 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17074 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17076 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17077 Andrew Dixie <andrewd@gentrack.com>
17079 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17082 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17084 * expmed.c (extract_bit_field_1): Pass the full width of the
17085 structure to get_best_reg_extraction_insn.
17087 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17090 * configure.ac (use_cxa_atexit): Add AIX.
17091 * configure: Regenerate.
17093 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17095 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17098 * final.c (need_profile_function): New variable.
17099 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17100 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17101 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17102 notes, targetm.asm_out.function_prologue doesn't emit anything,
17103 HAVE_prologue and profiler should be emitted before prologue,
17104 set need_profile_function instead of emitting it.
17105 (final_scan_insn): If need_profile_function, emit
17106 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17107 NOTE_INSN_FUNCTION_BEG note.
17109 2013-02-01 Richard Henderson <rth@redhat.com>
17111 * config/rs6000/rs6000.md (smulditi3): New.
17114 * config/alpha/alpha.md (umulditi3): New.
17116 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17118 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17119 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17121 2013-02-01 Richard Biener <rguenther@suse.de>
17123 PR tree-optimization/56113
17124 * tree-ssa-structalias.c (label_visit): Reduce work for
17125 single-predecessor nodes.
17127 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17129 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17130 range isn't testing for zero.
17132 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17134 PR middle-end/56113
17135 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17137 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17138 Nick Clifton <nickc@redhat.com>
17140 * config/v850/constraints.md (Q): Define as a memory constraint.
17141 * config/v850/predicates.md (label_ref_operand): New predicate.
17142 (e3v5_shift_operand): New predicate.
17143 (ior_operator): New predicate.
17144 * config/v850/t-v850: Add e3v5 multilib.
17145 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17146 (v850_gen_movdi): Prototype.
17147 * config/v850/v850.c: Add support for e3v5 architecture.
17148 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17150 (construct_save_jarl): Add e3v5 long JARL support.
17151 (v850_adjust_insn_length): New function. Adjust length of call
17152 insns when using e3v5 instructions.
17153 (v850_gen_movdi): New function: Generate instructions to move a
17155 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17156 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17157 (TARGET_USE_FPU): Enable for e3v5.
17158 (CONST_OK_FOR_W): New macro.
17159 (ADJUST_INSN_LENGTH): Define.
17160 * config/v850/v850.md (UNSPEC_LOOP): Define.
17161 (attr cpu): Add v850e3v5.
17162 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17163 (movdi): New pattern.
17164 (movdi_internal): New pattern.
17165 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17166 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17167 (cstoresf4): Likewise.
17168 (cstoredf4): Likewise.
17169 (insv): New pattern.
17170 (rotlso3_a): New pattern.
17171 (rotlsi3_b): New pattern
17172 (rotlsi3_v850e3v5): New pattern.
17173 (doloop_begin): New pattern.
17174 (fix_loop_counter): New pattern.
17175 (doloop_end): New pattern.
17176 (branch_normal): Add e3v5 long branch support.
17177 (branch_invert): Likewise.
17178 (branch_z_normal): Likewise.
17179 (branch_z_invert): Likewise.
17180 (branch_nz_normal): Likewise.
17181 (branch_nz_invert): Likewise.
17182 (call_internal_short): Add e3v5 register-indirect JARL support.
17183 (call_internal_long): Likewise.
17184 (call_value_internal_short): Likewise.
17185 (call_value_internal_long): Likewise.
17186 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17187 (mloop): New option.
17188 * config.gcc: Add support for configuring v840e3v5 target.
17189 * doc/invoke.texi: Document new v850 specific command line options.
17191 2013-01-31 Paul Koning <ni1d@arrl.net>
17195 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17196 children if parent is a class.
17197 (prune_unused_types_prune): Don't add DW_AT_declaration.
17199 2013-01-31 Richard Biener <rguenther@suse.de>
17201 PR tree-optimization/56157
17202 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17203 match up operand with SLP child.
17205 2013-01-31 Jason Merrill <jason@redhat.com>
17208 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17209 parameters the first time.
17210 (gen_scheduled_generic_parms_dies): Check completeness here.
17212 2013-01-31 Richard Biener <rguenther@suse.de>
17214 PR middle-end/53073
17215 * common.opt (faggressive-loop-optimizations): New flag,
17216 enabled by default.
17217 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17218 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17219 infer_loop_bounds_from_undefined by it.
17221 2013-01-31 Richard Biener <rguenther@suse.de>
17223 PR tree-optimization/56150
17224 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17225 visit virtual operands.
17226 (find_uses_to_rename_bb): Likewise.
17228 2013-01-31 Richard Biener <rguenther@suse.de>
17230 PR tree-optimization/56150
17231 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17232 mixed store non-store stmts.
17234 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17237 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17238 LIBASAN_EARLY_SPEC is defined.
17239 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17240 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17242 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17245 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17246 invalid args instead of ICEing on it.
17247 (ix86_valid_target_attribute_tree): Return error_mark_node if
17248 ix86_valid_target_attribute_inner_p failed.
17249 (ix86_valid_target_attribute_p): Return false only if
17250 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17251 target("default") attribute.
17252 (sorted_attr_string): Change argument from const char * to tree,
17253 merge in all target attribute arguments rather than just one.
17254 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17255 instead of free. Avoid using strcat.
17256 (ix86_mangle_function_version_assembler_name): Mangle
17257 target("default") as if no target attribute is present. Adjust
17258 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
17259 instead of xmalloc and XDELETEVEC instead of free.
17260 (ix86_function_versions): Don't return true if one of the decls
17261 doesn't have target attribute. If they don't and one of the decls
17262 is DECL_FUNCTION_VERSIONED, report an error. Adjust
17263 sorted_attr_string caller. Use XDELETEVEC instead of free.
17264 (ix86_supports_function_versions): Remove.
17265 (make_name): Fix up formatting.
17266 (make_dispatcher_decl): Remove resolver_name and its initialization.
17267 Avoid leaking memory.
17268 (is_function_default_version): Return true if there is
17269 target("default") attribute rather than no target attribute at all.
17270 (make_resolver_func): Avoid leaking memory.
17271 (ix86_generate_version_dispatcher_body): Likewise.
17272 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
17273 * target.def (supports_function_versions): Remove.
17274 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
17275 * doc/tm.texi: Regenerated.
17277 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
17279 PR rtl-optimization/56144
17280 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
17281 for values with side effects.
17283 2013-01-30 Richard Biener <rguenther@suse.de>
17285 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
17286 (sparseset_pop): Likewise.
17287 * cfganal.c (compute_idf): Likewise. Increase work-stack size
17288 to be able to use quick_push in the worker loop.
17290 2013-01-30 Marek Polacek <polacek@redhat.com>
17292 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
17294 2013-01-30 Richard Biener <rguenther@suse.de>
17297 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
17299 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
17301 PR tree-optimization/56064
17302 * fixed-value.c (fixed_from_double_int): New function.
17303 * fixed-value.h (fixed_from_double_int): New prototype.
17304 (const_fixed_from_double_int): New static inline function.
17305 * fold-const.c (native_interpret_fixed): New static function.
17306 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
17307 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
17308 (native_encode_fixed): New static function.
17309 (native_encode_expr) <FIXED_CST>: Use it.
17310 (native_interpret_int): Move double_int worker code to...
17311 * double-int.c (double_int::from_buffer): ...this new static method.
17312 * double-int.h (double_int::from_buffer): Prototype it.
17314 2013-01-30 Richard Biener <rguenther@suse.de>
17316 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
17317 New pointer-map and obstack.
17318 (init_alias_vars): Allocate pointer-map and obstack.
17319 (delete_points_to_sets): Free them.
17320 (find_what_var_points_to): Cache result.
17321 (find_what_p_points_to): Adjust for changed interface of
17322 find_what_var_points_to.
17323 (compute_points_to_sets): Likewise.
17324 (ipa_pta_execute): Likewise.
17326 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17328 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
17329 * configure: Regenerate.
17330 * config.in: Regenerate.
17331 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
17332 #nobits/#progbits if supported.
17334 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
17337 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
17338 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
17339 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
17341 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17343 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
17344 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
17346 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17348 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
17349 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
17351 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17353 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
17355 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
17356 * config/arm/cortex-a7.md: New bypasses using
17357 arm_mac_accumulator_is_result.
17359 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17361 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
17362 (cortex_a7_neon_mla): Likewise.
17363 (cortex_a7_fpfmad): New reservation.
17364 (cortex_a7_fpmacs): Use ffmas and update required units.
17365 (cortex_a7_fpmuld): Update required units and latency.
17366 (cortex_a7_fpmacd): Likewise.
17367 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
17368 (cortex_a7_neon). Likewise.
17369 (bypass) Update participating units.
17371 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17373 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
17374 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
17376 * config/arm/vfp11.md (vfp_farith): Use ffmas.
17377 (vfp_fmul): Use ffmad.
17378 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
17379 (cortex_r4_fmacd): Use ffmad.
17380 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
17381 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
17382 (cortex_a9_fmacd): Use ffmad.
17383 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
17384 (cortex_a8_vfp_macd): Use ffmad.
17385 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
17386 (cortex_a5_fpmacd): Use ffmad.
17387 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
17388 (cortex_a15_vfp_macd): Use ffmad.
17389 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
17391 2013-01-29 Jason Merrill <jason@redhat.com>
17394 * varasm.c (default_assemble_visibility): Don't warn about
17395 visibility on artificial decls.
17397 2013-01-29 Richard Biener <rguenther@suse.de>
17399 PR tree-optimization/56113
17400 * tree-ssa-structalias.c (equiv_class_lookup): Also return
17402 (label_visit): Free duplicate bitmaps and record the leader instead.
17403 (perform_var_substitution): Adjust.
17405 2013-01-29 Richard Biener <rguenther@suse.de>
17407 PR tree-optimization/55270
17408 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
17409 the CFG, schedule loops for fixup.
17411 2013-01-29 Nick Clifton <nickc@redhat.com>
17413 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
17416 2013-01-28 Leif Ekblad <leif@rdos.net>
17418 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
17419 * config/i386/i386.h (TARGET_RDOS): New macro.
17420 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
17421 * config/i386/i386.c (ix86_option_override_internal): For 64bit
17422 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
17423 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
17424 DEFAULT_LARGE_SECTION_THRESHOLD.
17425 * config/i386/i386.md (R14_REG, R15_REG): New constants.
17426 * config/i386/rdos.h: New file.
17427 * config/i386/rdos64.h: New file.
17429 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
17432 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
17435 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17437 PR rtl-optimization/56117
17438 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
17439 call cselib_lookup_from_insn on the MEM before calling
17440 add_insn_mem_dependence.
17442 2013-01-28 Richard Biener <rguenther@suse.de>
17444 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
17445 to a stmt that didn't have one.
17446 (copy_phis_for_bb): Likewise for PHI arguments.
17447 (copy_debug_stmt): Likewise for debug stmts.
17449 2013-01-28 Richard Biener <rguenther@suse.de>
17451 PR tree-optimization/56034
17452 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
17453 (partition_builtin_p): Adjust.
17454 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
17455 it is the last partition.
17456 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
17457 up the vertex for the definition.
17458 (classify_partition): Classify whether a partition is a
17459 PKIND_REDUCTION, thus has uses outside of the loop.
17460 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
17461 Merge all PKIND_REDUCTION partitions into the last partition.
17462 (tree_loop_distribution): Seed partitions from reductions as well.
17464 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17466 PR tree-optimization/56125
17467 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
17468 pow(x,c) into sqrt(x) * powi(x, n/2) or
17469 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
17470 optimizing for size.
17471 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
17472 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
17475 PR tree-optimization/56094
17476 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
17477 to UNKNOWN_LOCATION while gimplifying expr.
17479 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
17482 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
17483 operand 0 in movabs insn template for -masm=intel asm alternative.
17484 (*movabs<mode>_2): Ditto for operand 1.
17486 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
17489 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
17492 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17494 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
17497 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17499 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
17500 Undef to avoid warning.
17502 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17504 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
17505 * configure: Regenerate.
17507 2013-01-25 Jakub Jelinek <jakub@redhat.com>
17509 PR tree-optimization/56098
17510 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
17511 for stmts with volatile ops.
17512 (cond_store_replacement): Don't optimize if assign has volatile ops.
17513 (cond_if_else_store_replacement_1): Don't optimize if either
17514 then_assign or else_assign have volatile ops.
17515 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
17518 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17520 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
17522 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17524 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17525 missing ':' in asm example.
17527 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17529 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17530 entries into lane and laneq entries.
17531 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17532 Remove AdvSIMD scalar modes.
17533 (aarch64_sq<r>dmulh_laneq<mode>): New.
17534 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17536 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17537 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17538 * config/aarch64/iterators.md (VCOND): New.
17541 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17544 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17545 Add NULL LIBNAME argument to existing definitions.
17546 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17547 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17548 * config/avr/avr.c (DEF_BUILTIN): Same.
17549 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17550 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17551 implementation is available (DECL_ASSEMBLER_NAME is set).
17552 (avr_fold_absfx): New static function.
17553 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17554 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17555 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17556 AVR_BUILTIN_ABSLLK.
17557 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17558 (abshk, absk, abslk, absllk): Provide as static inline functions.
17560 2013-01-25 Marek Polacek <polacek@redhat.com>
17562 PR tree-optimization/56035
17563 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17565 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17567 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17568 (*movtf_internal_rex64): Add (!o,C) alternative
17569 (*movxf_internal_rex64): Ditto.
17570 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17572 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17574 * doc/invoke.texi: fix typo.
17575 * doc/objc.texi: fix typo.
17577 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17579 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17580 for the first two alternatives.
17582 2013-01-24 Diego Novillo <dnovillo@google.com>
17584 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17585 (ggc-zone.o): Remove.
17586 * configure.ac: Remove option --with-gc.
17587 * configure: Re-generate.
17588 * doc/install.texi: Remove documentation for --with-gc.
17589 * gengtype.c (write_enum_defn): Remove. Update all users.
17590 (write_Types_process_field): Remove generation of gt_e_* argument.
17591 (output_type_enum): Remove. Update all users.
17592 (write_enum_defn): Remove. Update all users.
17593 (enum alloc_zone): Remove. Update all users.
17594 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17595 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17596 Update all callers.
17597 (struct ptr_data): Remove field TYPE. Update all users.
17598 (gt_pch_note_object): Remove argument TYPE. Update all users.
17599 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17601 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17602 (struct alloc_zone): Remove.
17603 (ggc_internal_alloc_zone_stat): Remove.
17604 (ggc_internal_cleared_alloc_zone_stat): Remove.
17605 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17606 (ggc_pch_count_object): Remove last argument. Update all users.
17607 (ggc_pch_alloc_object): Remove last argument. Update all users.
17608 (struct alloc_zone): Remove.
17609 * ggc-zone.c: Remove.
17610 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17611 (struct alloc_zone): Remove.
17612 (ggc_alloc_typed_stat): Remove.
17613 (ggc_alloc_typed): Remove.
17614 (ggc_splay_alloc): Remove first argument.
17615 (rtl_zone): Remove. Update all users.
17616 (tree_zone): Remove. Update all users.
17617 (tree_id_zone): Remove. Update all users.
17618 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17619 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17620 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17621 * tree-ssanames.c: Remove references to zone allocator in comments.
17623 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17625 * config/avr/avr.c (avr_out_fract): Make register numbers that
17626 might be outside of source operand signed.
17628 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17630 * config/i386/constraints.md (Yf): New constraint.
17631 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17632 of f constraint to conditionaly disable x87 register preferences.
17633 (*movdf_internal): Ditto.
17634 (*movsf_internal): Ditto.
17636 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17638 PR inline-asm/55934
17639 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17640 that have operands with impossible constraints.
17641 Add a FIXME for a speed-up opportunity.
17642 * lra-constraints.c (process_alt_operands): Verify that a class
17643 selected from constraints on asms is valid for the operand mode.
17644 (curr_insn_transform): Remove incorrect comment.
17646 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17648 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17649 TOC operand is a valid symbol ref in the constant pool.
17651 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17653 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17655 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17658 * config/avr/stdfix.h: New file.
17659 * t-avr (stdfix-gcc.h): New rule to build it.
17660 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17662 2013-01-23 Kostya Serebryany <kcc@google.com>
17664 * config/darwin.h: remove dependency on
17665 CoreFoundation (asan on Mac OS).
17667 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17670 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17671 instead of cmpdi_operand for first comparison operand.
17672 Don't assert that comparison operands aren't both constants.
17674 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17676 * doc/install.texi (Downloading the Source): Update references to
17677 downloading separate components.
17679 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17681 * doc/extend.texi (__int128): Improve grammar.
17683 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17686 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17687 alternative to (o,r).
17688 (*movdi_internal_rex64): Remove (!o,n) alternative.
17689 (DImode immediate->memory splitter): Remove.
17690 (DImode immediate->memory peephole2): Remove.
17691 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17692 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17693 alternative to (!o,*r).
17694 (*movtf_internal_sse): New pattern.
17695 (*movxf_internal_rex64): New pattern.
17696 (*movxf_internal): Disable for TARGET_64BIT.
17697 (*movdf_internal_rex64): Remove (!o,F) alternative.
17699 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17701 PR middle-end/56074
17702 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17703 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17704 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17705 stmt locations where LOCATION_LOCUS of the stmt location is
17706 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17709 * config/i386/i386.md (UNSPEC_STOS): New.
17710 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17711 *strsetqi_1): Add UNSPEC_STOS.
17713 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17716 * doc/invoke.texi: Remove left over -Wsynth example.
17718 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17720 PR tree-optimization/56051
17721 * fold-const.c (fold_binary_loc): Don't fold
17722 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17723 a narrowing conversion, or widening conversion from signed
17726 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17728 PR rtl-optimization/56023
17729 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17730 dependent on debug instruction.
17732 2013-01-21 Martin Jambor <mjambor@suse.cz>
17734 PR middle-end/56022
17735 * function.c (allocate_struct_function): Call
17736 invoke_set_current_function_hook earlier.
17738 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17740 * reload1.c (init_reload): Only initialize reload_obstack
17741 during the first call.
17743 2013-01-21 Marek Polacek <polacek@redhat.com>
17745 * cfgloop.c (verify_loop_structure): Fix up grammar.
17747 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17749 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17750 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17752 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17755 * config/arm/marvell-pj4.md: Update copyright year.
17756 Fix up use of alu to alu_reg and simple_alu_imm.
17758 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17760 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17762 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17765 * lra-constraints.c (curr_insn_transform): Don't reuse original
17766 insn for secondary memory move when memory mode should be different.
17768 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17770 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17771 atomic_storedi_1): New patterns.
17773 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17775 btver2 pipeline descriptions.
17776 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17778 * config/i386/i386.md (btver2_decode): New type attributes.
17779 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17781 * config/i386/btver2.md: New file describing btver2 pipelines.
17783 2013-01-19 Andrew Pinski <apinski@cavium.com>
17785 PR tree-optimization/52631
17786 * tree-ssa-sccvn (visit_use): Before looking up the original
17787 statement, try looking up the simplified expression.
17789 2013-01-19 Anthony Green <green@moxielogic.com>
17791 * config/moxie/moxie.c (moxie_expand_prologue): Set
17792 current_function_static_stack_size.
17794 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17796 PR tree-optimization/56029
17797 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17798 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17800 2013-01-18 Sharad Singhai <singhai@google.com>
17802 PR tree-optimization/55995
17803 * dumpfile.c (dump_loc): Print location only if available.
17804 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17806 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17809 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17810 secondary memory move.
17811 (inherit_reload_reg): Use rclass instead of cl for
17812 check_secondary_memory_needed_p.
17814 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17816 PR middle-end/56015
17817 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17818 the case where writing real complex part of target modifies op1.
17820 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17822 * config/aarch64/aarch64-simd.md
17823 (aarch64_vcond_internal<mode>): Handle unordered cases.
17824 * config/aarch64/iterators.md (v_cmp_result): New.
17826 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17827 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17829 * config/arm/marvell-pj4.md: New file.
17830 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17831 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17832 (generic_vfp): Likewise.
17833 * config/arm/arm-cores.def: Add marvell-pj4.
17834 * config/arm/arm-tune.md: Regenerate.
17835 * config/arm/arm-tables.opt: Regenerate.
17836 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17837 * doc/invoke.texi: Document marvell-pj4.
17839 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17841 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17843 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17848 * var-tracking.c (negative_power_of_two_p): New.
17849 (global_get_addr_cache, local_get_addr_cache): New.
17850 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17851 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17853 (vt_stack_offset_p): Remove.
17854 (vt_canon_true_dep): Always canonicalize loc's address.
17855 (clobber_overlapping_mems): Make sure we have a MEM.
17856 (local_get_addr_clear_given_value): New.
17857 (val_reset): Clear local cached entries.
17858 (compute_bb_dataflow): Create and release the local cache.
17859 Disable duplicate MEMs clobbering.
17860 (emit_notes_in_bb): Clobber MEMs likewise.
17861 (vt_emit_notes): Create and release the local cache.
17862 (vt_initialize, vt_finalize): Create and release the global
17863 cache, respectively.
17864 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17866 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17868 PR libmudflap/53359
17869 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17870 not found in the symtab.
17872 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17875 PR rtl-optimization/55547
17876 PR rtl-optimization/53827
17879 * alias.c (offset_overlap_p): New, factored out of...
17880 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17881 the conservative special case for symbolic constants. Don't
17882 adjust zero sizes on alignment.
17884 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17886 PR rtl-optimization/52573
17887 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17888 REG_UNUSED for the same register.
17890 2013-01-17 Richard Biener <rguenther@suse.de>
17891 Marek Polacek <polacek@redhat.com>
17893 PR rtl-optimization/55833
17894 * loop-unswitch.c (unswitch_loops): Move loop verification...
17895 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17896 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17897 Set it to true when we're removing a loop from hierarchy tree in
17898 an irreducible region.
17899 (fix_bb_placements): Adjust caller.
17900 (fix_loop_placements): Likewise.
17902 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17904 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17905 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17906 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17907 Remove ID. Adjust comments.
17908 * config/avr/avr-c.c (avr_builtin_name): Remove.
17909 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17910 * config/avr/avr.c (avr_tolower): New static function.
17911 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17912 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17913 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17916 2013-01-17 Jan Hubicka <jh@suse.cz>
17918 PR tree-optimization/55273
17919 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17921 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17924 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17925 store through atomic_store<mode>_1.
17926 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17928 2013-01-17 Martin Jambor <mjambor@suse.cz>
17930 PR tree-optimizations/55264
17931 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17932 for virtual methods.
17933 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17934 virtual methods before inlining is over.
17935 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17937 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17940 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17942 PR rtl-optimization/56005
17943 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17944 pending reads for prefetch.
17946 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17948 * config/aarch64/aarch64.md
17949 (*cstoresi_neg_uxtw): New pattern.
17950 (*cmovsi_insn_uxtw): New pattern.
17951 (*<optab>si3_uxtw): New pattern.
17952 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17953 (*<optab>si3_insn_uxtw): New pattern.
17954 (*bswapsi2_uxtw): New pattern.
17956 2013-01-16 Richard Biener <rguenther@suse.de>
17958 * tree-inline.c (tree_function_versioning): Remove set but
17959 never used variable.
17961 2013-01-16 Richard Biener <rguenther@suse.de>
17963 PR tree-optimization/55964
17964 * tree-flow.h (rename_variables_in_loop): Remove.
17965 (rename_variables_in_bb): Likewise.
17966 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17967 (copy_loop_before): Adjust and delete update-ssa status.
17968 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17969 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17970 (rename_variables_in_loop): Remove.
17971 (slpeel_update_phis_for_duplicate_loop): Likewise.
17972 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17973 use available cfg machinery instead of duplicating it.
17974 Update PHI nodes and perform poor-mans SSA update here.
17975 (slpeel_tree_peel_loop_to_edge): Adjust.
17977 2013-01-16 Richard Biener <rguenther@suse.de>
17979 PR tree-optimization/54767
17980 PR tree-optimization/53465
17981 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17982 (vrp_visit_phi_node): For PHI arguments coming via backedges
17983 drop all symbolical range information.
17984 (execute_vrp): Compute backedges.
17986 2013-01-16 Richard Biener <rguenther@suse.de>
17988 * doc/install.texi: Update CLooG and ISL requirements to
17991 2013-01-16 Christian Bruel <christian.bruel@st.com>
17994 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17995 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17996 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17998 2013-01-16 DJ Delorie <dj@redhat.com>
18000 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
18001 (UNSPECV_SP_SWITCH_E): New.
18002 (sp_switch_1): Change to an unspec.
18003 (sp_switch_2): Change to an unspec. Don't use post-inc when we
18006 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
18008 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
18009 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
18010 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
18011 (expand_mem_thread_fence): Ditto.
18012 (expand_mem_signal_fence): Ditto.
18013 (expand_atomic_load): Ditto.
18014 (expand_atomic_store): Ditto.
18016 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
18018 PR rtl-optimization/55547
18019 PR rtl-optimization/53827
18022 * alias.c (memrefs_conflict_p): Set sizes to negative after
18025 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18028 * function.c (thread_prologue_and_epilogue_insns): Always
18029 add crtl->drap_reg to set_up_by_prologue.set, even if
18030 stack_realign_drap is false.
18032 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18034 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18035 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18036 *call): Fix indention.
18038 2013-01-15 Tom de Vries <tom@codesourcery.com>
18041 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18044 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18046 PR rtl-optimization/55153
18047 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18049 2013-01-15 Martin Jambor <mjambor@suse.cz>
18051 PR tree-optimization/55920
18052 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18053 accesses as grp_to_be_debug_replaced.
18055 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18057 PR tree-optimization/55920
18058 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18059 there is non-useless type conversion needed from debug rhs to lhs,
18060 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18062 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18063 Mikael Pettersson <mikpe@it.uu.se>
18066 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18068 (ASM_OUTPUT_CASE_LABEL): Remove.
18069 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18070 * final.c (shorten_branches): Update alignment of labels before
18071 jump tables if CASE_VECTOR_SHORTEN_MODE.
18073 2013-01-15 Richard Biener <rguenther@suse.de>
18076 * system.h: Do not include gmp.h for building host tools.
18078 2013-01-15 Richard Biener <rguenther@suse.de>
18080 PR middle-end/55882
18081 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18082 account for bitpos when computing alignment.
18084 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18086 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18087 (ix86_target_macros_internal): Likewise.
18089 * config/i386/i386.c (m_CORE2I7): Removed.
18090 (m_CORE_HASWELL): New macro.
18091 (m_CORE_ALL): Likewise.
18092 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18093 (initial_ix86_arch_features): Likewise.
18094 (processor_target_table): Initializations for Core avx2.
18095 (cpu_names): New names "core-avx2".
18096 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18097 PROCESSOR_CORE_HASWELL.
18098 (ix86_issue_rate): New case.
18099 (ia32_multipass_dfa_lookahead): Likewise.
18100 (ix86_sched_init_global): Likewise.
18102 * config/i386/i386.h (TARGET_HASWELL): New macro.
18103 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18104 (processor_type): New PROCESSOR_HASWELL.
18106 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18108 PR tree-optimization/55955
18109 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18110 *SHIFT_EXPR and *ROTATE_EXPR codes.
18112 PR tree-optimization/48766
18113 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18114 -ftrapv disable -fwrapv.
18116 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18119 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18120 etc. to 1 and not to __flash.
18121 Use LL suffix for __INT24_MAX__ with -mint8.
18122 Use ULL suffix for __UINT24_MAX__ with -mint8.
18124 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18126 * config/avr/avr-arch.h
18127 (struct base_arch_s): Use typedef avr_arch_t instead.
18128 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18129 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18130 * config/avr/avr.c: Same.
18131 * config/avr/avr-devices.c: Same.
18132 * config/avr/driver-avr.c: Same.
18133 * config/avr/gen-avr-mmcu-texi.c: Same.
18134 * config/avr/avr-mcus.def: Adjust comment.
18136 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18138 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18139 * config/aarch64/iterators.md (VALLDI): New.
18141 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18142 Andi Kleen <ak@linux.intel.com>
18145 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18146 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18149 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18151 * config/avr/avr-stdint.h: Remove trailing blanks.
18152 * config/avr/avr-log.h: Same.
18153 * config/avr/avr-arch.h: Same.
18154 * config/avr/avr-devices.c: Same.
18155 * config/avr/avr-dimode.md: Same.
18156 * config/avr/predicates.md: Same.
18157 * config/avr/avr-c.c: Same. And fix typo.
18159 * config/avr/avr-protos.h: Same. And:
18160 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18161 (init_cumulative_args): Rename to avr_init_cumulative_args.
18162 (expand_prologue): Rename to avr_expand_prologue.
18163 (expand_epilogue): Rename to avr_expand_epilogue.
18164 (adjust_insn_length): Rename to avr_adjust_insn_length.
18165 (notice_update_cc): Rename to avr_notice_update_cc.
18166 (final_prescan_insn): Rename to avr_final_prescan_insn.
18167 * config/avr/avr.c: Same.
18168 * config/avr/avr.h: Same.
18169 * config/avr/avr.md: Remove trailing blanks.
18170 (prologue): Use avr_expand_prologue.
18171 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18173 2013-01-14 Richard Biener <rguenther@suse.de>
18175 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18176 verify_location, collect_subblocks): New functions.
18177 (verify_gimple_in_cfg): Verify that locations only reference
18178 BLOCKs in the functions BLOCK tree.
18180 2013-01-14 Richard Biener <rguenther@suse.de>
18182 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18184 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18186 (insert_out_of_ssa_copy_on_edge): Likewise.
18187 (rewrite_close_phi_out_of_ssa): Likewise.
18188 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18190 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18191 propagated constants.
18192 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18195 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18197 * config/avr/avr-modes.def: Add GPL copyright notice.
18199 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18201 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18202 MEMMODEL_MASK to determine memory model.
18203 (atomic_store<mode>): Ditto from operands[2].
18204 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18206 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18209 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18210 (fold_gimple_assign): Don't call unshare_expr here.
18211 (fold_ctor_reference): Call unshare_expr.
18213 2013-01-13 Terry Guo <terry.guo@arm.com>
18215 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18216 * doc/fragments.texi: Document MULTILIB_REUSE.
18217 * gcc.c (multilib_reuse): New internal spec.
18218 (set_multilib_dir): Also search multilib from multilib_reuse.
18219 * genmultilib (tmpmultilib3): Refactor code.
18220 (tmpmultilib4): Ditto.
18221 (multilib_reuse): New multilib argument.
18223 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18225 * Makefile.in: Update copyright.
18227 2013-01-12 Tom de Vries <tom@codesourcery.com>
18229 PR middle-end/55890
18230 * calls.c (expand_call): Check if arg_nr is valid.
18232 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18234 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18235 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18236 documentation. Add missing '__' in front of
18237 __builtin_ia32_packssdw256.
18239 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18242 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18243 NO_REGS for larl operands.
18244 (s390_reload_larl_operand): Use s390_load_address instead of
18247 2013-01-11 Richard Biener <rguenther@suse.de>
18249 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18250 (verify_node_sharing): ... here.
18251 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18253 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18255 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18256 Substitute TREECHECKING.
18257 * configure: Regenerate.
18258 * Makefile.in (TREECHECKING): New.
18260 2013-01-11 Richard Guenther <rguenther@suse.de>
18262 PR tree-optimization/44061
18263 * tree-vrp.c (extract_range_basic): Compute zero as
18264 value-range for __builtin_constant_p of function parameters.
18266 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
18268 Update copyright years.
18270 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
18272 PR rtl-optimization/55672
18273 * lra-eliminations.c (mark_not_eliminable): Permit addition with
18274 const to be eliminable.
18276 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
18278 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
18279 * configure: Regenerate.
18281 2013-01-10 Richard Biener <rguenther@suse.de>
18283 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
18285 2013-01-10 Richard Biener <rguenther@suse.de>
18288 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
18289 locations for virtual PHI arguments.
18290 (rewrite_update_phi_arguments): Likewise.
18292 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
18294 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
18297 2013-01-10 Jakub Jelinek <jakub@redhat.com>
18299 PR tree-optimization/55921
18300 * tree-complex.c (expand_complex_asm): New function.
18301 (expand_complex_operations_1): Call it for GIMPLE_ASM.
18303 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18306 * config/s390/s390.c (s390_symref_operand_p)
18307 (s390_loadrelative_operand_p): Merge the two functions.
18308 (s390_check_qrst_address, print_operand_address): Add parameters
18309 to s390_loadrelative_operand_p invokation.
18310 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
18311 (s390_reload_larl_operand, s390_secondary_reload): Use
18312 s390_loadrelative_operand_p instead of s390_symref_operand_p.
18313 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
18315 2013-01-09 Mike Stump <mikestump@comcast.net>
18317 * dse.c (record_store): Remove unnecessary assert.
18319 2013-01-09 Jan Hubicka <jh@suse.cz>
18321 PR tree-optimization/55569
18322 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
18323 * cfgloop.h (scale_loop_profile): Likewise.
18325 2013-01-09 Jan Hubicka <jh@suse.cz>
18328 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
18330 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
18332 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
18334 PR middle-end/55114
18335 * expr.h (maybe_emit_group_store): Declare.
18336 * expr.c (maybe_emit_group_store): New function.
18337 * builtins.c (expand_builtin_int_roundingfn): Call it.
18338 (expand_builtin_int_roundingfn_2): Likewise.
18340 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
18342 PR rtl-optimization/55829
18343 * lra-constraints.c (match_reload): Add code for absent output.
18344 (curr_insn_transform): Add code for reloads of matched inputs
18347 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
18349 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
18350 attribute of movddup insn to DF.
18351 (*vec_interleave_lowv2df): Ditto.
18352 (vec_dupv2df): Ditto.
18354 2013-01-09 Jan Hubicka <jh@suse.cz>
18356 PR tree-optimiation/55875
18357 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
18358 EVERY_ITERATION parameter.
18359 (number_of_iterations_exit): Check if exit is executed every iteration.
18360 (idx_infer_loop_bounds): Similarly here.
18361 (n_of_executions_at_most): Simplify
18362 to only test for cases where statement is dominated by the
18363 particular bound; handle correctly the "postdominance" test.
18364 (scev_probably_wraps_p): Use max loop iterations info
18365 as a global bound first.
18367 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
18368 Nick Clifton <nickc@redhat.com>
18370 * config/v850/v850.md (cbranchsf4): New pattern.
18371 (cstoresf4): New pattern.
18372 (cbranchdf4): New pattern.
18373 (cstoredf4): New pattern.
18374 (movsicc): Disallow floating point comparisons.
18375 (cmpsf_le_insn): Fix order of operators.
18376 (cmpsf_lt_insn): Likewise.
18377 (cmpsf_eq_insn): Likewise.
18378 (cmpdf_le_insn): Likewise.
18379 (cmpdf_lt_insn): Likewise.
18380 (cmpdf_eq_insn): Likewise.
18381 (cmpsf_ge_insn): Use LE comparison.
18382 (cmpdf_ge_insn): Likewise.
18383 (cmpsf_gt_insn): Use LT comparison.
18384 (cmpdf_gt_insn): Likewise.
18385 (cmpsf_ne_insn): Delete pattern.
18386 (cmpdf_ne_insn): Delete pattern.
18387 * config/v850/v850.c (v850_gen_float_compare): Use
18388 gen_cmpdf_eq_insn for NE comparison.
18389 (v850_float_z_comparison_operator)
18390 (v850_float_nz_comparison_operator): Move from here ...
18391 * config/v850/predicates.md: ... to here. Move GT and GE
18392 comparisons into v850_float_z_comparison_operator.
18393 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
18395 (v850_float_nz_comparison_operator): Likewise.
18397 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18399 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
18400 with calls to gen_insvsi/gen_insvdi.
18402 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18404 * config/i386/i386.c (initial_ix86_tune_features): Set up
18405 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
18407 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
18408 Jakub Jelinek <jakub@redhat.com>
18410 PR tree-optimization/48189
18411 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
18412 If nitercst is 0, don't predict the exit edge.
18414 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18416 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
18417 in asm_fprintf with reg_names.
18418 (aarch64_print_operand_address): Likewise.
18419 (aarch64_return_addr): Likewise.
18420 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
18422 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18424 * config/pa/pa.h (VAL_U6_BITS_P): Define.
18425 (INT_U6_BITS): Likewise.
18426 * config/pa/predicates.md (uint6_operand): New predicate.
18427 (shift5_operand, shift6_operand): Likewise.
18428 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
18430 (lshrdi3): Use shift6_operand.
18431 (shrpsi4, shrpdi4): New insn patterns.
18432 (extzv): Delete expander.
18433 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
18434 predicates in unamed zero extract patterns. Tighten common constraint.
18435 (extv): Delete expander.
18436 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
18437 predicates in unamed sign extract patterns. Tighten common constraint.
18438 (insv): Delete expander.
18439 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
18440 predicates in unamed insert patterns. Tighten common constraint.
18441 Change uint32_operand predicate to uint6_operand predicate in unamed
18442 DImode pattern to insert constant values of type 1...1xxxx.
18444 2013-01-04 Jan Hubicka <jh@suse.cz>
18446 PR tree-optimization/55823
18447 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
18450 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18451 Uros Bizjak <ubizjak@gmail.com>
18453 PR rtl-optimization/55845
18454 * df-problems.c (can_move_insns_across): Stop scanning at
18455 volatile_insn_p source instruction or give up if
18456 across_from .. across_to range contains any volatile_insn_p
18459 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18461 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
18462 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
18464 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
18465 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
18467 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18470 * asan.c (asan_clear_shadow): New function.
18471 (asan_emit_stack_protection): Use it.
18473 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18475 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
18476 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
18477 with tab instead of space.
18479 2013-01-08 Nick Clifton <nickc@redhat.com>
18481 * config/rl78/rl78.c (rl78_expand_prologue): Always select
18482 register bank 0 at the start of an interrupt handler.
18483 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
18486 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18488 * config/aarch64/aarch64-simd.md
18489 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
18490 (aarch64_simd_bsl): Likewise.
18491 (aarch64_vcond_internal<mode>): Likewise.
18492 (vcond<mode><mode>): Likewise.
18493 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
18494 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
18496 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18498 * config/aarch64/aarch64-builtins.c
18499 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
18501 2013-01-08 Martin Jambor <mjambor@suse.cz>
18504 * tree-sra.c (analyze_access_subtree): Return true also after
18505 potentially creating a debug-only replacement.
18507 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18509 PR middle-end/55890
18510 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
18512 PR tree-optimization/54120
18513 * tree-vrp.c (range_fits_type_p): Don't allow
18514 src_precision < precision from signed vr to unsigned_p
18515 if vr->min or vr->max is negative.
18516 (simplify_float_conversion_using_ranges): Test can_float_p
18517 against CODE_FOR_nothing.
18519 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18520 Richard Biener <rguenther@suse.de>
18522 PR middle-end/55851
18523 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
18524 types instead of just INTEGER_TYPE types.
18526 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18528 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18531 2013-01-07 Steve Ellcey <sellcey@mips.com>
18534 * config/mips/mips.opt: Change mad to mmad to match documentation.
18536 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18539 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18540 .progmemx.data now.
18542 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18545 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18546 (avr_addrspace_t): Add .section_name field.
18547 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18549 (avr_addrspace): Same. Initialize .section_name. Remove last
18550 NULL entry. Put __memx into .progmemx.data.
18551 (progmem_section_prefix): Remove.
18552 (avr_asm_init_sections): No need to initialize progmem_section.
18553 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18554 section name prefix.
18555 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18556 retrieve the progmem section.
18557 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18558 boundary to run over avr_addrspace[].
18559 (avr_register_target_pragmas): Ditto.
18561 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18563 * varasm.c (output_constant_def_contents): For asan_protect_global
18564 protected strings, adjust DECL_ALIGN if needed, before testing for
18566 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18567 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18569 (output_object_block): For asan protected decls, emit asan padding
18570 after their contents.
18571 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18572 (asan_finish_file): Test it here instead.
18574 2013-01-07 Nick Clifton <nickc@redhat.com>
18575 Matthias Klose <doko@debian.org>
18576 Doug Kwan <dougkwan@google.com>
18577 H.J. Lu <hongjiu.lu@intel.com>
18580 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18582 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18584 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18586 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18588 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18590 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18593 * doc/install.texi (Cross-Compiler-Specific Options): Document
18596 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18598 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18599 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18600 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18601 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18602 vqmovun_high_s64): Fix source operand number and update copyright.
18604 2013-01-07 Richard Biener <rguenther@suse.de>
18606 PR middle-end/55890
18607 * gimple.h (gimple_call_builtin_p): New overload.
18608 * gimple.c (validate_call): New function.
18609 (gimple_call_builtin_p): Likewise.
18610 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18611 Use gimple_call_builtin_p.
18612 (find_func_clobbers): Likewise.
18613 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18614 (strlen_optimize_stmt): Likewise.
18616 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18618 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18619 (vld1q_dup_*): Likewise.
18620 (vld1_*): Likewise.
18621 (vld1q_*): Likewise.
18622 (vld1_lane_*): Likewise.
18623 (vld1q_lane_*): Likewise.
18625 2013-01-07 Richard Biener <rguenther@suse.de>
18627 * lto-streamer.h (LTO_minor_version): Bump to 2.
18629 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18631 * config/aarch64/aarch64-protos.h
18632 (aarch64_const_double_zero_rtx_p): Rename to...
18633 (aarch64_float_const_zero_rtx_p): ...this.
18634 (aarch64_float_const_representable_p): New.
18635 (aarch64_output_simd_mov_immediate): Likewise.
18636 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18637 move immediate case.
18638 * config/aarch64/aarch64.c
18639 (aarch64_const_double_zero_rtx_p): Rename to...
18640 (aarch64_float_const_zero_rtx_p): ...this.
18641 (aarch64_print_operand): Allow printing of new constants.
18642 (aarch64_valid_floating_const): New.
18643 (aarch64_legitimate_constant_p): Check for valid floating-point
18645 (aarch64_simd_valid_immediate): Likewise.
18646 (aarch64_vect_float_const_representable_p): New.
18647 (aarch64_float_const_representable_p): Likewise.
18648 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18649 (aarch64_output_simd_mov_immediate): New.
18650 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18651 (*movdf_aarch64): Likewise.
18652 * config/aarch64/constraints.md (Ufc): New.
18653 (Y): call aarch64_float_const_zero_rtx.
18654 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18656 2013-01-07 Richard Biener <rguenther@suse.de>
18658 PR tree-optimization/55888
18659 PR tree-optimization/55862
18660 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18661 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18662 not if it is contained therein.
18664 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18666 * config/avr/t-avr: Typo.
18668 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18671 * config/avr/t-avr: Don't automatically rebuild
18672 $(srcdir)/config/avr/t-multilib
18673 $(srcdir)/config/avr/avr-tables.opt
18674 $(srcdir)/doc/avr-mmcu.texi
18675 (avr-mcus): New phony target to build them on request.
18676 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18677 * avr/avr-mcus.def: Adjust comments.
18679 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18681 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18683 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18685 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18687 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18689 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18691 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18693 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18694 to generate profiling.
18695 * config/rs6000/aix64.h (LIB_SPEC): Same.
18697 2013-01-04 Andrew Pinski <apinski@cavium.com>
18699 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18701 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18703 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18705 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18707 (ix86_expand_move): Ditto.
18708 (ix86_zero_extend_to_Pmode): Ditto.
18709 (ix86_expand_call): Ditto.
18710 (ix86_expand_special_args_builtin): Ditto.
18711 (ix86_expand_builtin): Ditto.
18713 2013-01-04 Richard Biener <rguenther@suse.de>
18715 PR tree-optimization/55862
18716 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18717 translating them through PHI nodes.
18719 2013-01-04 Martin Jambor <mjambor@suse.cz>
18721 PR tree-optimization/55755
18722 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18723 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18725 2013-01-04 Marek Polacek <polacek@redhat.com>
18727 PR middle-end/55859
18728 * opts.c (default_options_optimization): Clarify error message.
18730 2013-01-04 Richard Biener <rguenther@suse.de>
18732 PR middle-end/55863
18733 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18736 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18739 * config/pa/pa.md (movsi): Revert previous change.
18740 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18743 2013-01-03 Richard Henderson <rth@redhat.com>
18745 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18746 after eliminating TLS symbols.
18748 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18751 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18752 * graphite-poly.c (debug_gmp_value): Likewise.
18754 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18757 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18758 selected code model, define __code_mode_small__, __code_model_medium__,
18759 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18760 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18761 xchg temporary register with %k. Declare temporary register as
18763 [__x86_64__]: For medium and large code models, preserve %rbx register.
18765 2013-01-03 Richard Biener <rguenther@suse.de>
18767 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18768 (dump_subscript): Adjust.
18769 (finalize_ddr_dependent): Do not dump redundant info.
18770 (analyze_siv_subscript): Adjust.
18771 (subscript_dependence_tester): Likewise.
18772 (compute_affine_dependence): Likewise.
18774 2013-01-03 Richard Biener <rguenther@suse.de>
18777 2013-01-03 Richard Biener <rguenther@suse.de>
18779 PR tree-optimization/55857
18780 * tree-vect-stmts.c (vectorizable_load): Do not setup
18781 re-alignment for invariant loads.
18783 2013-01-02 Richard Biener <rguenther@suse.de>
18785 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18786 invariant load do not generate a vector load from the scalar location.
18788 2013-01-03 Richard Biener <rguenther@suse.de>
18790 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18791 for not vectorizing.
18792 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18793 not build INDIRECT_REFs, call get_name once only.
18794 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18795 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18797 2013-01-03 Richard Biener <rguenther@suse.de>
18799 PR tree-optimization/55857
18800 * tree-vect-stmts.c (vectorizable_load): Do not setup
18801 re-alignment for invariant loads.
18803 2013-01-03 Richard Biener <rguenther@suse.de>
18806 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18807 prefer a built-in decl.
18809 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18811 * gcc.c (process_command): Update copyright notice dates.
18812 * gcov.c (print_version): Likewise.
18813 * gcov-dump.c (print_version): Likewise.
18815 PR rtl-optimization/55838
18816 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18817 iv0.step, iv1.step and step.
18819 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18820 Marc Glisse <marc.glisse@inria.fr>
18822 PR tree-optimization/55832
18823 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18824 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18825 integer_{one,zero}_node.
18827 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18830 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18831 * var-tracking.c (reverse_op): Don't add reverse ops to
18832 VALUEs that have already
18833 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18835 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18837 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18839 2013-01-02 Teresa Johnson <tejohnson@google.com>
18841 * dumpfile.c (dump_loc): Print filename with location.
18842 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18843 new location_t parameter to emit complete unroll message with
18844 new dump framework.
18845 (canonicalize_loop_induction_variables): Compute loops location
18846 and pass to try_unroll_loop_completely.
18847 * loop-unroll.c (report_unroll_peel): New function.
18848 (peel_loops_completely): Use new dump format with location
18849 for main dumpfile message, and invoke report_unroll_peel on success.
18850 (decide_unrolling_and_peeling): Ditto.
18851 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18852 by report_unroll_peel.
18853 (decide_peel_completely): Ditto.
18854 (decide_unroll_constant_iterations): Ditto.
18855 (decide_unroll_runtime_iterations): Ditto.
18856 (decide_peel_simple): Ditto.
18857 (decide_unroll_stupid): Ditto.
18858 * cfgloop.c (get_loop_location): New function.
18859 * cfgloop.h (get_loop_location): Declare.
18861 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18863 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18866 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18868 PR middle-end/55198
18869 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18870 BLKmode objects when EXPAND_MEMORY is specified.
18872 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18874 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18876 (fold_builtin_cpu): Do not share cpu model decls across statements.
18878 2013-01-02 Jason Merrill <jason@redhat.com>
18881 * tree.c (build_array_type_1): Revert earlier change.
18883 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18885 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18887 * config/aarch64/aarch64-tune.md: Re-generate.
18889 2013-01-02 Richard Biener <rguenther@suse.de>
18891 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18892 invariant load do not generate a vector load from the scalar location.
18894 2013-01-02 Richard Biener <rguenther@suse.de>
18897 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18898 * configure: Regenerate.
18900 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18902 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18903 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18904 (expand_builtin_int_roundingfn_2): Keep the original target around
18905 for the fallback case.
18907 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18909 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18910 to be clear for sign changes.
18912 2013-01-01 Jan Hubicka <jh@suse.cz>
18914 * ipa-inline-analysis.c: Fix formatting.
18916 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18918 PR tree-optimization/55831
18919 * tree-vect-loop.c (get_initial_def_for_induction): Use
18920 gsi_after_labels instead of gsi_start_bb.
18922 Copyright (C) 2013 Free Software Foundation, Inc.
18924 Copying and distribution of this file, with or without modification,
18925 are permitted in any medium without royalty provided the copyright
18926 notice and this notice are preserved.