1 2013-08-20 Nick Clifton <nickc@redhat.com>
3 * target.def (narrow_volatile_bitfield): Note that the default
4 value is false, not !TARGET_STRICT_ALIGN.
5 * doc/tm.texi: Regenerate.
7 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9 Fix LIB_SPEC for systems without libpthread.
11 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
14 * config/i386/linux-common.h: Likewise.
15 * config/mips/linux-common.h: Likewise.
17 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
19 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
22 2013-08-20 David Malcolm <dmalcolm@redhat.com>
24 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
25 instances can own GC refs.
27 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
28 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
29 (gcc::context::gt_pch_nx): Likewise.
30 (gcc::context::gt_pch_nx): Likewise.
31 * ggc.h (gt_ggc_mx <T>): New.
32 (gt_pch_nx_with_op <T>): New.
34 * passes.c (opt_pass::gt_ggc_mx): New.
35 (opt_pass::gt_pch_nx): New.
36 (opt_pass::gt_pch_nx_with_op): New.
37 (pass_manager::gt_ggc_mx): New.
38 (pass_manager::gt_pch_nx): New.
39 (pass_manager::gt_pch_nx_with_op): New.
40 (pass_manager::operator new): Use
41 ggc_internal_cleared_alloc_stat rather than xcalloc.
42 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
43 (pass_manager::gt_ggc_mx): New.
44 (pass_manager::gt_pch_nx): New.
45 (pass_manager::gt_pch_nx_with_op): New.
46 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
47 (opt_pass::operator new): New.
48 (opt_pass::gt_ggc_mx): New.
49 (opt_pass::gt_pch_nx): New.
50 (opt_pass::gt_pch_nx_with_op): New.
52 2013-08-20 David Malcolm <dmalcolm@redhat.com>
54 * Makefile.in (GTFILES): Add context.h.
55 * context.c (gcc::context::operator new): New.
56 (gcc::context::gt_ggc_mx): New.
57 (gcc::context::gt_pch_nx): New.
58 (gcc::context::gt_pch_nx): New.
59 * context.h (gcc::context): Add GTY((user)) marking.
60 (gcc::context::operator new): New.
61 (gcc::context::gt_ggc_mx): New.
62 (gcc::context::gt_pch_nx): New.
63 (gcc::context::gt_pch_nx): New.
65 (gt_ggc_mx (gcc::context *)): New.
66 (gt_pch_nx (gcc::context *)): New.
67 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
69 * gengtype.c (open_base_files) <ifiles>: Add context.h.
71 2013-08-20 Alan Modra <amodra@gmail.com>
74 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
75 (rs6000_emit_epilogue): Likewise.
77 2013-08-19 Dehao Chen (dehao@google.com)
79 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
81 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
82 Jakub Jelinek <jakub@redhat.com>
84 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
85 (BUILT_IN_FABSD64): Likewise.
86 (BUILT_IN_FABSD128): Likewise.
87 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
88 (fold_builtin_1): Likewise.
89 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping destination
91 (*abstd2_fpr): Likewise.
92 (*nabstd2_fpr): Likewise.
94 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
96 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
99 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
101 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
104 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
106 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
107 * builtins.c (default_libc_has_function): New.
108 (gnu_libc_has_function): Ditto.
109 (no_c99_libc_has_function): Ditto.
110 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
111 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
112 (fold_builtin_sincos): Likewise.
113 (fold_builtin_cexp): Likewise.
114 * builtins.def (DEF_C94_BUILTIN): Likewise.
115 (DEF_C99_BUILTIN): Likewise.
116 (DEF_C99_C90RES_BUILTIN): Likewise.
117 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
118 definitions to using this define.
119 * config/darwin-protos.h (darwin_libc_has_function): New.
120 * config/darwin.c: (darwin_libc_has_function: Ditto.
121 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
122 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
123 * config/darwin.h: Ditto.
124 * config/elfos.h: Ditto.
125 * config/freebsd.h: Ditto.
126 * config/i386/cygming.h: Ditto.
127 * config/i386/djgpp.h: Ditto.
128 * config/i386/i386-interix.h: Ditto.
129 * config/microblaze/microblaze.h: Ditto.
130 * config/mmix/mmix.h: Ditto.
131 * config/gnu-user.h: Ditto.
132 * config/ia64/hpux.h: Ditto.
133 * config/pa/pa-hpux.h: Ditto.
134 * config/pdp11/pdp11.h: Ditto.
135 * config/picochip/picochip.h: Ditto.
136 * config/linux.h: Ditto.
137 * config/netbsd.h: Ditto.
138 * config/openbsd.h: Ditto.
139 * config/rs6000/aix43.h: Ditto.
140 * config/rs6000/aix51.h: Ditto.
141 * config/rs6000/aix52.h: Ditto.
142 * config/rs6000/aix53.h: Ditto.
143 * config/rs6000/aix61.h: Ditto.
144 * config/rs6000/darwin.h: Ditto.
145 * config/rs6000/linux.h: Ditto.
146 * config/rs6000/linux64.h: Ditto.
147 * config/s390/tpf.h: Ditto.
148 * config/sol2-10.h: Ditto.
149 * config/sol2.h: Ditto.
150 * config/vms/vms.h: Ditto.
151 * config/vxworks.h: Ditto.
152 * config/linux-android.c (linux_android_libc_has_function):
153 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
154 * config/linux-protos.h (linux_android_libc_has_function):
156 * config/i386/i386.c (ix86_libc_has_function): New.
157 * config/i386/i386-protos.h
158 (ix86_libc_has_function): New declaration.
159 * config/i386/i386.md
160 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
161 ("isinf<mode>2): Likewise.
162 * convert.c (convert_to_integer): Using new target hook
163 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
164 TARGET_C99_FUNCTIONS.
165 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
166 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
167 * coretypes.h (function_class): New enum for different
168 classes of functions.
169 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
170 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
171 (TARGET_HAS_SINCOS): Likewise.
172 (TARGET_LIBC_HAS_FUNCTION): New.
173 * doc/tm.texi: Regenerated.
174 * targhooks.h (default_libc_has_function): New declaration.
175 (no_c99_libc_has_function): Ditto.
176 (gnu_libc_has_function): Ditto.
177 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
178 and TARGET_HAS_SINCOS.
180 2013-08-18 Jan Hubicka <jh@suse.cz>
182 * Makeifle-in (ipa-devirt.o): New.
183 (GTFILES): Add ipa-utils.h and ipa-devirt.c
184 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
185 (analyze_functions): Look into possible targets of polymorphic call.
186 * dumpfile.c (dump_files): Add type-inheritance dump.
187 * dumpfile.h (TDI_inheritance): New.
188 * ipa-devirt.c: New file.
189 * ipa-utils.h (odr_type_d): Forward declare.
190 (odr_type): New type.
191 (build_type_inheritance_graph): Declare.
192 (possible_polymorphic_call_targets): Declare and introduce inline
193 variant when only edge is pased.
194 (dump_possible_polymorphic_call_targets): Likewise.
195 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
196 * tree.c (type_in_anonymous_namespace_p): Break out from ...
197 (types_same_for_odr): ... here.
198 * tree.h (type_in_anonymous_namespace_p): Declare.
200 2013-08-18 Jakub Jelinek <jakub@redhat.com>
202 PR tree-optimization/58006
203 * tree-parloops.c (take_address_of): Don't ICE if get_name
205 (eliminate_local_variables_stmt): Remove clobber stmts.
207 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
209 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
210 error message is issued for an alias to undefined symbol.
212 2013-08-18 Jan Hubicka <jh@suse.cz>
214 * cgraph.c (cgraph_create_indirect_edge): Discover
215 polymorphic calls and record basic info into indirect_info.
216 * gimple-fold.c (gimple_fold_call): When doing BINFO based
217 devirtualization, ignore objc function calls.
218 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
219 call with no parm index info.
220 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
221 * tree.c (virtual_method_call_p): New function.
222 * tree.h (virtual_method_call_p): Declare.
224 2013-08-16 Jan Hubicka <jh@suse.cz>
227 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
229 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
231 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
234 2013-08-16 David Malcolm <dmalcolm@redhat.com>
236 * gengtype.c (type_for_name): Add special-case support for
237 locating types within the "gcc::" namespace.
238 (open_base_files): Emit a "using namespace gcc" directive.
240 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
243 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
244 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
246 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
247 array instead of each individual operand as a separate argument.
248 (emit_fusion_gpr_load): Likewise.
249 (expand_fusion_gpr_load): Add new function declaration.
251 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
252 signature to have the operands passed as an array, instead of as
253 separate arguments. Allow ZERO_EXTEND to be in the memory
254 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
255 depend on the register live/dead flags when peepholes are run.
256 (expand_fusion_gpr_load): New function to be called from the
257 peephole2 pass, to change the register that addis sets to be the
259 (emit_fusion_gpr_load): Change the calling signature to have the
260 operands passed as an array, instead of as separate arguments.
261 Allow ZERO_EXTEND to be in the memory address, and also
262 SIGN_EXTEND if -mpower8-fusion-sign.
264 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
266 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
267 adjust the register addis loads up in the peephole2 pass. Do not
268 depend on the register live/dead state when the peephole pass is
271 2013-08-16 David Malcolm <dmalcolm@redhat.com>
273 * gengtype.c (create_user_defined_type): Ensure that the kind
274 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
275 declaration is seen before the GTY((user)) marking.
277 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
280 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
282 2013-08-16 Jan Hubicka <jh@suse.cz>
284 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
285 arugment expected_type.
286 (gimple_fold_call): Use it.
287 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
288 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
289 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
290 (try_make_edge_direct_virtual_call): Likewise.
291 * tree.c (obj_type_ref_class): New.
292 * tree.h (obj_type_ref_class): Use it.
294 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
296 * sched-vis.c (rtl_slim_pp_initialized): Remove.
297 (rtl_slim_pp): Likewise.
298 (init_rtl_slim_pretty_print): Likewise.
299 (dump_value_slim): Don't call it. Use local pretty printer.
300 (dump_insn_slim): Likewise.
301 (dump_rtl_slim): Likewise.
302 (str_pattern_slim): Likewise.
303 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
306 2013-08-16 Jakub Jelinek <jakub@redhat.com>
308 PR tree-optimization/58164
309 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
310 walk gimple_goto_dest of GIMPLE_GOTO.
312 PR tree-optimization/58165
313 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
314 bi_call must be the last stmt in a bb, don't split_block, instead
315 use fallthru edge from it and give up if there is none.
316 Release conds vector when returning early.
318 2013-08-14 Xinliang David Li <davidxl@google.com>
320 * config/i386/i386.c (ix86_option_override_internal):
321 Remove unused variable and field.
323 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
326 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
327 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
328 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
329 and Linux, correct BLKmode alignment when 128-bit alignment is
330 required and compatibility flag is not set.
331 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
332 for zero-size arguments when compatibility flag is not set.
334 2013-08-14 Jakub Jelinek <jakub@redhat.com>
336 PR tree-optimization/58145
337 * tree-sra.c (build_ref_for_offset): If prev_base has
338 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
340 2013-08-14 Xinliang David Li <davidxl@google.com>
342 * config/i386/i386.c (ix86_option_override_internal):
343 Fix uninitialized variable error.
345 2013-08-14 Xinliang David Li <davidxl@google.com>
347 * config/i386/i386.opt: Define two new options.
348 * config/i386/x86-tune.def: Add arch selector field in macros.
349 * config/i386/i386.h: Adjust macro definition.
350 * config/i386/i386.c (ix86_option_override_internal):
352 (parse_mtune_ctrl_str): New function.
353 (set_ix86_tune_features): New function.
354 (ix86_function_specific_restore): Call the new helper function.
356 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
358 PR rtl-optimization/57662
359 * sel-sched.c (code_motion_process_successors): When the current insn
360 is removed after the recursive traversal, break from the loop.
361 Add comments and debug printouts.
363 2013-08-14 Jakub Jelinek <jakub@redhat.com>
364 Alexandre Oliva <aoliva@redhat.com>
367 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
368 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
369 there also UNSPEC_PLTOFF.
371 2013-08-14 Marek Polacek <polacek@redhat.com>
373 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
374 NUM_CONDITIONS bit positions.
376 2013-08-13 Cary Coutant <ccoutant@google.com>
378 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
379 (attr_checksum): Hash vector contents instead of pointer.
380 (attr_checksum_ordered): Likewise.
382 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
384 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
385 when Pmode != word_mode. Add length_address attribute.
386 (sse3_monitor_<mode>): Merge from sse3_monitor and
387 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
388 Pmode != word_mode. Update insn length attribute.
389 * config/i386/i386.c (ix86_option_override_internal): Update
390 ix86_gen_monitor selection for merged sse3_monitor insn.
392 2013-08-13 Julian Brown <julian@codesourcery.com>
394 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
395 perform invalid legitimization on greater-than-word-size modes for
398 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
400 * ira.c (setup_class_translate_array): Use aclass instead of cl
401 for classes not fully covered by allocno classes.
403 2013-08-13 Jakub Jelinek <jakub@redhat.com>
405 PR tree-optimization/57661
406 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
407 * tree-inline.c (tree_function_versioning): Initialize it.
408 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
409 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
410 that is not being copied.
413 * asan.c (instrument_strlen_call): Fix typo in comment.
414 Use char * type even for the lhs of POINTER_PLUS_EXPR.
416 2013-08-13 Steve Ellcey <sellcey@mips.com>
418 * config/mips/mips.md (prefetch): Use lw instead of ld on
419 loongson in 32bit mode.
421 2013-08-13 Nick Clifton <nickc@redhat.com>
423 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
425 2013-08-13 Jan Hubicka <jh@suse.cz>
427 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
428 introduced edge; fix typo in sanity check.
429 (cgraph_resolve_speculation): Export; improve diagnostic.
430 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
431 speculation at type mismatch.
432 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
433 (cgraph_resolve_speculation): Declare.
434 (symtab_can_be_discarded): New function.
435 * value-prof.c (gimple_ic_transform): Remove actual transform code.
436 * ipa-inline-transform.c (speculation_removed): New global var.
437 (clone_inlined_nodes): See if speculation can be removed.
438 (inline_call): If speculations was removed, we growths may not match.
439 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
440 (speculation_useful_p): New function.
441 (resolve_noninline_speculation): New function.
442 (inline_small_functions): Resolve useless speculations.
443 * ipa-inline.h (speculation_useful_p): Declare
444 * ipa.c (can_replace_by_local_alias): Simplify.
445 (ipa_profile): Produce speculative calls in non-lto, too;
446 add simple cost model; produce local aliases.
448 2013-08-13 David Malcolm <dmalcolm@redhat.com>
450 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
453 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
455 * config/i386/i386.c (ix86_function_versions): Use error + inform.
457 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
459 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
460 iterator instead of X87MODEF.
462 2013-08-12 Perez Read <netfirewall@gmail.com>
465 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
466 operand 0 for intel asm alternative.
467 (*movabs<mode>_2): Ditto for operand 1.
469 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
471 * config/aarch64/arm_none.h
472 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
474 2013-08-12 Nick Clifton <nickc@redhat.com>
476 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
478 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
480 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
481 expand for QI/HImode operand to produce more effictive code for
482 unsigned char(short) --> float(double) conversion.
484 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
486 * doc/invoke.texi: Mention that -ftls-model does not force the final
489 2013-08-12 Marek Polacek <polacek@redhat.com>
490 Marc Glisse <marc.glisse@inria.fr>
492 PR tree-optimization/57980
493 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
494 when creating -1 constant.
496 2013-08-10 Jan Hubicka <jh@suse.cz>
498 Workaround binutils PR14342.
499 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
500 (gimple_init_edge_profiler): Likewise.
501 (gimple_gen_ic_func_profiler): Likewise.
503 2013-08-09 Jan Hubicka <jh@suse.cz>
505 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
507 2013-08-09 Xinliang David Li <davidxl@google.com>
509 * config/i386/stringop.def: New file.
510 * config/i386/stringop.opt: New file.
511 * config/i386/i386-opts.h: Include stringopt.def.
512 * config/i386/i386.opt: Include stringopt.opt.
513 * config/i386/i386.c (ix86_option_override_internal):
514 Override default size based stringop inline strategies with options.
515 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
518 2013-08-09 Jan Hubicka <jh@suse.cz>
520 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
522 2013-08-09 Jan Hubicka <jh@suse.cz>
524 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
526 (dump_cgraph_node): Dump profile-id.
527 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
528 and common_target_probability.
529 * lto-cgraph.c (lto_output_edge): Stream common targets.
530 (lto_output_node): Stream profile ids.
531 (input_node): Stream profile ids.
532 (input_edge): Stream common targets.
533 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
534 * ipa.c: Include value-prof.h
535 (ipa_profile_generate_summary): Turn indirect call statement histograms
537 (ipa_profile): Turn common targets into speculative edges.
539 2013-08-09 Jan Hubicka <jh@suse.cz>
541 * cgraph.h (cgraph_node): Add profile_id.
542 * value-prof.c (cgraph_node_map): Turn into pointer_map.
543 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
544 (del_node_map): Update.
545 (find_func_by_funcdef_no): Replace by ...
546 (find_func_by_profile_id): ... this one.
547 (gimple_ic_transform): Do not remove useful histograms when
548 speculation is not done; dump info when indirect call removal
550 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
551 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
552 (__gcov_indirect_call_profiler_v2): .. this one.
553 * profile.h (init_node_map): Update.
554 * coverage.c (coverage_compute_profile_id): New function.
555 * coverage.h (coverage_compute_profile_id): Declare.
556 * tree-profile.c (init_ic_make_global_vars): Make
557 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
558 (gimple_init_edge_profiler): Update prototype of
559 __gcov_indirect_call_profiler.
560 (gimple_gen_ic_func_profiler): Simplify.
561 (tree_profiling): Use init_node_map
563 2013-08-09 Jan Hubicka <jh@suse.cz>
565 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
566 non-speculative refs.
567 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
568 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
569 (cgraph_set_call_stmt): Likewise.
570 (cgraph_create_edge_1): Fix release checking compilatoin;
572 (cgraph_free_edge): Free indirect info.
573 (cgraph_turn_edge_to_speculative): New function.
574 (cgraph_speculative_call_info): New function.
575 (cgraph_make_edge_direct): Return direct edge; handle speculation.
576 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
577 (dump_cgraph_node): Dump speculation.
578 (verify_edge_count_and_frequency): Accept speculative edges.
579 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
580 (verify_cgraph_node): Handle speculation.
581 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
582 (cgraph_set_call_stmt): Update prototype.
583 (cgraph_make_edge_direct): Update prototype.
584 (cgraph_speculative_call_info): Declare.
585 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
586 to change; update call of ipa_find_references.
587 * ipa-ref.c (ipa_record_reference): Fix return value; clear
588 lto_stmt_uid and speculative flags.
589 (ipa_dump_references): Dump speculation.
590 (ipa_clone_references): Clone speculative flag.
591 (ipa_clone_referring): Likewise.
592 (ipa_clone_ref): New function.
593 (ipa_find_reference): Look into lto_stmt_uids
594 (ipa_clear_stmts_in_references): Do not clear speculative calls.
595 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
596 (ipa_find_reference): Update declaration.
597 (ipa_clone_ref): Declare.
598 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
599 stream speculative flag.
600 (lto_output_ref): Stream statements uids and speculation.
601 (input_ref): Likewise.
602 (input_edge): Stream speuclation.
603 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
604 (cgraph_set_call_stmt_including_clones): Handle speculation.
605 * ipa-inline.c (heap_edge_removal_hook): New function.
606 (inline_small_functions): Register it.
607 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
608 also initialize refs.
609 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
611 (try_make_edge_direct_simple_call): Likewise.
612 (try_make_edge_direct_simple_call): Likewise.
613 (update_indirect_edges_after_inlining): Likewise.
614 (remove_described_reference): Look proper lto_stmt_uid.
615 (propagate_controlled_uses): Likewise.
616 (propagate_controlled_uses): Liekwise.
617 * tree-inline.c (copy_bb): Copy speculative edges.
618 (redirect_all_calls): New function.
619 (copy_cfg_body): Do redirection after loop info is updated.
620 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
622 2013-08-09 Jan Hubicka <jh@suse.cz>
624 * lto-streamer-out.c (output_function): Renumber PHIs.
625 * lto-streamer-in.c (input_function): Likewise.
627 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
629 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
630 (get_lane_unsigned): Likewise.
631 (dup_lane_scalar): Likewise.
632 (get_lane): enable for VALL.
633 * config/aarch64/aarch64-simd.md
634 (aarch64_dup_lane_scalar<mode>): Remove.
635 (aarch64_get_lane_signed<mode>): Likewise.
636 (aarch64_get_lane_unsigned<mode>): Likewise.
637 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
638 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
639 (aarch64_get_lane<mode>): Enable for all vector modes.
640 (aarch64_get_lanedi): Remove misleading constraints.
641 * config/aarch64/arm_neon.h
642 (__aarch64_vget_lane_any): Define.
643 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
644 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
645 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
646 * config/aarch64/iterators.md (VDQQH): New.
650 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
652 * configure.ac: Add GAS check for LEON instructions on SPARC.
653 * configure: Regenerate.
654 * config.in: Likewise.
655 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
657 * config/sparc/sparc.opt (LEON, LEON3): New masks.
658 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
660 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
661 (AS_LEON_FLAG): New macro.
662 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
663 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
664 Deal with LEON and LEON3 for the memory model.
665 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
666 (atomic_compare_and_swap<mode>_1): Likewise.
667 (*atomic_compare_and_swap<mode>_1): Likewise.
669 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
671 * config/arm/neon.md (vcond): Fix floating-point vector
672 comparisons against 0.
674 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
676 * lra-constraints.c (emit_spill_move): Remove assert.
677 (process_alt_operands): Add more debugging
678 output. Increase reject for spilling into memory. Decrease
679 reject for reloading scratch.
680 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
682 2013-08-08 Steve Ellcey <sellcey@mips.com>
684 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
685 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
686 micromips incompatible. Add nan2008.
687 (MULTILIB_DIRNAMES): Add nan2008.
688 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
689 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
690 and micromips incompatible. Add nan2008.
691 (MULTILIB_DIRNAMES): Add nan2008.
692 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
694 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
696 PR rtl-optimization/58079
697 * combine.c (combine_simplify_rtx): Avoid using SUBST if
698 simplify_comparison has widened a comparison with an integer.
700 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
702 * config/arm/neon.md (movmisalign<mode>): Disable when we
703 don't allow unaligned accesses.
704 (*movmisalign<mode>_neon_store): Likewise.
705 (*movmisalign<mode>_neon_load): Likewise.
706 (*movmisalign<mode>_neon_store): Likewise.
707 (*movmisalign<mode>_neon_load): Likewise.
709 2013-08-08 Jan Hubicka <jh@suse.cz>
711 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
712 (make_pass_rebuild_cgraph_edges): Also clear references.
713 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
714 * ipa-inline-transform.c (inline_transform): Remove all references
716 * cgraphunit.c (expand_function): Remove all references after
718 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
719 (ipa_find_reference): Rewrite to iterator.
720 (remove_stmt_references): Likewise.
721 (ipa_clear_stmts_in_references): New function.
722 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
723 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
725 * ipa-split.c (split_function): Remove references in split function.
727 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
730 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
731 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
733 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
736 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
737 suggested mode for the assignment isn't compatible with the
740 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
743 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
745 2013-08-07 Xinliang David Li <davidxl@google.com>
747 * config/i386/i386.opt: New option -mtune-ctrl=.
748 * config/i386/x86-tune.def: New file.
749 * config/i386/i386.h: include x86-tune.def.
750 * config/i386/i386.c (ix86_option_override_internal):
751 Parsing -mtune-ctrl= option and set tune features.
753 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
756 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
757 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
760 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
762 * rtl.h (update_alignments): Declare.
763 * final.c (grow_label_align): New function extracted from...
764 (shorten_branches): ...here. Call it.
765 (update_alignments): New function.
766 * reorg.c (sibling_labels): New variable.
767 (get_label_before): Add SIBLING parameter. If it is non-zero, push
768 the new label along with it onto the sibling_labels vector.
769 (fill_simple_delay_slots): Adjust call to get_label_before.
770 (fill_slots_from_thread): Likewise.
771 (relax_delay_slots): Likewise.
772 (make_return_insns): Likewise.
773 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
775 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
777 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
778 document its semantics.
779 (diagnostic_report_diagnostic): Adjust accordingly.
781 2013-08-07 David Malcolm <dmalcolm@redhat.com>
783 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
784 (sparc_option_override): ...and port to new C++ pass API.
785 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
787 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
789 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
791 2013-08-06 Caroline Tice <cmtice@google.com>
793 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
794 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
795 * tree-pass.h: Add pass_vtable_verify.
796 * varasm.c (assemble_variable): Add code to properly set the comdat
797 section and name for the .vtable_map_vars section.
798 (assemble_vtyv_preinit_initializer): New function.
799 (default_sectin_type_flags): Make sure .vtable_map_vars section has
801 * output.h: Add function decl for assemble_vtv_preinit_initializer.
802 * vtable-verify.c: New file.
803 * vtable-verify.h: New file.
804 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
805 initialiation levels.
806 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
807 * passes.def: Insert pass_vtable_verify.
808 * aclocal.m4: Reorder includes.
809 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
810 -fvtv-counts options.
811 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
812 as appropriate, if -fvtable-verify=... is used.
813 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
814 -fvtable-verify=... is used.
815 * Makefile.in (OBJS): Add vtable-verify.o to list.
816 (vtable-verify.o): Add new build rule.
817 (GTFILES): Add vtable-verify.c to list.
818 * common.opt (fvtable-verify=): New flag.
819 (vtv_priority): Values for fvtable-verify= flag.
820 (fvtv-counts): New flag.
821 (fvtv-debug): New flag.
822 * tree.h (save_vtable_map_decl): New extern function decl.
824 2013-08-07 David Malcolm <dmalcolm@redhat.com>
826 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
827 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
828 (pass_data_rl78_devirt): ...new pass_data instance and...
829 (make_pass_rl78_devirt): ...new function.
830 (rl78_asm_file_start): Port pass registration to new C++ API.
832 2013-08-07 David Malcolm <dmalcolm@redhat.com>
834 * coretypes.h (rtl_opt_pass): Add.
836 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
837 (epiphany_init): Port to new C++ pass API.
838 (epiphany_optimize_mode_switching): Likewise.
839 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
840 (pass_manager::get_pass_mode_switching): New.
841 (pass_manager::get_pass_peephole2): New.
842 * mode-switching.c (pass_mode_switching): Add clone method.
843 * recog.c (pass_peephole2): Add clone method.
844 (pass_split_all_insns): Add clone method.
846 2013-08-06 David Malcolm <dmalcolm@redhat.com>
848 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
849 (mips_option_override): ...here, porting to new C++ API for passes.
851 2013-08-06 Jan Hubicka <jh@suse.cz>
853 * cgraph.c (cgraph_get_body): New function based on lto.c
855 * cgraph.h (cgraph_get_body): Declare.
856 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
858 * cgraphunit.c (expand_function): Get body prior expanding.
859 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
860 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
862 * passes.c (do_per_function_toporder): Get body.
863 * tree-inline.c (expand_call_inline): Get body prior inlining it.
864 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
866 2013-08-06 Martin Jambor <mjambor@suse.cz>
869 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
870 is not re-finalized. Rename second parameter to no_collect.
872 2013-08-06 Martin Jambor <mjambor@suse.cz>
875 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
876 MEM_REF has proper alignment information.
878 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
881 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
883 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
884 rtx (*) (rtx, ...) with insn_gen_fn.
885 * genoutput.c (output_insn_data): Cast gen_? function pointers to
886 insn_gen_fn::stored_funcptr. Add initializer braces.
888 2013-08-05 David Malcolm <dmalcolm@redhat.com>
890 Rewrite how instances of passes are cloned to remove assumptions
891 about their sizes (thus allowing pass subclasses to have
892 additional data fields, albeit non-GC-managed ones at this point).
894 * passes.c (make_pass_instance): Now that passes have clone
895 methods, rewrite this function to eliminate XNEW and memcpy
896 calls that used hardcoded sizes. Since this function no longer
897 creates pass instances, rename it to...
898 (add_pass_instance): ...this. Document the old way that passes were
899 numbered and flagged, and rework this function to continue using it.
900 (next_pass_1): Add an initial_pass argument for use by
902 (position_pass): When adding multiple instances of a pass, use
903 the pass's clone method, rather than relying on the XNEW/memcpy
904 within the former make_pass_instance (now add_pass_instance).
905 (pass_manager::pass_manager): When invoking next_pass_1, also supply
906 the initial instance of the current pass within the pass manager.
908 2013-08-05 David Malcolm <dmalcolm@redhat.com>
910 This is the automated part of the conversion of passes from C
911 structs to C++ classes.
913 Patch autogenerated by refactor_passes.py from
914 https://github.com/davidmalcolm/gcc-refactoring-scripts
915 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
917 * asan.c (pass_asan): Convert from a global struct to a subclass of
918 gimple_opt_pass along with...
919 (pass_data_asan): ...new pass_data instance and...
920 (make_pass_asan): ...new function.
921 (pass_asan_O0): Convert from a global struct to a subclass of
922 gimple_opt_pass along with...
923 (pass_data_asan_O0): ...new pass_data instance and...
924 (make_pass_asan_O0): ...new function.
925 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
926 subclass of rtl_opt_pass along with...
927 (pass_data_inc_dec): ...new pass_data instance and...
928 (make_pass_inc_dec): ...new function.
929 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
930 a subclass of rtl_opt_pass along with...
931 (pass_data_reorder_blocks): ...new pass_data instance and...
932 (make_pass_reorder_blocks): ...new function.
933 (pass_duplicate_computed_gotos): Convert from a global struct to a
934 subclass of rtl_opt_pass along with...
935 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
936 (make_pass_duplicate_computed_gotos): ...new function.
937 (pass_partition_blocks): Convert from a global struct to a subclass of
938 rtl_opt_pass along with...
939 (pass_data_partition_blocks): ...new pass_data instance and...
940 (make_pass_partition_blocks): ...new function.
941 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
942 struct to a subclass of rtl_opt_pass along with...
943 (pass_data_branch_target_load_optimize1): ...new pass_data instance
945 (make_pass_branch_target_load_optimize1): ...new function.
946 (pass_branch_target_load_optimize2): Convert from a global struct to a
947 subclass of rtl_opt_pass along with...
948 (pass_data_branch_target_load_optimize2): ...new pass_data instance
950 (make_pass_branch_target_load_optimize2): ...new function.
951 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
952 of rtl_opt_pass along with...
953 (pass_data_jump): ...new pass_data instance and...
954 (make_pass_jump): ...new function.
955 (pass_jump2): Convert from a global struct to a subclass of
956 rtl_opt_pass along with...
957 (pass_data_jump2): ...new pass_data instance and...
958 (make_pass_jump2): ...new function.
959 * cfgexpand.c (pass_expand): Convert from a global struct to a
960 subclass of rtl_opt_pass along with...
961 (pass_data_expand): ...new pass_data instance and...
962 (make_pass_expand): ...new function.
963 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
964 of rtl_opt_pass along with...
965 (pass_data_free_cfg): ...new pass_data instance and...
966 (make_pass_free_cfg): ...new function.
967 (pass_into_cfg_layout_mode): Convert from a global struct to a
968 subclass of rtl_opt_pass along with...
969 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
970 (make_pass_into_cfg_layout_mode): ...new function.
971 (pass_outof_cfg_layout_mode): Convert from a global struct to a
972 subclass of rtl_opt_pass along with...
973 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
974 (make_pass_outof_cfg_layout_mode): ...new function.
975 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
976 struct to a subclass of gimple_opt_pass along with...
977 (pass_data_build_cgraph_edges): ...new pass_data instance and...
978 (make_pass_build_cgraph_edges): ...new function.
979 (pass_rebuild_cgraph_edges): Convert from a global struct to a
980 subclass of gimple_opt_pass along with...
981 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
982 (make_pass_rebuild_cgraph_edges): ...new function.
983 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
984 subclass of gimple_opt_pass along with...
985 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
987 (make_pass_remove_cgraph_callee_edges): ...new function.
988 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
989 struct to a subclass of rtl_opt_pass along with...
990 (pass_data_stack_adjustments): ...new pass_data instance and...
991 (make_pass_stack_adjustments): ...new function.
992 * combine.c (pass_combine): Convert from a global struct to a subclass
993 of rtl_opt_pass along with...
994 (pass_data_combine): ...new pass_data instance and...
995 (make_pass_combine): ...new function.
996 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
997 global struct to a subclass of rtl_opt_pass along with...
998 (pass_data_compare_elim_after_reload): ...new pass_data instance
1000 (make_pass_compare_elim_after_reload): ...new function.
1001 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1002 of rtl_opt_pass along with...
1003 (pass_data_rtl_cprop): ...new pass_data instance and...
1004 (make_pass_rtl_cprop): ...new function.
1005 * cse.c (pass_cse): Convert from a global struct to a subclass of
1006 rtl_opt_pass along with...
1007 (pass_data_cse): ...new pass_data instance and...
1008 (make_pass_cse): ...new function.
1009 (pass_cse2): Convert from a global struct to a subclass of
1010 rtl_opt_pass along with...
1011 (pass_data_cse2): ...new pass_data instance and...
1012 (make_pass_cse2): ...new function.
1013 (pass_cse_after_global_opts): Convert from a global struct to a
1014 subclass of rtl_opt_pass along with...
1015 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1016 (make_pass_cse_after_global_opts): ...new function.
1017 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1018 of rtl_opt_pass along with...
1019 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1020 (make_pass_ud_rtl_dce): ...new function.
1021 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1022 rtl_opt_pass along with...
1023 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1024 (make_pass_fast_rtl_dce): ...new function.
1025 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1026 a subclass of rtl_opt_pass along with...
1027 (pass_data_df_initialize_opt): ...new pass_data instance and...
1028 (make_pass_df_initialize_opt): ...new function.
1029 (pass_df_initialize_no_opt): Convert from a global struct to a
1030 subclass of rtl_opt_pass along with...
1031 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1032 (make_pass_df_initialize_no_opt): ...new function.
1033 (pass_df_finish): Convert from a global struct to a subclass of
1034 rtl_opt_pass along with...
1035 (pass_data_df_finish): ...new pass_data instance and...
1036 (make_pass_df_finish): ...new function.
1037 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1038 rtl_opt_pass along with...
1039 (pass_data_rtl_dse1): ...new pass_data instance and...
1040 (make_pass_rtl_dse1): ...new function.
1041 (pass_rtl_dse2): Convert from a global struct to a subclass of
1042 rtl_opt_pass along with...
1043 (pass_data_rtl_dse2): ...new pass_data instance and...
1044 (make_pass_rtl_dse2): ...new function.
1045 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1046 subclass of rtl_opt_pass along with...
1047 (pass_data_dwarf2_frame): ...new pass_data instance and...
1048 (make_pass_dwarf2_frame): ...new function.
1049 * except.c (pass_set_nothrow_function_flags): Convert from a global
1050 struct to a subclass of rtl_opt_pass along with...
1051 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1053 (make_pass_set_nothrow_function_flags): ...new function.
1054 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1055 subclass of rtl_opt_pass along with...
1056 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1058 (make_pass_convert_to_eh_region_ranges): ...new function.
1059 * final.c (pass_compute_alignments): Convert from a global struct to a
1060 subclass of rtl_opt_pass along with...
1061 (pass_data_compute_alignments): ...new pass_data instance and...
1062 (make_pass_compute_alignments): ...new function.
1063 (pass_final): Convert from a global struct to a subclass of
1064 rtl_opt_pass along with...
1065 (pass_data_final): ...new pass_data instance and...
1066 (make_pass_final): ...new function.
1067 (pass_shorten_branches): Convert from a global struct to a subclass of
1068 rtl_opt_pass along with...
1069 (pass_data_shorten_branches): ...new pass_data instance and...
1070 (make_pass_shorten_branches): ...new function.
1071 (pass_clean_state): Convert from a global struct to a subclass of
1072 rtl_opt_pass along with...
1073 (pass_data_clean_state): ...new pass_data instance and...
1074 (make_pass_clean_state): ...new function.
1075 * function.c (pass_instantiate_virtual_regs): Convert from a global
1076 struct to a subclass of rtl_opt_pass along with...
1077 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1078 (make_pass_instantiate_virtual_regs): ...new function.
1079 (pass_leaf_regs): Convert from a global struct to a subclass of
1080 rtl_opt_pass along with...
1081 (pass_data_leaf_regs): ...new pass_data instance and...
1082 (make_pass_leaf_regs): ...new function.
1083 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1084 subclass of rtl_opt_pass along with...
1085 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1087 (make_pass_thread_prologue_and_epilogue): ...new function.
1088 (pass_match_asm_constraints): Convert from a global struct to a
1089 subclass of rtl_opt_pass along with...
1090 (pass_data_match_asm_constraints): ...new pass_data instance and...
1091 (make_pass_match_asm_constraints): ...new function.
1092 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1093 subclass of rtl_opt_pass along with...
1094 (pass_data_rtl_fwprop): ...new pass_data instance and...
1095 (make_pass_rtl_fwprop): ...new function.
1096 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1097 rtl_opt_pass along with...
1098 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1099 (make_pass_rtl_fwprop_addr): ...new function.
1100 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1101 rtl_opt_pass along with...
1102 (pass_data_rtl_pre): ...new pass_data instance and...
1103 (make_pass_rtl_pre): ...new function.
1104 (pass_rtl_hoist): Convert from a global struct to a subclass of
1105 rtl_opt_pass along with...
1106 (pass_data_rtl_hoist): ...new pass_data instance and...
1107 (make_pass_rtl_hoist): ...new function.
1108 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1109 subclass of gimple_opt_pass along with...
1110 (pass_data_lower_cf): ...new pass_data instance and...
1111 (make_pass_lower_cf): ...new function.
1112 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1113 from a global struct to a subclass of gimple_opt_pass along with...
1114 (pass_data_strength_reduction): ...new pass_data instance and...
1115 (make_pass_strength_reduction): ...new function.
1116 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1117 of rtl_opt_pass along with...
1118 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1119 (make_pass_rtl_ifcvt): ...new function.
1120 (pass_if_after_combine): Convert from a global struct to a subclass of
1121 rtl_opt_pass along with...
1122 (pass_data_if_after_combine): ...new pass_data instance and...
1123 (make_pass_if_after_combine): ...new function.
1124 (pass_if_after_reload): Convert from a global struct to a subclass of
1125 rtl_opt_pass along with...
1126 (pass_data_if_after_reload): ...new pass_data instance and...
1127 (make_pass_if_after_reload): ...new function.
1128 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1129 a subclass of rtl_opt_pass along with...
1130 (pass_data_initialize_regs): ...new pass_data instance and...
1131 (make_pass_initialize_regs): ...new function.
1132 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1133 of ipa_opt_pass_d along with...
1134 (pass_data_ipa_cp): ...new pass_data instance and...
1135 (make_pass_ipa_cp): ...new function.
1136 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1137 global struct to a subclass of gimple_opt_pass along with...
1138 (pass_data_inline_parameters): ...new pass_data instance and...
1139 (make_pass_inline_parameters): ...new function.
1140 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1141 subclass of gimple_opt_pass along with...
1142 (pass_data_early_inline): ...new pass_data instance and...
1143 (make_pass_early_inline): ...new function.
1144 (pass_ipa_inline): Convert from a global struct to a subclass of
1145 ipa_opt_pass_d along with...
1146 (pass_data_ipa_inline): ...new pass_data instance and...
1147 (make_pass_ipa_inline): ...new function.
1148 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1149 struct to a subclass of gimple_opt_pass along with...
1150 (pass_data_local_pure_const): ...new pass_data instance and...
1151 (make_pass_local_pure_const): ...new function.
1152 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1153 ipa_opt_pass_d along with...
1154 (pass_data_ipa_pure_const): ...new pass_data instance and...
1155 (make_pass_ipa_pure_const): ...new function.
1156 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1157 to a subclass of ipa_opt_pass_d along with...
1158 (pass_data_ipa_reference): ...new pass_data instance and...
1159 (make_pass_ipa_reference): ...new function.
1160 * ipa-split.c (pass_split_functions): Convert from a global struct to
1161 a subclass of gimple_opt_pass along with...
1162 (pass_data_split_functions): ...new pass_data instance and...
1163 (make_pass_split_functions): ...new function.
1164 (pass_feedback_split_functions): Convert from a global struct to a
1165 subclass of gimple_opt_pass along with...
1166 (pass_data_feedback_split_functions): ...new pass_data instance and...
1167 (make_pass_feedback_split_functions): ...new function.
1168 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1169 global struct to a subclass of simple_ipa_opt_pass along with...
1170 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1172 (make_pass_ipa_function_and_variable_visibility): ...new function.
1173 (pass_ipa_free_inline_summary): Convert from a global struct to a
1174 subclass of simple_ipa_opt_pass along with...
1175 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1176 (make_pass_ipa_free_inline_summary): ...new function.
1177 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1178 subclass of ipa_opt_pass_d along with...
1179 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1181 (make_pass_ipa_whole_program_visibility): ...new function.
1182 (pass_ipa_profile): Convert from a global struct to a subclass of
1183 ipa_opt_pass_d along with...
1184 (pass_data_ipa_profile): ...new pass_data instance and...
1185 (make_pass_ipa_profile): ...new function.
1186 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1187 ipa_opt_pass_d along with...
1188 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1189 (make_pass_ipa_cdtor_merge): ...new function.
1190 * ira.c (pass_ira): Convert from a global struct to a subclass of
1191 rtl_opt_pass along with...
1192 (pass_data_ira): ...new pass_data instance and...
1193 (make_pass_ira): ...new function.
1194 (pass_reload): Convert from a global struct to a subclass of
1195 rtl_opt_pass along with...
1196 (pass_data_reload): ...new pass_data instance and...
1197 (make_pass_reload): ...new function.
1198 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1199 subclass of rtl_opt_pass along with...
1200 (pass_data_cleanup_barriers): ...new pass_data instance and...
1201 (make_pass_cleanup_barriers): ...new function.
1202 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1203 of rtl_opt_pass along with...
1204 (pass_data_loop2): ...new pass_data instance and...
1205 (make_pass_loop2): ...new function.
1206 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1207 rtl_opt_pass along with...
1208 (pass_data_rtl_loop_init): ...new pass_data instance and...
1209 (make_pass_rtl_loop_init): ...new function.
1210 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1211 rtl_opt_pass along with...
1212 (pass_data_rtl_loop_done): ...new pass_data instance and...
1213 (make_pass_rtl_loop_done): ...new function.
1214 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1215 subclass of rtl_opt_pass along with...
1216 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1217 (make_pass_rtl_move_loop_invariants): ...new function.
1218 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1219 rtl_opt_pass along with...
1220 (pass_data_rtl_unswitch): ...new pass_data instance and...
1221 (make_pass_rtl_unswitch): ...new function.
1222 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1223 subclass of rtl_opt_pass along with...
1224 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1226 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1227 (pass_rtl_doloop): Convert from a global struct to a subclass of
1228 rtl_opt_pass along with...
1229 (pass_data_rtl_doloop): ...new pass_data instance and...
1230 (make_pass_rtl_doloop): ...new function.
1231 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1232 a subclass of rtl_opt_pass along with...
1233 (pass_data_lower_subreg): ...new pass_data instance and...
1234 (make_pass_lower_subreg): ...new function.
1235 (pass_lower_subreg2): Convert from a global struct to a subclass of
1236 rtl_opt_pass along with...
1237 (pass_data_lower_subreg2): ...new pass_data instance and...
1238 (make_pass_lower_subreg2): ...new function.
1239 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1240 struct to a subclass of ipa_opt_pass_d along with...
1241 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1242 (make_pass_ipa_lto_gimple_out): ...new function.
1243 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1244 of ipa_opt_pass_d along with...
1245 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1246 (make_pass_ipa_lto_finish_out): ...new function.
1247 * mode-switching.c (pass_mode_switching): Convert from a global struct
1248 to a subclass of rtl_opt_pass along with...
1249 (pass_data_mode_switching): ...new pass_data instance and...
1250 (make_pass_mode_switching): ...new function.
1251 * modulo-sched.c (pass_sms): Convert from a global struct to a
1252 subclass of rtl_opt_pass along with...
1253 (pass_data_sms): ...new pass_data instance and...
1254 (make_pass_sms): ...new function.
1255 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1256 subclass of gimple_opt_pass along with...
1257 (pass_data_expand_omp): ...new pass_data instance and...
1258 (make_pass_expand_omp): ...new function.
1259 (pass_lower_omp): Convert from a global struct to a subclass of
1260 gimple_opt_pass along with...
1261 (pass_data_lower_omp): ...new pass_data instance and...
1262 (make_pass_lower_omp): ...new function.
1263 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1264 of gimple_opt_pass along with...
1265 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1266 (make_pass_diagnose_omp_blocks): ...new function.
1267 * passes.c (pass_early_local_passes): Convert from a global struct to
1268 a subclass of simple_ipa_opt_pass along with...
1269 (pass_data_early_local_passes): ...new pass_data instance and...
1270 (make_pass_early_local_passes): ...new function.
1271 (pass_all_early_optimizations): Convert from a global struct to a
1272 subclass of gimple_opt_pass along with...
1273 (pass_data_all_early_optimizations): ...new pass_data instance and...
1274 (make_pass_all_early_optimizations): ...new function.
1275 (pass_all_optimizations): Convert from a global struct to a subclass
1276 of gimple_opt_pass along with...
1277 (pass_data_all_optimizations): ...new pass_data instance and...
1278 (make_pass_all_optimizations): ...new function.
1279 (pass_all_optimizations_g): Convert from a global struct to a subclass
1280 of gimple_opt_pass along with...
1281 (pass_data_all_optimizations_g): ...new pass_data instance and...
1282 (make_pass_all_optimizations_g): ...new function.
1283 (pass_rest_of_compilation): Convert from a global struct to a subclass
1284 of rtl_opt_pass along with...
1285 (pass_data_rest_of_compilation): ...new pass_data instance and...
1286 (make_pass_rest_of_compilation): ...new function.
1287 (pass_postreload): Convert from a global struct to a subclass of
1288 rtl_opt_pass along with...
1289 (pass_data_postreload): ...new pass_data instance and...
1290 (make_pass_postreload): ...new function.
1291 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1292 subclass of rtl_opt_pass along with...
1293 (pass_data_gcse2): ...new pass_data instance and...
1294 (make_pass_gcse2): ...new function.
1295 * postreload.c (pass_postreload_cse): Convert from a global struct to
1296 a subclass of rtl_opt_pass along with...
1297 (pass_data_postreload_cse): ...new pass_data instance and...
1298 (make_pass_postreload_cse): ...new function.
1299 * predict.c (pass_profile): Convert from a global struct to a subclass
1300 of gimple_opt_pass along with...
1301 (pass_data_profile): ...new pass_data instance and...
1302 (make_pass_profile): ...new function.
1303 (pass_strip_predict_hints): Convert from a global struct to a subclass
1304 of gimple_opt_pass along with...
1305 (pass_data_strip_predict_hints): ...new pass_data instance and...
1306 (make_pass_strip_predict_hints): ...new function.
1307 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1308 of rtl_opt_pass along with...
1309 (pass_data_peephole2): ...new pass_data instance and...
1310 (make_pass_peephole2): ...new function.
1311 (pass_split_all_insns): Convert from a global struct to a subclass of
1312 rtl_opt_pass along with...
1313 (pass_data_split_all_insns): ...new pass_data instance and...
1314 (make_pass_split_all_insns): ...new function.
1315 (pass_split_after_reload): Convert from a global struct to a subclass
1316 of rtl_opt_pass along with...
1317 (pass_data_split_after_reload): ...new pass_data instance and...
1318 (make_pass_split_after_reload): ...new function.
1319 (pass_split_before_regstack): Convert from a global struct to a
1320 subclass of rtl_opt_pass along with...
1321 (pass_data_split_before_regstack): ...new pass_data instance and...
1322 (make_pass_split_before_regstack): ...new function.
1323 (pass_split_before_sched2): Convert from a global struct to a subclass
1324 of rtl_opt_pass along with...
1325 (pass_data_split_before_sched2): ...new pass_data instance and...
1326 (make_pass_split_before_sched2): ...new function.
1327 (pass_split_for_shorten_branches): Convert from a global struct to a
1328 subclass of rtl_opt_pass along with...
1329 (pass_data_split_for_shorten_branches): ...new pass_data instance
1331 (make_pass_split_for_shorten_branches): ...new function.
1332 * ree.c (pass_ree): Convert from a global struct to a subclass of
1333 rtl_opt_pass along with...
1334 (pass_data_ree): ...new pass_data instance and...
1335 (make_pass_ree): ...new function.
1336 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1337 subclass of rtl_opt_pass along with...
1338 (pass_data_stack_regs): ...new pass_data instance and...
1339 (make_pass_stack_regs): ...new function.
1340 (pass_stack_regs_run): Convert from a global struct to a subclass of
1341 rtl_opt_pass along with...
1342 (pass_data_stack_regs_run): ...new pass_data instance and...
1343 (make_pass_stack_regs_run): ...new function.
1344 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1345 subclass of rtl_opt_pass along with...
1346 (pass_data_cprop_hardreg): ...new pass_data instance and...
1347 (make_pass_cprop_hardreg): ...new function.
1348 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1349 subclass of rtl_opt_pass along with...
1350 (pass_data_reginfo_init): ...new pass_data instance and...
1351 (make_pass_reginfo_init): ...new function.
1352 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1353 of rtl_opt_pass along with...
1354 (pass_data_regmove): ...new pass_data instance and...
1355 (make_pass_regmove): ...new function.
1356 * regrename.c (pass_regrename): Convert from a global struct to a
1357 subclass of rtl_opt_pass along with...
1358 (pass_data_regrename): ...new pass_data instance and...
1359 (make_pass_regrename): ...new function.
1360 * reorg.c (pass_delay_slots): Convert from a global struct to a
1361 subclass of rtl_opt_pass along with...
1362 (pass_data_delay_slots): ...new pass_data instance and...
1363 (make_pass_delay_slots): ...new function.
1364 (pass_machine_reorg): Convert from a global struct to a subclass of
1365 rtl_opt_pass along with...
1366 (pass_data_machine_reorg): ...new pass_data instance and...
1367 (make_pass_machine_reorg): ...new function.
1368 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1369 of rtl_opt_pass along with...
1370 (pass_data_sched): ...new pass_data instance and...
1371 (make_pass_sched): ...new function.
1372 (pass_sched2): Convert from a global struct to a subclass of
1373 rtl_opt_pass along with...
1374 (pass_data_sched2): ...new pass_data instance and...
1375 (make_pass_sched2): ...new function.
1376 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1377 to a subclass of rtl_opt_pass along with...
1378 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1379 (make_pass_stack_ptr_mod): ...new function.
1380 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1381 to a subclass of rtl_opt_pass along with...
1382 (pass_data_rtl_store_motion): ...new pass_data instance and...
1383 (make_pass_rtl_store_motion): ...new function.
1384 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1385 of gimple_opt_pass along with...
1386 (pass_data_tracer): ...new pass_data instance and...
1387 (make_pass_tracer): ...new function.
1388 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1389 to a subclass of gimple_opt_pass along with...
1390 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1391 (make_pass_diagnose_tm_blocks): ...new function.
1392 (pass_lower_tm): Convert from a global struct to a subclass of
1393 gimple_opt_pass along with...
1394 (pass_data_lower_tm): ...new pass_data instance and...
1395 (make_pass_lower_tm): ...new function.
1396 (pass_tm_init): Convert from a global struct to a subclass of
1397 gimple_opt_pass along with...
1398 (pass_data_tm_init): ...new pass_data instance and...
1399 (make_pass_tm_init): ...new function.
1400 (pass_tm_mark): Convert from a global struct to a subclass of
1401 gimple_opt_pass along with...
1402 (pass_data_tm_mark): ...new pass_data instance and...
1403 (make_pass_tm_mark): ...new function.
1404 (pass_tm_edges): Convert from a global struct to a subclass of
1405 gimple_opt_pass along with...
1406 (pass_data_tm_edges): ...new pass_data instance and...
1407 (make_pass_tm_edges): ...new function.
1408 (pass_tm_memopt): Convert from a global struct to a subclass of
1409 gimple_opt_pass along with...
1410 (pass_data_tm_memopt): ...new pass_data instance and...
1411 (make_pass_tm_memopt): ...new function.
1412 (pass_ipa_tm): Convert from a global struct to a subclass of
1413 simple_ipa_opt_pass along with...
1414 (pass_data_ipa_tm): ...new pass_data instance and...
1415 (make_pass_ipa_tm): ...new function.
1416 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1417 subclass of gimple_opt_pass along with...
1418 (pass_data_call_cdce): ...new pass_data instance and...
1419 (make_pass_call_cdce): ...new function.
1420 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1421 subclass of gimple_opt_pass along with...
1422 (pass_data_build_cfg): ...new pass_data instance and...
1423 (make_pass_build_cfg): ...new function.
1424 (pass_split_crit_edges): Convert from a global struct to a subclass of
1425 gimple_opt_pass along with...
1426 (pass_data_split_crit_edges): ...new pass_data instance and...
1427 (make_pass_split_crit_edges): ...new function.
1428 (pass_warn_function_return): Convert from a global struct to a
1429 subclass of gimple_opt_pass along with...
1430 (pass_data_warn_function_return): ...new pass_data instance and...
1431 (make_pass_warn_function_return): ...new function.
1432 (pass_warn_function_noreturn): Convert from a global struct to a
1433 subclass of gimple_opt_pass along with...
1434 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1435 (make_pass_warn_function_noreturn): ...new function.
1436 (pass_warn_unused_result): Convert from a global struct to a subclass
1437 of gimple_opt_pass along with...
1438 (pass_data_warn_unused_result): ...new pass_data instance and...
1439 (make_pass_warn_unused_result): ...new function.
1440 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1441 a subclass of gimple_opt_pass along with...
1442 (pass_data_merge_phi): ...new pass_data instance and...
1443 (make_pass_merge_phi): ...new function.
1444 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1445 a subclass of gimple_opt_pass along with...
1446 (pass_data_lower_complex): ...new pass_data instance and...
1447 (make_pass_lower_complex): ...new function.
1448 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1449 gimple_opt_pass along with...
1450 (pass_data_lower_complex_O0): ...new pass_data instance and...
1451 (make_pass_lower_complex_O0): ...new function.
1452 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1453 subclass of gimple_opt_pass along with...
1454 (pass_data_lower_eh): ...new pass_data instance and...
1455 (make_pass_lower_eh): ...new function.
1456 (pass_refactor_eh): Convert from a global struct to a subclass of
1457 gimple_opt_pass along with...
1458 (pass_data_refactor_eh): ...new pass_data instance and...
1459 (make_pass_refactor_eh): ...new function.
1460 (pass_lower_resx): Convert from a global struct to a subclass of
1461 gimple_opt_pass along with...
1462 (pass_data_lower_resx): ...new pass_data instance and...
1463 (make_pass_lower_resx): ...new function.
1464 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1465 of gimple_opt_pass along with...
1466 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1467 (make_pass_lower_eh_dispatch): ...new function.
1468 (pass_cleanup_eh): Convert from a global struct to a subclass of
1469 gimple_opt_pass along with...
1470 (pass_data_cleanup_eh): ...new pass_data instance and...
1471 (make_pass_cleanup_eh): ...new function.
1472 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1473 to a subclass of simple_ipa_opt_pass along with...
1474 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1475 (make_pass_ipa_lower_emutls): ...new function.
1476 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1477 a subclass of gimple_opt_pass along with...
1478 (pass_data_if_conversion): ...new pass_data instance and...
1479 (make_pass_if_conversion): ...new function.
1480 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1481 subclass of gimple_opt_pass along with...
1482 (pass_data_build_ssa): ...new pass_data instance and...
1483 (make_pass_build_ssa): ...new function.
1484 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1485 global struct to a subclass of gimple_opt_pass along with...
1486 (pass_data_loop_distribution): ...new pass_data instance and...
1487 (make_pass_loop_distribution): ...new function.
1488 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1489 subclass of gimple_opt_pass along with...
1490 (pass_data_mudflap_1): ...new pass_data instance and...
1491 (make_pass_mudflap_1): ...new function.
1492 (pass_mudflap_2): Convert from a global struct to a subclass of
1493 gimple_opt_pass along with...
1494 (pass_data_mudflap_2): ...new pass_data instance and...
1495 (make_pass_mudflap_2): ...new function.
1496 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1497 subclass of gimple_opt_pass along with...
1498 (pass_data_mudflap_1): ...new pass_data instance and...
1499 (make_pass_mudflap_1): ...new function.
1500 (pass_mudflap_2): Convert from a global struct to a subclass of
1501 gimple_opt_pass along with...
1502 (pass_data_mudflap_2): ...new pass_data instance and...
1503 (make_pass_mudflap_2): ...new function.
1504 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1505 gimple_opt_pass along with...
1506 (pass_data_nrv): ...new pass_data instance and...
1507 (make_pass_nrv): ...new function.
1508 (pass_return_slot): Convert from a global struct to a subclass of
1509 gimple_opt_pass along with...
1510 (pass_data_return_slot): ...new pass_data instance and...
1511 (make_pass_return_slot): ...new function.
1512 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1513 to a subclass of gimple_opt_pass along with...
1514 (pass_data_object_sizes): ...new pass_data instance and...
1515 (make_pass_object_sizes): ...new function.
1516 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1517 global struct to a subclass of gimple_opt_pass along with...
1518 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1520 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1521 (pass_fixup_cfg): Convert from a global struct to a subclass of
1522 gimple_opt_pass along with...
1523 (pass_data_fixup_cfg): ...new pass_data instance and...
1524 (make_pass_fixup_cfg): ...new function.
1525 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1526 (make_pass_mudflap_1): ...new function.
1527 (pass_mudflap_2): Replace declaration with that of...
1528 (make_pass_mudflap_2): ...new function.
1529 (pass_asan): Replace declaration with that of...
1530 (make_pass_asan): ...new function.
1531 (pass_asan_O0): Replace declaration with that of...
1532 (make_pass_asan_O0): ...new function.
1533 (pass_tsan): Replace declaration with that of...
1534 (make_pass_tsan): ...new function.
1535 (pass_tsan_O0): Replace declaration with that of...
1536 (make_pass_tsan_O0): ...new function.
1537 (pass_lower_cf): Replace declaration with that of...
1538 (make_pass_lower_cf): ...new function.
1539 (pass_refactor_eh): Replace declaration with that of...
1540 (make_pass_refactor_eh): ...new function.
1541 (pass_lower_eh): Replace declaration with that of...
1542 (make_pass_lower_eh): ...new function.
1543 (pass_lower_eh_dispatch): Replace declaration with that of...
1544 (make_pass_lower_eh_dispatch): ...new function.
1545 (pass_lower_resx): Replace declaration with that of...
1546 (make_pass_lower_resx): ...new function.
1547 (pass_build_cfg): Replace declaration with that of...
1548 (make_pass_build_cfg): ...new function.
1549 (pass_early_tree_profile): Replace declaration with that of...
1550 (make_pass_early_tree_profile): ...new function.
1551 (pass_cleanup_eh): Replace declaration with that of...
1552 (make_pass_cleanup_eh): ...new function.
1553 (pass_sra): Replace declaration with that of...
1554 (make_pass_sra): ...new function.
1555 (pass_sra_early): Replace declaration with that of...
1556 (make_pass_sra_early): ...new function.
1557 (pass_early_ipa_sra): Replace declaration with that of...
1558 (make_pass_early_ipa_sra): ...new function.
1559 (pass_tail_recursion): Replace declaration with that of...
1560 (make_pass_tail_recursion): ...new function.
1561 (pass_tail_calls): Replace declaration with that of...
1562 (make_pass_tail_calls): ...new function.
1563 (pass_tree_loop): Replace declaration with that of...
1564 (make_pass_tree_loop): ...new function.
1565 (pass_tree_loop_init): Replace declaration with that of...
1566 (make_pass_tree_loop_init): ...new function.
1567 (pass_lim): Replace declaration with that of...
1568 (make_pass_lim): ...new function.
1569 (pass_tree_unswitch): Replace declaration with that of...
1570 (make_pass_tree_unswitch): ...new function.
1571 (pass_predcom): Replace declaration with that of...
1572 (make_pass_predcom): ...new function.
1573 (pass_iv_canon): Replace declaration with that of...
1574 (make_pass_iv_canon): ...new function.
1575 (pass_scev_cprop): Replace declaration with that of...
1576 (make_pass_scev_cprop): ...new function.
1577 (pass_empty_loop): Replace declaration with that of...
1578 (make_pass_empty_loop): ...new function.
1579 (pass_record_bounds): Replace declaration with that of...
1580 (make_pass_record_bounds): ...new function.
1581 (pass_graphite): Replace declaration with that of...
1582 (make_pass_graphite): ...new function.
1583 (pass_graphite_transforms): Replace declaration with that of...
1584 (make_pass_graphite_transforms): ...new function.
1585 (pass_if_conversion): Replace declaration with that of...
1586 (make_pass_if_conversion): ...new function.
1587 (pass_loop_distribution): Replace declaration with that of...
1588 (make_pass_loop_distribution): ...new function.
1589 (pass_vectorize): Replace declaration with that of...
1590 (make_pass_vectorize): ...new function.
1591 (pass_slp_vectorize): Replace declaration with that of...
1592 (make_pass_slp_vectorize): ...new function.
1593 (pass_complete_unroll): Replace declaration with that of...
1594 (make_pass_complete_unroll): ...new function.
1595 (pass_complete_unrolli): Replace declaration with that of...
1596 (make_pass_complete_unrolli): ...new function.
1597 (pass_parallelize_loops): Replace declaration with that of...
1598 (make_pass_parallelize_loops): ...new function.
1599 (pass_loop_prefetch): Replace declaration with that of...
1600 (make_pass_loop_prefetch): ...new function.
1601 (pass_iv_optimize): Replace declaration with that of...
1602 (make_pass_iv_optimize): ...new function.
1603 (pass_tree_loop_done): Replace declaration with that of...
1604 (make_pass_tree_loop_done): ...new function.
1605 (pass_ch): Replace declaration with that of...
1606 (make_pass_ch): ...new function.
1607 (pass_ccp): Replace declaration with that of...
1608 (make_pass_ccp): ...new function.
1609 (pass_phi_only_cprop): Replace declaration with that of...
1610 (make_pass_phi_only_cprop): ...new function.
1611 (pass_build_ssa): Replace declaration with that of...
1612 (make_pass_build_ssa): ...new function.
1613 (pass_build_alias): Replace declaration with that of...
1614 (make_pass_build_alias): ...new function.
1615 (pass_build_ealias): Replace declaration with that of...
1616 (make_pass_build_ealias): ...new function.
1617 (pass_dominator): Replace declaration with that of...
1618 (make_pass_dominator): ...new function.
1619 (pass_dce): Replace declaration with that of...
1620 (make_pass_dce): ...new function.
1621 (pass_dce_loop): Replace declaration with that of...
1622 (make_pass_dce_loop): ...new function.
1623 (pass_cd_dce): Replace declaration with that of...
1624 (make_pass_cd_dce): ...new function.
1625 (pass_call_cdce): Replace declaration with that of...
1626 (make_pass_call_cdce): ...new function.
1627 (pass_merge_phi): Replace declaration with that of...
1628 (make_pass_merge_phi): ...new function.
1629 (pass_split_crit_edges): Replace declaration with that of...
1630 (make_pass_split_crit_edges): ...new function.
1631 (pass_pre): Replace declaration with that of...
1632 (make_pass_pre): ...new function.
1633 (pass_profile): Replace declaration with that of...
1634 (make_pass_profile): ...new function.
1635 (pass_strip_predict_hints): Replace declaration with that of...
1636 (make_pass_strip_predict_hints): ...new function.
1637 (pass_lower_complex_O0): Replace declaration with that of...
1638 (make_pass_lower_complex_O0): ...new function.
1639 (pass_lower_complex): Replace declaration with that of...
1640 (make_pass_lower_complex): ...new function.
1641 (pass_lower_vector): Replace declaration with that of...
1642 (make_pass_lower_vector): ...new function.
1643 (pass_lower_vector_ssa): Replace declaration with that of...
1644 (make_pass_lower_vector_ssa): ...new function.
1645 (pass_lower_omp): Replace declaration with that of...
1646 (make_pass_lower_omp): ...new function.
1647 (pass_diagnose_omp_blocks): Replace declaration with that of...
1648 (make_pass_diagnose_omp_blocks): ...new function.
1649 (pass_expand_omp): Replace declaration with that of...
1650 (make_pass_expand_omp): ...new function.
1651 (pass_expand_omp_ssa): Replace declaration with that of...
1652 (make_pass_expand_omp_ssa): ...new function.
1653 (pass_object_sizes): Replace declaration with that of...
1654 (make_pass_object_sizes): ...new function.
1655 (pass_strlen): Replace declaration with that of...
1656 (make_pass_strlen): ...new function.
1657 (pass_fold_builtins): Replace declaration with that of...
1658 (make_pass_fold_builtins): ...new function.
1659 (pass_stdarg): Replace declaration with that of...
1660 (make_pass_stdarg): ...new function.
1661 (pass_early_warn_uninitialized): Replace declaration with that of...
1662 (make_pass_early_warn_uninitialized): ...new function.
1663 (pass_late_warn_uninitialized): Replace declaration with that of...
1664 (make_pass_late_warn_uninitialized): ...new function.
1665 (pass_cse_reciprocals): Replace declaration with that of...
1666 (make_pass_cse_reciprocals): ...new function.
1667 (pass_cse_sincos): Replace declaration with that of...
1668 (make_pass_cse_sincos): ...new function.
1669 (pass_optimize_bswap): Replace declaration with that of...
1670 (make_pass_optimize_bswap): ...new function.
1671 (pass_optimize_widening_mul): Replace declaration with that of...
1672 (make_pass_optimize_widening_mul): ...new function.
1673 (pass_warn_function_return): Replace declaration with that of...
1674 (make_pass_warn_function_return): ...new function.
1675 (pass_warn_function_noreturn): Replace declaration with that of...
1676 (make_pass_warn_function_noreturn): ...new function.
1677 (pass_cselim): Replace declaration with that of...
1678 (make_pass_cselim): ...new function.
1679 (pass_phiopt): Replace declaration with that of...
1680 (make_pass_phiopt): ...new function.
1681 (pass_forwprop): Replace declaration with that of...
1682 (make_pass_forwprop): ...new function.
1683 (pass_phiprop): Replace declaration with that of...
1684 (make_pass_phiprop): ...new function.
1685 (pass_tree_ifcombine): Replace declaration with that of...
1686 (make_pass_tree_ifcombine): ...new function.
1687 (pass_dse): Replace declaration with that of...
1688 (make_pass_dse): ...new function.
1689 (pass_nrv): Replace declaration with that of...
1690 (make_pass_nrv): ...new function.
1691 (pass_rename_ssa_copies): Replace declaration with that of...
1692 (make_pass_rename_ssa_copies): ...new function.
1693 (pass_sink_code): Replace declaration with that of...
1694 (make_pass_sink_code): ...new function.
1695 (pass_fre): Replace declaration with that of...
1696 (make_pass_fre): ...new function.
1697 (pass_check_data_deps): Replace declaration with that of...
1698 (make_pass_check_data_deps): ...new function.
1699 (pass_copy_prop): Replace declaration with that of...
1700 (make_pass_copy_prop): ...new function.
1701 (pass_vrp): Replace declaration with that of...
1702 (make_pass_vrp): ...new function.
1703 (pass_uncprop): Replace declaration with that of...
1704 (make_pass_uncprop): ...new function.
1705 (pass_return_slot): Replace declaration with that of...
1706 (make_pass_return_slot): ...new function.
1707 (pass_reassoc): Replace declaration with that of...
1708 (make_pass_reassoc): ...new function.
1709 (pass_rebuild_cgraph_edges): Replace declaration with that of...
1710 (make_pass_rebuild_cgraph_edges): ...new function.
1711 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
1712 (make_pass_remove_cgraph_callee_edges): ...new function.
1713 (pass_build_cgraph_edges): Replace declaration with that of...
1714 (make_pass_build_cgraph_edges): ...new function.
1715 (pass_local_pure_const): Replace declaration with that of...
1716 (make_pass_local_pure_const): ...new function.
1717 (pass_tracer): Replace declaration with that of...
1718 (make_pass_tracer): ...new function.
1719 (pass_warn_unused_result): Replace declaration with that of...
1720 (make_pass_warn_unused_result): ...new function.
1721 (pass_diagnose_tm_blocks): Replace declaration with that of...
1722 (make_pass_diagnose_tm_blocks): ...new function.
1723 (pass_lower_tm): Replace declaration with that of...
1724 (make_pass_lower_tm): ...new function.
1725 (pass_tm_init): Replace declaration with that of...
1726 (make_pass_tm_init): ...new function.
1727 (pass_tm_mark): Replace declaration with that of...
1728 (make_pass_tm_mark): ...new function.
1729 (pass_tm_memopt): Replace declaration with that of...
1730 (make_pass_tm_memopt): ...new function.
1731 (pass_tm_edges): Replace declaration with that of...
1732 (make_pass_tm_edges): ...new function.
1733 (pass_split_functions): Replace declaration with that of...
1734 (make_pass_split_functions): ...new function.
1735 (pass_feedback_split_functions): Replace declaration with that of...
1736 (make_pass_feedback_split_functions): ...new function.
1737 (pass_strength_reduction): Replace declaration with that of...
1738 (make_pass_strength_reduction): ...new function.
1739 (pass_ipa_lower_emutls): Replace declaration with that of...
1740 (make_pass_ipa_lower_emutls): ...new function.
1741 (pass_ipa_function_and_variable_visibility): Replace declaration with
1743 (make_pass_ipa_function_and_variable_visibility): ...new function.
1744 (pass_ipa_tree_profile): Replace declaration with that of...
1745 (make_pass_ipa_tree_profile): ...new function.
1746 (pass_early_local_passes): Replace declaration with that of...
1747 (make_pass_early_local_passes): ...new function.
1748 (pass_ipa_whole_program_visibility): Replace declaration with that
1750 (make_pass_ipa_whole_program_visibility): ...new function.
1751 (pass_ipa_lto_gimple_out): Replace declaration with that of...
1752 (make_pass_ipa_lto_gimple_out): ...new function.
1753 (pass_ipa_increase_alignment): Replace declaration with that of...
1754 (make_pass_ipa_increase_alignment): ...new function.
1755 (pass_ipa_inline): Replace declaration with that of...
1756 (make_pass_ipa_inline): ...new function.
1757 (pass_ipa_free_lang_data): Replace declaration with that of...
1758 (make_pass_ipa_free_lang_data): ...new function.
1759 (pass_ipa_free_inline_summary): Replace declaration with that of...
1760 (make_pass_ipa_free_inline_summary): ...new function.
1761 (pass_ipa_cp): Replace declaration with that of...
1762 (make_pass_ipa_cp): ...new function.
1763 (pass_ipa_reference): Replace declaration with that of...
1764 (make_pass_ipa_reference): ...new function.
1765 (pass_ipa_pure_const): Replace declaration with that of...
1766 (make_pass_ipa_pure_const): ...new function.
1767 (pass_ipa_pta): Replace declaration with that of...
1768 (make_pass_ipa_pta): ...new function.
1769 (pass_ipa_lto_finish_out): Replace declaration with that of...
1770 (make_pass_ipa_lto_finish_out): ...new function.
1771 (pass_ipa_tm): Replace declaration with that of...
1772 (make_pass_ipa_tm): ...new function.
1773 (pass_ipa_profile): Replace declaration with that of...
1774 (make_pass_ipa_profile): ...new function.
1775 (pass_ipa_cdtor_merge): Replace declaration with that of...
1776 (make_pass_ipa_cdtor_merge): ...new function.
1777 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
1779 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1780 (pass_init_datastructures): Replace declaration with that of...
1781 (make_pass_init_datastructures): ...new function.
1782 (pass_fixup_cfg): Replace declaration with that of...
1783 (make_pass_fixup_cfg): ...new function.
1784 (pass_expand): Replace declaration with that of...
1785 (make_pass_expand): ...new function.
1786 (pass_instantiate_virtual_regs): Replace declaration with that of...
1787 (make_pass_instantiate_virtual_regs): ...new function.
1788 (pass_rtl_fwprop): Replace declaration with that of...
1789 (make_pass_rtl_fwprop): ...new function.
1790 (pass_rtl_fwprop_addr): Replace declaration with that of...
1791 (make_pass_rtl_fwprop_addr): ...new function.
1792 (pass_jump): Replace declaration with that of...
1793 (make_pass_jump): ...new function.
1794 (pass_jump2): Replace declaration with that of...
1795 (make_pass_jump2): ...new function.
1796 (pass_lower_subreg): Replace declaration with that of...
1797 (make_pass_lower_subreg): ...new function.
1798 (pass_cse): Replace declaration with that of...
1799 (make_pass_cse): ...new function.
1800 (pass_fast_rtl_dce): Replace declaration with that of...
1801 (make_pass_fast_rtl_dce): ...new function.
1802 (pass_ud_rtl_dce): Replace declaration with that of...
1803 (make_pass_ud_rtl_dce): ...new function.
1804 (pass_rtl_dce): Replace declaration with that of...
1805 (make_pass_rtl_dce): ...new function.
1806 (pass_rtl_dse1): Replace declaration with that of...
1807 (make_pass_rtl_dse1): ...new function.
1808 (pass_rtl_dse2): Replace declaration with that of...
1809 (make_pass_rtl_dse2): ...new function.
1810 (pass_rtl_dse3): Replace declaration with that of...
1811 (make_pass_rtl_dse3): ...new function.
1812 (pass_rtl_cprop): Replace declaration with that of...
1813 (make_pass_rtl_cprop): ...new function.
1814 (pass_rtl_pre): Replace declaration with that of...
1815 (make_pass_rtl_pre): ...new function.
1816 (pass_rtl_hoist): Replace declaration with that of...
1817 (make_pass_rtl_hoist): ...new function.
1818 (pass_rtl_store_motion): Replace declaration with that of...
1819 (make_pass_rtl_store_motion): ...new function.
1820 (pass_cse_after_global_opts): Replace declaration with that of...
1821 (make_pass_cse_after_global_opts): ...new function.
1822 (pass_rtl_ifcvt): Replace declaration with that of...
1823 (make_pass_rtl_ifcvt): ...new function.
1824 (pass_into_cfg_layout_mode): Replace declaration with that of...
1825 (make_pass_into_cfg_layout_mode): ...new function.
1826 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1827 (make_pass_outof_cfg_layout_mode): ...new function.
1828 (pass_loop2): Replace declaration with that of...
1829 (make_pass_loop2): ...new function.
1830 (pass_rtl_loop_init): Replace declaration with that of...
1831 (make_pass_rtl_loop_init): ...new function.
1832 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1833 (make_pass_rtl_move_loop_invariants): ...new function.
1834 (pass_rtl_unswitch): Replace declaration with that of...
1835 (make_pass_rtl_unswitch): ...new function.
1836 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1837 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1838 (pass_rtl_doloop): Replace declaration with that of...
1839 (make_pass_rtl_doloop): ...new function.
1840 (pass_rtl_loop_done): Replace declaration with that of...
1841 (make_pass_rtl_loop_done): ...new function.
1842 (pass_web): Replace declaration with that of...
1843 (make_pass_web): ...new function.
1844 (pass_cse2): Replace declaration with that of...
1845 (make_pass_cse2): ...new function.
1846 (pass_df_initialize_opt): Replace declaration with that of...
1847 (make_pass_df_initialize_opt): ...new function.
1848 (pass_df_initialize_no_opt): Replace declaration with that of...
1849 (make_pass_df_initialize_no_opt): ...new function.
1850 (pass_reginfo_init): Replace declaration with that of...
1851 (make_pass_reginfo_init): ...new function.
1852 (pass_inc_dec): Replace declaration with that of...
1853 (make_pass_inc_dec): ...new function.
1854 (pass_stack_ptr_mod): Replace declaration with that of...
1855 (make_pass_stack_ptr_mod): ...new function.
1856 (pass_initialize_regs): Replace declaration with that of...
1857 (make_pass_initialize_regs): ...new function.
1858 (pass_combine): Replace declaration with that of...
1859 (make_pass_combine): ...new function.
1860 (pass_if_after_combine): Replace declaration with that of...
1861 (make_pass_if_after_combine): ...new function.
1862 (pass_ree): Replace declaration with that of...
1863 (make_pass_ree): ...new function.
1864 (pass_partition_blocks): Replace declaration with that of...
1865 (make_pass_partition_blocks): ...new function.
1866 (pass_match_asm_constraints): Replace declaration with that of...
1867 (make_pass_match_asm_constraints): ...new function.
1868 (pass_regmove): Replace declaration with that of...
1869 (make_pass_regmove): ...new function.
1870 (pass_split_all_insns): Replace declaration with that of...
1871 (make_pass_split_all_insns): ...new function.
1872 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1873 (make_pass_fast_rtl_byte_dce): ...new function.
1874 (pass_lower_subreg2): Replace declaration with that of...
1875 (make_pass_lower_subreg2): ...new function.
1876 (pass_mode_switching): Replace declaration with that of...
1877 (make_pass_mode_switching): ...new function.
1878 (pass_sms): Replace declaration with that of...
1879 (make_pass_sms): ...new function.
1880 (pass_sched): Replace declaration with that of...
1881 (make_pass_sched): ...new function.
1882 (pass_ira): Replace declaration with that of...
1883 (make_pass_ira): ...new function.
1884 (pass_reload): Replace declaration with that of...
1885 (make_pass_reload): ...new function.
1886 (pass_clean_state): Replace declaration with that of...
1887 (make_pass_clean_state): ...new function.
1888 (pass_branch_prob): Replace declaration with that of...
1889 (make_pass_branch_prob): ...new function.
1890 (pass_value_profile_transformations): Replace declaration with that
1892 (make_pass_value_profile_transformations): ...new function.
1893 (pass_postreload_cse): Replace declaration with that of...
1894 (make_pass_postreload_cse): ...new function.
1895 (pass_gcse2): Replace declaration with that of...
1896 (make_pass_gcse2): ...new function.
1897 (pass_split_after_reload): Replace declaration with that of...
1898 (make_pass_split_after_reload): ...new function.
1899 (pass_branch_target_load_optimize1): Replace declaration with that
1901 (make_pass_branch_target_load_optimize1): ...new function.
1902 (pass_thread_prologue_and_epilogue): Replace declaration with that
1904 (make_pass_thread_prologue_and_epilogue): ...new function.
1905 (pass_stack_adjustments): Replace declaration with that of...
1906 (make_pass_stack_adjustments): ...new function.
1907 (pass_peephole2): Replace declaration with that of...
1908 (make_pass_peephole2): ...new function.
1909 (pass_if_after_reload): Replace declaration with that of...
1910 (make_pass_if_after_reload): ...new function.
1911 (pass_regrename): Replace declaration with that of...
1912 (make_pass_regrename): ...new function.
1913 (pass_cprop_hardreg): Replace declaration with that of...
1914 (make_pass_cprop_hardreg): ...new function.
1915 (pass_reorder_blocks): Replace declaration with that of...
1916 (make_pass_reorder_blocks): ...new function.
1917 (pass_branch_target_load_optimize2): Replace declaration with that
1919 (make_pass_branch_target_load_optimize2): ...new function.
1920 (pass_leaf_regs): Replace declaration with that of...
1921 (make_pass_leaf_regs): ...new function.
1922 (pass_split_before_sched2): Replace declaration with that of...
1923 (make_pass_split_before_sched2): ...new function.
1924 (pass_compare_elim_after_reload): Replace declaration with that of...
1925 (make_pass_compare_elim_after_reload): ...new function.
1926 (pass_sched2): Replace declaration with that of...
1927 (make_pass_sched2): ...new function.
1928 (pass_stack_regs): Replace declaration with that of...
1929 (make_pass_stack_regs): ...new function.
1930 (pass_stack_regs_run): Replace declaration with that of...
1931 (make_pass_stack_regs_run): ...new function.
1932 (pass_df_finish): Replace declaration with that of...
1933 (make_pass_df_finish): ...new function.
1934 (pass_compute_alignments): Replace declaration with that of...
1935 (make_pass_compute_alignments): ...new function.
1936 (pass_duplicate_computed_gotos): Replace declaration with that of...
1937 (make_pass_duplicate_computed_gotos): ...new function.
1938 (pass_variable_tracking): Replace declaration with that of...
1939 (make_pass_variable_tracking): ...new function.
1940 (pass_free_cfg): Replace declaration with that of...
1941 (make_pass_free_cfg): ...new function.
1942 (pass_machine_reorg): Replace declaration with that of...
1943 (make_pass_machine_reorg): ...new function.
1944 (pass_cleanup_barriers): Replace declaration with that of...
1945 (make_pass_cleanup_barriers): ...new function.
1946 (pass_delay_slots): Replace declaration with that of...
1947 (make_pass_delay_slots): ...new function.
1948 (pass_split_for_shorten_branches): Replace declaration with that of...
1949 (make_pass_split_for_shorten_branches): ...new function.
1950 (pass_split_before_regstack): Replace declaration with that of...
1951 (make_pass_split_before_regstack): ...new function.
1952 (pass_convert_to_eh_region_ranges): Replace declaration with that
1954 (make_pass_convert_to_eh_region_ranges): ...new function.
1955 (pass_shorten_branches): Replace declaration with that of...
1956 (make_pass_shorten_branches): ...new function.
1957 (pass_set_nothrow_function_flags): Replace declaration with that of...
1958 (make_pass_set_nothrow_function_flags): ...new function.
1959 (pass_dwarf2_frame): Replace declaration with that of...
1960 (make_pass_dwarf2_frame): ...new function.
1961 (pass_final): Replace declaration with that of...
1962 (make_pass_final): ...new function.
1963 (pass_rtl_seqabstr): Replace declaration with that of...
1964 (make_pass_rtl_seqabstr): ...new function.
1965 (pass_release_ssa_names): Replace declaration with that of...
1966 (make_pass_release_ssa_names): ...new function.
1967 (pass_early_inline): Replace declaration with that of...
1968 (make_pass_early_inline): ...new function.
1969 (pass_inline_parameters): Replace declaration with that of...
1970 (make_pass_inline_parameters): ...new function.
1971 (pass_update_address_taken): Replace declaration with that of...
1972 (make_pass_update_address_taken): ...new function.
1973 (pass_convert_switch): Replace declaration with that of...
1974 (make_pass_convert_switch): ...new function.
1975 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1976 to a subclass of simple_ipa_opt_pass along with...
1977 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1978 (make_pass_ipa_tree_profile): ...new function.
1979 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1980 subclass of gimple_opt_pass along with...
1981 (pass_data_sra_early): ...new pass_data instance and...
1982 (make_pass_sra_early): ...new function.
1983 (pass_sra): Convert from a global struct to a subclass of
1984 gimple_opt_pass along with...
1985 (pass_data_sra): ...new pass_data instance and...
1986 (make_pass_sra): ...new function.
1987 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1988 gimple_opt_pass along with...
1989 (pass_data_early_ipa_sra): ...new pass_data instance and...
1990 (make_pass_early_ipa_sra): ...new function.
1991 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1992 subclass of gimple_opt_pass along with...
1993 (pass_data_ccp): ...new pass_data instance and...
1994 (make_pass_ccp): ...new function.
1995 (pass_fold_builtins): Convert from a global struct to a subclass of
1996 gimple_opt_pass along with...
1997 (pass_data_fold_builtins): ...new pass_data instance and...
1998 (make_pass_fold_builtins): ...new function.
1999 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2000 subclass of gimple_opt_pass along with...
2001 (pass_data_copy_prop): ...new pass_data instance and...
2002 (make_pass_copy_prop): ...new function.
2003 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2004 global struct to a subclass of gimple_opt_pass along with...
2005 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2006 (make_pass_rename_ssa_copies): ...new function.
2007 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2008 subclass of gimple_opt_pass along with...
2009 (pass_data_dce): ...new pass_data instance and...
2010 (make_pass_dce): ...new function.
2011 (pass_dce_loop): Convert from a global struct to a subclass of
2012 gimple_opt_pass along with...
2013 (pass_data_dce_loop): ...new pass_data instance and...
2014 (make_pass_dce_loop): ...new function.
2015 (pass_cd_dce): Convert from a global struct to a subclass of
2016 gimple_opt_pass along with...
2017 (pass_data_cd_dce): ...new pass_data instance and...
2018 (make_pass_cd_dce): ...new function.
2019 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2020 subclass of gimple_opt_pass along with...
2021 (pass_data_dominator): ...new pass_data instance and...
2022 (make_pass_dominator): ...new function.
2023 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2024 gimple_opt_pass along with...
2025 (pass_data_phi_only_cprop): ...new pass_data instance and...
2026 (make_pass_phi_only_cprop): ...new function.
2027 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2028 subclass of gimple_opt_pass along with...
2029 (pass_data_dse): ...new pass_data instance and...
2030 (make_pass_dse): ...new function.
2031 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2032 a subclass of gimple_opt_pass along with...
2033 (pass_data_forwprop): ...new pass_data instance and...
2034 (make_pass_forwprop): ...new function.
2035 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2036 struct to a subclass of gimple_opt_pass along with...
2037 (pass_data_tree_ifcombine): ...new pass_data instance and...
2038 (make_pass_tree_ifcombine): ...new function.
2039 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2040 subclass of gimple_opt_pass along with...
2041 (pass_data_ch): ...new pass_data instance and...
2042 (make_pass_ch): ...new function.
2043 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2044 subclass of gimple_opt_pass along with...
2045 (pass_data_tree_loop): ...new pass_data instance and...
2046 (make_pass_tree_loop): ...new function.
2047 (pass_tree_loop_init): Convert from a global struct to a subclass of
2048 gimple_opt_pass along with...
2049 (pass_data_tree_loop_init): ...new pass_data instance and...
2050 (make_pass_tree_loop_init): ...new function.
2051 (pass_lim): Convert from a global struct to a subclass of
2052 gimple_opt_pass along with...
2053 (pass_data_lim): ...new pass_data instance and...
2054 (make_pass_lim): ...new function.
2055 (pass_tree_unswitch): Convert from a global struct to a subclass of
2056 gimple_opt_pass along with...
2057 (pass_data_tree_unswitch): ...new pass_data instance and...
2058 (make_pass_tree_unswitch): ...new function.
2059 (pass_predcom): Convert from a global struct to a subclass of
2060 gimple_opt_pass along with...
2061 (pass_data_predcom): ...new pass_data instance and...
2062 (make_pass_predcom): ...new function.
2063 (pass_vectorize): Convert from a global struct to a subclass of
2064 gimple_opt_pass along with...
2065 (pass_data_vectorize): ...new pass_data instance and...
2066 (make_pass_vectorize): ...new function.
2067 (pass_graphite): Convert from a global struct to a subclass of
2068 gimple_opt_pass along with...
2069 (pass_data_graphite): ...new pass_data instance and...
2070 (make_pass_graphite): ...new function.
2071 (pass_graphite_transforms): Convert from a global struct to a subclass
2072 of gimple_opt_pass along with...
2073 (pass_data_graphite_transforms): ...new pass_data instance and...
2074 (make_pass_graphite_transforms): ...new function.
2075 (pass_check_data_deps): Convert from a global struct to a subclass of
2076 gimple_opt_pass along with...
2077 (pass_data_check_data_deps): ...new pass_data instance and...
2078 (make_pass_check_data_deps): ...new function.
2079 (pass_iv_canon): Convert from a global struct to a subclass of
2080 gimple_opt_pass along with...
2081 (pass_data_iv_canon): ...new pass_data instance and...
2082 (make_pass_iv_canon): ...new function.
2083 (pass_scev_cprop): Convert from a global struct to a subclass of
2084 gimple_opt_pass along with...
2085 (pass_data_scev_cprop): ...new pass_data instance and...
2086 (make_pass_scev_cprop): ...new function.
2087 (pass_record_bounds): Convert from a global struct to a subclass of
2088 gimple_opt_pass along with...
2089 (pass_data_record_bounds): ...new pass_data instance and...
2090 (make_pass_record_bounds): ...new function.
2091 (pass_complete_unroll): Convert from a global struct to a subclass of
2092 gimple_opt_pass along with...
2093 (pass_data_complete_unroll): ...new pass_data instance and...
2094 (make_pass_complete_unroll): ...new function.
2095 (pass_complete_unrolli): Convert from a global struct to a subclass of
2096 gimple_opt_pass along with...
2097 (pass_data_complete_unrolli): ...new pass_data instance and...
2098 (make_pass_complete_unrolli): ...new function.
2099 (pass_parallelize_loops): Convert from a global struct to a subclass
2100 of gimple_opt_pass along with...
2101 (pass_data_parallelize_loops): ...new pass_data instance and...
2102 (make_pass_parallelize_loops): ...new function.
2103 (pass_loop_prefetch): Convert from a global struct to a subclass of
2104 gimple_opt_pass along with...
2105 (pass_data_loop_prefetch): ...new pass_data instance and...
2106 (make_pass_loop_prefetch): ...new function.
2107 (pass_iv_optimize): Convert from a global struct to a subclass of
2108 gimple_opt_pass along with...
2109 (pass_data_iv_optimize): ...new pass_data instance and...
2110 (make_pass_iv_optimize): ...new function.
2111 (pass_tree_loop_done): Convert from a global struct to a subclass of
2112 gimple_opt_pass along with...
2113 (pass_data_tree_loop_done): ...new pass_data instance and...
2114 (make_pass_tree_loop_done): ...new function.
2115 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2116 struct to a subclass of gimple_opt_pass along with...
2117 (pass_data_cse_reciprocals): ...new pass_data instance and...
2118 (make_pass_cse_reciprocals): ...new function.
2119 (pass_cse_sincos): Convert from a global struct to a subclass of
2120 gimple_opt_pass along with...
2121 (pass_data_cse_sincos): ...new pass_data instance and...
2122 (make_pass_cse_sincos): ...new function.
2123 (pass_optimize_bswap): Convert from a global struct to a subclass of
2124 gimple_opt_pass along with...
2125 (pass_data_optimize_bswap): ...new pass_data instance and...
2126 (make_pass_optimize_bswap): ...new function.
2127 (pass_optimize_widening_mul): Convert from a global struct to a
2128 subclass of gimple_opt_pass along with...
2129 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2130 (make_pass_optimize_widening_mul): ...new function.
2131 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2132 subclass of gimple_opt_pass along with...
2133 (pass_data_phiopt): ...new pass_data instance and...
2134 (make_pass_phiopt): ...new function.
2135 (pass_cselim): Convert from a global struct to a subclass of
2136 gimple_opt_pass along with...
2137 (pass_data_cselim): ...new pass_data instance and...
2138 (make_pass_cselim): ...new function.
2139 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2140 subclass of gimple_opt_pass along with...
2141 (pass_data_phiprop): ...new pass_data instance and...
2142 (make_pass_phiprop): ...new function.
2143 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2144 subclass of gimple_opt_pass along with...
2145 (pass_data_pre): ...new pass_data instance and...
2146 (make_pass_pre): ...new function.
2147 (pass_fre): Convert from a global struct to a subclass of
2148 gimple_opt_pass along with...
2149 (pass_data_fre): ...new pass_data instance and...
2150 (make_pass_fre): ...new function.
2151 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2152 subclass of gimple_opt_pass along with...
2153 (pass_data_reassoc): ...new pass_data instance and...
2154 (make_pass_reassoc): ...new function.
2155 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2156 subclass of gimple_opt_pass along with...
2157 (pass_data_sink_code): ...new pass_data instance and...
2158 (make_pass_sink_code): ...new function.
2159 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2160 subclass of gimple_opt_pass along with...
2161 (pass_data_strlen): ...new pass_data instance and...
2162 (make_pass_strlen): ...new function.
2163 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2164 struct to a subclass of gimple_opt_pass along with...
2165 (pass_data_build_alias): ...new pass_data instance and...
2166 (make_pass_build_alias): ...new function.
2167 (pass_build_ealias): Convert from a global struct to a subclass of
2168 gimple_opt_pass along with...
2169 (pass_data_build_ealias): ...new pass_data instance and...
2170 (make_pass_build_ealias): ...new function.
2171 (pass_ipa_pta): Convert from a global struct to a subclass of
2172 simple_ipa_opt_pass along with...
2173 (pass_data_ipa_pta): ...new pass_data instance and...
2174 (make_pass_ipa_pta): ...new function.
2175 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2176 subclass of gimple_opt_pass along with...
2177 (pass_data_uncprop): ...new pass_data instance and...
2178 (make_pass_uncprop): ...new function.
2179 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2180 global struct to a subclass of gimple_opt_pass along with...
2181 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2182 (make_pass_late_warn_uninitialized): ...new function.
2183 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2184 to a subclass of gimple_opt_pass along with...
2185 (pass_data_init_datastructures): ...new pass_data instance and...
2186 (make_pass_init_datastructures): ...new function.
2187 (pass_early_warn_uninitialized): Convert from a global struct to a
2188 subclass of gimple_opt_pass along with...
2189 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2190 (make_pass_early_warn_uninitialized): ...new function.
2191 (pass_update_address_taken): Convert from a global struct to a
2192 subclass of gimple_opt_pass along with...
2193 (pass_data_update_address_taken): ...new pass_data instance and...
2194 (make_pass_update_address_taken): ...new function.
2195 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2196 struct to a subclass of gimple_opt_pass along with...
2197 (pass_data_release_ssa_names): ...new pass_data instance and...
2198 (make_pass_release_ssa_names): ...new function.
2199 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2200 subclass of gimple_opt_pass along with...
2201 (pass_data_stdarg): ...new pass_data instance and...
2202 (make_pass_stdarg): ...new function.
2203 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2204 global struct to a subclass of gimple_opt_pass along with...
2205 (pass_data_convert_switch): ...new pass_data instance and...
2206 (make_pass_convert_switch): ...new function.
2207 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2208 to a subclass of gimple_opt_pass along with...
2209 (pass_data_tail_recursion): ...new pass_data instance and...
2210 (make_pass_tail_recursion): ...new function.
2211 (pass_tail_calls): Convert from a global struct to a subclass of
2212 gimple_opt_pass along with...
2213 (pass_data_tail_calls): ...new pass_data instance and...
2214 (make_pass_tail_calls): ...new function.
2215 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2216 struct to a subclass of gimple_opt_pass along with...
2217 (pass_data_lower_vector): ...new pass_data instance and...
2218 (make_pass_lower_vector): ...new function.
2219 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2220 gimple_opt_pass along with...
2221 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2222 (make_pass_lower_vector_ssa): ...new function.
2223 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2224 to a subclass of gimple_opt_pass along with...
2225 (pass_data_slp_vectorize): ...new pass_data instance and...
2226 (make_pass_slp_vectorize): ...new function.
2227 (pass_ipa_increase_alignment): Convert from a global struct to a
2228 subclass of simple_ipa_opt_pass along with...
2229 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2230 (make_pass_ipa_increase_alignment): ...new function.
2231 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2232 gimple_opt_pass along with...
2233 (pass_data_vrp): ...new pass_data instance and...
2234 (make_pass_vrp): ...new function.
2235 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2236 subclass of simple_ipa_opt_pass along with...
2237 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2238 (make_pass_ipa_free_lang_data): ...new function.
2239 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2240 gimple_opt_pass along with...
2241 (pass_data_tsan): ...new pass_data instance and...
2242 (make_pass_tsan): ...new function.
2243 (pass_tsan_O0): Convert from a global struct to a subclass of
2244 gimple_opt_pass along with...
2245 (pass_data_tsan_O0): ...new pass_data instance and...
2246 (make_pass_tsan_O0): ...new function.
2247 * var-tracking.c (pass_variable_tracking): Convert from a global
2248 struct to a subclass of rtl_opt_pass along with...
2249 (pass_data_variable_tracking): ...new pass_data instance and...
2250 (make_pass_variable_tracking): ...new function.
2251 * web.c (pass_web): Convert from a global struct to a subclass of
2252 rtl_opt_pass along with...
2253 (pass_data_web): ...new pass_data instance and...
2254 (make_pass_web): ...new function.
2255 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2256 declaration with that of...
2257 (make_pass_mode_switch_use): ...new function.
2258 (pass_resolve_sw_modes): Replace declaration with that of...
2259 (make_pass_resolve_sw_modes): ...new function.
2260 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2261 from a global struct to a subclass of rtl_opt_pass along with...
2262 (pass_data_mode_switch_use): ...new pass_data instance and...
2263 (make_pass_mode_switch_use): ...new function.
2264 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2265 from a global struct to a subclass of rtl_opt_pass along with...
2266 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2267 (make_pass_resolve_sw_modes): ...new function.
2268 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2269 struct to a subclass of rtl_opt_pass along with...
2270 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2271 (make_pass_insert_vzeroupper): ...new function.
2272 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2273 global struct to a subclass of rtl_opt_pass along with...
2274 (pass_data_work_around_errata): ...new pass_data instance and...
2275 (make_pass_work_around_errata): ...new function.
2276 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2277 struct to a subclass of rtl_opt_pass along with...
2278 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2279 (make_pass_mips_machine_reorg2): ...new function.
2281 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2283 * passes.c (pass_manager::operator new): New.
2285 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2287 Handwritten part of conversion of passes to C++ classes.
2289 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2290 (toplev.o): Add dep on PASS_MANAGER_H.
2291 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2292 of early local pases to reflect this moving from a global to a
2293 member of gcc::pass_manager.
2294 (cgraph_add_new_function): Likewise.
2295 * lto-cgraph.c (lto_output_node): Update for conversion of
2296 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2297 * passes.c (opt_pass::clone): New.
2298 (opt_pass::gate): New.
2299 (opt_pass::execute): New.
2300 (opt_pass::opt_pass): New.
2301 (pass_manager::execute_early_local_passes): New.
2302 (pass_manager::execute_pass_mode_switching): new.
2303 (finish_optimization_passes): Convert to...
2304 (pass_manager::finish_optimization_passes): ...this.
2305 (finish_optimization_passes): Update for conversion of passes to
2307 (register_dump_files_1): Use has_gate since we cannot portably
2308 check a vtable entry against NULL.
2309 (dump_one_pass): Likewise.
2310 (ipa_write_summaries_2): Likewise.
2311 (ipa_write_optimization_summaries_1): Likewise.
2312 (ipa_read_summaries_1): Likewise.
2313 (ipa_read_optimization_summaries_1): Likewise.
2314 (execute_ipa_stmt_fixups): Likewise.
2315 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2316 pass-creation functions rather than wiring up globals, and
2317 storing the results in fields of pass_manager generated using
2319 (pass_manager::dump_profile_report): Update for conversion of
2320 passes to C++ classes.
2321 (pass_manager::execute_ipa_summary_passes): Likewise.
2322 (execute_one_ipa_transform_pass): Likewise.
2323 (execute_one_pass): Use has_gate and has_execute since we cannot
2324 portably check a vtable entry against NULL.
2325 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2326 (pass_manager): Use pass-instances.def to add fields for the
2327 various pass instances.
2328 * toplev.c (finalize): Update for move of
2329 finish_optimization_passes to a method of gcc::pass_manager.
2330 * toplev.h (finish_optimization_passes): Move to method of class
2332 * tree-pass.h (struct pass_data): New.
2333 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2334 (opt_pass::gate): Convert to virtual function.
2335 (opt_pass::~opt_pass): New.
2336 (opt_pass::clone): New.
2337 (opt_pass::execute): Convert to virtual function.
2338 (opt_pass::opt_pass): New.
2339 (opt_pass::ctxt_): new.
2340 (gimple_opt_pass): Convert to subclass of opt_pass.
2341 (gimple_opt_pass::gimple_opt_pass): New.
2342 (rtl_opt_pass): Convert to subclass of opt_pass.
2343 (rtl_opt_pass::rtl_opt_pass): New.
2344 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2345 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2346 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2347 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2348 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2349 invocation of pass_mode_switching to reflect this moving from a
2350 global to a member of gcc::pass_manager.
2351 (ix86_option_override): Rework how pass_insert_vzeroupper is
2352 added to the pass_manager to reflect autogenerated changes.
2353 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2355 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2357 PR rtl-optimization/57708
2358 * recog.c (peep2_find_free_register): Validate all regs in a
2361 2013-08-05 Jan Hubicka <jh@suse.cz>
2364 * cgraph.c (verify_cgraph_node): Accept local flags from other
2366 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2367 (function_and_variable_visibility): Likewise.
2368 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2370 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2372 * graph.c (init_graph_slim_pretty_print): Remove.
2373 (print_graph_cfg): Do not call it. Use local pretty printer.
2374 (start_graph_dump): Likewise.
2376 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2378 * gimple-pretty-print.c (buffer): Remove.
2379 (initialized): Likewise.
2380 (maybe_init_pretty_print): Likewise.
2381 (print_gimple_stmt): Do not call it. Use non-static local
2382 pretty_printer variable.
2383 (print_gimple_expr): Likewise.
2384 (print_gimple_seq): Likewise.
2385 (gimple_dump_bb): Likewise.
2387 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2389 * asan.c (asan_pp): Remove.
2390 (asan_pp_initialized): Likewise.
2391 (asan_pp_initialize): Likewise.
2392 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2393 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2394 (asan_add_global): Likewise.
2396 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2398 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2399 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2400 * pretty-print.c (pp_formatted_text_data): Likewise.
2401 (pp_indent): Rename from pp_base_indent.
2402 (pp_format): Rename from pp_base_format.
2403 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2404 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2405 (pp_flush): Rename from pp_base_flush.
2406 (pp_set_line_maximum_length): Rename from
2407 pp_base_set_line_maximum_length.
2408 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2409 (pp_set_prefix): Rename from pp_base_set_prefix.
2410 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2411 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2412 (pp_append_text): Rename from pp_base_append_text.
2413 (pp_formatted_text): Rename from pp_base_formatted_text.
2414 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2415 (pp_remaining_character_count_for_line): Rename from
2416 pp_base_remaining_character_count_for_line.
2417 (pp_newline): Rename from pp_base_newline.
2418 (pp_character): Rename from pp_base_character.
2419 (pp_string): Rename from pp_base_string.
2420 (pp_maybe_space): Rename from pp_base_maybe_space.
2421 * asan.c (asan_pp_string): Adjust.
2422 (asan_emit_stack_protection): Likewise.
2423 (asan_add_global): Likewise.
2424 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2425 * tree-mudflap.c (mf_varname_tree): Likewise.
2426 * tree-pretty-print.c (pp_tree_identifier): Rename from
2427 pp_base_tree_identifier.
2428 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2429 Declare as function.
2431 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2433 * pretty-print.h (pp_bar_bar): New.
2434 (pp_ampersand_ampersand): Likewise.
2435 (pp_less_equal): Likewise.
2436 (pp_greater_equal): Likewise.
2437 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2438 printer functions instead of pp_string or operators and punctuators.
2439 (dump_gimple_call): Likewise.
2440 (dump_gimple_omp_for): Likewise.
2441 (dump_gimple_transaction): Likewise.
2442 (dump_gimple_phi): Likewise.
2443 (pp_gimple_stmt_1): Likewise.
2444 * sched-vis.c (print_insn): Likewise.
2445 * tree-mudflap.c (mf_varname_tree): Likewise.
2446 * tree-pretty-print.c (dump_block_node): Likewise.
2447 (dump_generic_node): Likewise.
2449 2013-08-02 Jan Hubicka <jh@suse.cz>
2451 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2453 * lto-streamer-out.c (tree_is_indexable): Results decls and
2454 parm decls are not indexable.
2455 (DFS_write_tree_body): Do not follow args and results.
2456 (hash_tree): Likewise.
2457 (output_functions): Rearrange so struct function is needed
2458 only when real body is output; be able to also ouptut abstract
2459 functions; output DECL_ARGUMENTS and DECL_RESULT.
2460 (lto_output): When not in WPA, ale store abstract functions.
2461 (write_symbol): Do not care about RESULT_DECL.
2462 (output_symbol_p): Handle correctly sbtract decls.
2463 * lto-streamer-in.c (input_function): Rearrange so struct
2464 function can be NULL at entry; allow streaming of
2465 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2466 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2467 sanity check during LTO.
2468 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2469 RESULT_DECl and DECL_ARGUMENTS.
2470 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2473 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2475 * pretty-print.h (pp_underscore): New.
2477 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2478 printer functions instead of pp_character.
2479 (dump_binary_rhs): Likewise.
2480 (dump_ternary_rhs): Likewise.
2481 (dump_gimple_call_args): Likewise.
2482 (pp_points_to_solution): Likewise.
2483 (dump_gimple_call): Likewise.
2484 (dump_gimple_switch): Likewise.
2485 (dump_gimple_cond): Likewise.
2486 (dump_gimple_bind): Likewise.
2487 (dump_gimple_try): Likewise.
2488 (dump_gimple_omp_for): Likewise.
2489 (dump_gimple_omp_continue): Likewise.
2490 (dump_gimple_omp_single): Likewise.
2491 (dump_gimple_omp_sections): Likewise.
2492 (dump_gimple_omp_block): Likewise.
2493 (dump_gimple_omp_critical): Likewise.
2494 (dump_gimple_transaction): Likewise.
2495 (dump_gimple_asm): Likewise.
2496 (dump_gimple_phi): Likewise.
2497 (dump_gimple_omp_parallel): Likewise.
2498 (dump_gimple_omp_task): Likewise.
2499 (dump_gimple_omp_atomic_load): Likewise.
2500 (dump_gimple_omp_atomic_store): Likewise.
2501 (dump_gimple_mem_ops): Likewise.
2502 (pp_gimple_stmt_1): Likewise.
2503 (pp_cfg_jump): Likewise.
2504 (dump_implicit_edges): Likewise.
2505 (gimple_dump_bb_for_graph): Likewise.
2506 * graph.c (draw_cfg_node): Likewise.
2507 * langhooks.c (lhd_print_error_function): Likewise.
2508 * sched-vis.c (print_exp): Likewise.
2509 (print_value): Likewise.
2510 (print_pattern): Likewise.
2511 (print_insn): Likewise.
2512 (rtl_dump_bb_for_graph): Likewise.
2513 * tree-pretty-print.c (dump_function_declaration): Likewise.
2514 (dump_array_domain): Likewise.
2515 (dump_omp_clause): Likewise.
2516 (dump_location): Likewise.
2517 (dump_generic_node): Likewise.
2518 (print_struct_decl): Likewise.
2519 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2521 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2523 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2524 candidate table when replacing a candidate statement.
2525 (replace_rhs_if_not_dup): Likewise.
2526 (replace_one_candidate): Likewise.
2528 2013-08-02 Jan Hubicka <jh@suse.cz>
2529 Martin Liska <marxin.liska@gmail.com>
2531 * cgraphunit.c (add_new_function): Fix logic when adding from
2533 (assemble_thunk): Rename to ...
2534 (expand_thunk); .. this one; export; get it working with
2535 general functions; make produced gimple valid.
2536 * cgraph.h (expand_thunk): Declare.
2538 2013-08-02 Jan Hubicka <jh@suse.cz>
2540 * ipa-cp.c (gather_context_independent_values): Use
2541 ipa_get_param_move_cost.
2542 (get_replacement_map): Remove PARAM; move parameter folding
2544 (create_specialized_node): Update.
2545 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2546 assert that we have gimple body; update move_cost.
2547 (count_formal_params): Assert that we have gimple body.
2548 (ipa_dump_param): New function.
2549 (ipa_alloc_node_params): Break out from ...
2550 (ipa_initialize_node_params): ... here.
2551 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2552 (ipa_write_node_info): Stream move costs.
2553 (ipa_read_node_info): Read move costs.
2554 (ipa_update_after_lto_read): Do not recompute node params.
2555 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2556 (ipa_get_param): Check we are not in WPA.
2557 (ipa_get_param_move_cost): New.
2558 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2559 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2560 parm numbers to be present.
2562 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2564 PR rtl-optimization/58048
2565 * lra-constraints.c (process_alt_operands): Don't check asm
2566 operand on register.
2568 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2570 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2571 the implied StoreLoad barrier for atomic operations if before.
2573 2013-08-02 Jan Hubicka <jh@suse.cz>
2574 Martin Liska <marxin.liska@gmail.com>
2576 * cgraph.c (cgraph_function_body_availability): Do not check
2578 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2579 symtab_node_availability): Declare.
2580 * ipa.c (can_replace_by_local_alias): New.
2581 (function_and_variable_visibility): Use it.
2582 * symtab.c (symtab_for_node_and_aliases,
2583 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2585 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2587 PR rtl-optimization/57963
2588 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2589 (lra_constraints): Use them.
2591 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2593 * config/arm/types.md (define_attr "type"): Add "load_acq"
2595 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2597 (cortex_a53_store1): Likewise.
2599 2013-08-01 Jan Hubicka <jh@suse.cz>
2601 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2602 partitions are not needed.
2604 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2606 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2607 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2608 MAYBE_NON_Q_CLASS_P where appropriate.
2610 2013-08-01 Jan Hubicka <jh@suse.cz>
2612 * cgraph.h (release_function_body): Declare.
2613 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2614 of unused delcarations.
2616 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2618 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
2619 RTL form when subtracting a constant.
2621 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2623 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
2624 Generate canonical plus rtx with negated immediate instead of minus
2626 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
2628 2013-08-01 Jan Hubicka <jh@suse.cz>
2630 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
2631 (cgraph_release_function_body): Likewise.
2632 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2633 * cgraph.h (cgrpah_node): Rename abstract_and_needed
2634 to used_as_abstract_origin.
2635 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
2636 symbols used as abstract origins.
2637 * cgraphunit.c (analyze_functions): Update.
2638 * ipa.c (symtab_remove_unreachable_nodes): Recompute
2639 used_as_abstract_origin.
2640 * tree-inline.c (tree_function_versioning): Update
2641 used_as_abstract_origin; be ready for DECL_RESULT and
2642 DECL_ARGUMENTS to be NULL.
2644 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
2645 for abstract functions.
2646 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
2649 2013-08-01 Jan Hubicka <jh@suse.cz>
2651 * profile.c (compute_value_histograms): Fix thinko.
2653 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2655 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
2656 aarch-common-protos.h to extra_headers.
2657 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2658 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
2659 * config/aarch64/t-aarch64 (aarch-common.o): Define.
2661 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2663 * config/aarch64/aarch64.md (define_attr "type"): Delete.
2664 Include "../arm/types.md". Define "type" attribute for all patterns.
2665 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
2668 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2670 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
2671 to support power8 load fusion.
2672 (fusion_gpr_mem_load): Likewise.
2674 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
2676 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
2677 declarations for power8 load fusion.
2678 (emit_fusion_gpr_load): Likewise.
2680 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
2681 tuning for power8, turn on fusion mode by default. Turn on sign
2682 extending fusion mode if normal fusion mode is on, and we are at
2684 (fusion_gpr_load_p): New function, return true if we can fuse an
2685 addis instruction with a dependent load to a GPR.
2686 (emit_fusion_gpr_load): Emit the instructions for power8 load
2689 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
2690 (VSX load fusion peepholes): New peepholes to fuse together an
2691 addi instruction with a VSX load instruction.
2693 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
2694 peepholes to fuse an addis instruction with a load to a GPR base
2695 register. If we are supporting sign extending fusions, convert
2696 sign extending loads to zero extending loads and add an explicit
2699 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2701 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
2702 aarch-common-protos.h to extra_headers.
2703 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2704 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
2705 (arm_early_store_addr_dep): Likewise.
2706 (arm_no_early_alu_shift_dep): Likewise.
2707 (arm_no_early_alu_shift_value_dep): Likewise.
2708 (arm_no_early_mul_dep): Likewise.
2709 (arm_no_early_store_addr_dep): Likewise.
2710 (arm_mac_accumulator_is_mul_result): Likewise.
2711 (arm_mac_accumulator_is_result): Likewise.
2712 * config/arm/aarch-common.c: ... here. New file.
2713 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
2715 (arm_early_store_addr_dep): Likewise.
2716 (arm_no_early_alu_shift_dep): Likewise.
2717 (arm_no_early_alu_shift_value_dep): Likewise.
2718 (arm_no_early_mul_dep): Likewise.
2719 (arm_no_early_store_addr_dep): Likewise.
2720 (arm_mac_accumulator_is_mul_result): Likewise.
2721 (arm_mac_accumulator_is_result): Likewise.
2722 * config/arm/aarch-common-protos.h: ... here. New file.
2723 * config/arm/t-arm (aarch-common.o): Define.
2725 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2727 * config/arm/arm.md: Include new file "types.md".
2728 (define_attr "type"): Move from here to ...
2729 (define_attr "mul32"): Likewise.
2730 (define_attr "mul64"): Likewise.
2731 * config/arm/types.md: ... here. New file.
2733 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
2735 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
2736 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
2738 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2740 * gen-pass-instances.awk: Fix offset of substr().
2742 2013-07-31 David Malcolm <dmalcolm@redhat.com>
2744 * Makefile.in (pass-instances.def): New.
2745 (passes.o): Replace dependency on passes.def with one on
2748 * gen-pass-instances.awk: New.
2750 * passes.c (pass_manager::pass_manager): Use pass-instances.def
2751 rather than passes.def, updating local definition of NEXT_PASS
2752 macro to add an extra NUM parameter (currently unused).
2754 2013-07-30 David Malcolm <dmalcolm@redhat.com>
2756 * Makefile.in (PASS_MANAGER_H): New.
2757 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
2758 (passes.o): Likewise.
2759 (statistics.o): Likewise.
2760 (cgraphunit.o): Likewise.
2761 (context.o): Depend on PASS_MANAGER_H.
2763 * pass_manager.h: New.
2765 * cgraphunit.c (cgraph_add_new_function): Update for moves
2766 of globals to fields of pass_manager.
2767 (analyze_function): Likewise.
2768 (expand_function): Likewise.
2769 (ipa_passes): Likewise.
2770 (compile): Likewise.
2772 * context.c (context::context): New.
2773 * context.h (context::context): New.
2774 (context::get_passes): New.
2775 (context::passes_): New.
2777 * lto-cgraph.c (input_node): Update for moves of globals to
2778 fields of pass_manager.
2780 * passes.c (all_passes): Remove, in favor of a field of the
2781 same name within the new class pass_manager.
2782 (all_small_ipa_passes): Likewise.
2783 (all_lowering_passes): Likewise.
2784 (all_regular_ipa_passes): Likewise.
2785 (all_late_ipa_passes): Likewise.
2786 (all_lto_gen_passes): Likewise.
2787 (passes_by_id): Likewise.
2788 (passes_by_id_size): Likewise.
2789 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2790 the new class pass_manager.
2791 (set_pass_for_id): Convert to...
2792 (pass_manager::set_pass_for_id): ...method.
2793 (get_pass_for_id): Convert to...
2794 (pass_manager::get_pass_for_id): ...method.
2795 (register_one_dump_file): Move body of implementation into...
2796 (pass_manager::register_one_dump_file): ...here.
2797 (register_dump_files_1): Convert to...
2798 (pass_manager::register_dump_files_1): ...method.
2799 (register_dump_files): Convert to...
2800 (pass_manager::register_dump_files): ...method.
2801 (create_pass_tab): Update for moves of globals to fields of
2803 (dump_passes): Move body of implementation into...
2804 (pass_manager::dump_passes): ...here.
2805 (register_pass): Move body of implementation into...
2806 (pass_manager::register_pass): ...here.
2807 (init_optimization_passes): Convert into...
2808 (pass_manager::pass_manager): ...constructor for new
2809 pass_manager class, and initialize the pass_lists array.
2810 (check_profile_consistency): Update for moves of globals to
2811 fields of pass_manager.
2812 (dump_profile_report): Move body of implementation into...
2813 (pass_manager::dump_profile_report): ...here.
2814 (ipa_write_summaries_1): Update for moves of pass lists from
2815 being globals to fields of pass_manager.
2816 (ipa_write_optimization_summaries): Likewise.
2817 (ipa_read_summaries): Likewise.
2818 (ipa_read_optimization_summaries): Likewise.
2819 (execute_all_ipa_stmt_fixups): Likewise.
2821 * statistics.c (statistics_fini): Update for moves of globals to
2822 fields of pass_manager.
2824 * toplev.c (general_init): Replace call to
2825 init_optimization_passes with construction of the pass_manager
2828 * tree-pass.h (all_passes): Remove, in favor of a field of the
2829 same name within the new class pass_manager.
2830 (all_small_ipa_passes): Likewise.
2831 (all_lowering_passes): Likewise.
2832 (all_regular_ipa_passes): Likewise.
2833 (all_lto_gen_passes): Likewise.
2834 (all_late_ipa_passes): Likewise.
2835 (passes_by_id): Likewise.
2836 (passes_by_id_size): Likewise.
2837 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2838 the new class pass_manager.
2839 (get_pass_for_id): Remove.
2841 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2843 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2846 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2848 * arm.md (mulhi3): New expand pattern.
2850 2013-07-30 Jan Hubicka <jh@suse.cz>
2851 Martin Liska <marxin.liska@gmail.com>
2853 * profile.c (compute_value_histograms): Do not ICE when
2854 there is mismatch only on some counters.
2856 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2858 PR rtl-optimization/57637
2859 * function.c (move_insn_for_shrink_wrap): Also check the
2860 GEN set of the LIVE problem for the liveness analysis
2861 if it exists, otherwise give up.
2863 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2865 PR tree-optimization/57993
2866 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2867 replaced statement in the candidate table.
2868 (phi_add_costs): Return infinite cost when the hidden basis does
2869 not dominate all phis on which the candidate is dependent.
2870 (replace_one_candidate): Record replaced statement in the
2873 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2875 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2876 (ashlv2si3): New expander.
2877 (*ashlv2si3_i): New define_insn_and_split.
2878 * predicates.md (float_operation): Allow patterns with three
2881 PR rtl-optimization/58021
2882 * mode-switching.c (create_pre_exit): Always split off preceding
2883 insns if we are not at the basic block head.
2885 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2887 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2888 (UCLIBC_DYNAMIC_LINKER): New macro.
2889 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2891 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2892 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2894 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2895 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2896 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2897 for SF and DF modes. Use ieee_quad_format for TF mode.
2898 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2899 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2900 (mips_option_override): Handle `-mnan=legacy'.
2901 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2902 `-mabs=2008' and `-mnan=2008'.
2903 (OPTION_DEFAULT_SPECS): Add "nan" default.
2904 (ASM_SPEC): Handle `-mnan='.
2905 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2906 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2907 comment accordingly.
2908 (neg<mode>2): Likewise.
2909 * config/mips/mips.opt (mabs, mnan): New options.
2910 * doc/install.texi (Configuration): Document `--with-nan=' option.
2911 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2913 (MIPS Options): Document them.
2914 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2915 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2916 * configure: Regenerate.
2917 * config.in: Regenerate.
2919 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2921 * config/i386/i386.md (float post-reload splitters): Do not check
2922 for subregs of SSE registers.
2924 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2925 H.J. Lu <hongjiu.lu@intel.com>
2929 * config/i386/i386.md (post-reload splitter
2930 to avoid partial SSE reg dependency stalls): New pattern.
2932 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2934 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2936 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2938 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2939 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2941 2013-07-26 Cary Coutant <ccoutant@google.com>
2943 * dwarf2out.c (die_checksum_ordered): Don't include template
2944 instantiations in signature.
2945 (is_template_parameter): New function.
2946 (is_template_instantiation): New function.
2947 (generate_skeleton_bottom_up): Don't include template instantiations
2949 (generate_skeleton): Likewise.
2950 (break_out_comdat_types): Move recursive call to break out nested
2952 (prune_unused_types_mark_generic_parms_dies): Call
2953 is_template_parameter.
2955 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2957 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2958 uses vector registers.
2959 * config/aarch64/iterators.md: Add attributes rtn and vas.
2961 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2962 Richard Earnshaw <richard.earnshaw@arm.com>
2964 * combine.c (simplify_comparison): Re-canonicalize operands
2966 * config/arm/arm.md (movcond_addsi): New splitter.
2968 2013-07-25 Sterling Augustine <saugustine@google.com>
2970 * dwarf2out.c (size_of_pubnames): Move code to...
2971 (include_pubname_in_output): ...here. New.
2972 (want_pubnames): Rearrange.
2973 (output_pubnames): Call include_pubname_in_output. Move assertion.
2975 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2977 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2979 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2982 * doc/extend.texi: Remove obsolete builtins. Fix
2983 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2985 2013-07-25 Jan Hubicka <jh@suse.cz>
2987 * cgraph.c (release_function_body): Break out from ...
2988 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2990 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2991 old_tree in the map.
2992 (create_specialized_node): Update.
2993 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2995 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2996 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2997 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2998 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3001 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3003 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3004 addsi3_carryin_alt2_<optab>): Correct output template.
3006 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3008 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3009 Adjust for arm_restrict_it.
3010 Remove trailing whitespace.
3012 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3014 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3015 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3017 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3019 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3021 PR rtl-optimization/57960
3022 * lra-constraints.c (process_alt_operands): Use the right mode
3023 when checking strict_low.
3025 2013-07-25 Jan Hubicka <jh@suse.cz>
3027 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3028 * cgraph.c (cgraph_remove_node): Do not release function body
3029 when in cgraph streaming.
3030 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3031 in other partitions are not considered reachable; fix handling of
3034 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3036 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3038 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3043 * config/arm/arm.md ("*sibcall_insn): Replace use of
3044 Ss with US. Adjust output for v5 and v4t.
3045 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3047 * config/arm/constraints.md ("Ss"): Rename to US.
3049 2013-07-25 Terry Guo <terry.guo@arm.com>
3051 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3052 shift_add/shift_sub0/shift_sub1 RTXs.
3054 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3055 Anton Blanchard <anton@au1.ibm.com>
3057 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3058 (altivec_vpks<VI_char>ss): Likewise.
3059 (altivec_vpks<VI_char>us): Likewise.
3060 (altivec_vpku<VI_char>us): Likewise.
3061 (altivec_vpku<VI_char>um): Likewise.
3063 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3065 Introduce context class.
3067 * Makefile.in (CONTEXT_H): New.
3068 (OBJS): Add context.o.
3069 (toplev.o): Add CONTEXT_H to dependencies.
3072 * toplev.c (general_init): Create the singleton gcc::context instance.
3078 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3080 PR rtl-optimization/57968
3081 * mode-switching.c (create_pre_exit): Allow instructions that
3082 don't set a return register to need a non-exit mode.
3084 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3085 Anton Blanchard <anton@au1.ibm.com>
3087 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3088 operands to vperm for little endian.
3089 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3090 of lvsl to create the control mask for a vperm for little endian.
3092 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3093 Anton Blanchard <anton@au1.ibm.com>
3095 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3096 two operands for little-endian.
3098 2013-07-23 Steve Ellcey <sellcey@mips.com>
3100 * config/mips/mips.c (mips_case_values_threshold): New.
3101 (TARGET_CASE_VALUES_THRESHOLD): Define.
3103 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3104 Anton Blanchard <anton@au1.ibm.com>
3106 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3107 selection of field for vector splat in little endian mode.
3109 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3111 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3112 expanders to rs6000.md.
3113 (ior<mode>3): Likewise.
3114 (and<mode>3): Likewise.
3115 (one_cmpl<mode>2): Likewise.
3116 (nor<mode>3): Likewise.
3117 (andc<mode>3): Likewise.
3118 (eqv<mode>3): Likewise.
3119 (nand<mode>3): Likewise.
3120 (orc<mode>3): Likewise.
3122 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3125 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3126 to split multi-word logical operations.
3127 (rs6000_split_logical_di): Likewise.
3128 (rs6000_split_logical): Likewise.
3130 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3131 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3132 and allow TImode operations in 32-bit.
3133 (vsx_and<mode>3_64bit): Likewise.
3134 (vsx_ior<mode>3_32bit): Likewise.
3135 (vsx_ior<mode>3_64bit): Likewise.
3136 (vsx_xor<mode>3_32bit): Likewise.
3137 (vsx_xor<mode>3_64bit): Likewise.
3138 (vsx_one_cmpl<mode>2_32bit): Likewise.
3139 (vsx_one_cmpl<mode>2_64bit): Likewise.
3140 (vsx_nor<mode>3_32bit): Likewise.
3141 (vsx_nor<mode>3_64bit): Likewise.
3142 (vsx_andc<mode>3_32bit): Likewise.
3143 (vsx_andc<mode>3_64bit): Likewise.
3144 (vsx_eqv<mode>3_32bit): Likewise.
3145 (vsx_eqv<mode>3_64bit): Likewise.
3146 (vsx_nand<mode>3_32bit): Likewise.
3147 (vsx_nand<mode>3_64bit): Likewise.
3148 (vsx_orc<mode>3_32bit): Likewise.
3149 (vsx_orc<mode>3_64bit): Likewise.
3151 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3152 logical types in GPRs.
3154 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3155 logical insns to rs6000.md, and allow TImode operations in
3157 (altivec_ior<mode>3): Likewise.
3158 (altivec_xor<mode>3): Likewise.
3159 (altivec_one_cmpl<mode>2): Likewise.
3160 (altivec_nor<mode>3): Likewise.
3161 (altivec_andc<mode>3): Likewise.
3163 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3164 attributes for moving the 128-bit logical operations into
3166 (BOOL_REGS_OUTPUT): Likewise.
3167 (BOOL_REGS_OP1): Likewise.
3168 (BOOL_REGS_OP2): Likewise.
3169 (BOOL_REGS_UNARY): Likewise.
3170 (BOOL_REGS_AND_CR0): Likewise.
3171 (one_cmpl<mode>2): Add support for DI logical operations on
3172 32-bit, splitting the operations to 32-bit.
3176 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3177 changes to combine the 32/64-bit code, allow logical operations on
3178 TI mode in 32-bit, and to use similar match_operator patterns like
3179 scalar mode uses. Combine the Altivec and VSX code for logical
3180 operations, and move it here.
3181 (ior<mode>3, 128-bit types): Likewise.
3182 (xor<mode>3, 128-bit types): Likewise.
3183 (one_cmpl<mode>3, 128-bit types): Likewise.
3184 (nor<mode>3, 128-bit types): Likewise.
3185 (andc<mode>3, 128-bit types): Likewise.
3186 (eqv<mode>3, 128-bit types): Likewise.
3187 (nand<mode>3, 128-bit types): Likewise.
3188 (orc<mode>3, 128-bit types): Likewise.
3189 (and<mode>3_internal): Likewise.
3190 (bool<mode>3_internal): Likewise.
3191 (boolc<mode>3_internal1): Likewise.
3192 (boolc<mode>3_internal2): Likewise.
3193 (boolcc<mode>3_internal1): Likewise.
3194 (boolcc<mode>3_internal2): Likewise.
3195 (eqv<mode>3_internal1): Likewise.
3196 (eqv<mode>3_internal2): Likewise.
3197 (one_cmpl1<mode>3_internal): Likewise.
3199 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3201 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3202 Rename flag_stack_usage to flag_stack_usage_info.
3204 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3206 * config/microblaze/sync.md: New file.
3207 * config/microblaze/microblaze.md: Include sync.md
3208 * config/microblaze/microblaze.c: Add print_operand 'y'.
3209 * config/microblaze/constraints.md: Add memory_contraint
3210 'Q' which is a single register.
3212 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3214 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3216 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3218 * reload.c (find_reloads): Exit loop once we find this operand
3219 cannot be reloaded somehow for this alternative.
3221 * reload.c (find_reloads): Exit loop once we find a hard register.
3223 * rtlanal.c (computed_jump_p): Exit loop once we find label
3226 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3228 * cfgloopmanip.c (remove_path): Exit loop after setting
3231 * gensupport.c (subst_dup): Avoid loop if code is not
3232 MATCH_DUP nor MATCH_OP_DUP.
3234 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3236 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3238 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3240 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3241 true for SP_REGNUM if mode == ptr_mode.
3242 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3243 with value R0_REGNUM + 31.
3245 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3247 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3248 pad pointer-typed argument downward.
3250 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3252 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3253 and __ILP32__ when the ILP32 model is in use.
3255 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3257 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3258 (aarch64_load_symref_appropriately): In the case of
3259 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3260 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3261 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3262 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3263 if 'mode' doesn't equal to 'ptr_mode'.
3264 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3265 'vcall_offset'; change to call aarch64_emit_move differently depending
3266 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3267 to calculate the upper bound of 'vcall_offset'.
3268 (aarch64_cannot_force_const_mem): Change to also return true if
3270 (aarch64_legitimize_reload_address): In the case of large
3271 displacements, add new local variable 'xmode' and an assertion
3272 based on it; change to use 'xmode' to generate the new rtx and
3274 (aarch64_asm_trampoline_template): Change to generate the template
3275 differently depending on TARGET_ILP32 or not; change to use
3276 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3277 (aarch64_trampoline_size): Removed.
3278 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3279 and replace immediate literals with it. Change to use 'ptr_mode'
3280 instead of 'DImode' and call convert_memory_address if the mode
3281 of 'fnaddr' doesn't equal to 'ptr_mode'.
3282 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3284 (aarch64_elf_asm_destructor): Likewise.
3285 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3286 on TARGET_ILP32 instead of aarch64_trampoline_size.
3287 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3288 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3289 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3290 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3291 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3292 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3293 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3294 depending on the value of 'mode'.
3295 (add_losym_<mode>): New.
3296 (ldr_got_small_<mode>): New, based on ldr_got_small.
3297 (ldr_got_small): Remove.
3298 (ldr_got_small_sidi): New.
3299 * config/aarch64/iterators.md (P): New.
3300 (PTR): Change to 'ptr_mode' in the condition.
3302 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3304 * config.gcc (aarch64*-*-*): Support --with-abi.
3305 (aarch64*-*-elf): Support --with-multilib-list.
3306 (aarch64*-*-linux*): Likewise.
3307 (supported_defaults): Add abi to aarch64*-*-*.
3308 * configure.ac: Mention AArch64 for --with-multilib-list.
3309 * configure: Re-generated.
3310 * config/aarch64/biarchilp32.h: New file.
3311 * config/aarch64/biarchlp64.h: New file.
3312 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3314 (MULTILIB_DEFAULTS): Ditto.
3315 (DRIVER_SELF_SPECS): Ditto.
3316 (ASM_SPEC): Update to also substitute -mabi.
3317 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3318 file whose name depends on -mabi= and -mbig-endian.
3319 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3321 (POINTER_SIZE): New define.
3322 (POINTERS_EXTEND_UNSIGNED): Ditto.
3323 (enum aarch64_abi_type): New enumeration tag.
3324 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3325 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3326 (TARGET_ILP32): New define.
3327 * config/aarch64/aarch64.opt (mabi): New.
3329 (ilp32, lp64): New values for -mabi.
3330 * config/aarch64/t-aarch64 (comma): New define.
3331 (MULTILIB_OPTIONS): Ditto.
3332 (MULTILIB_DIRNAMES): Ditto.
3333 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3334 * doc/invoke.texi: Document -mabi for AArch64.
3336 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3338 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3340 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3341 Anton Blanchard <anton@au1.ibm.com>
3343 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3344 endianness when selecting field to splat.
3346 2013-07-22 Eric Christopher <echristo@gmail.com>
3348 * dwarf2out.c (die_odr_checksum): New function to use
3349 CHECKSUM_ macros and ULEB128 for DIE tag.
3350 (generate_type_signature): Use.
3352 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3354 * config.gcc (sparc*-*-*): Accept leon3 processor.
3355 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3356 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3357 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3358 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3359 (mfix-ut699): Adjust comment.
3360 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3361 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3362 (CPP_CPU_SPEC): Likewise.
3363 (ASM_CPU_SPEC): Likewise.
3364 * config/sparc/sparc.c (leon3_cost): New constant.
3365 (sparc_option_override): Add leon3 support.
3366 (mem_ref): New function.
3367 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3368 (sparc_do_work_around_errata): Look into the instruction in the delay
3369 slot and adjust accordingly. Add fix for the data cache nullify issues
3370 of the UT699. Change insertion position for the NOP.
3371 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3372 (leon3_load): New reservation.
3373 (leon_store): Bump latency to 2.
3374 (grfpu): New automaton.
3375 (grfpu_alu): New unit.
3376 (grfpu_ds): Likewise.
3377 (leon_fp_alu): Adjust.
3378 (leon_fp_mult): Delete.
3379 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3380 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3381 * config/sparc/sparc.md (cpu): Add leon3.
3382 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3384 (atomic_test_and_set): Likewise.
3387 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3389 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3390 default for R5900 targets.
3391 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3392 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3393 * config/mips/mips.c (mips_option_override): Report an error for
3394 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3395 for -march=r5900 -mhard-float.
3397 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3399 * df-problems.c (can_move_insns_across): Exit loop once we
3400 find a non-fixed, non-global register.
3402 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3405 * omega.c (omega_eliminate_red): Break after setting red_found.
3406 (omega_problem_has_red_equations): Similarly after setting found.
3407 (omega_query_variable): Similarly after setting coupled.
3409 2013-07-22 Marek Polacek <polacek@redhat.com>
3411 * gimplify.c: Don't include gimple.h twice.
3413 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3415 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3416 instead of TARGET_THUMB1.
3417 (Pz): New constraint.
3418 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3420 (compare_negsi_si): Likewise.
3421 (compare_addsi2_op0): Likewise.
3422 (compare_addsi2_op1): Likewise.
3423 (addsi3_carryin_<optab>): Likewise.
3424 (addsi3_carryin_alt2_<optab>): Likewise.
3425 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3426 for arm_restrict_it.
3427 (subsi3_carryin): Likewise.
3428 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3429 (minmax_arithsi): Disable for arm_restrict_it.
3430 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3431 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3432 (satsi_<SAT:code>_shift): Likewise.
3433 (arm_shiftsi3): Add alternative for 16-bit encoding.
3434 (arm32_movhf): Disable for arm_restrict_it.
3435 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3436 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3438 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3440 * config/arm/arm.md (attribute "insn"): Delete.
3441 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3442 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3443 (not_shiftsi): Update for attribute change.
3444 (not_shiftsi_compare0): Likewise.
3445 (not_shiftsi_compare0_scratch): Likewise.
3446 (arm_one_cmplsi2): Likewise.
3447 (thumb1_one_cmplsi2): Likewise.
3448 (notsi_compare0): Likewise.
3449 (notsi_compare0_scratch): Likewise.
3450 (thumb1_movdi_insn): Likewise.
3451 (arm_movsi_insn): Likewise.
3452 (movhi_insn_arch4): Likewise.
3453 (movhi_bytes): Likewise.
3454 (arm_movqi_insn): Likewise.
3455 (thumb1_movqi_insn): Likewise.
3456 (arm32_movhf): Likewise.
3457 (thumb1_movhf): Likewise.
3458 (arm_movsf_soft_insn): Likewise.
3459 (thumb1_movsf_insn): Likewise.
3460 (thumb_movdf_insn): Likewise.
3461 (movsicc_insn): Likewise.
3462 (movsfcc_soft_insn): Likewise.
3463 (and_scc): Likewise.
3464 (cond_move): Likewise.
3465 (if_move_not): Likewise.
3466 (if_not_move): Likewise.
3467 (if_shift_move): Likewise.
3468 (if_move_shift): Likewise.
3469 (if_shift_shift): Likewise.
3470 (if_not_arith): Likewise.
3471 (if_arith_not): Likewise.
3472 (cond_move_not): Likewise.
3473 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3474 (neon_mov<mode>): Likewise.
3475 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3476 (thumb2_movsi_vfp): Likewise.
3477 (movsf_vfp): Likewise.
3478 (thumb2_movsf_vfp): Likewise.
3479 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3481 (cortexa7_older_only): Likewise.
3482 (cortexa7_younger): Likewise.
3483 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3484 (1020alu_shift_op): Likewise.
3485 (1020alu_shift_reg_op): Likewise.
3486 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3487 (alu_shift_op): Likewise.
3488 (alu_shift_reg_op): Likewise.
3489 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3490 (11_alu_shift_op): Likewise.
3491 (11_alu_shift_reg_op): Likewise.
3492 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3493 (9_alu_shift_reg_op): Likewise.
3494 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3496 (cortex_a15_alu_shift): Likewise.
3497 (cortex_a15_alu_shift_reg): Likewise.
3498 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3499 (cortex_a5_alu_shift): Likewise.
3500 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3502 (cortex_a53_alu_shift): Likewise.
3503 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3505 (cortex_a7_alu_reg): Likewise.
3506 (cortex_a7_alu_shift): Likewise.
3507 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3508 (cortex_a8_alu_shift): Likewise.
3509 (cortex_a8_alu_shift_reg): Likewise.
3510 (cortex_a8_mov): Likewise.
3511 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3512 (cortex_a9_dp_shift): Likewise.
3513 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3514 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3515 (cortex_r4_mov): Likewise.
3516 (cortex_r4_alu_shift): Likewise.
3517 (cortex_r4_alu_shift_reg): Likewise.
3518 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3519 (526_alu_shift_op): Likewise.
3520 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3521 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3522 (626te_alu_shift_op): Likewise.
3523 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3524 (726te_alu_op): Likewise.
3525 (726te_alu_shift_op): Likewise.
3526 (726te_alu_shift_reg_op): Likewise.
3527 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3528 (mp626_alu_shift_op): Likewise.
3529 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3530 (pj4_alu_e1_conds): Likewise.
3531 (pj4_alu): Likewise.
3532 (pj4_alu_conds): Likewise.
3533 (pj4_shift): Likewise.
3534 (pj4_shift_conds): Likewise.
3535 (pj4_alu_shift): Likewise.
3536 (pj4_alu_shift_conds): Likewise.
3538 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3540 * config/arm/predicates.md (shiftable_operator_strict_it):
3542 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3543 Disable cond_exec version for arm_restrict_it.
3544 (thumb2_smaxsi3): Convert to generate cond_exec.
3545 (thumb2_sminsi3): Likewise.
3546 (thumb32_umaxsi3): Likewise.
3547 (thumb2_uminsi3): Likewise.
3548 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3549 (thumb2_neg_abssi2): Likewise.
3550 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3551 (thumb2_movsicc_insn): Adjust alternatives.
3552 (thumb2_mov_negscc): Disable for arm_restrict_it.
3553 (thumb2_mov_negscc_strict_it): New pattern.
3554 (thumb2_mov_notscc_strict_it): New pattern.
3555 (thumb2_mov_notscc): Disable for arm_restrict_it.
3556 (thumb2_ior_scc): Likewise.
3557 (thumb2_ior_scc_strict_it): New pattern.
3558 (thumb2_cond_move): Adjust for arm_restrict_it.
3559 (thumb2_cond_arith): Disable for arm_restrict_it.
3560 (thumb2_cond_arith_strict_it): New pattern.
3561 (thumb2_cond_sub): Adjust for arm_restrict_it.
3562 (thumb2_movcond): Likewise.
3563 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3564 (thumb2_zero_extendhisi2_v6): Likewise.
3565 (thumb2_zero_extendqisi2_v6): Likewise.
3566 (orsi_notsi_si): Likewise.
3567 (orsi_not_shiftsi_si): Likewise.
3569 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3571 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3572 instruction sequence is 1 byte shorter.
3574 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3576 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3577 it is not needed after split.
3579 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3582 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3583 second label for nonlocal goto receivers. Don't output pic base labels
3584 unless we're producing PIC; mark that action unreachable().
3585 (ix86_save_reg): If the function contains a nonlocal label, save the
3587 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3588 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3589 (update_pic_label_number_if_needed): New.
3590 (machopic_output_function_base_name): Adjust for nonlocal receiver
3592 (machopic_should_output_picbase_label): New.
3593 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3594 (nonlocal_goto_receiver): New insn and split.
3596 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3598 * config/aarch64/aarch64-builtins.c
3599 (aarch64_fold_builtin): Fold abs in all modes.
3600 * config/aarch64/aarch64-simd-builtins.def
3601 (abs): Enable for all modes.
3602 * config/aarch64/arm_neon.h
3603 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3604 (vabs_f64): Add missing intrinsic.
3606 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3608 * config/aarch64/arm_neon.h (vabs_s64): New function
3610 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3613 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3614 * config/avr/avr.md (adjust_len): Add `round'.
3615 * config/avr/avr-protos.h (avr_out_round): New prototype.
3616 (avr_out_plus): Add `out_label' argument.
3617 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
3618 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
3619 Handle the case where `insn' is just a pattern.
3620 (avr_out_bitop): Handle the case where `insn' is just a pattern.
3621 (avr_out_round): New function.
3622 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
3624 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
3626 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3627 Add check for flag_stack_usage to handle -fstack-usage support
3629 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
3631 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
3632 interaction for new Power8 flags and VSX.
3634 2013-07-18 Sriraman Tallam <tmsriram@google.com>
3637 * tree-inline.c (expand_call_inline): Emit errors during
3638 early_inlining only if optimization is not turned on.
3640 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3644 * passes.c (init_optimization_passes): Move the construction of
3645 the pass hierarchy into a new passes.def file.
3647 * Makefile.in (passes.o): Add dependency on passes.def.
3649 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3651 * passes.c (init_optimization_passes): Introduce macros for
3652 constructing the tree of passes (INSERT_PASSES_AFTER,
3653 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
3654 TERMINATE_PASS_LIST).
3656 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
3657 Wei Mi <wmi@google.com>
3659 PR rtl-optimization/57878
3660 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
3662 (reload_pseudo_compare_func): Check nregs first for reload
3665 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3667 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
3669 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
3671 * read-rtl.c (validate_const_int): Once an invalid character is
3672 seen, quit the loop.
3674 * gengtype.c (write_roots): Similarly once we find the "deletable"
3675 or "if_marked" option.
3677 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3679 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
3680 "xtab" and "sat". Move value "clz" from here to ...
3681 (attriubte "type"): ... here.
3682 (satsi_<SAT:code>): Delete "insn" attribute.
3683 (satsi_<SAT:code>_shift): Likewise.
3684 (arm_zero_extendqisi2addsi): Likewise.
3685 (arm_extendqisi2addsi): Likewise.
3686 (clzsi2): Update for attribute changes.
3687 (rbitsi2): Likewise.
3688 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
3690 (arm_usatsihi): Likewise.
3691 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3693 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3695 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
3696 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
3697 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
3698 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
3699 in alphabetical order.
3700 (attribute "core_cycles"): Update for attribute changes.
3701 (arm_addsi3): Likewise.
3702 (addsi3_compare0): Likewise.
3703 (addsi3_compare0_scratch): Likewise.
3704 (addsi3_compare_op1): Likewise.
3705 (addsi3_compare_op2): Likewise.
3706 (compare_addsi2_op0): Likewise.
3707 (compare_addsi2_op1): Likewise.
3708 (addsi3_carryin_shift_<optab>): Likewise.
3709 (subsi3_carryin_shift): Likewise.
3710 (rsbsi3_carryin_shift): Likewise.
3711 (arm_subsi3_insn): Likewise.
3712 (subsi3_compare0): Likewise.
3713 (subsi3_compare): Likewise.
3714 (arm_andsi3_insn): Likewise.
3715 (thumb1_andsi3_insn): Likewise.
3716 (andsi3_compare0): Likewise.
3717 (andsi3_compare0_scratch): Likewise.
3718 (zeroextractsi_compare0_scratch
3719 (andsi_not_shiftsi_si): Likewise.
3720 (iorsi3_insn): Likewise.
3721 (iorsi3_compare0): Likewise.
3722 (iorsi3_compare0_scratch): Likewise.
3723 (arm_xorsi3): Likewise.
3724 (thumb1_xorsi3_insn): Likewise.
3725 (xorsi3_compare0): Likewise.
3726 (xorsi3_compare0_scratch): Likewise.
3727 (satsi_<SAT:code>_shift): Likewise.
3729 (arm_shiftsi3): Likewise.
3730 (shiftsi3_compare0): Likewise.
3731 (not_shiftsi): Likewise.
3732 (not_shiftsi_compare0): Likewise.
3733 (not_shiftsi_compare0_scratch): Likewise.
3734 (arm_one_cmplsi2): Likewise.
3735 (thumb_one_complsi2): Likewise.
3736 (notsi_compare0): Likewise.
3737 (notsi_compare0_scratch): Likewise.
3738 (thumb1_zero_extendhisi2): Likewise.
3739 (arm_zero_extendhisi2): Likewise.
3740 (arm_zero_extendhisi2_v6): Likewise.
3741 (arm_zero_extendhisi2addsi): Likewise.
3742 (thumb1_zero_extendqisi2): Likewise.
3743 (thumb1_zero_extendqisi2_v6): Likewise.
3744 (arm_zero_extendqisi2): Likewise.
3745 (arm_zero_extendqisi2_v6): Likewise.
3746 (arm_zero_extendqisi2addsi): Likewise.
3747 (thumb1_extendhisi2): Likewise.
3748 (arm_extendhisi2): Likewise.
3749 (arm_extendhisi2_v6): Likewise.
3750 (arm_extendqisi): Likewise.
3751 (arm_extendqisi_v6): Likewise.
3752 (arm_extendqisi2addsi): Likewise.
3753 (thumb1_extendqisi2): Likewise.
3754 (thumb1_movdi_insn): Likewise.
3755 (arm_movsi_insn): Likewise.
3756 (movsi_compare0): Likewise.
3757 (movhi_insn_arch4): Likewise.
3758 (movhi_bytes): Likewise.
3759 (arm_movqi_insn): Likewise.
3760 (thumb1_movqi_insn): Likewise.
3761 (arm32_movhf): Likewise.
3762 (thumb1_movhf): Likewise.
3763 (arm_movsf_soft_insn): Likewise.
3764 (thumb1_movsf_insn): Likewise.
3765 (movdf_soft_insn): Likewise.
3766 (thumb_movdf_insn): Likewise.
3767 (arm_cmpsi_insn): Likewise.
3768 (cmpsi_shiftsi): Likewise.
3769 (cmpsi_shiftsi_swp): Likewise.
3770 (arm_cmpsi_negshiftsi_si): Likewise.
3771 (movsicc_insn): Likewise.
3772 (movsfcc_soft_insn): Likewise.
3773 (arith_shiftsi): Likewise.
3774 (arith_shiftsi_compare0
3775 (arith_shiftsi_compare0_scratch
3776 (sub_shiftsi): Likewise.
3777 (sub_shiftsi_compare0
3778 (sub_shiftsi_compare0_scratch
3779 (and_scc): Likewise.
3780 (cond_move): Likewise.
3781 (if_plus_move): Likewise.
3782 (if_move_plus): Likewise.
3783 (if_move_not): Likewise.
3784 (if_not_move): Likewise.
3785 (if_shift_move): Likewise.
3786 (if_move_shift): Likewise.
3787 (if_shift_shift): Likewise.
3788 (if_not_arith): Likewise.
3789 (if_arith_not): Likewise.
3790 (cond_move_not): Likewise.
3791 (thumb1_ashlsi3): Set type attribute.
3792 (thumb1_ashrsi3): Likewise.
3793 (thumb1_lshrsi3): Likewise.
3794 (thumb1_rotrsi3): Likewise.
3795 (shiftsi3_compare0_scratch): Likewise.
3796 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3797 (neon_mov<mode>): Likewise.
3798 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3800 (thumb2_movsi_insn): Likewise.
3801 (thumb2_cmpsi_neg_shiftsi): Likewise.
3802 (thumb2_extendqisi_v6): Likewise.
3803 (thumb2_zero_extendhisi2_v6): Likewise.
3804 (thumb2_zero_extendqisi2_v6): Likewise.
3805 (thumb2_shiftsi3_short): Likewise.
3806 (thumb2_addsi3_compare0_scratch): Likewise.
3807 (orsi_not_shiftsi_si): Likewise.
3808 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3809 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3811 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3812 (1020alu_shift_op): Likewise.
3813 (1020alu_shift_reg_op): Likewise.
3814 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3815 (alu_shift_op): Likewise.
3816 (alu_shift_reg_op): Likewise.
3817 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3818 (11_alu_shift_op): Likewise.
3819 (11_alu_shift_reg_op): Likewise.
3820 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3821 (9_alu_shift_reg_op): Likewise.
3822 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3824 (cortex_a15_alu_shift): Likewise.
3825 (cortex_a15_alu_shift_reg): Likewise.
3826 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3828 (cortex_a5_alu_shift): Likewise.
3829 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3831 (cortex_a53_alu_shift): Likewise.
3832 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3834 (cortex_a7_alu_reg): Likewise.
3835 (cortex_a7_alu_shift): Likewise.
3836 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3838 (cortex_a8_alu_shift): Likewise.
3839 (cortex_a8_alu_shift_reg): Likewise.
3840 (cortex_a8_mov): Likewise.
3841 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3842 (cortex_a9_dp_shift): Likewise.
3843 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3845 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3847 (cortex_r4_mov): Likewise.
3848 (cortex_r4_alu_shift): Likewise.
3849 (cortex_r4_alu_shift_reg): Likewise.
3850 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3851 (526_alu_shift_op): Likewise.
3852 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3853 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3854 (626te_alu_shift_op): Likewise.
3855 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3856 (726te_alu_op): Likewise.
3857 (726te_alu_shift_op): Likewise.
3858 (726te_alu_shift_reg_op): Likewise.
3859 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3860 (mp626_alu_shift_op): Likewise.
3861 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3862 (pj4_alu_e1_conds): Likewise.
3863 (pj4_alu): Likewise.
3864 (pj4_alu_conds): Likewise.
3865 (pj4_shift): Likewise.
3866 (pj4_shift_conds): Likewise.
3867 (pj4_alu_shift): Likewise.
3868 (pj4_alu_shift_conds): Likewise.
3869 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3871 (cortexa7_older_only): Likewise.
3872 (cortexa7_younger): Likewise.
3874 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3876 * ipa-pure-const.c (generate_summary): Rename to...
3877 (pure_const_generate_summary): ... this.
3879 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3881 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3883 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3886 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3889 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3891 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3892 enabled without -march=zEC12.
3893 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3896 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3898 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3899 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3900 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3901 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3902 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3903 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3904 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3906 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3907 and ISA_HAS_NMADD3_NMSUB3.
3908 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3909 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3910 (nmsub4<mode>, nmsub3<mode>): Likewise.
3911 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3913 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3915 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3916 TARGET_MIPS5400 checking.
3918 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3919 Peter Bergner <bergner@vnet.ibm.com>
3921 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3922 registers in the comment.
3923 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3924 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3925 rather than FIRST_PSEUDO_REGISTERS.
3927 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3929 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3930 enable extra ISA flags with TARGET_HTM.
3932 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3934 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3937 2013-07-15 Cong Hou <congh@google.com>
3939 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3940 in compare function for sorting.
3942 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3944 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3945 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3946 * config/rs6000/rs6000.opt: Add -mhtm option.
3947 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3948 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3949 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3950 __HTM__ if the HTM instructions are available.
3951 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3952 htm_spr_reg_operand): New define_predicates.
3953 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3954 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3956 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3957 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3958 HTM builtin functions.
3959 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3960 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3961 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3962 (rs6000_builtin_mask_calculate): Likewise.
3963 (rs6000_option_override_internal): Likewise.
3964 (bdesc_htm): Add new HTM builtin support.
3965 (htm_spr_num): New function.
3966 (htm_spr_regno): Likewise.
3967 (rs6000_htm_spr_icode): Likewise.
3968 (htm_expand_builtin): Likewise.
3969 (htm_init_builtins): Likewise.
3970 (rs6000_expand_builtin): Add support for HTM builtin functions.
3971 (rs6000_init_builtins): Likewise.
3972 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3974 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3975 (TARGET_HTM, MASK_HTM): Define macros.
3976 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3977 (FIXED_REGISTERS): Likewise.
3978 (CALL_USED_REGISTERS): Likewise.
3979 (CALL_REALLY_USED_REGISTERS): Likewise.
3980 (REG_ALLOC_ORDER): Likewise.
3981 (enum reg_class): Likewise.
3982 (REG_CLASS_NAMES): Likewise.
3983 (REG_CLASS_CONTENTS): Likewise.
3984 (REGISTER_NAMES): Likewise.
3985 (ADDITIONAL_REGISTER_NAMES): Likewise.
3986 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3987 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3988 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3989 * config/rs6000/htm.md: New file.
3990 * config/rs6000/htmintrin.h: New file.
3991 * config/rs6000/htmxlintrin.h: New file.
3993 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3995 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3996 Define SYMBOL_TINY_GOT, update comment.
3997 * config/aarch64/aarch64.c
3998 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3999 (aarch64_expand_mov_immediate): Likewise.
4000 (aarch64_print_operand): Likewise.
4001 (aarch64_classify_symbol): Likewise.
4002 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4003 (ldr_got_tiny): Define.
4005 2013-07-13 Tobias Grosser <tobias@grosser.es>
4007 PR tree-optimization/54094
4008 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4009 scheduling dimension for the parallelism check from the polyhedral
4010 information in the AST.
4011 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4014 2013-07-13 Jason Merrill <jason@redhat.com>
4016 * print-tree.c (debug_vec_tree): Use debug_raw.
4017 (debug_raw (vec<tree, va_gc> &)): New.
4018 (debug_raw (vec<tree, va_gc> *)): New.
4019 * tree.h: Declare them.
4021 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4023 * ifcvt.c (ifcvt_after_combine): New static variable.
4024 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4026 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4027 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4028 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4030 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4032 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4034 2013-07-12 Michael Matz <matz@suse.de>
4037 * convert.c (convert_to_real): Reject non-float inner types.
4039 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4041 * config/aarch64/aarch64-protos.h
4042 (aarch64_simd_immediate_valid_for_move): Remove.
4043 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4044 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4046 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4048 2013-07-11 Steve Ellcey <sellcey@mips.com>
4050 * config/mips/mips.c (mips_conditional_register_usage): Do not
4051 use t[0-7] registers in MIPS16 mode when optimizing for size.
4053 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4055 * config/i386/i386.c (dispatch_function_versions): Fix array
4056 indexing of function_version_info to match actual_versions.
4058 2013-07-11 Teresa Johnson <tejohnson@google.com>
4060 * vec.h (struct va_gc): Move release out-of-line.
4061 (va_gc::release): Call ggc_free on released vec.
4063 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4065 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4066 Require GOT register as additional operand in UNSPEC.
4067 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4068 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4069 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4070 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4071 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4072 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4073 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4075 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4078 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4079 name seen by assembler/linker rather if available.
4081 2013-07-11 Andreas Schwab <schwab@suse.de>
4083 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4085 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4087 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4089 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4091 * doc/tm.texi.in: Move hook documentation to ....
4092 * target.def: ... here.
4094 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4096 * doc/tm.texi: Regenerate.
4098 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4101 * doc/invoke.texi: Document -Wconditionally-supported.
4103 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4106 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4109 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4112 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4113 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4114 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4115 Remove duplicate devices.
4116 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4117 * config/avr/t-multilib: Regenerate.
4118 * config/avr/avr-tables.opt: Regenerate.
4119 * doc/avr-mmcu.texi: Regenerate.
4121 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4124 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4126 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4128 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4129 the cost of MULT when optimizing for size.
4131 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4133 * config/cr16/cr16-protos.h: Don't include target.h.
4135 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4137 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4138 adjust register size for TDmode and TFmode for VSX registers.
4140 2013-07-08 Kai Tietz <ktietz@redhat.com>
4143 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4144 hook_bool_const_tree_true.
4146 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4148 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4149 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4150 * config/s390/s390.md: Define FPR*_REGNUM constants.
4151 Fix FPR2_REGNUM constant (18 -> 17).
4152 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4153 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4154 ("trunc<BFP:mode><DFP_ALL:mode>2")
4155 ("trunc<DFP_ALL:mode><BFP:mode>2")
4156 ("*extend<BFP:mode><DFP_ALL:mode>2")
4157 ("*extend<DFP_ALL:mode><BFP:mode>2")
4158 ("extend<BFP:mode><DFP_ALL:mode>2")
4159 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4162 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4164 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4166 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4168 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4169 and cfun_fpr_bit_p to cfun_fpr_save_p.
4170 (s390_frame_area, s390_register_info, s390_frame_info)
4171 (s390_emit_prologue, s390_emit_epilogue)
4172 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4174 * config/s390/s390.h: Define *_REGNUM macros for floating point
4177 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4179 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4181 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4183 PR rtl-optimization/57786
4184 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4185 and break out of the loop when it is set to false.
4187 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4190 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4191 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4192 (const_int 63)) 0)).
4193 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4194 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4195 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4197 PR rtl-optimization/57829
4198 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4199 mask bits outside of mode are just sign-extension from mode to HWI.
4201 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4203 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4204 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4205 adjust_address instead of change_address to keep info about alignment.
4206 (emit_strmov): Remove.
4207 (emit_memmov): New function.
4208 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4209 (expand_movmem_epilogue): Likewise and return updated rtx for
4211 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4212 destination and source.
4213 (decide_alignment): Refactor, handle vector_loop.
4214 (ix86_expand_movmem): Likewise.
4215 (ix86_expand_setmem): Likewise.
4216 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4218 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4220 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4221 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4223 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4225 * config/i386/sse.md (sse_movlhps): Change alternative 3
4226 of operand 2 to "m".
4228 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4231 * config/i386/sse.md (iptr): New mode attribute.
4232 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4233 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4234 (<sse>_vmmul<mode>3): Ditto.
4235 (<sse>_vmdiv<mode>3): Ditto.
4236 (sse_vmrcpv4sf2): Ditto.
4237 (<sse>_vmsqrt<mode>2): Ditto.
4238 (sse_vmrsqrtv4sf2): Ditto.
4239 (<sse>_vm<code><mode>3): Ditto.
4240 (avx_vmcmp<mode>3): Ditto.
4241 (<sse>_vmmaskcmp<mode>3): Ditto.
4242 (<sse>_comi): Ditto.
4243 (<sse>_ucomi): Ditto.
4244 (*xop_vmfrcz_<mode>): Ditto.
4245 (*fmai_fmadd_<mode>): Ditto.
4246 (*fmai_fmsub_<mode>): Ditto.
4247 (*fmai_fnmadd_<mode>): Ditto.
4248 (*fmai_fnmsub_<mode>): Ditto.
4249 (*fma4i_vmfmadd_<mode>): Ditto.
4250 (*fma4i_vmfmsub_<mode>): Ditto.
4251 (*fma4i_vmfnmadd_<mode>): Ditto.
4252 (*fma4i_vmfnmsub_<mode>): Ditto.
4253 (*xop_vmfrcz_<mode>): Ditto.
4254 (sse_cvtps2pi): Ditto.
4255 (sse_cvttps2pi): Ditto.
4256 (sse_cvtss2si): Ditto.
4257 (sse_cvtss2si_2): Ditto.
4258 (sse_cvtss2siq_2): Ditto.
4259 (sse_cvttss2si): Ditto.
4260 (sse_cvttss2siq): Ditto.
4261 (sse_cvtsd2si): Ditto.
4262 (sse_cvtsd2si_2): Ditto.
4263 (sse_cvtsd2siq_2): Ditto.
4264 (sse_cvttsd2si): Ditto.
4265 (sse_cvttsd2siq): Ditto.
4266 (sse_cvtsd2ss): Ditto.
4267 (sse_cvtss2sd): Ditto.
4268 (avx2_pbroadcast<mode>): Ditto.
4269 (avx2_pbroadcast<mode>_1): Ditto.
4270 (*avx_vperm_broadcast_v4sf): Ditto.
4272 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4273 (sse_movlhps): Ditto.
4274 (sse_storehps): Ditto.
4275 (sse_loadhps): Ditto.
4276 (sse_storelps): Ditto.
4277 (sse_loadlps): Ditto.
4278 (*vec_concatv4sf): Ditto.
4279 (*vec_interleave_highv2df): Ditto.
4280 (*vec_interleave_lowv2df): Ditto.
4281 (*vec_extractv2df_1_sse): Ditto.
4282 (*vec_extractv2df_0_sse): Ditto.
4283 (sse2_storelpd): Ditto.
4284 (sse2_loadlpd): Ditto.
4285 (sse2_movsd): Ditto.
4286 (*vec_concatv4si): Ditto.
4287 (vec_concatv2di): Ditto.
4289 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4290 for Intel asm dialect.
4291 (mmx_punpcklwd): Ditto.
4292 (mmx_punpckldq): Ditto.
4294 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4295 for intel assembler dialect.
4297 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4300 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4301 for BUILT_IN_C{LZ,LRSB}*.
4302 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4303 * tree-vrp.c (extract_range_basic): Handle
4304 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4305 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4306 fall thru to code calling set_value*.
4307 * builtins.c (expand_builtin): Remove *IMAX cases.
4308 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4309 if width is bigger than 2*HWI.
4311 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4313 PR rtl-optimization/55342
4314 * lra-int.h (lra_subreg_reload_pseudos): New.
4315 * lra.c: Add undoing optional reloads to the block diagram.
4316 (lra_subreg_reload_pseudos): New.
4317 (lra_optional_reload_pseudos): Change comments.
4318 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4319 lra_optional_reload_pseudos after undo transformations.
4320 * lra-assigns.c (pseudo_prefix_title): New.
4321 (lra_setup_reg_renumber): Use it.
4322 (spill_for): Ditto. Check subreg reload pseudos too.
4323 (assign_by_spills): Consider subreg reload pseudos too.
4324 * lra-constraints.c (simplify_operand_subreg): Use
4325 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4326 (curr_insn_transform): Recognize and do optional reloads.
4327 (undo_optional_reloads): New.
4328 (lra_undo_inheritance): Call undo_optional_reloads.
4330 2013-07-05 Thomas Quinot <quinot@adacore.com>
4332 * tree-complex.c (expand_complex_operations_1): Fix typo.
4334 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4336 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4337 (tune_params): New member 'const vec_costs'.
4338 * config/aarch64/aarch64.c (generic_vector_cost): New.
4339 (generic_tunings): New member 'generic_vector_cost'.
4340 (aarch64_builtin_vectorization_cost): New.
4341 (aarch64_add_stmt_cost): New.
4342 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4343 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4345 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4348 * config/i386/predicates.md (vsib_address_operand): Disallow
4349 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4351 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4354 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4355 expand_builtin_setjmp_receiver.
4356 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4357 with NULL for the label parameter.
4358 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4359 the frame-pointer. Adjust comments.
4360 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4361 only if LABEL is non-NULL.
4363 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4365 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4366 (ARM_ABI_AAPCS64): Ditto.
4368 (ARM_DEFAULT_ABI): Ditto.
4370 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4372 * config/aarch64/aarch64-builtins.c
4373 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4374 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4376 * config/aarch64/aarch64-simd.md
4377 (aarch64_ld1<VALL:mode>): New.
4378 (aarch64_st1<VALL:mode>): Likewise.
4379 * config/aarch64/arm_neon.h
4380 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4382 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4384 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4386 (ix86_option_override_internal):Turn on all -mavx target flags by
4387 default as they are dependent on AVX anyway.
4389 2013-07-02 Cary Coutant <ccoutant@google.com>
4391 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4393 (loc_checksum_ordered): Likewise.
4394 (hash_loc_operands): Remove inline keyword.
4396 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4398 PR tree-optimization/57741
4399 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4400 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4401 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4402 Allow REAL_CST step_exprs if flag_associative_math.
4403 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4405 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4407 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4409 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4411 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4413 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4415 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4417 (iorsi3_insn): Likewise.
4418 (arm_xorsi3): Likewise.
4420 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4422 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4424 (attribute "type"): ... here, and prefix with "wmmx_".
4425 (attribute "core_cycles"): Update for attribute changes.
4426 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4427 (tbcstv4hi): Likewise.
4428 (tbcstv2si): Likewise.
4429 (iwmmxt_iordi3): Likewise.
4430 (iwmmxt_xordi3): Likewise.
4431 (iwmmxt_anddi3): Likewise.
4432 (iwmmxt_nanddi3): Likewise.
4433 (iwmmxt_arm_movdi): Likewise.
4434 (iwmmxt_movsi_insn): Likewise.
4435 (mov<mode>_internal): Likewise.
4436 (and<mode>3_iwmmxt): Likewise.
4437 (ior<mode>3_iwmmxt): Likewise.
4438 (xor<mode>3_iwmmxt): Likewise.
4439 (add<mode>3_iwmmxt): Likewise.
4440 (ssaddv8qi3): Likewise.
4441 (ssaddv4hi3): Likewise.
4442 (ssaddv2si3): Likewise.
4443 (usaddv8qi3): Likewise.
4444 (usaddv4hi3): Likewise.
4445 (usaddv2si3): Likewise.
4446 (sub<mode>3_iwmmxt): Likewise.
4447 (sssubv8qi3): Likewise.
4448 (sssubv4hi3): Likewise.
4449 (sssubv2si3): Likewise.
4450 (ussubv8qi3): Likewise.
4451 (ussubv4hi3): Likewise.
4452 (ussubv2si3): Likewise.
4453 (mulv4hi3_iwmmxt): Likewise.
4454 (smulv4hi3_highpart): Likewise.
4455 (umulv4hi3_highpart): Likewise.
4456 (iwmmxt_wmacs): Likewise.
4457 (iwmmxt_wmacsz): Likewise.
4458 (iwmmxt_wmacu): Likewise.
4459 (iwmmxt_wmacuz): Likewise.
4460 (iwmmxt_clrdi): Likewise.
4461 (iwmmxt_clrv8qi): Likewise.
4462 (iwmmxt_clr4hi): Likewise.
4463 (iwmmxt_clr2si): Likewise.
4464 (iwmmxt_uavgrndv8qi3): Likewise.
4465 (iwmmxt_uavgrndv4hi3): Likewise.
4466 (iwmmxt_uavgv8qi3): Likewise.
4467 (iwmmxt_uavgv4hi3): Likewise.
4468 (iwmmxt_tinsrb): Likewise.
4469 (iwmmxt_tinsrh): Likewise.
4470 (iwmmxt_tinsrw): Likewise.
4471 (iwmmxt_textrmub): Likewise.
4472 (iwmmxt_textrmsb): Likewise.
4473 (iwmmxt_textrmuh): Likewise.
4474 (iwmmxt_textrmsh): Likewise.
4475 (iwmmxt_textrmw): Likewise.
4476 (iwmxxt_wshufh): Likewise.
4477 (eqv8qi3): Likewise.
4478 (eqv4hi3): Likewise.
4479 (eqv2si3): Likewise.
4480 (gtuv8qi3): Likewise.
4481 (gtuv4hi3): Likewise.
4482 (gtuv2si3): Likewise.
4483 (gtv8qi3): Likewise.
4484 (gtv4hi3): Likewise.
4485 (gtv2si3): Likewise.
4486 (smax<mode>3_iwmmxt): Likewise.
4487 (umax<mode>3_iwmmxt): Likewise.
4488 (smin<mode>3_iwmmxt): Likewise.
4489 (umin<mode>3_iwmmxt): Likewise.
4490 (iwmmxt_wpackhss): Likewise.
4491 (iwmmxt_wpackwss): Likewise.
4492 (iwmmxt_wpackdss): Likewise.
4493 (iwmmxt_wpackhus): Likewise.
4494 (iwmmxt_wpackwus): Likewise.
4495 (iwmmxt_wpackdus): Likewise.
4496 (iwmmxt_wunpckihb): Likewise.
4497 (iwmmxt_wunpckihh): Likewise.
4498 (iwmmxt_wunpckihw): Likewise.
4499 (iwmmxt_wunpckilb): Likewise.
4500 (iwmmxt_wunpckilh): Likewise.
4501 (iwmmxt_wunpckilw): Likewise.
4502 (iwmmxt_wunpckehub): Likewise.
4503 (iwmmxt_wunpckehuh): Likewise.
4504 (iwmmxt_wunpckehuw): Likewise.
4505 (iwmmxt_wunpckehsb): Likewise.
4506 (iwmmxt_wunpckehsh): Likewise.
4507 (iwmmxt_wunpckehsw): Likewise.
4508 (iwmmxt_wunpckelub): Likewise.
4509 (iwmmxt_wunpckeluh): Likewise.
4510 (iwmmxt_wunpckeluw): Likewise.
4511 (iwmmxt_wunpckelsb): Likewise.
4512 (iwmmxt_wunpckelsh): Likewise.
4513 (iwmmxt_wunpckelsw): Likewise.
4514 (ror<mode>3): Likewise.
4515 (ashr<mode>3_iwmmxt): Likewise.
4516 (lshr<mode>3_iwmmxt): Likewise.
4517 (ashl<mode>3_iwmmxt): Likewise.
4518 (ror<mode>3_di): Likewise.
4519 (ashr<mode>3_di): Likewise.
4520 (lshr<mode>3_di): Likewise.
4521 (ashl<mode>3_di): Likewise.
4522 (iwmmxt_wmadds): Likewise.
4523 (iwmmxt_wmaddu): Likewise.
4524 (iwmmxt_tmia): Likewise.
4525 (iwmmxt_tmiaph): Likewise.
4526 (iwmmxt_tmiabb): Likewise.
4527 (iwmmxt_tmiatb): Likewise.
4528 (iwmmxt_tmiabt): Likewise.
4529 (iwmmxt_tmiatt): Likewise.
4530 (iwmmxt_tmovmskb): Likewise.
4531 (iwmmxt_tmovmskh): Likewise.
4532 (iwmmxt_tmovmskw): Likewise.
4533 (iwmmxt_waccb): Likewise.
4534 (iwmmxt_wacch): Likewise.
4535 (iwmmxt_waccw): Likewise.
4536 (iwmmxt_waligni): Likewise.
4537 (iwmmxt_walignr): Likewise.
4538 (iwmmxt_walignr0): Likewise.
4539 (iwmmxt_walignr1): Likewise.
4540 (iwmmxt_walignr2): Likewise.
4541 (iwmmxt_walignr3): Likewise.
4542 (iwmmxt_wsadb): Likewise.
4543 (iwmmxt_wsadh): Likewise.
4544 (iwmmxt_wsadbz): Likewise.
4545 (iwmmxt_wsadhz): Likewise.
4546 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4547 (iwmmxt_wabsdiffb): Likewise.
4548 (iwmmxt_wabsdiffh): Likewise.
4549 (iwmmxt_wabsdiffw): Likewise.
4550 (iwmmxt_waddsubhx): Likewise
4551 (iwmmxt_wsubaddhx): Likewise.
4552 (addc<mode>3): Likewise.
4553 (iwmmxt_avg4): Likewise.
4554 (iwmmxt_avg4r): Likewise.
4555 (iwmmxt_wmaddsx): Likewise.
4556 (iwmmxt_wmaddux): Likewise.
4557 (iwmmxt_wmaddsn): Likewise.
4558 (iwmmxt_wmaddun): Likewise.
4559 (iwmmxt_wmulwsm): Likewise.
4560 (iwmmxt_wmulwum): Likewise.
4561 (iwmmxt_wmulsmr): Likewise.
4562 (iwmmxt_wmulumr): Likewise.
4563 (iwmmxt_wmulwsmr): Likewise.
4564 (iwmmxt_wmulwumr): Likewise.
4565 (iwmmxt_wmulwl): Likewise.
4566 (iwmmxt_wqmulm): Likewise.
4567 (iwmmxt_wqmulwm): Likewise.
4568 (iwmmxt_wqmulmr): Likewise.
4569 (iwmmxt_wqmulwmr): Likewise.
4570 (iwmmxt_waddbhusm): Likewise.
4571 (iwmmxt_waddbhusl): Likewise.
4572 (iwmmxt_wqmiabb): Likewise.
4573 (iwmmxt_wqmiabt): Likewise.
4574 (iwmmxt_wqmiatb): Likewise.
4575 (iwmmxt_wqmiatt): Likewise.
4576 (iwmmxt_wqmiabbn): Likewise.
4577 (iwmmxt_wqmiabtn): Likewise.
4578 (iwmmxt_wqmiatbn): Likewise.
4579 (iwmmxt_wqmiattn): Likewise.
4580 (iwmmxt_wmiabb): Likewise.
4581 (iwmmxt_wmiabt): Likewise.
4582 (iwmmxt_wmiatb): Likewise.
4583 (iwmmxt_wmiatt): Likewise.
4584 (iwmmxt_wmiabbn): Likewise.
4585 (iwmmxt_wmiabtn): Likewise.
4586 (iwmmxt_wmiatbn): Likewise.
4587 (iwmmxt_wmiattn): Likewise.
4588 (iwmmxt_wmiawbb): Likewise.
4589 (iwmmxt_wmiawbt): Likewise.
4590 (iwmmxt_wmiawtb): Likewise.
4591 (iwmmxt_wmiawtt): Likewise.
4592 (iwmmxt_wmiawbbn): Likewise.
4593 (iwmmxt_wmiawbtn): Likewise.
4594 (iwmmxt_wmiawtbn): Likewise.
4595 (iwmmxt_wmiawttn): Likewise.
4596 (iwmmxt_wmerge): Likewise.
4597 (iwmmxt_tandc<mode>3): Likewise.
4598 (iwmmxt_torc<mode>3): Likewise.
4599 (iwmmxt_torvsc<mode>3): Likewise.
4600 (iwmmxt_textrc<mode>3): Likewise.
4601 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4602 (wmmxt_pack): Likewise.
4603 (wmmxt_mult_c1): Likewise.
4604 (wmmxt_mult_c2): Likewise.
4605 (wmmxt_alu_c1): Likewise.
4606 (wmmxt_alu_c2): Likewise.
4607 (wmmxt_alu_c3): Likewise.
4608 (wmmxt_transfer_c1): Likewise.
4609 (wmmxt_transfer_c2): Likewise.
4610 (wmmxt_transfer_c3): Likewise.
4611 (marvell_f_iwmmxt_wstr): Likewise.
4612 (marvell_f_iwmmxt_wldr): Likewise.
4614 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
4616 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
4618 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4621 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4622 * lra-constraints.c (need_for_split_p): Check call used hard regs
4623 living through calls.
4625 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
4626 call used regs for call insn.
4628 2013-06-28 Jakub Jelinek <jakub@redhat.com>
4631 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
4632 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
4634 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
4636 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
4638 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4640 * lra-constraints.c (need_for_split_p): Check call used hard regs
4641 living through calls.
4643 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
4646 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
4647 to tie with any other modes. Eliminate Altivec vector mode tests,
4648 since these are a subset of ALTIVEC or VSX vector modes. Simplify
4649 code, to return 0 if testing MODE2 for a condition, if we've
4650 already tested MODE1 for the same condition.
4652 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4654 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
4657 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4659 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4660 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
4662 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4664 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
4666 (aarch64_symbolic_constant_p): Remove.
4667 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
4668 static. Fix line length and white space.
4669 (aarch64_symbolic_constant_p): Remove.
4670 * config/aarch64/predicates.md (aarch64_valid_symref):
4671 Use aarch64_classify_symbol_expression.
4673 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4675 * config/arm/constraints.md (Ts): New constraint.
4676 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
4678 (compare_scc): Use "Ts" constraint for operand 0.
4679 (ior_scc_scc): Likewise.
4680 (and_scc_scc): Likewise.
4681 (and_scc_scc_nodom): Likewise.
4682 (ior_scc_scc_cmp): Likewise for operand 7.
4683 (and_scc_scc_cmp): Likewise.
4684 * config/arm/thumb2.md (thumb2_movsi_insn):
4685 Add alternatives for 16-bit encodings.
4686 (thumb2_movhi_insn): Likewise.
4687 (thumb2_movsicc_insn): Likewise.
4688 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
4689 (thumb2_negscc): Use "Ts" constraint.
4690 Move mvn instruction outside cond_exec block.
4691 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
4692 for 16-bit encodings.
4694 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4696 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
4698 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
4699 (mulsi3subsi): Likewise.
4700 (mulsidi3adddi): Likewise.
4701 (mulsidi3_v6): Likewise.
4702 (umulsidi3_v6): Likewise.
4703 (umulsidi3adddi_v6): Likewise.
4704 (smulsi3_highpart_v6): Likewise.
4705 (umulsi3_highpart_v6): Likewise.
4706 (mulhisi3tb): Likewise.
4707 (mulhisi3bt): Likewise.
4708 (mulhisi3tt): Likewise.
4709 (maddhisi4): Likewise.
4710 (maddhisi4tb): Likewise.
4711 (maddhisi4tt): Likewise.
4712 (maddhidi4): Likewise.
4713 (maddhidi4tb): Likewise.
4714 (maddhidi4tt): Likewise.
4715 (zeroextractsi_compare0_scratch): Likewise.
4716 (insv_zero): Likewise.
4717 (insv_t2): Likewise.
4718 (anddi_notzesidi_di): Likewise.
4719 (anddi_notsesidi_di): Likewise.
4720 (andsi_notsi_si): Likewise.
4721 (iordi_zesidi_di): Likewise.
4722 (xordi_zesidi_di): Likewise.
4723 (andsi_iorsi3_notsi): Likewise.
4725 (smax_m1): Likewise.
4727 (not_shiftsi): Likewise.
4728 (unaligned_loadsi): Likewise.
4729 (unaligned_loadhis): Likewise.
4730 (unaligned_loadhiu): Likewise.
4731 (unaligned_storesi): Likewise.
4732 (unaligned_storehi): Likewise.
4733 (extv_reg): Likewise.
4734 (extzv_t2): Likewise.
4736 (udivsi3): Likewise.
4737 (arm_zero_extendhisi2addsi): Likewise.
4738 (arm_zero_extendqisi2addsi): Likewise.
4739 (compareqi_eq0): Likewise.
4740 (arm_extendhisi2_v6): Likewise.
4741 (arm_extendqisi2addsi): Likewise.
4742 (arm_movt): Likewise.
4743 (thumb2_ldrd): Likewise.
4744 (thumb2_ldrd_base): Likewise.
4745 (thumb2_ldrd_base_neg): Likewise.
4746 (thumb2_strd): Likewise.
4747 (thumb2_strd_base): Likewise.
4748 (thumb2_strd_base_neg): Likewise.
4749 (arm_negsi2): Add alternative for 16-bit encoding.
4750 (arm_one_cmplsi2): Likewise.
4752 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4754 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
4755 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
4756 (movdfcc): Likewise.
4757 * config/arm/vfp.md (*thumb2_movsf_vfp):
4758 Disable predication for arm_restrict_it.
4759 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
4760 (*thumb2_movdfcc_vfp): Likewise.
4761 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
4762 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
4763 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
4764 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
4765 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
4766 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
4767 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
4768 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
4769 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
4770 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
4771 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
4772 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
4773 Disable predication for arm_restrict_it.
4775 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
4777 * config/i386/bmiintrin.h (_bextr_u32): New.
4778 (_bextr_u64): Ditto.
4780 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
4782 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
4783 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
4784 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
4785 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
4786 * config/mips/n32-elf.h: ...this new file.
4788 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4791 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4792 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4794 2013-06-27 Catherine Moore <clm@codesourcery.com>
4796 * config/mips/mips-tables.opt: Regenerate.
4797 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4798 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4799 * doc/invoke.texi: Add -m14kc.
4801 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4804 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4805 constraints of operand 1 and 2.
4808 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4809 to match RTL canonicalization. Swap predicates and
4810 constraints of operand 1 and 2.
4812 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4814 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4815 Process OP_INOUT regs for splitting too.
4817 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4819 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4820 decl before the loop, initialize to NULL.
4821 (vectorizable_load): Initialize ptr_incr to NULL.
4823 2013-06-27 Martin Jambor <mjambor@suse.cz>
4826 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4827 * ipa-ref.c (ipa_maybe_record_reference): New function.
4828 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4829 * ipa-cp.c (create_specialized_node): Record potential references from
4831 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4833 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4835 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4836 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4837 'mode' to force_reg.
4838 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4839 (aarch64_expand_mov_immediate): Likewise.
4841 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4843 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4844 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4846 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4848 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4849 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4850 (struct machine_function): Add tbegin_p.
4851 (s390_canonicalize_comparison): Fold CC mode compares to
4852 conditional jump if possible.
4853 (s390_emit_jump): Return the emitted jump.
4854 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4855 Handle CCRAWmode compares.
4856 (s390_option_override): Default to -mhtm if available.
4857 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4858 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4859 FPRs instead of df_regs_ever_live_p.
4860 (s390_optimize_nonescaping_tx): New function.
4861 (s390_init_frame_layout): Extend clobbered_regs array to cover
4863 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4864 (s390_expand_tbegin): New function.
4865 (enum s390_builtin): New enum definition.
4866 (code_for_builtin): New array definition.
4867 (s390_init_builtins): New function.
4868 (s390_expand_builtin): New function.
4869 (TARGET_INIT_BUILTINS): Define.
4870 (TARGET_EXPAND_BUILTIN): Define.
4871 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4872 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4873 (s390_alc_comparison): Likewise.
4874 * config/s390/s390-modes.def: Add CCRAWmode.
4875 * config/s390/s390.h (processor_flags): Add PF_TX.
4876 (TARGET_CPU_HTM): Define macro.
4877 (TARGET_HTM): Define macro.
4878 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4879 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4880 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4881 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4882 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4883 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4884 ("*cc_to_int"): Move up.
4885 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4886 constants other than 0.
4887 ("*ccraw_to_int"): New insn and splitter definition.
4888 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4889 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4890 ("tx_assist"): New expander.
4891 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4892 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4893 * config/s390/s390.opt: Add -mhtm option.
4894 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4895 * config/s390/htmxlintrin.h: New file.
4896 * config/s390/htmintrin.h: New file.
4897 * config/s390/s390intrin.h: New file.
4898 * doc/extend.texi: Document htm builtins.
4899 * config.gcc: Add the new header files to extra_headers.
4901 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4903 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4904 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4906 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4907 Pat Haugen <pthaugen@us.ibm.com>
4908 Peter Bergner <bergner@vnet.ibm.com>
4910 * config/rs6000/power8.md: New.
4911 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4912 setting for power8 entry.
4913 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4914 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4915 test for Power4/Power5 only.
4916 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4918 (force_new_group): Adjust comment.
4919 * config/rs6000/rs6000.md: Include power8.md.
4921 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4923 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4924 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4926 (tune_params): Update comment.
4927 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4928 (arm_max_conditional_execute): New function.
4929 (thumb2_final_prescan_insn): Use max_insn_skipped and
4930 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4932 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4934 PR tree-optimization/57705
4935 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4936 SSA_NAME step, provided that it is not defined inside the loop.
4937 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4938 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4940 2013-06-25 Martin Jambor <mjambor@suse.cz>
4943 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4944 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4946 (ipa_note_param_call): Initialize member_ptr flag.
4947 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4948 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4949 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4950 (ipa_read_indirect_edge_info): Likewise.
4952 2013-06-25 Richard Biener <rguenther@suse.de>
4955 * passes.c (init_optimization_passes): Move pass_fold_builtins
4956 and pass_dce earlier with -Og.
4958 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4960 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4961 <BIT_FIELD_REF>: Remove trailing TAB.
4962 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4965 2013-06-24 Martin Jambor <mjambor@suse.cz>
4967 PR tree-optimization/57358
4968 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4969 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4970 (ipa_analyze_params_uses): Generate pessimistic info when true.
4972 2013-06-24 Martin Jambor <mjambor@suse.cz>
4974 PR tree-optimization/57539
4975 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4976 global.inlined_to of the new node to it. All callers changed.
4977 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4978 inlining_into, pass it to cgraph_clone_node.
4979 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4980 ipa_free_edge_args_substructures.
4981 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4982 rdesc linked list. Do not assert rdesc edges have inlined caller.
4983 Assert we have found an rdesc in the rdesc list.
4985 2013-06-24 Richard Biener <rguenther@suse.de>
4987 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4988 (pointer_set_lookup): Declare.
4989 (class pointer_map): New template class implementing a
4990 generic pointer to T map.
4991 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4992 pointer_map<T>::contains, pointer_map<T>::insert,
4993 pointer_map<T>::traverse): New functions.
4994 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4995 (pointer_set_lookup): New function.
4996 (pointer_set_contains): Use pointer_set_lookup.
4997 (pointer_set_insert): Likewise.
4998 (insert_aux): Remove.
4999 (struct pointer_map_t): Embed a pointer_set_t.
5000 (pointer_map_create): Adjust.
5001 (pointer_map_destroy): Likewise.
5002 (pointer_map_contains): Likewise.
5003 (pointer_map_insert): Likewise.
5004 (pointer_map_traverse): Likewise.
5005 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5006 pointer_map<unsigned> instead of a pointer_map_t.
5007 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5008 (streamer_tree_cache_lookup): Likewise.
5009 (streamer_tree_cache_create): Likewise.
5010 (streamer_tree_cache_delete): Likewise.
5011 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5012 pointer_map<unsigned> instead of a pointer_map_t.
5013 (lto_init_tree_ref_encoder): Adjust.
5014 (lto_destroy_tree_ref_encoder): Likewise.
5015 * lto-section-out.c (lto_output_decl_index): Likewise.
5016 (lto_record_function_out_decl_state): Likewise.
5017 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5019 2013-06-24 Richard Biener <rguenther@suse.de>
5021 PR tree-optimization/57488
5022 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5024 2013-06-24 Alan Modra <amodra@gmail.com>
5026 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5027 (gen_easy_altivec_constant): Likewise.
5028 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5029 easy_vector_constant_msb): Likewise.
5031 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5034 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5035 add missing return true.
5037 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5040 * config/sh/predicates.md (general_extend_operand): Invoke
5041 general_movsrc_operand for memory operands.
5042 (general_movsrc_operand): Allow reg+reg addressing, do not use
5043 general_operand for memory operands.
5045 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5047 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5048 when current target options does not apply.
5049 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5050 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5051 * config/i386/bmiintrin.h: Pass appropriate target
5052 attributes to header.
5053 * config/i386/mmintrin.h: Ditto.
5054 * config/i386/nmmintrin.h: Ditto.
5055 * config/i386/avx2intrin.h: Ditto.
5056 * config/i386/fxsrintrin.h: Ditto.
5057 * config/i386/tbmintrin.h: Ditto.
5058 * config/i386/xsaveintrin.h: Ditto.
5059 * config/i386/f16cintrin.h: Ditto.
5060 * config/i386/xtestintrin.h: Ditto.
5061 * config/i386/xsaveoptintrin.h: Ditto.
5062 * config/i386/bmi2intrin.h: Ditto.
5063 * config/i386/lzcntintrin.h: Ditto.
5064 * config/i386/smmintrin.h: Ditto.
5065 * config/i386/wmmintrin.h: Ditto.
5066 * config/i386/x86intrin.h: Remove all header include guards.
5067 * config/i386/prfchwintrin.h: Ditto.
5068 * config/i386/pmmintrin.h: Ditto.
5069 * config/i386/tmmintrin.h: Ditto.
5070 * config/i386/xmmintrin.h: Ditto.
5071 * config/i386/popcntintrin.h: Ditto.
5072 * config/i386/rdseedintrin.h: Ditto.
5073 * config/i386/ammintrin.h: Ditto.
5074 * config/i386/emmintrin.h: Ditto.
5075 * config/i386/immintrin.h: Remove all header include guards.
5076 * config/i386/fma4intrin.h: Ditto.
5077 * config/i386/lwpintrin.h: Ditto.
5078 * config/i386/xopintrin.h: Ditto.
5079 * config/i386/ia32intrin.h: Ditto.
5080 * config/i386/avxintrin.h: Ditto.
5081 * config/i386/rtmintrin.h: Ditto.
5082 * config/i386/fmaintrin.h: Ditto.
5083 * config/i386/mm3dnow.h: Ditto.
5085 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5087 * common/config/i386/i386-common.c: Handle LZCNT.
5089 2013-06-22 Andi Kleen <ak@linux.intel.com>
5091 * doc/extend.texi: Use __atomic_store_n instead of
5092 __atomic_store in HLE example.
5094 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5096 * config/sh/sh.c: Remove <cstdlib> workaround.
5098 2013-06-21 Andi Kleen <ak@linux.intel.com>
5100 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5102 2013-06-21 Andi Kleen <ak@linux.intel.com>
5104 * doc/extend.texi: Document that __atomic_clear and
5105 __atomic_test_and_set should only be used with bool.
5107 2013-06-20 Jan Hubicka <jh@suse.cz>
5109 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5111 * tree.c (decls_same_for_odr): New function.
5112 (same_for_edr): New function.
5113 (types_same_for_odr): New function.
5114 (get_binfo_at_offset): Use it.
5115 * tree.h (types_same_for_odr): Declare.
5117 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5118 Jason Merrill <jason@redhat.com>
5120 * system.h: Include <cstdlib> as well as <stdlib.h>.
5122 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5125 * config/i386/i386.c (construct_container): Report error if
5126 long double is used with disabled x87 float returns.
5128 2013-06-20 Jan Hubicka <jh@suse.cz>
5130 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5132 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5134 PR rtl-optimization/57425
5135 PR rtl-optimization/57569
5136 * alias.c (write_dependence_p): Remove parameters mem_mode and
5137 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5138 Changed all callers.
5139 (canon_anti_dependence): Get comments and semantics in sync.
5140 Add parameter mem_canonicalized. Changed all callers.
5141 * rtl.h (canon_anti_dependence): Update prototype.
5143 2013-06-20 Richard Biener <rguenther@suse.de>
5145 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5146 byte case, inline streamer_read_uchar and defer section
5149 2013-06-20 Richard Biener <rguenther@suse.de>
5151 PR tree-optimization/57584
5152 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5153 SSA names into the expanded expression that take part in
5154 abnormal coalescing.
5156 2013-06-19 Sharad Singhai <singhai@google.com>
5158 * gcov.c (print_usage): Handle new option.
5159 (process_args): Ditto.
5160 (get_gcov_intermediate_filename): New function.
5161 (output_intermediate_file): New function.
5162 (output_gcov_file): New function
5163 (generate_results): Handle new option.
5164 (release_function): Relase demangled name.
5165 (read_graph_file): Handle demangled name.
5166 (output_lines): Ditto.
5167 * doc/gcov.texi: Document gcov intermediate format.
5169 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5172 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5173 (lra_emit_add): Use the functions. Add comment about Y as an
5176 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5179 * collect2.c (collect_atexit): New.
5180 (collect_exit): Delete.
5181 (main): Register collect_atexit with atexit.
5182 (collect_wait): Change collect_exit to exit.
5184 * collect2.h (collect_exit): Delete.
5185 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5187 2013-06-19 Wei Mi <wmi@google.com>
5189 PR rtl-optimization/57518
5190 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5191 if regno is used in paradoxical subreg.
5192 (update_equiv_regs): Check pdx_subregs[regno] before
5193 set a reg to be equivalent with a mem.
5195 2013-06-19 Matthias Klose <doko@ubuntu.com>
5198 * file-find.h (find_a_file): Add a mode parameter.
5199 * file-find.c (find_a_file): Likewise.
5200 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5201 with X_OK for the executables.
5202 * collect2.c (main): Call find_a_file with X_OK.
5204 2013-06-19 Steve Ellcey <sellcey@mips.com>
5207 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5208 Use NEXT_INSN instead of next_real_insn.
5210 2013-06-19 Jan Hubicka <jh@suse.cz>
5212 * cgraph.h (const_value_known_p): Replace by ...
5213 (ctor_for_folding): .. this one.
5214 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5215 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5216 * expr.c (expand_expr_real_1): Likewise.
5217 (string_constant): Likewise.
5218 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5219 * ipa.c (process_references): Likewise.
5220 (symtab_remove_unreachable_nodes): Likewise.
5221 * ipa-inline-analysis.c (param_change_prob): Likewise.
5222 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5223 (get_base_constructor): Likwise.
5224 * varpool.c (varpool_remove_node): Likewise.
5225 (varpool_remove_initializer): LIkewise.
5226 (dump_varpool_node): LIkwise.
5227 (const_value_known_p): Rewrite to ...
5228 (ctor_for_folding): ... this one.
5230 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5233 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5234 PERSONALITY in $PATH derived prefixes.
5236 2013-06-19 Jeff Law <law@redhat.com>
5238 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5241 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5242 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5245 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5247 * config/arm/vfp.md: Move VFP instruction classification documentation
5249 * config/arm/arm.md: ... here. Update instruction classification
5252 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5254 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5255 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5256 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5258 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5261 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5262 with NEXT_INSN. Use tablejump_p to check for jump table data
5265 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5268 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5269 that now in C++ the value is correct per the C++ standards.
5271 2013-06-19 Richard Biener <rguenther@suse.de>
5273 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5276 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5279 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5282 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5283 with next_active_insn.
5285 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5287 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5288 functions are inlined during failures to flag an error.
5289 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5290 in early inline pass.
5292 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5294 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5297 2013-06-18 Julian Brown <julian@codesourcery.com>
5299 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5300 Permit virtual register pre-reload if !strict.
5301 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5303 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5305 * config/arm/neon.md (movmisalign<mode>): Use
5306 neon_perm_struct_or_reg_operand instead of
5307 neon_struct_or_register_operand.
5308 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5309 neon_permissive_struct_operand instead of neon_struct_operand.
5310 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5311 neon_vector_mem_operand.
5312 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5313 neon_vector_mem_operand.
5314 (neon_permissive_struct_operand): New.
5315 (neon_struct_or_register_operand): Rename to...
5316 (neon_perm_struct_or_reg_operand): This. Adjust call to
5317 neon_vector_mem_operand.
5319 2013-06-18 Richard Biener <rguenther@suse.de>
5321 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5322 * lto-streamer.h: Include pointer-set.h.
5323 (struct lto_decl_slot): Remove.
5324 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5325 Remove next_index entry.
5326 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5327 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5328 (lto_init_tree_ref_encoder): Adjust.
5329 (lto_destroy_tree_ref_encoder): Likewise.
5330 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5331 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5332 (lto_output_decl_index): Adjust.
5333 (lto_new_out_decl_state): Likewise.
5334 (lto_record_function_out_decl_state): Likewise.
5335 * lto-streamer-out.c (copy_function): Likewise.
5337 2013-06-18 Richard Biener <rguenther@suse.de>
5339 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5340 * cgraphunit.c: Include cfgloop.h.
5341 (init_lowered_empty_function): Initialize the loop tree.
5342 (assemble_thunk): Insert new BBs into loops.
5344 2013-06-18 Richard Biener <rguenther@suse.de>
5346 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5347 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5348 the map from cache entry to cache index optional.
5349 (streamer_tree_cache_replace_tree): Adjust accordingly.
5350 (streamer_tree_cache_append): Likewise.
5351 (streamer_tree_cache_delete): Likewise.
5352 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5353 streamer cache map from cache entry to cache index.
5354 * lto-streamer-out.c (create_output_block): Adjust.
5356 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5358 * config/arm/arm.md (attribute "insn"): Move multiplication and
5359 division attributes to...
5360 (attribute "type"): ... here. Remove mult.
5361 (attribute "mul32"): New attribute.
5362 (attribute "mul64"): Add umaal.
5363 (*arm_mulsi3): Update attributes.
5364 (*arm_mulsi3_v6): Likewise.
5365 (*thumb_mulsi3): Likewise.
5366 (*thumb_mulsi3_v6): Likewise.
5367 (*mulsi3_compare0): Likewise.
5368 (*mulsi3_compare0_v6): Likewise.
5369 (*mulsi_compare0_scratch): Likewise.
5370 (*mulsi_compare0_scratch_v6): Likewise.
5371 (*mulsi3addsi): Likewise.
5372 (*mulsi3addsi_v6): Likewise.
5373 (*mulsi3addsi_compare0): Likewise.
5374 (*mulsi3addsi_compare0_v6): Likewise.
5375 (*mulsi3addsi_compare0_scratch): Likewise.
5376 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5377 (*mulsi3subsi): Likewise.
5378 (*mulsidi3adddi): Likewise.
5379 (*mulsi3addsi_v6): Likewise.
5380 (*mulsidi3adddi_v6): Likewise.
5381 (*mulsidi3_nov6): Likewise.
5382 (*mulsidi3_v6): Likewise.
5383 (*umulsidi3_nov6): Likewise.
5384 (*umulsidi3_v6): Likewise.
5385 (*umulsidi3adddi): Likewise.
5386 (*umulsidi3adddi_v6): Likewise.
5387 (*smulsi3_highpart_nov6): Likewise.
5388 (*smulsi3_highpart_v6): Likewise.
5389 (*umulsi3_highpart_nov6): Likewise.
5390 (*umulsi3_highpart_v6): Likewise.
5391 (mulhisi3): Likewise.
5392 (*mulhisi3tb): Likewise.
5393 (*mulhisi3bt): Likewise.
5394 (*mulhisi3tt): Likewise.
5395 (maddhisi4): Likewise.
5396 (*maddhisi4tb): Likewise.
5397 (*maddhisi4tt): Likewise.
5398 (maddhidi4): Likewise.
5399 (*maddhidi4tb): Likewise.
5400 (*maddhidi4tt): Likewise.
5402 (udivsi3): Likewise.
5403 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5404 (thumb2_mulsi_short_compare0): Likewise.
5405 (thumb2_mulsi_short_compare0_scratch): Likewise.
5406 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5407 (1020mult2): Likewise.
5408 (1020mult3): Likewise.
5409 (1020mult4): Likewise.
5410 (1020mult5): Likewise.
5411 (1020mult6): Likewise.
5412 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5414 (cortex_a15_mult64): Likewise.
5415 (cortex_a15_sdiv): Likewise.
5416 (cortex_a15_udiv): Likewise.
5417 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5423 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5424 (pj4_ir_div): Likewise.
5425 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5426 (11_mult2): Likewise.
5427 (11_mult3): Likewise.
5428 (11_mult4): Likewise.
5429 (11_mult5): Likewise.
5430 (11_mult6): Likewise.
5431 (11_mult7): Likewise.
5432 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5433 (cortex_a8_mla): Likewise.
5434 (cortex_a8_mull): Likewise.
5435 (cortex_a8_smulwy): Likewise.
5436 (cortex_a8_smlald): Likewise.
5437 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5438 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5439 (cortex_r4_mul_3): Likewise.
5440 (cortex_r4_mla_4): Likewise.
5441 (cortex_r4_mla_3): Likewise.
5442 (cortex_r4_smlald): Likewise.
5443 (cortex_r4_mull): Likewise.
5444 (cortex_r4_sdiv): Likewise.
5445 (cortex_r4_udiv): Likewise.
5446 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5447 (cortex_a7_idiv): Likewise.
5448 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5449 (9_mult2): Likewise.
5450 (9_mult3): Likewise.
5451 (9_mult4): Likewise.
5452 (9_mult5): Likewise.
5453 (9_mult6): Likewise.
5454 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5455 (cortex_a53_sdiv): Likewise.
5456 (cortex_a53_udiv): Likewise.
5457 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5458 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5459 (mp626_mult2): Likewise.
5460 (mp626_mult3): Likewise.
5461 (mp626_mult4): Likewise.
5462 * config/arm/fa526.md (526_mult1): Update attribute change.
5463 (526_mult2): Likewise.
5464 * config/arm/arm-generic.md (mult): Update attribute change.
5465 (mult_ldsched_strongarm): Likewise.
5466 (mult_ldsched): Likewise.
5467 (multi_cycle): Likewise.
5468 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5469 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5470 (606te_mult2): Likewise.
5471 (606te_mult3): Likewise.
5472 (606te_mult4): Likewise.
5473 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5474 (cortex_a9_mac16): Likewise.
5475 (cortex_a9_multiply): Likewise.
5476 (cortex_a9_mac): Likewise.
5477 (cortex_a9_multiply_long): Likewise.
5478 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5479 (626te_mult2): Likewise.
5480 (626te_mult3): Likewise.
5481 (626te_mult4): Likewise.
5483 2013-06-18 Richard Biener <rguenther@suse.de>
5486 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5488 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5491 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5492 with next_active_insn.
5494 2013-06-18 Alan Modra <amodra@gmail.com>
5496 * config/rs6000/rs6000.h (enum data_align): New.
5497 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5498 (DATA_ABI_ALIGNMENT): Define.
5499 (CONSTANT_ALIGNMENT): Correct comment.
5500 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5501 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5503 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5505 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5506 ATTRIBUTE_UNUSED marking.
5508 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5510 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5511 alternative and update.
5512 (aarch64_dup_lanedi): Delete.
5513 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5514 * config/aarch64/aarch64-simd-builtins.def: Update.
5516 2013-06-17 Richard Biener <rguenther@suse.de>
5518 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5519 (lto_input_scc): Declare.
5520 (lto_input_tree_1): Likewise.
5521 (struct lto_stats_d): Add num_tree_bodies_output and
5522 num_pickle_refs_output.
5523 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5524 (lto_read_tree_1): Split out from ...
5525 (lto_read_tree): ... this.
5526 (lto_input_scc): New function.
5527 (lto_input_tree_1): Split out from ...
5528 (lto_input_tree): ... this. Handle LTO_tree_scc.
5529 (lto_data_in_create): Create the streamer cache without hashes.
5530 * lto-streamer-out.c (create_output_block): Create the streamer
5531 cache with hashes when not doing WPA.
5532 (lto_write_tree_1): Split out from ...
5533 (lto_write_tree): ... this.
5534 (get_symbol_initial_value): New function.
5535 (lto_output_tree_1): Split out from ...
5536 (lto_output_tree): ... this. Write trees as series of SCCs
5537 using a DFS walk via DFS_write_tree.
5538 (struct sccs, struct scc_entry): New types.
5539 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5540 (DFS_write_tree_body): New function.
5541 (DFS_write_tree): Likewise.
5542 (hash_tree): Likewise.
5543 (scc_entry_compare): Likewise.
5544 (hash_scc): Likewise.
5545 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5546 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5547 TREE_CHAIN as regular reference.
5548 (streamer_read_integer_cst): Remove.
5549 (streamer_get_pickled_tree): Adjust.
5550 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5551 of DECL_EXTERNALs in BLOCK_VARS for now.
5552 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5554 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5555 Add hash value argument and record that if hashes are recorded
5557 (streamer_tree_cache_insert_1): Adjust.
5558 (streamer_tree_cache_insert): Likewise.
5559 (streamer_tree_cache_insert_at): Rename to ...
5560 (streamer_tree_cache_replace_tree): ... this and adjust.
5561 (streamer_tree_cache_append): Adjust.
5562 (record_common_node): Likewise.
5563 (streamer_tree_cache_create): Add argument whether to
5564 record hash values together with trees.
5565 (streamer_tree_cache_delete): Adjust.
5566 * tree-streamer.h (struct streamer_tree_cache_d): Add
5568 (streamer_read_integer_cst): Remove.
5569 (streamer_tree_cache_insert): Adjust.
5570 (streamer_tree_cache_append): Likewise.
5571 (streamer_tree_cache_insert_at): Rename to ...
5572 (streamer_tree_cache_replace_tree): ... this and adjust.
5573 (streamer_tree_cache_create): Add argument whether to record hashes.
5574 (streamer_tree_cache_get): Rename to ...
5575 (streamer_tree_cache_get_tree): ... this.
5576 (streamer_tree_cache_get_hash): New function.
5577 * tree.c (cache_integer_cst): New function.
5578 * tree.h (cache_integer_cst): Declare.
5579 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5580 * lto-symtab.c (lto_varpool_replace_node): Only release
5581 DECL_INITIAL of non-prevailing decls.
5582 * varpool.c (varpool_remove_initializer): Do not release
5583 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5585 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5587 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5588 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5589 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5590 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5591 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5592 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5593 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5594 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5595 instead of TARGET_64BIT.
5596 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5597 Require ISA_HAS_<D>DIV.
5599 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5601 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5602 (mips*-*-linux*): Move default with_llsc setting to where other
5604 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5606 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5607 Likewise. Remove default with_tune setting. Move default float
5608 setting to its own block. Handle with_llsc in the same block as above.
5610 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5612 PR rtl-optimization/57425
5613 PR rtl-optimization/57569
5614 * alias.c (write_dependence_p): Add new parameters mem_mode,
5615 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
5616 Changed all callers.
5617 (canon_anti_dependence): New function.
5618 * cse.c (check_dependence): Use canon_anti_dependence.
5619 * cselib.c (cselib_invalidate_mem): Likewise.
5620 * rtl.h (canon_anti_dependence): Declare.
5622 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5624 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
5625 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
5626 ".set mips3" for 64-bit targets.
5628 2013-06-15 Dehao Chen <dehao@google.com>
5630 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
5631 * gimple-low.c (gimple_check_call_matching_types): Likewise.
5632 (gimple_check_call_args): Likewise.
5633 * value-prof.c (check_ic_target): Likewise.
5634 * ipa-inline.c (early_inliner): Likewise.
5635 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
5636 * cgraph.c (cgraph_create_edge_1): Likewise.
5637 (cgraph_make_edge_direct): Likewise.
5639 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
5642 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
5643 rs6000_output_move_128bit to handle emitting quad memory
5644 operations. Set attribute length to 8 bytes.
5646 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
5648 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
5650 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
5651 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
5652 (aarch64_<su>mlsl<mode>): Likewise.
5654 2013-06-14 Mike Stump <mikestump@comcast.net>
5656 * Makefile.in (TARGET_H): Add insn-codes.h.
5658 2013-06-14 Alan Modra <amodra@gmail.com>
5662 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
5663 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
5664 bitfield expansion when EXPAND_MEMORY.
5665 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
5667 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
5669 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
5670 test for clearing quad memory on 32-bit later.
5672 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5674 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
5675 (fold_negate_expr): Likewise.
5676 (fold_real_zero_addition_p): Handle vectors.
5677 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
5679 2013-06-14 Alan Modra <amodra@gmail.com>
5681 * varasm.c (force_const_mem): Revert 2013-06-07 change.
5683 2013-06-13 Jan Hubicka <jh@suse.cz>
5685 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
5686 Local comdats are not externally visible.
5687 * symtab.c (dump_symtab_base): Dump externally visible.
5688 (verify_symtab_base): Verify back links in the symtab hash.
5690 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5692 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
5693 CONVERT_EXPR as equal nodes.
5695 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5697 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
5699 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5701 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
5702 Generalize to complex and vector.
5703 * tree.c (build_all_ones_cst): New function.
5704 * tree.h (build_all_ones_cst): Declare it.
5706 2013-06-13 Alan Modra <amodra@gmail.com>
5708 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
5709 * config/rs6000/rs6000.md (signbittf2): New insn.
5710 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
5711 (abstf2_internal, cmptf_internal2): Likewise.
5712 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
5714 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5715 Pat Haugen <pthaugen@us.ibm.com>
5716 Peter Bergner <bergner@vnet.ibm.com>
5718 * config/rs6000/rs6000.c (emit_load_locked): Add support for
5719 power8 byte, half-word, and quad-word atomic instructions.
5720 (emit_store_conditional): Likewise.
5721 (rs6000_expand_atomic_compare_and_swap): Likewise.
5722 (rs6000_expand_atomic_op): Likewise.
5724 * config/rs6000/sync.md (larx): Add new modes for power8.
5726 (AINT): New mode iterator to include TImode as well as normal
5727 integer modes on power8.
5728 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
5729 that VSX registers are not considered. Use AINT mode iterator
5730 instead of INT1 to allow inclusion of quad word atomic operations
5732 (load_locked<mode>): Likewise.
5733 (store_conditional<mode>): Likewise.
5734 (atomic_compare_and_swap<mode>): Likewise.
5735 (atomic_exchange<mode>): Likewise.
5736 (atomic_nand<mode>): Likewise.
5737 (atomic_fetch_<fetchop_name><mode>): Likewise.
5738 (atomic_nand_fetch<mode>): Likewise.
5739 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
5741 (ATOMIC): On power8, add QImode, HImode modes.
5742 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
5743 modes that promote to SImode.
5744 (load_lockedti): Convert TImode arguments to PTImode, so that we
5745 get a guaranteed even/odd register pair.
5746 (load_lockedpti): Likewise.
5747 (store_conditionalti): Likewise.
5748 (store_conditionalpti): Likewise.
5750 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
5751 atomic load/store instructions.
5754 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
5756 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
5758 (insn_count): New attribute, with most cases extracted from...
5759 (length): ...here. Redefine most cases in terms of insn_count.
5760 (single_insn): Delete.
5761 (can_delay): Use insn_count to check for single instructions.
5762 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
5763 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
5764 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
5765 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
5766 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
5767 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
5768 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
5769 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
5770 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
5771 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
5772 rather than "length".
5773 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
5774 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
5775 Use "insn_count" rather than "length".
5776 * config/mips/mips-dsp.md
5777 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
5778 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
5781 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
5783 PR tree-optimization/57361
5784 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
5786 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
5788 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5790 (aarch64_simd_combine<mode>): New instruction expansion.
5791 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5793 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5794 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5796 2013-06-12 Jan Hubicka <jh@suse.cz>
5798 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5799 decl has when in streaming stage.
5800 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5801 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5803 2013-06-12 Roland Stigge <stigge@antcom.de>
5806 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5808 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5810 PR tree-optimization/57537
5811 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5812 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5814 2013-06-12 Richard Biener <rguenther@suse.de>
5816 * data-streamer.h (streamer_write_char_stream): CSE
5817 obs->current_pointer.
5818 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5819 streamer_write_char_stream manually and optimize the resulting loop.
5820 (streamer_write_hwi_stream): Likewise.
5822 2013-06-12 Jan Hubicka <jh@suse.cz>
5824 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5825 * cgraph.h (varpool_create_empty_node): Declare.
5826 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5828 * symtab.c (symtab_unregister_node): Be lax about missin entries
5830 (symtab_get_node): Update comment.
5831 * varpool.c (varpool_create_empty_node): Break out from ...
5832 (varpool_node_for_decl): ... here.
5833 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5835 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5837 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5838 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5839 part. Use straight-line flow at the end.
5840 <COMPONENT_REF>: Remove superfluous else.
5841 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5843 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5846 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5847 target hook even for !TREE_PUBLIC decls. If no resolution info
5848 is available, return false for common and external decls.
5850 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5852 * config/rl78/constraints.md (U): New constraint.
5853 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5856 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5859 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5860 to allow returning address to AT_PLATFORM name.
5862 2013-06-11 Jan Hubicka <jh@suse.cz>
5864 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5865 * cgraph.h (symtab_node_base): Add weakref flag.
5866 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5867 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5868 (output_weakrefs): Use weakref flag.
5869 * fold-const.c (simple_operand_p): Handle WEAK.
5870 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5871 * ipa.c (varpool_externally_visible_p): Drop weakref.
5872 (function_and_variable_visibility): Update comment; fix weakref
5873 sanity checks; do not clear DECL_WEAK on them.
5874 * lto-cgraph.c (lto_output_node): update.
5875 (lto_output_varpool_node): Update.
5876 (input_overwrite_node): Update.
5877 (input_node): Update.
5878 (input_varpool_node): Update.
5879 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5880 (lto_symtab_merge_symbols): Add sanity check.
5881 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5882 * passes.c (rest_of_decl_compilation): Set static flag, too.
5883 * symtab.c (dump_symtab_base): Dump weakref.
5884 (verify_symtab_base): Sanity check weakrefs.
5885 (symtab_make_decl_local): Remove duplicated code.
5886 (symtab_alias_ultimate_target): Simplify.
5887 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5889 2013-06-11 DJ Delorie <dj@redhat.com>
5891 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5892 (rl78_unwind_word_mode): New.
5894 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5896 * final.c (debug_prefix_maps): Make static.
5898 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5900 * function.c (initial_trampoline): Remove stray copy.
5902 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5904 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5906 2013-06-11 Martin Jambor <mjambor@suse.cz>
5908 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5909 within bounds at the beginning of the function.
5911 2013-06-11 Alan Modra <amodra@gmail.com>
5913 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5915 (get_named_section): Don't NULL !DECL_P decl.
5917 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5919 * doc/invoke.texi (core-avx2): Document.
5921 (atom): Updated with MOVBE.
5923 2013-06-11 Richard Biener <rguenther@suse.de>
5925 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5927 2013-06-11 Anton Blanchard <anton@samba.org>
5929 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5930 correct shift value in little-endian mode.
5932 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5935 * varasm.c (get_variable_align): Move #endif to the right place.
5937 2013-06-10 Cary Coutant <ccoutant@google.com>
5939 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5940 for hash so that hash table traversal order is deterministic.
5942 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5943 Pat Haugen <pthaugen@us.ibm.com>
5944 Peter Bergner <bergner@vnet.ibm.com>
5946 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5947 of vectors in GPRS if they are direct moves or quad word load or
5950 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5952 (direct_move_p): Likewise.
5953 (quad_load_store_p): Likewise.
5955 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5956 classes into bins based on the physical register type.
5957 (reg_class_to_reg_type): Likewise.
5958 (IS_STD_REG_TYPE): Likewise.
5959 (IS_FP_VECT_REG_TYPE): Likewise.
5960 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5961 use direct move instructions.
5962 (reload_gpr_vsx): Likewise.
5963 (reload_vsx_gpr): Likewise.
5964 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5965 information that is a simplification of register classes. Also
5966 precalculate direct move reload helpers.
5967 (direct_move_p): New function to return true if the operation can
5968 be done as a direct move instruciton.
5969 (quad_load_store_p): New function to return true if the operation
5970 is a quad memory operation.
5971 (rs6000_legitimize_address): If quad memory, only allow register
5972 indirect for TImode addresses.
5973 (rs6000_legitimate_address_p): Likewise.
5974 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5975 (rs6000_reload_register_type): Likewise.
5976 (register_to_reg_type): Return register type.
5977 (rs6000_secondary_reload_simple_move): New helper function for
5978 secondary reload and secondary memory needed to identify anything
5979 that is a simple move, and does not need reloading.
5980 (rs6000_secondary_reload_direct_move): New helper function for
5981 secondary reload to identify cases that can be done with several
5982 instructions via the direct move instructions.
5983 (rs6000_secondary_reload_move): New helper function for secondary
5984 reload to identify moves between register types that can be done.
5985 (rs6000_secondary_reload): Add support for quad memory operations
5986 and for direct move.
5987 (rs6000_secondary_memory_needed): Likewise.
5988 (rs6000_debug_secondary_memory_needed): Change argument names.
5989 (rs6000_output_move_128bit): New function to return the move to
5990 use for 128-bit moves, including knowing about the various
5991 limitations of quad memory operations.
5993 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5994 memory operations. call rs6000_output_move_128bit for the actual
5995 instruciton(s) to generate.
5996 (vsx_movti_64bit): Likewise.
5998 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5999 (UNSPEC_P8V_MTVSRWZ): Likewise.
6000 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6001 (UNSPEC_P8V_MTVSRD): Likewise.
6002 (UNSPEC_P8V_XXPERMDI): Likewise.
6003 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6004 (UNSPEC_FUSION_GPR): Likewise.
6005 (FMOVE128_GPR): New iterator for direct move.
6006 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6009 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6010 loads and direct move instructions.
6011 (zero_extendsidi2_lfiwzx): Likewise.
6012 (extendsidi2_lfiwax): Likewise.
6013 (extendsidi2_nocell): Likewise.
6014 (floatsi<mode>2_lfiwax): Likewise.
6016 (floatunssi<mode>2_lfiwzx): Likewise.
6018 (fix_trunc<mode>_stfiwx): Likewise.
6019 (fixuns_trunc<mode>_stfiwx): Likewise.
6020 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6021 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6022 (parity<mode>2_cmpb): Set length/type attr.
6023 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6024 for 'mr.' to fast_compare.
6025 (bpermd_<mode>): Change type attr to popcnt.
6026 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6027 (p8_mtvsrwz_1): Likewise.
6028 (p8_mtvsrwz_2): Likewise.
6029 (reload_fpr_from_gpr<mode>): Likewise.
6030 (p8_mtvsrd_1): Likewise.
6031 (p8_mtvsrd_2): Likewise.
6032 (p8_xxpermdi_<mode>): Likewise.
6033 (reload_vsx_from_gpr<mode>): Likewise.
6034 (reload_vsx_from_gprsf): Likewise.
6035 (p8_mfvsrd_3_<mode>): LIkewise.
6036 (reload_gpr_from_vsx<mode>): Likewise.
6037 (reload_gpr_from_vsxsf): Likewise.
6038 (p8_mfvsrd_4_disf): Likewise.
6039 (multi-word GPR splits): Do not split direct moves or quad memory
6042 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6044 * tree-into-ssa.c (interesting_blocks): Make static.
6046 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6049 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6050 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6051 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6052 (get_variable_align): New function.
6053 (get_variable_section, emit_bss, emit_common,
6054 assemble_variable_contents, place_block_symbol): Use
6055 get_variable_align instead of DECL_ALIGN.
6056 (assemble_noswitch_variable): Add align argument, use it
6057 instead of DECL_ALIGN.
6058 (assemble_variable): Adjust caller. Use get_variable_align
6059 instead of DECL_ALIGN.
6060 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6062 (DATA_ABI_ALIGNMENT): Define.
6063 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6064 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6065 opt is false, only return the psABI mandated alignment increase.
6066 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6067 (DATA_ABI_ALIGNMENT): ... this.
6068 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6069 (DATA_ABI_ALIGNMENT): ... this.
6070 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6071 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6072 (DATA_ABI_ALIGNMENT): ... this.
6073 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6074 * doc/tm.texi: Regenerated.
6076 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6078 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6079 cmp_code to construct REG_EQUAL note.
6081 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6084 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6085 that operands[2] doesn't overlap with operands[0].
6087 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6088 Jan Hubicka <jh@suse.cz>
6090 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6091 hack to mark symbols as used.
6093 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6095 PR rtl-optimization/57559
6096 * lra-constraints.c (process_alt_operands): Don't discourage
6097 memory with known offset for offsetable memory constraint.
6098 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6100 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6102 * varasm.c (struct oc_local_state): Reorder fields.
6103 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6104 and adjust accordingly.
6105 (output_constructor): Reorder initialization code and adjust call to
6106 output_constructor_bitfield.
6108 2013-06-07 Jan Hubicka <jh@suse.cz>
6110 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6112 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6114 * tree-object-size.c (unknown): Make const.
6116 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6118 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6119 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6120 for last alternative in the cpu_facility attribute.
6122 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6125 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6126 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6127 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6128 * config/arm/constraints.md (Dg): New constraint.
6129 * config/arm/neon.md (xordi3_neon): Remove.
6130 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6131 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6133 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6135 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6136 Clean up alternatives.
6138 2013-06-07 Alan Modra <amodra@gmail.com>
6140 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6143 2013-06-07 Alan Modra <amodra@gmail.com>
6145 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6147 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6149 * config/arm/constraints.md (Df): New constraint.
6150 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6151 Correct length attribute for last two alternatives.
6153 2013-06-07 Alan Modra <amodra@gmail.com>
6155 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6156 override user -mfp-in-toc.
6157 (offsettable_ok_by_alignment): Consider just the current access
6158 rather than the whole object, unless BLKmode. Handle
6159 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6160 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6161 for -mcmodel=medium.
6162 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6163 override user -mfp-in-toc or -msum-in-toc. Default to
6164 -mno-fp-in-toc for -mcmodel=medium.
6166 2013-06-06 DJ Delorie <dj@redhat.com>
6168 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6169 TARGET_VALID_POINTER_MODE.
6171 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6172 Pat Haugen <pthaugen@us.ibm.com>
6173 Peter Bergner <bergner@vnet.ibm.com>
6175 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6176 Document new power8 builtins.
6178 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6179 condition code register, to allow 128-bit logical operations to be
6180 done in the VSX or GPR registers.
6181 (nor<mode>3): Use the canonical form for nor.
6182 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6183 vclz*, and vpopcnt* vector instructions.
6184 (nand<mode>3): Likewise.
6185 (orc<mode>3): Likewise.
6186 (clz<mode>2): LIkewise.
6187 (popcount<mode>2): Likewise.
6189 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6190 that only the GPRs are recognized.
6192 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6193 support for new power8 builtins.
6195 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6197 (xscvdpspn): Likewise.
6203 (vpopcnt): Likewise.
6204 (vpopcntb): Likewise.
6205 (vpopcnth): Likewise.
6206 (vpopcntw): Likewise.
6207 (vpopcntd): Likewise.
6212 (eqv_v16qi3): Likewise.
6213 (eqv_v8hi3): Likewise.
6214 (eqv_v4si3): Likewise.
6215 (eqv_v2di3): Likewise.
6216 (eqv_v4sf3): Likewise.
6217 (eqv_v2df3): Likewise.
6219 (nand_v16qi3): Likewise.
6220 (nand_v8hi3): Likewise.
6221 (nand_v4si3): Likewise.
6222 (nand_v2di3): Likewise.
6223 (nand_v4sf3): Likewise.
6224 (nand_v2df3): Likewise.
6226 (orc_v16qi3): Likewise.
6227 (orc_v8hi3): Likewise.
6228 (orc_v4si3): Likewise.
6229 (orc_v2di3): Likewise.
6230 (orc_v4sf3): Likewise.
6231 (orc_v2df3): Likewise.
6233 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6234 allow power8 quad mode in 64-bit.
6235 (rs6000_builtin_vectorized_function): Add support to vectorize
6236 ISA 2.07 count leading zeros, population count builtins.
6237 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6238 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6239 (builtin_function_type): Add vgbbd builtin function which takes an
6241 (altivec_expand_vec_perm_const): Add support for new power8 merge
6244 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6245 that does not include TImdoe for use with 32-bit.
6246 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6248 (UNSPEC_VSX_CVDPSPN): Likewise.
6249 (vsx_xscvdpspn): Likewise.
6250 (vsx_xscvspdpn): Likewise.
6251 (vsx_xscvdpspn_scalar): Likewise.
6252 (vsx_xscvspdpn_directmove): Likewise.
6253 (vsx_and<mode>3): Split logical operations into 32-bit and
6254 64-bit. Add support to do logical operations on TImode as well as
6255 VSX vector types. Allow logical operations to be done in either
6256 VSX registers or in general purpose registers in 64-bit mode. Add
6257 splitters if GPRs were used. For AND, add clobber of CCmode to
6258 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6260 (vsx_and<mode>3_32bit): Likewise.
6261 (vsx_and<mode>3_64bit): Likewise.
6262 (vsx_ior<mode>3): Likewise.
6263 (vsx_ior<mode>3_32bit): Likewise.
6264 (vsx_ior<mode>3_64bit): Likewise.
6265 (vsx_xor<mode>3): Likewise.
6266 (vsx_xor<mode>3_32bit): Likewise.
6267 (vsx_xor<mode>3_64bit): Likewise.
6268 (vsx_one_cmpl<mode>2): Likewise.
6269 (vsx_one_cmpl<mode>2_32bit): Likewise.
6270 (vsx_one_cmpl<mode>2_64bit): Likewise.
6271 (vsx_nor<mode>3): Likewise.
6272 (vsx_nor<mode>3_32bit): Likewise.
6273 (vsx_nor<mode>3_64bit): Likewise.
6274 (vsx_andc<mode>3): Likewise.
6275 (vsx_andc<mode>3_32bit): Likewise.
6276 (vsx_andc<mode>3_64bit): Likewise.
6277 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6278 and xxlorc instructions.
6279 (vsx_eqv<mode>3_64bit): Likewise.
6280 (vsx_nand<mode>3_32bit): Likewise.
6281 (vsx_nand<mode>3_64bit): Likewise.
6282 (vsx_orc<mode>3_32bit): Likewise.
6283 (vsx_orc<mode>3_64bit): Likewise.
6285 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6287 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6289 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6290 (p8_vmrgow): Likewise.
6291 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6292 GPRs to be split under VSX.
6293 (p8v_clz<mode>2): Add power8 count leading zero support.
6294 (p8v_popcount<mode>2): Add power8 population count support.
6295 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6298 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6301 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6303 (vec_nand): Likewise.
6304 (vec_vclz): Likewise.
6305 (vec_vclzb): Likewise.
6306 (vec_vclzd): Likewise.
6307 (vec_vclzh): Likewise.
6308 (vec_vclzw): Likewise.
6309 (vec_vgbbd): Likewise.
6310 (vec_vmrgew): Likewise.
6311 (vec_vmrgow): Likewise.
6312 (vec_vpopcnt): Likewise.
6313 (vec_vpopcntb): Likewise.
6314 (vec_vpopcntd): Likewise.
6315 (vec_vpopcnth): Likewise.
6316 (vec_vpopcntw): Likewise.
6318 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6320 PR rtl-optimization/57468
6321 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6324 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6326 PR rtl-optimization/57459
6327 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6328 type when setting live regs.
6330 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6332 * config/s390/s390.opt (mlra): New option.
6333 * config/s390/s390.c (s390_decompose_address): Check displacement
6334 for all registers for LRA.
6335 (s390_secondary_reload): Don't used secondary reloads for LRA.
6336 (s390_lra_p): New function.
6337 (TARGET_LRA_P): Define.
6338 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6339 of attribute cpu_facility to zarch for the last alternative.
6340 (*cmpmem_short): Ditto.
6342 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6344 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6345 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6346 (arm_expand_prologue): Likewise.
6348 2013-06-06 Teresa Johnson <tejohnson@google.com>
6351 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6352 as this is now done by redirect_edge_and_branch_force.
6353 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6354 barriers, and fix interaction with splitting.
6355 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6356 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6357 reflect changes made in the routine.
6358 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6359 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6360 since this is called in cfglayout mode, and replace partition fixup
6361 with assert as that is now done by force_nonfallthru_and_redirect.
6362 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6363 already be marked with region crossing note.
6364 (insert_section_boundary_note): Make non-static, gate on flag
6365 has_bb_partition, rewrite to also check for multiple partitions.
6366 (rest_of_handle_reorder_blocks): Remove call to
6367 insert_section_boundary_note, now done later during free_cfg.
6368 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6369 * bb-reorder.h (insert_section_boundary_note): Declare.
6370 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6371 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6372 invoke insert_section_boundary_note.
6373 (try_redirect_by_replacing_jump): Remove unnecessary
6374 check for region crossing note.
6375 (fixup_partition_crossing): New function.
6376 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6377 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6378 in non-cfglayout mode.
6379 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6380 remove old code that tried to do this. Emit barrier correctly
6381 when we are in cfglayout mode.
6382 (last_bb_in_partition): New function.
6383 (rtl_split_edge): Correctly fixup partition boundaries.
6384 (commit_one_edge_insertion): Remove old code that tried to
6385 fixup region crossing edge since this is now handled in
6386 split_block, and set up insertion point correctly since
6387 block may now end in a jump.
6388 (verify_hot_cold_block_grouping): Guard against checking when not in
6389 linearized RTL mode.
6390 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6392 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6393 rtl_verify_flow_info, so not called in cfglayout mode.
6394 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6395 (fixup_reorder_chain): Remove old code that attempted to fixup region
6396 crossing note as this is now handled in force_nonfallthru_and_redirect.
6397 (duplicate_insn_chain): Don't duplicate switch section notes.
6398 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6400 * basic-block.h (emit_barrier_after_bb): Declare.
6402 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6404 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6405 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6406 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6408 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6410 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6412 * config/arm/ldmstm.md: Regenerate.
6414 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6416 * config/arm/sync.md (atomic_loaddi_1):
6417 Disable predication for arm_restrict_it.
6418 (arm_load_exclusive<mode>): Likewise.
6419 (arm_load_exclusivesi): Likewise.
6420 (arm_load_exclusivedi): Likewise.
6421 (arm_load_acquire_exclusive<mode>): Likewise.
6422 (arm_load_acquire_exclusivesi): Likewise.
6423 (arm_load_acquire_exclusivedi): Likewise.
6424 (arm_store_exclusive<mode>): Likewise.
6425 (arm_store_exclusive<mode>): Likewise.
6426 (arm_store_release_exclusivedi): Likewise.
6427 (arm_store_release_exclusive<mode>): Likewise.
6429 2013-06-06 Richard Biener <rguenther@suse.de>
6431 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6433 (lto_tag_is_tree_code_p): Adjust.
6434 (lto_tag_is_gimple_code_p): Likewise.
6435 (lto_gimple_code_to_tag): Likewise.
6436 (lto_tag_to_gimple_code): Likewise.
6437 (lto_tree_code_to_tag): Likewise.
6438 (lto_tag_to_tree_code): Likewise.
6439 * data-streamer.h (streamer_write_hwi_in_range): Use
6440 uhwi streaming to stream the normalized range.
6441 (streamer_read_hwi_in_range): Likewise.
6443 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6445 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6446 (predicable_short_it): Likewise.
6447 (predicated): Likewise.
6448 (enabled): Handle above.
6449 (define_cond_exec): Set predicated attribute to yes.
6451 2013-06-05 Mike Stump <mikestump@comcast.net>
6453 * gdbinit.in (__FUNCTION__): Add.
6455 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6457 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6458 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6460 2013-06-05 Jan Hubicka <jh@suse.cz>
6462 * varasm.c (mark_decl_referenced): Revert the removal until targets
6465 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6467 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6468 instead of mark_decl_referenced.
6470 2013-06-05 Jan Hubicka <jh@suse.cz>
6472 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6473 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6474 and symtab_used_from_object_file_p.
6475 (cgraph_make_node_local_1): Clear forced_by_abi.
6476 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6477 * cgraph.h (symtab_node_base): Add forced_by_abi.
6478 (decide_is_variable_needed): Remove.
6479 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6480 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6481 (decide_is_symbol_needed): ... this one; handle symbols in general;
6482 always analyze virtuals; honnor forced_by_abi.
6483 (cgraph_finalize_function): Update.
6484 (varpool_finalize_decl): Update.
6485 (symbol_defined_and_needed): Remove.
6486 (analyze_functions): Update.
6487 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6488 output_refs, input_overwrite_node): Handle forced_by_abi.
6489 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6490 (address_taken_from_non_vtable_p): ... this one.
6491 (comdat_can_be_unshared_p_1): New function.
6492 (cgraph_comdat_can_be_unshared_p): Rename to ...
6493 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6494 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6495 (function_and_variable_visibility): Clear forced_by_abi as needed.
6496 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6497 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6498 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6499 * varpool.c (decide_is_variable_needed): Remove.
6501 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6503 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6504 (arm_option_override): Override arm_restrict_it where appropriate.
6505 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6506 * config/arm/arm.opt (mrestrict-it): New command-line option.
6507 * doc/invoke.texi: Document -mrestrict-it.
6509 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6511 * tsan.c (tsan_atomic_table): Make const.
6513 2013-06-05 Richard Biener <rguenther@suse.de>
6515 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6516 index associated with the tree we are supposed to replace.
6517 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6518 TREE_ASM_WRITTEN as zero for everything but SSA names.
6520 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6522 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6524 2013-06-04 Jan Hubicka <jh@suse.cz>
6526 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6527 (input_node, input_varpool_node): Handle correctly external same
6529 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6530 nodes at ltrans stage.
6532 2013-06-04 Jan Hubicka <jh@suse.cz>
6534 * ipa-inline.c (update_caller_keys): Fix availability test.
6535 (update_callee_keys): Likewise.
6536 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6537 to follow ELF standard.
6539 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6541 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6542 (mips64r5900el-*-elf*): New configurations.
6543 * config/mips/mips-cpus.def (r5900): New processor.
6544 * config/mips/mips-tables.opt: Regenerate.
6545 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6546 (mips_issue_rate): Handle PROCESSOR_R5900.
6547 (mips_reorg_process_insns): Force reorder mode for the R5900.
6548 * config/mips/mips.h (TARGET_MIPS5900): Define.
6549 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6551 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6553 * config/mips/mips.md (processor): Add r5900.
6554 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6556 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6558 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6559 into function to generate MOVI instruction.
6560 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6561 (aarch64_preferred_simd_mode): Turn into wrapper.
6562 (aarch64_output_scalar_simd_mov_immediate): New function.
6563 * config/aarch64/aarch64-protos.h: Add prototype for above.
6565 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6567 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6568 element_char member.
6569 (sizetochar): Return signed char.
6570 (aarch64_simd_valid_immediate): Remove elchar and other
6571 unnecessary variables.
6572 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6573 Calculate element_char as required.
6574 * config/aarch64/aarch64-protos.h: Update and move prototype
6575 for aarch64_output_simd_mov_immediate.
6576 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6579 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6581 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6582 information completed by aarch64_simd_valid_immediate.
6583 (aarch64_legitimate_constant_p): Update arguments.
6584 (aarch64_simd_valid_immediate): Work with struct rather than many
6586 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6587 (aarch64_simd_make_constant): Update arguments.
6588 (aarch64_output_simd_mov_immediate): Work with struct rather than
6589 many pointers. Output immediate directly rather than as operand.
6590 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6592 * config/aarch64/constraints.md (Dn): Update arguments.
6594 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6596 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6598 (aarch64_simd_immediate_valid_for_move): Remove.
6599 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6600 (aarch64_simd_make_constant): Update call.
6601 (aarch64_output_simd_mov_immediate): Update call.
6602 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6604 * config/aarch64/constraints.md (Dn): Update call.
6606 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6608 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6609 return type to bool for prototype.
6610 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6611 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6612 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6614 2013-06-04 Catherine Moore <clm@codesourcery.com>
6616 * config/mips/mips.opt (meva): New.
6617 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
6618 (ASM_SPEC): Handle -meva.
6619 * doc/invoke.texi (meva): Document.
6621 2013-06-04 Alan Modra <amodra@gmail.com>
6623 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
6626 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6628 * rtl.def: Add extra fourth optional field to define_cond_exec.
6629 * gensupport.c (process_one_cond_exec): Process attributes from
6631 * doc/md.texi: Document fourth field in define_cond_exec.
6633 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
6635 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
6636 out the processing order as in store_bit_field_1.
6638 2013-06-04 Jan Hubicka <jh@suse.cz>
6641 * cgraphunit.c (cgraph_process_same_body_aliases): Create
6642 non-VAR_DECL node if it does not exist yet.
6644 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
6646 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
6647 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
6648 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
6649 target_cpu_default setting.
6651 2013-06-03 Teresa Johnson <tejohnson@google.com>
6653 * dumpfile.c (opt_info_switch_p): Change -fopt-info
6654 default to -fopt-info=optimized instead of all.
6655 * doc/invoke.texi: Ditto.
6656 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
6657 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
6658 (execute_vect_slp): Emit BB vectorization success under
6659 MSG_OPTIMIZED_LOCATIONS.
6660 * tree-vect-slp.c (vect_slp_transform_bb): Change
6661 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
6662 * tree-vect-loop.c (vect_transform_loop): Ditto.
6664 2013-06-03 Jason Merrill <jason@redhat.com>
6667 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
6668 Use TARGET_EXPR for C++.
6670 2013-06-03 Jakub Jelinek <jakub@redhat.com>
6672 PR rtl-optimization/57268
6673 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
6674 if DEBUG_INSN_P (insn).
6677 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6679 PR rtl-optimization/57268
6680 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6681 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6683 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6685 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
6686 (ix86_avoid_lea_for_addr): Likewise.
6687 (exact_dependency_1): Likewise.
6688 (ix86_adjust_cost): Likewise.
6689 (swap_top_of_ready_list): Fix formatting and !reload_completed check
6691 (do_reorder_for_imul): Fix typo, formatting and
6692 !reload_completed check removed.
6693 (ix86_sched_reorder): Fix typo and formatting.
6694 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
6697 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
6699 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
6701 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
6703 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
6704 <STRING_CST>: Likewise.
6705 <VECTOR_CST>: Likewise.
6707 2013-06-01 Janus Weil <janus@gcc.gnu.org>
6708 Mikael Morin <mikael@gcc.gnu.org>
6710 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
6711 * config.in: Regenerated.
6712 * configure: Regenerated.
6714 2013-06-01 Jan Hubicka <jh@suse.cz>
6717 * cgraphunit.c (compile): When weakref is not supported,
6718 set up transparent aliases before final output pass.
6719 * varasm.c (assemble_alias): Do not try to do it here.
6721 2013-06-01 Jan Hubicka <jh@suse.cz>
6724 * passes.c (for_per_function): Skip unanalyzed functions.
6726 2013-06-01 Jan Hubicka <jh@suse.cz>
6728 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
6729 (lto_symtab_merge_symbols_1): ... this one.
6730 (lto_symtab_merge_cgraph_nodes): Rename to ...
6731 (lto_symtab_merge_symbols): ... this one; simplify.
6732 * cgraph.c (same_body_aliases_done): Rename to ...
6733 (cpp_implicit_aliases_done): ... this one.
6734 (cgraph_create_function_alias): Update.
6735 (cgraph_same_body_alias): Update.
6736 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
6737 (verify_edge_corresponds_to_fndecl): Simplify.
6738 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
6739 (cgraph_node): Remove same_body_alias.
6740 (varpool_node): Remove alias_of and extra_name_alias.
6741 (same_body_aliases_done): Rename to ..
6742 (cpp_implicit_aliases_done): ... this one.
6743 (symtab_alias_ultimate_target): Add default parameter.
6744 (symtab_resolve_alias): New function.
6745 (fixup_same_cpp_alias_visibility): Declare.
6746 (cgraph_function_node): Add default parameter.
6747 (cgraph_node_asm_name): Likewise.
6748 (cgraph_function_or_thunk_node): Add default parameter; do
6749 not ICE when it is NULL.
6750 (varpool_variable_node): Likewise.
6751 * tree-emutls.c (create_emultls_var): Update.
6752 (ipa_lower_emutls): Update.
6753 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
6754 (cgraph_reset_node): Reset alias info.
6755 (cgraph_finalize_function): Update.
6756 (fixup_same_cpp_alias_visibility): Move to symtab.c.
6757 (analyze_function): Simplify.
6758 (cgraph_process_same_body_aliases): Simplify.
6759 (analyze_functions): Fixup same body aliases.
6760 (handle_alias_pairs): Simplify.
6761 (assemble_thunk): Update.
6762 (assemble_thunks_and_aliases): Update.
6763 (output_weakrefs): Rewrite.
6764 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
6765 (lto_output_varpool_node): Likewise.
6766 (compute_ltrans_boundary): Remve assert.
6767 (get_alias_symbol): New functoin.
6768 (input_node): Rewrite alias handling.
6769 (input_varpool_node): Likewise.
6770 * ipa-pure-const.c (propagate_pure_const): Fix formating.
6771 * ipa.c (process_references): Handle weakrefs correctly.
6772 (symtab_remove_unreachable_nodes): Likewise.
6773 * trans-mem.c (get_cg_data): Update.
6774 (ipa_tm_create_version_alias): Update.
6775 (ipa_tm_execute): Update.
6776 * symtab.c (dump_symtab_base): Dump aliases.
6777 (verify_symtab_base): Verify aliases.
6778 (symtab_node_availability): New function.
6779 (symtab_alias_ultimate_target): Simplify.
6780 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
6781 handle all the fixup cases.
6782 (symtab_resolve_alias): New function.
6783 * passes.c (ipa_write_summaries): Handle weakrefs.
6784 * varpool.c (varpool_analyze_node): Simplify.
6785 (assemble_aliases): Update.
6786 (varpool_create_variable_alias): Simplify.
6787 (varpool_extra_name_alias): Simplify.
6788 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
6789 (lto_symtab_merge_symbols): ... this one.
6791 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6794 PR rtl-optimization/57268
6795 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6796 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6798 2013-06-01 Tobias Burnus <burnus@net-b.de>
6801 2013-05-31 Tobias Burnus <burnus@net-b.de>
6804 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6807 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6809 PR rtl-optimization/57268
6810 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6811 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6813 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6815 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6816 unordered comparison operators when -fno-trapping-math is in effect
6818 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6819 and implement unordered comparison operators properly on the e500.
6821 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6823 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6824 for constant scalar integers.
6825 (simplify_relational_operation_1): Likewise.
6827 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6829 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6830 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6833 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6834 Igor Zamyatin <igor.zamyatin@intel.com>
6836 Silvermont (SLM) architecture performance tuning.
6837 * config/i386/i386.h (enum ix86_tune_indices): Add
6838 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6839 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6841 * config/i386/i386.c (initial_ix86_tune_features)
6842 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6843 (ix86_lea_outperforms): Handle Silvermont tuning.
6844 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6846 (ix86_use_lea_for_mov): Likewise.
6847 (ix86_avoid_lea_for_addr): Likewise.
6848 (ix86_lea_for_add_ok): Likewise.
6849 (exact_dependency_1): New function.
6850 (exact_store_load_dependency): Likewise.
6851 (ix86_adjust_cost): Handle Silvermont tuning.
6852 (do_reoder_for_imul): Likewise.
6853 (swap_top_of_ready_list): New function.
6854 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6856 * config/i386/i386.md (peepholes that split memory operand in fp
6859 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6861 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6862 Remove un-necessary braces.
6864 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6866 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6867 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6869 2013-05-31 Tobias Burnus <burnus@net-b.de>
6872 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6874 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6877 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6878 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6879 * config/arm/neon.md (iordi3_neon): Remove.
6880 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6881 * config/arm/predicates.md (imm_for_neon_logic_operand):
6882 Move to earlier in the file.
6883 (neon_logic_op2): Likewise.
6884 (arm_iordi_operand_neon): New predicate.
6886 2013-05-31 Richard Biener <rguenther@suse.de>
6888 PR tree-optimization/57478
6889 PR tree-optimization/57453
6890 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6893 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6895 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6896 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6898 2013-05-30 Tobias Burnus <burnus@net-b.de>
6899 Thomas Koenig <tkoenig@gcc.gnu.org>
6902 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6903 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6905 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6907 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6909 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6911 * target.def (register_usage_leveling_p): New hook.
6912 * targhooks.c (default_register_usage_leveling_p): New.
6913 * targhooks.h (default_register_usage_leveling_p): New prototype.
6914 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6915 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6916 * doc/tm.texi: Update.
6917 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6919 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6921 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6922 (*insv_reg<mode>): New define_insn.
6924 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6926 PR rtl-optimization/57439
6927 * postreload.c (move2add_valid_value_p): Check that we have
6928 a zero subreg_regno_offset when accessing the register in
6931 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6932 Igor Zamyatin <igor.zamyatin@intel.com>
6934 Silvermont (SLM) architecture pipeline model, tuning and
6936 * config.gcc: Add slm config options and target.
6938 * config/i386/slm.md: New.
6940 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6942 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6944 (ix86_target_macros_internal): Likewise.
6946 * gcc/config/i386/i386.c (slm_cost): New cost.
6947 (m_SLM): New macro flag.
6948 (initial_ix86_tune_features): Set m_SLM.
6949 (x86_accumulate_outgoing_args): Likewise.
6950 (x86_arch_always_fancy_math_387): Likewise.
6951 (processor_target_table): Add slm cost.
6952 (cpu_names): Add slm cpu name.
6953 (x86_option_override_internal): Set SLM ISA.
6954 (ix86_issue_rate): New case PROCESSOR_SLM.
6955 (ia32_multipass_dfa_lookahead): Likewise.
6956 (fold_builtin_cpu): Add slm.
6958 * config/i386/i386.h (TARGET_SLM): New target macro.
6959 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6960 (processor_type): Add PROCESSOR_SLM.
6962 * config/i386/i386.md (cpu): Add new value "slm".
6963 (slm.md): Include slm.md.
6965 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6966 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6968 * config/arm/arm-protos.h: Add and update function protos.
6969 * config/arm/arm.c (use_simple_return_p): New added.
6970 (thumb2_expand_return): Check simple_return flag.
6971 * config/arm/arm.md: Add simple_return and conditional simple_return.
6972 * config/arm/iterators.md: Add iterator for return and simple_return.
6974 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6976 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6977 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6978 (arm_emit_vfp_multi_reg_pop): Likewise.
6979 (thumb2_emit_ldrd_pop): Likewise.
6980 (arm_expand_epilogue): Add misc REG_CFA notes.
6981 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6983 2013-05-29 Lawrence Crowl <crowl@google.com>
6985 * config/arm/t-arm: Update for below.
6987 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6988 Change type to hash_table. Update dependent calls and types.
6990 * config/i386/t-cygming: Update for below.
6992 * config/i386/t-interix: Update for below.
6994 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6995 Change type to hash_table. Update dependent calls and types.
6996 (i386_find_on_wrapper_list::wrappers): Likewise.
6998 * config/ia64/t-ia64: Update for below.
7000 * config/ia64/ia64.c (bundle_state_table):
7001 Change type to hash_table. Update dependent calls and types.
7003 * config/mips/mips.c (mips_reorg_process_insns::htab):
7004 Change type to hash_table. Update dependent calls and types.
7006 * config/sol2.c (solaris_comdat_htab):
7007 Change type to hash_table. Update dependent calls and types.
7009 * config/t-sol2: Update for above.
7011 2013-05-29 Teresa Johnson <tejohnson@google.com>
7013 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7014 functions are not yet marked as defined.
7016 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7017 Pat Haugen <pthaugen@us.ibm.com>
7018 Peter Bergner <bergner@vnet.ibm.com>
7020 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7024 (vrotl<mode>3): Likewise.
7025 (vashl<mode>3): Likewise.
7026 (vlshr<mode>3): Likewise.
7027 (vashr<mode>3): Likewise.
7029 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7030 support for power8 V2DI builtins.
7032 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7033 power8 V2DI builtins.
7034 (vupkhsw): Likewise.
7035 (vupklsw): Likewise.
7036 (vaddudm): Likewise.
7041 (vpkudum): Likewise.
7042 (vpksdss): Likewise.
7043 (vpkudus): Likewise.
7044 (vpksdus): Likewise.
7049 (vsubudm): Likewise.
7050 (vcmpequd): Likewise.
7051 (vcmpgtsd): Likewise.
7052 (vcmpgtud): Likewise.
7053 (vcmpequd_p): Likewise.
7054 (vcmpgtsd_p): Likewise.
7055 (vcmpgtud_p): Likewise.
7056 (vupkhsw): Likewise.
7057 (vupklsw): Likewise.
7058 (vaddudm): Likewise.
7063 (vpksdss): Likewise.
7064 (vpksdus): Likewise.
7065 (vpkudum): Likewise.
7066 (vpkudus): Likewise.
7071 (vsubudm): Likewise.
7073 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7074 support for power8 V2DI instructions.
7076 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7077 power8 V2DI instructions. Combine pack and unpack insns to use an
7078 iterator for each mode. Check whether a particular mode supports
7079 Altivec instructions instead of just checking TARGET_ALTIVEC.
7080 (UNSPEC_VPKUWUM): Likewise.
7081 (UNSPEC_VPKSHSS): Likewise.
7082 (UNSPEC_VPKSWSS): Likewise.
7083 (UNSPEC_VPKUHUS): Likewise.
7084 (UNSPEC_VPKSHUS): Likewise.
7085 (UNSPEC_VPKUWUS): Likewise.
7086 (UNSPEC_VPKSWUS): Likewise.
7087 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7088 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7089 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7090 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7091 (UNSPEC_VUPKHSB): Likewise.
7092 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7093 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7094 (UNSPEC_VUPKHSH): Likewise.
7095 (UNSPEC_VUPKLSB): Likewise.
7096 (UNSPEC_VUPKLSH): Likewise.
7098 (VI_char): Likewise.
7099 (VI_scalar): Likewise.
7100 (VI_unit): Likewise.
7102 (VP_small): Likewise.
7103 (VP_small_lc): Likewise.
7104 (VU_char): Likewise.
7105 (add<mode>3): Likewise.
7106 (altivec_vaddcuw): Likewise.
7107 (altivec_vaddu<VI_char>s): Likewise.
7108 (altivec_vadds<VI_char>s): Likewise.
7109 (sub<mode>3): Likewise.
7110 (altivec_vsubcuw): Likewise.
7111 (altivec_vsubu<VI_char>s): Likewise.
7112 (altivec_vsubs<VI_char>s): Likewise.
7113 (altivec_vavgs<VI_char>): Likewise.
7114 (altivec_vcmpbfp): Likewise.
7115 (altivec_eq<mode>): Likewise.
7116 (altivec_gt<mode>): Likewise.
7117 (altivec_gtu<mode>): Likewise.
7118 (umax<mode>3): Likewise.
7119 (smax<mode>3): Likewise.
7120 (umin<mode>3): Likewise.
7121 (smin<mode>3): Likewise.
7122 (altivec_vpkuhum): Likewise.
7123 (altivec_vpkuwum): Likewise.
7124 (altivec_vpkshss): Likewise.
7125 (altivec_vpkswss): Likewise.
7126 (altivec_vpkuhus): Likewise.
7127 (altivec_vpkshus): Likewise.
7128 (altivec_vpkuwus): Likewise.
7129 (altivec_vpkswus): Likewise.
7130 (altivec_vpks<VI_char>ss): Likewise.
7131 (altivec_vpks<VI_char>us): Likewise.
7132 (altivec_vpku<VI_char>us): Likewise.
7133 (altivec_vpku<VI_char>um): Likewise.
7134 (altivec_vrl<VI_char>): Likewise.
7135 (altivec_vsl<VI_char>): Likewise.
7136 (altivec_vsr<VI_char>): Likewise.
7137 (altivec_vsra<VI_char>): Likewise.
7138 (altivec_vsldoi_<mode>): Likewise.
7139 (altivec_vupkhsb): Likewise.
7140 (altivec_vupkhs<VU_char>): Likewise.
7141 (altivec_vupkls<VU_char>): Likewise.
7142 (altivec_vupkhsh): Likewise.
7143 (altivec_vupklsb): Likewise.
7144 (altivec_vupklsh): Likewise.
7145 (altivec_vcmpequ<VI_char>_p): Likewise.
7146 (altivec_vcmpgts<VI_char>_p): Likewise.
7147 (altivec_vcmpgtu<VI_char>_p): Likewise.
7148 (abs<mode>2): Likewise.
7149 (vec_unpacks_hi_v16qi): Likewise.
7150 (vec_unpacks_hi_v8hi): Likewise.
7151 (vec_unpacks_lo_v16qi): Likewise.
7152 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7153 (vec_unpacks_lo_v8hi): Likewise.
7154 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7155 (vec_pack_trunc_v8h): Likewise.
7156 (vec_pack_trunc_v4si): Likewise.
7157 (vec_pack_trunc_<mode>): Likewise.
7159 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7161 (vec_vmaxsd): Likewise.
7162 (vec_vmaxud): Likewise.
7163 (vec_vminsd): Likewise.
7164 (vec_vminud): Likewise.
7165 (vec_vpksdss): Likewise.
7166 (vec_vpksdus): Likewise.
7167 (vec_vpkudum): Likewise.
7168 (vec_vpkudus): Likewise.
7169 (vec_vrld): Likewise.
7170 (vec_vsld): Likewise.
7171 (vec_vsrad): Likewise.
7172 (vec_vsrd): Likewise.
7173 (vec_vsubudm): Likewise.
7174 (vec_vupkhsw): Likewise.
7175 (vec_vupklsw): Likewise.
7177 2013-05-29 Jan Hubicka <jh@suse.cz>
7179 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7180 flags; reorder rest of fields in more consistent way.
7181 (varpool_node): Remove analyzed, finalized and alias.
7182 (cgraph_ndoe): Likewise.
7183 (symtab_alias_ultimate_target): New function.
7184 (cgraph_function_node): Move offline.
7185 (cgraph_reset_node): Declare.
7186 (cgraph_comdat_can_be_unshared_p): Remove.
7187 (varpool_remove_initializer): Declare.
7188 (varpool_first_defined_variable, varpool_next_defined_variable
7189 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7190 (cgraph_function_with_gimple_body_p): Update.
7191 (varpool_all_refs_explicit_p): Update.
7192 (symtab_alias_target): New function.
7193 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7194 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7195 (cgraph_function_or_thunk_node): Simplify using
7196 symtab_alias_ultimate_target.
7197 (varpool_variable_node): Likewise.
7198 * cgraph.c (cgraph_create_function_alias): Update.
7199 (cgraph_add_thunk): Update.
7200 (cgraph_remove_node): Update.
7201 (dump_cgraph_node): Do not dump removed flags.
7202 (cgraph_function_body_availability): Update.
7203 (cgraph_propagate_frequency): Update.
7204 (verify_cgraph_node): Check sanity of local flag.
7205 (cgraph_function_node): Move here from cgraph.h; revamp for
7206 cgraph_function_or_thunk_node.
7207 * lto-symtab.c (lto_varpool_replace_node): Update.
7208 (lto_symtab_resolve_can_prevail_p): Update.
7209 (lto_symtab_merge_cgraph_nodes): Update.
7210 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7211 propagate_constants_accross_call, devirtualization_time_bonus,
7212 ipcp_propagate_stage): Update.
7213 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7214 * ipa-inline-transform.c (clone_inlined_nodes,
7215 preserve_function_body_p): Update.
7216 * ipa-reference.c (propagate): Update.
7217 (write_node_summary_p): Update.
7218 * toplev.c (wrapup_global_declaration_2): Update.
7219 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7220 (analyze_function) ... this one.
7221 (cgraph_process_new_functions): Update.
7222 (cgraph_reset_node): Export.
7223 (cgraph_finalize_function): Update.
7224 (cgraph_add_new_function): Update.
7225 (process_function_and_variable_attributes): Update.
7226 (varpool_finalize_decl): Update.
7227 (symbol_finalized): Remove.
7228 (symbol_finalized_and_needed): Rename to ...
7229 (symbol_defined_and_needed): ... update.
7230 (cgraph_analyze_functions): Update.
7231 (handle_alias_pairs): Update.
7232 (mark_functions_to_output): Update.
7233 (assemble_thunk): Update.
7234 (output_in_order): Update.
7235 (output_weakrefs): Update.
7236 (finalize_compilation_unit): Update.
7237 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7238 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7239 input_node, input_varpool_node): Update.
7240 * dbxout.c (dbxout_expand_expr): Update.
7241 * cgraphclones.c (cgraph_clone_node): Update.
7242 (cgraph_copy_node_for_versioning): Update.
7243 (cgraph_materialize_clone): Update.
7244 (cgraph_materialize_all_clones): Update.
7245 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7246 propagate_pure_const, propagate_nothrow): Update.
7247 * lto-streamer-out.c (lto_output, write_symbol): Update.
7248 * ipa-utils.c (ipa_reverse_postorder): Update.
7249 * ipa-inline.c (can_inline_edge_p): Update.
7250 (update_caller_keys, ipa_inline): Update.
7251 * dwarf2out.c (reference_to_unused,
7252 premark_types_used_by_global_vars_helper): Update.
7253 * tree-eh.c (tree_could_trap_p): Update.
7254 * ipa-split.c (consider_split, execute_split_functions): Update.
7255 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7256 has_addr_references_p): Update; move ahead in file for better
7258 (process_references): Simplify.
7259 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7261 (cgraph_comdat_can_be_unshared_p): Make static.
7262 (cgraph_externally_visible_p): Update.
7263 (varpool_externally_visible_p): Update.
7264 (function_and_variable_visibility): Update.
7265 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7266 ipa_tm_mark_force_output_node): Update.
7267 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7268 estimate_edge_devirt_benefit, inline_generate_summary,
7269 inline_write_summary): Update.
7270 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7271 * ipa-prop.c (ipa_compute_jump_functions): Update.
7272 (ipa_print_node_params, ipa_prop_read_section,
7273 ipa_update_after_lto_read, read_replacements_section): Update.
7274 * varasm.c (mark_decl_referenced): Update.
7275 (assemble_alias, dump_tm_clone_pairs): Update.
7276 * tree-inline.c (copy_bb): Update.
7277 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7279 * symtab.c (dump_symtab_base): Print new flags.
7280 (verify_symtab_base): Verify new flags.
7281 (symtab_alias_ultimate_target): New function.
7282 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7283 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7285 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7287 * i386.c (ix86_get_function_versions_dispatcher,
7288 ix86_generate_version_dispatcher_body): Update.
7289 (fold_builtin_cpu): Use varpool_add_new_variable.
7290 * varpool.c (varpool_remove_initializer): Break out from ...
7291 (varpool_remove_node): ... this one.
7292 (dump_varpool_node, varpool_node_for_asm,
7293 cgraph_variable_initializer_availability, varpool_analyze_node,
7294 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7295 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7298 2013-05-29 Jan Hubicka <jh@suse.cz>
7300 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7302 2013-05-29 Easwaran Raman <eraman@google.com>
7304 PR tree-optimization/57442
7305 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7306 when control exits the main loop.
7308 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7310 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7312 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7313 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7314 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7317 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7319 PR tree-optimization/57441
7320 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7321 Don't limit size of incr_vec to number of candidates.
7323 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7325 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7326 and mips16 directories.
7327 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7328 (MULTILIB_DIRNAMES): Ditto.
7329 (MULTILIB_EXCEPTIONS): Add new exceptions.
7330 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7331 (MULTILIB_DIRNAMES): Ditto.
7332 (MULTILIB_EXCEPTIONS): Add new exceptions.
7334 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7335 Marcus Shawcroft <marcus.shawcroft@arm.com>
7337 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7338 SYMBOL_TINY_ABSOLUTE.
7339 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7340 SYMBOL_TINY_ABSOLUTE.
7341 (aarch64_expand_mov_immediate): Likewise.
7342 (aarch64_classify_symbol): Likewise.
7343 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7344 Permit SYMBOL_TINY_ABSOLUTE.
7345 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7347 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7348 Marcus Shawcroft <marcus.shawcroft@arm.com>
7350 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7351 Refactor if/switch. Replace gcc_assert with if.
7353 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7355 * config/i386/i386.c (initial_ix86_tune_features): Enable
7356 FP Reassociation for AMD bdver1 and bdver2.
7358 2013-05-29 Martin Jambor <mjambor@suse.cz>
7360 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7361 and IMAGPART_EXPR do not occur within other handled_components.
7363 2013-05-29 Richard Biener <rguenther@suse.de>
7365 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7366 access on whether the use is in the BB we currently try to
7368 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7369 vectorize to vect_bb_slp_scalar_cost.
7371 2013-05-29 Richard Biener <rguenther@suse.de>
7373 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7374 computing scalar cost offsetted by stmts that are kept live
7376 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7377 for computation of scalar cost.
7379 2013-05-28 Steve Ellcey <sellcey@mips.com>
7381 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7383 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7385 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7386 array notation built-in reduction functions.
7387 * doc/passes.texi (Passes): Added documentation about changes done
7389 * doc/invoke.texi (C Dialect Options): Added documentation about
7390 the -fcilkplus flag.
7391 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7392 (BUILTINS_DEF): Depend on cilkplus.def.
7393 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7394 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7395 * cilkplus.def: New file.
7397 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7399 PR rtl-optimization/57439
7400 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7402 2013-05-28 Easwaran Raman <eraman@google.com>
7404 PR tree-optimization/57337
7405 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7406 (find_insert_point): Correctly identify the insertion point
7407 when two statements with the same UID is compared.
7409 2013-05-28 Richard Biener <rguenther@suse.de>
7411 PR tree-optimization/56787
7412 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7413 from the list of data references.
7414 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7416 (vect_analyze_loop_operations): Likewise.
7417 (vect_transform_loop): Remove clobbers.
7419 2013-05-28 Martin Jambor <mjambor@suse.cz>
7421 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7422 and REALPART_EXPRs have scalar type.
7424 2013-05-28 Richard Biener <rguenther@suse.de>
7426 PR tree-optimization/57411
7427 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7429 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7430 virtual operand propagation.
7432 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7434 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7435 destination register for bmasksi_vis.
7436 (vector_init_bshuffle): Likewise.
7437 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7439 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7441 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7442 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7443 mode if the instruction isn't available in the original mode.
7444 * config/sparc/sparc.opt (mfix-ut699): New option.
7445 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7446 (divdf3): Turn into expander.
7447 (divdf3_nofix): New insn.
7448 (divdf3_fix): Likewise.
7449 (divsf3): Disable if -mfix-ut699.
7450 (sqrtdf2): Turn into expander.
7451 (sqrtdf2_nofix): New insn.
7452 (sqrtdf2_fix): Likewise.
7453 (sqrtsf2): Disable if -mfix-ut699.
7455 2013-05-27 Richard Biener <rguenther@suse.de>
7458 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7459 block for the new loop.
7461 2013-05-27 Richard Biener <rguenther@suse.de>
7463 PR tree-optimization/57343
7464 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7465 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7466 (number_of_iterations_cond): Do not build the folded tree.
7468 2013-05-27 Richard Biener <rguenther@suse.de>
7472 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7473 OEP_CONSTANT_ADDRESS_OF retained.
7475 PR tree-optimization/57417
7476 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7478 (set_ssa_val_to): Compare addresses using
7479 get_addr_base_and_unit_offset.
7481 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7483 PR rtl-optimization/56833
7484 * postreload.c (move2add_record_mode): New function.
7485 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7486 (move2add_use_add2_insn): Use move2add_record_sym_value.
7487 (move2add_use_add3_insn): Likewise.
7488 (reload_cse_move2add): Use move2add_valid_value_p and
7489 move2add_record_mode. Invalidate call-clobbered and REG_INC
7490 affected regs by setting reg_mode to VOIDmode.
7491 (move2add_note_store): Don't pretend the inside of a SUBREG is
7492 the actual destination. Invalidate single/leading registers by
7493 setting reg_mode to VOIDmode.
7494 Use move2add_record_sym_value, move2add_valid_value_p and
7495 move2add_record_mode.
7497 2013-05-27 Richard Biener <rguenther@suse.de>
7499 PR tree-optimization/57396
7500 * tree-affine.c (double_int_constant_multiple_p): Properly
7501 return false for val == 0 and div != 0.
7503 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7505 * config/mips/mips.h: Use #elif in preprocessor conditions.
7507 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7510 * config/mips/constraints.md (kl): New constraint.
7511 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7512 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7513 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7515 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7517 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7520 * config/mips/mips.c (mips_can_inline_p): New function.
7521 (TARGET_CAN_INLINE_P): Define.
7523 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7525 * sched-int.h (ds_t, dw_t): Make unsigned int.
7526 Fix documentation that describes how all the ds_t bits are used.
7527 Reserve the last bit for delayed-branch scheduling.
7528 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7529 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7530 (gen_dep_weak_1): Remove prototype.
7531 * sched-deps.c (get_dep_weak_1): Make static.
7532 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7533 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7534 * doc/tm.texi: Regenerate.
7535 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7537 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7540 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7542 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7543 Sandra Loosemore <sandra@codesourcery.com>
7545 * config.gcc (powerpc-*): Allow native for with-cpu.
7547 2013-05-24 Jeff Law <law@redhat.com>
7549 PR tree-optimization/57124
7550 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7551 conversion feeding a condition if the range has an overflow
7552 if -fstrict-overflow. Add warnings for when we do make the
7555 2013-05-24 Dehao Chen <dehao@google.com>
7557 * tree-cfg.c (locus_discrim_map): Fix the typo.
7558 (locus_discrim_hasher): Likewise.
7559 (locus_discrim_hasher::hash): Likewise.
7560 (locus_discrim_hasher::equal): Likewise.
7562 2013-05-24 Martin Jambor <mjambor@suse.cz>
7564 PR tree-optimization/57294
7565 * cgraph.h (ipa_record_stmt_references): Declare.
7566 * cgraphbuild.c (ipa_record_stmt_references): New function.
7567 (build_cgraph_edges): Use ipa_record_stmt_references.
7568 (rebuild_cgraph_edges): Likewise.
7569 (cgraph_rebuild_references): Likewise.
7570 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7571 associated with the old statement and build references from the
7572 newly built statements.
7573 * ipa-ref.c (ipa_remove_stmt_references): New function.
7574 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7576 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7578 * lra-constraints.c (emit_spill_move): Use smaller mode for
7580 (check_and_process_move): Consider mem-reg moves for secondary
7582 (curr_insn_transform): Don't lose insns emitted before for
7583 secondary memory moves.
7584 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7585 reg set up in the current insn.
7587 2013-05-24 Dehao Chen <dehao@google.com>
7589 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7591 (locus_descrim_hasher::equal): Likewise.
7592 (build_gimple_cfg): New discriminator assignment algorithm.
7593 (make_edges): Likewise.
7594 (next_discriminator_for_locus): Likewise.
7595 (same_line_p): Likewise.
7596 (assign_discriminators): Likewise.
7597 (make_cond_expr_edges): Likewise.
7598 (make_gimple_switch_edges): Likewise.
7599 (make_goto_expr_edges): Likewise.
7600 (make_gimple_asm_edges): Likewise.
7602 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7604 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7605 X format specifier to only display bottom 16 bits.
7606 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7607 immediate to match for operand 2, since it will be masked.
7609 2013-05-24 Richard Biener <rguenther@suse.de>
7611 PR tree-optimization/57287
7612 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7613 all SSA names that occur in abnormal PHIs.
7615 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
7618 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
7619 that index is not negative.
7621 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
7623 PR rtl-optimization/55177
7624 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
7625 (simplify_byte_swapping_operation): New.
7626 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
7627 (simplify_relational_operation_1): Deal with BSWAP.
7629 2013-05-23 Richard Henderson <rth@redhat.com>
7632 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
7633 (ix86_reorg): Call it.
7635 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
7638 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
7639 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
7640 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
7642 2013-05-23 Christian Bruel <christian.bruel@st.com>
7645 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
7647 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7648 Marcus Shawcroft <marcus.shawcroft@arm.com>
7650 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
7651 * config/aarch64/constraints.md (Usa): Remove.
7652 * doc/md.texi (AArch64 Usa): Remove.
7654 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7655 Marcus Shawcroft <marcus.shawcroft@arm.com>
7657 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
7658 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
7659 * config/aarch64/predicates.md (aarch64_const_address): Remove.
7660 (aarch64_mov_operand): Use aarch64_mov_operand_p.
7662 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
7664 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
7665 instruction (AdvSIMD).
7666 * config/aarch64/aarch64-builtins.c
7667 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
7668 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
7670 2013-05-23 Martin Jambor <mjambor@suse.cz>
7673 * tree.h (contains_bitfld_component_ref_p): Declare.
7674 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
7675 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
7677 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
7678 not access a bit-field. Assert all final offsets are byte-aligned.
7680 2013-05-23 Richard Biener <rguenther@suse.de>
7682 PR tree-optimization/57380
7683 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
7684 least one invariant or re-used load.
7685 * passes.c (init_optimization_passes): Move pass_phiprop before
7688 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
7690 * config/aarch64/aarch64-simd.md
7691 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
7693 2013-05-23 Richard Biener <rguenther@suse.de>
7696 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7697 OEP_CONSTANT_ADDRESS_OF retained.
7699 2013-05-23 Jakub Jelinek <jakub@redhat.com>
7702 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
7703 don't lower unit. Handle unit not being always BITS_PER_WORD.
7705 2013-05-23 Richard Biener <rguenther@suse.de>
7707 PR rtl-optimization/57341
7708 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
7709 instead of true_dependence.
7711 2013-05-22 David Malcolm <dmalcolm@redhat.com>
7713 * bb-reorder.c (branch_threshold): Make const.
7714 (exec_threshold): Ditto.
7716 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7717 Pat Haugen <pthaugen@us.ibm.com>
7718 Peter Bergner <bergner@vnet.ibm.com>
7720 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
7721 documentation for the power8 crypto builtins.
7723 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
7725 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
7726 macros for defining power8 builtin functions.
7727 (BU_P8V_AV_2): Likewise.
7728 (BU_P8V_AV_P): Likewise.
7729 (BU_P8V_VSX_1): Likewise.
7730 (BU_P8V_OVERLOAD_1): Likewise.
7731 (BU_P8V_OVERLOAD_2): Likewise.
7732 (BU_CRYPTO_1): Likewise.
7733 (BU_CRYPTO_2): Likewise.
7734 (BU_CRYPTO_3): Likewise.
7735 (BU_CRYPTO_OVERLOAD_1): Likewise.
7736 (BU_CRYPTO_OVERLOAD_2): Likewise.
7737 (XSCVSPDP): Fix typo, point to the correct instruction.
7738 (VCIPHER): Add power8 crypto builtins.
7739 (VCIPHERLAST): Likewise.
7740 (VNCIPHER): Likewise.
7741 (VNCIPHERLAST): Likewise.
7742 (VPMSUMB): Likewise.
7743 (VPMSUMH): Likewise.
7744 (VPMSUMW): Likewise.
7745 (VPERMXOR_V2DI): Likewise.
7746 (VPERMXOR_V4SI: Likewise.
7747 (VPERMXOR_V8HI: Likewise.
7748 (VPERMXOR_V16QI: Likewise.
7749 (VSHASIGMAW): Likewise.
7750 (VSHASIGMAD): Likewise.
7752 (VPERMXOR): Likewise.
7753 (VSHASIGMA): Likewise.
7755 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7756 __CRYPTO__ if the crypto instructions are available.
7757 (altivec_overloaded_builtins): Add support for overloaded power8
7760 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
7761 support for power8 crypto builtins.
7762 (builtin_function_type): Likewise.
7763 (altivec_init_builtins): Add support for builtins that take vector
7764 long long (V2DI) arguments.
7766 * config/rs6000/crypto.md: New file, define power8 crypto
7769 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7770 Pat Haugen <pthaugen@us.ibm.com>
7771 Peter Bergner <bergner@vnet.ibm.com>
7773 * doc/invoke.texi (Option Summary): Add power8 options.
7774 (RS/6000 and PowerPC Options): Likewise.
7776 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
7777 constraints.md instead of rs6000.h. Reorder w* constraints. Add
7778 wm, wn, wr documentation.
7780 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
7781 registers if direct move instructions are enabled.
7782 (wn): New constraint for no registers.
7783 (wq): New constraint for quad word even GPR registers.
7784 (wr): New constraint if 64-bit instructions are enabled.
7785 (wv): New constraint if power8 vector instructions are enabled.
7786 (wQ): New constraint for quad word memory locations.
7788 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
7789 constraint for 0..15 for crypto instructions.
7790 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7791 as GPR and floating point registers.
7792 (int_reg_operand): New predicate to match only GPR registers.
7793 (base_reg_operand): New predicate to match base registers.
7794 (quad_int_reg_operand): New predicate to match even GPR registers
7795 for quad memory operations.
7796 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7797 operations in both GPR and VSX registers.
7798 (quad_memory_operand): New predicate for quad memory operations.
7799 (reg_or_indexed_operand): New predicate for direct move support.
7801 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7802 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7803 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7804 (POWERPC_MASKS): Add power8 options.
7805 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7808 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7809 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7811 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7812 (-mpower8-fusion): New power8 options.
7813 (-mpower8-fusion-sign): Likewise.
7814 (-mpower8-vector): Likewise.
7815 (-mcrypto): Likewise.
7816 (-mdirect-move): Likewise.
7817 (-mquad-memory): Likewise.
7819 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7821 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7823 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7824 (rs6000_debug_vector_unit): Add p8_vector.
7825 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7826 definitions. Also print fusion state.
7827 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7828 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7829 (rs6000_option_override_internal): Add support for power8.
7830 (rs6000_common_init_builtins): Add debugging for skipped builtins
7832 (rs6000_adjust_cost): Add power8 support.
7833 (rs6000_issue_rate): Likewise.
7834 (insn_must_be_first_in_group): Likewise.
7835 (insn_must_be_last_in_group): Likewise.
7836 (force_new_group): Likewise.
7837 (rs6000_register_move_cost): Likewise.
7838 (rs6000_opt_masks): Likewise.
7840 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7841 power8 capable assembler, default to power7 options.
7842 (TARGET_DIRECT_MOVE): Likewise.
7843 (TARGET_CRYPTO): Likewise.
7844 (TARGET_P8_VECTOR): Likewise.
7845 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7846 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7847 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7848 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7849 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7850 (TARGET_XSCVDPSPN): Likewise.
7851 (TARGET_XSCVSPDPN): Likewsie.
7852 (TARGET_SYNC_HI_QI): Likewise.
7853 (TARGET_SYNC_TI): Likewise.
7854 (MASK_CRYPTO): Likewise.
7855 (MASK_DIRECT_MOVE): Likewise.
7856 (MASK_P8_FUSION): Likewise.
7857 (MASK_P8_VECTOR): Likewise.
7858 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7859 temporary used by some of the direct move instructions to get two FP
7860 temporary registers does not force creation of a stack frame.
7861 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7862 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7863 that any VSX registers are tieable, even if they are also an
7864 Altivec vector mode.
7865 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7866 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7867 (RS6000_BTM_CRYPTO): Likewise.
7868 (RS6000_BTM_COMMON): Likewise.
7870 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7871 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7872 (enum rs6000_vector): Add power8 vector support.
7874 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7878 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7879 (any_sibcall_could_use_r3): this and handle indirect calls.
7880 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7882 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7884 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7886 2013-05-22 Richard Biener <rguenther@suse.de>
7889 * profile.c (branch_prob): Do not split blocks that are
7890 abnormally receiving from ECF_RETURNS_TWICE functions.
7892 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7894 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7895 address mode. Move pointer mode initialization to the same place.
7897 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7899 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7900 while it has any effect.
7902 2013-05-21 Easwaran Raman <eraman@google.com>
7904 PR tree-optimization/57322
7905 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7906 UID of the statement added to the BB to be 1.
7908 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7910 PR tree-optimization/57331
7911 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7912 of conversion from pointer type to integral type with integer.
7914 2013-05-21 Martin Jambor <mjambor@suse.cz>
7917 * ipa-prop.c (ipa_read_node_info): Process param_used and
7918 controlled_uses in the same order as when writing.
7920 2013-05-21 Magnus Granberg <baldrick@free.fr>
7923 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7925 2013-05-21 Richard Biener <rguenther@suse.de>
7927 PR tree-optimization/57318
7928 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7929 estimate stmts with side-effects as likely eliminated.
7931 2013-05-21 Richard Biener <rguenther@suse.de>
7933 PR tree-optimization/57330
7934 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7935 preserve the call stmts fntype.
7937 2013-05-21 Richard Biener <rguenther@suse.de>
7939 PR tree-optimization/57303
7940 * tree-ssa-sink.c (statement_sink_location): Improve killing
7941 stmt detection and properly handle self-assignments.
7943 2013-05-21 Christian Bruel <christian.bruel@st.com>
7945 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7946 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7947 registers. Set register size out of the PARALLEL loop.
7949 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7952 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7953 (*fmasf4, *fmasf4_media): New insns.
7955 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7957 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7958 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7959 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7960 (mips_idiv_insns): Update the comments to say that the returned
7961 instruction counts are in units of BASE_INSN_LENGTH.
7962 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7963 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7964 using 2 rather than 4 as the length of indirect MIPS16 and
7965 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7966 length of a NOP. Don't divide MIPS16 lengths by 2.
7967 (mips16_split_long_branches): Assume a branch is long if the
7968 length is greater than 4 rather than 8.
7969 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7970 rather than multiplying them by 2. Multiply instruction counts
7971 by BASE_INSN_LENGTH rather than 4.
7972 (*jump_mips16, tls_get_tp_mips16_<mode>)
7973 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7975 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7977 * config/mips/mips.md (extended_mips16): Remove branch case.
7978 (length): Remove duplicated extended_mips16 test.
7980 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7982 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7984 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7986 * recog.h (Recog_data): Rename to...
7987 (recog_data_d): ...this.
7988 (recog_data): Update accordingly.
7989 * recog.c (recog_data): Likewise.
7990 * reload.c (save_recog_data): Likewise.
7991 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7992 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7994 2013-05-17 Julian Brown <julian@codesourcery.com>
7996 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7997 found in a REG_EQUAL note, invalidate it.
7999 2013-05-17 Easwaran Raman <eraman@google.com>
8001 * tree-ssa-reassoc.c (find_insert_point): New function.
8002 (insert_stmt_after): Likewise.
8003 (get_def_stmt): Likewise.
8004 (ensure_ops_are_available): Likewise.
8005 (not_dominated_by): Likewise.
8006 (rewrite_expr_tree): Do not move statements beyond what is
8007 necessary. Remove call to swap_ops_for_binary_stmt...
8008 (reassociate_bb): ... and move it here.
8009 (build_and_add_sum): Assign UIDs for new statements.
8010 (linearize_expr): Likewise.
8011 (do_reassoc): Renumber gimple statement UIDs.
8013 2013-05-17 Jan Hubicka <jh@suse.cz>
8015 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8017 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8018 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8020 (output_weakrefs): Update.
8022 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8023 Martin Jambor <mjambor@suse.cz>
8026 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8027 value that corresponds to the given aggval is found in values vector.
8029 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8031 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8032 sse, sse2, sse3, ssse3 and sse4a flags to options.
8034 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8036 * gengtype-state.c: (s_expr_writer): New class, to handle
8037 prettifying of output layout of s-expressions.
8038 (state_writer): New class, to write out gtype.state.
8039 (state_written_type_count): Move this variable into member data of
8041 (s_expr_writer::s_expr_writer): New code: constructor for new class
8042 (state_writer::state_writer(): ditto
8043 (s_expr_writer::write_new_line): New function
8044 (s_expr_writer::write_any_indent): ditto
8045 (s_expr_writer::begin_s_expr): ditto
8046 (s_expr_writer::end_s_expr): ditto
8047 (write_state_fileloc): convert to method of state_writer...
8048 (state_writer:: write_state_fileloc): ...and use methods of
8049 s_expr_writer to write indentation into the gtype.state output file
8050 to visually represent the hierarchical structure of the list
8052 (write_state_fields): ditto, renaming to...
8053 (state_writer::write_state_fields)
8054 (write_state_a_string): ditto, renaming to...
8055 (state_writer::write_state_a_string)
8056 (write_state_string_option): ditto, renaming to...
8057 (state_writer::write_state_string_option)
8058 (write_state_type_option): ditto, renaming to...
8059 (state_writer::write_state_type_option)
8060 (write_state_nested_option): ditto, renaming to...
8061 (state_writer::write_state_nested_option)
8062 (write_state_option): ditto, renaming to...
8063 (state_writer::write_state_option)
8064 (write_state_options): ditto, renaming to...
8065 (state_writer::write_state_options)
8066 (write_state_lang_bitmap): ditto, renaming to...
8067 (state_writer::write_state_lang_bitmap)
8068 (write_state_version): ditto, renaming to...
8069 (state_writer::write_state_version)
8070 (write_state_scalar_type): ditto, renaming to...
8071 (state_writer::write_state_scalar_type)
8072 (write_state_string_type): ditto, renaming to...
8073 (state_writer::write_state_string_type)
8074 (write_state_undefined_type): ditto, renaming to...
8075 (state_writer::write_state_undefined_type)
8076 (write_state_struct_union_type): ditto, renaming to...
8077 (state_writer::write_state_struct_union_type)
8078 (write_state_struct_type): ditto, renaming to...
8079 (state_writer::write_state_struct_type)
8080 (write_state_user_struct_type): ditto, renaming to...
8081 (state_writer::write_state_user_struct_type)
8082 (write_state_lang_struct_type): ditto, renaming to...
8083 (state_writer::write_state_lang_struct_type)
8084 (write_state_param_struct_type): ditto, renaming to...
8085 (state_writer::write_state_param_struct_type)
8086 (write_state_pointer_type): ditto, renaming to...
8087 (state_writer::write_state_pointer_type)
8088 (write_state_array_type): ditto, renaming to...
8089 (state_writer::write_state_array_type)
8090 (write_state_gc_used): ditto, renaming to...
8091 (state_writer::write_state_gc_used)
8092 (write_state_common_type_content): ditto, renaming to...
8093 (state_writer::write_state_common_type_content)
8094 (write_state_type): ditto, renaming to...
8095 (state_writer::write_state_type)
8096 (write_state_pair_list): ditto, renaming to...
8097 (state_writer::write_state_pair_list)
8098 (write_state_pair): ditto, renaming to...
8099 (state_writer::write_state_pair)
8100 (write_state_typedefs): ditto, renaming to...
8101 (state_writer::write_state_typedefs)
8102 (write_state_structures): ditto, renaming to...
8103 (state_writer::write_state_structures)
8104 (write_state_param_structs): ditto, renaming to...
8105 (state_writer::write_state_param_structs)
8106 (write_state_variables): ditto, renaming to...
8107 (state_writer::write_state_variables)
8108 (write_state_srcdir): ditto, renaming to...
8109 (state_writer::write_state_srcdir)
8110 (write_state_files_list): ditto, renaming to...
8111 (state_writer::write_state_files_list)
8112 (write_state_languages): ditto, renaming to...
8113 (state_writer::write_state_languages)
8114 (write_state): create a state_writer instance and use it when
8115 writing out the state file
8117 2013-05-17 Mike Stump <mikestump@comcast.net>
8119 PR rtl-optimization/57304
8120 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8121 accessing DF_REF_REAL_LOC.
8123 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8125 PR rtl-optimization/57281
8126 PR rtl-optimization/57300
8127 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8128 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8129 what the other splitter did if the registers are dead.
8131 2013-05-17 Richard Biener <rguenther@suse.de>
8133 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8136 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8138 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8141 2013-05-17 Marek Polacek <polacek@redhat.com>
8143 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8144 length when doing non-zero store of storing '\0' to '\0'.
8146 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8148 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8149 vect_external_def oprnd1 with loop_vinfo, try to emit
8150 optional cast, negation and and stmts on the loop preheader
8151 edge instead of into the pattern def seq.
8153 PR tree-optimization/57051
8154 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8155 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8157 2013-05-16 Nick Clifton <nickc@redhat.com>
8159 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8160 (rl78_is_naked_func): New function.
8161 (rl78_expand_prologue): Skip prologue generation for naked functions.
8162 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8163 * doc/extend.texi (naked): Add RL78 to the list of processors
8164 that supports this attribute.
8166 2013-05-16 Jeff Law <law@redhat.com>
8168 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8170 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8172 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8173 cache parameters using detect_caches_amd also for CYRIX,
8174 NSC and TM2 signatures.
8176 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8177 Dzianis Kahanovich <mahatma@eu.by>
8181 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8182 VIA/Centaur processors and determine their cache parameters
8183 using detect_caches_amd.
8185 2013-05-16 Teresa Johnson <tejohnson@google.com>
8187 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8188 (rtl_verify_edges): New function.
8189 (rtl_verify_bb_insns): Ditto.
8190 (rtl_verify_bb_pointers): Ditto.
8191 (rtl_verify_bb_insn_chain): Ditto.
8192 (rtl_verify_fallthru): Ditto.
8193 (rtl_verify_bb_layout): Ditto.
8194 (rtl_verify_flow_info_1): Outline checks into new functions.
8195 (rtl_verify_flow_info): Ditto.
8197 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8199 * cfghooks.c (copy_bbs): Add update_dominance argument.
8200 * cfghooks.h (copy_bbs): Update prototype.
8201 * tree-cfg.c (gimple_duplicate_sese_region):
8202 Add update_dominance argument.
8203 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8204 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8205 gimple_duplicate_sese_region call.
8206 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8207 Update copy_bbs call.
8208 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8209 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8211 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8213 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8214 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8215 vect_recog_rotate_pattern.
8216 (vect_recog_rotate_pattern): New function.
8218 2013-05-16 Jason Merrill <jason@redhat.com>
8220 * Makefile.in (LLINKER): New variable.
8221 (mostlyclean): Remove link mutex.
8222 * configure.ac: Handle --enable-link-mutex.
8223 * lock-and-run.sh: New script.
8225 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8228 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8231 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8233 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8235 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8237 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8238 * config/arm/arm.c (next_consecutive_mem): New function.
8239 (gen_movmem_ldrd_strd): Likewise.
8240 * config/arm/arm.md (movmemqi): Update condition and code.
8241 (unaligned_loaddi, unaligned_storedi): New patterns.
8243 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8245 * config.gcc: Obsolete *-*-solaris2.9*.
8246 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8248 2013-05-16 Richard Biener <rguenther@suse.de>
8250 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8251 earlier, after GRAPHITE transforms and IV canonicalization.
8253 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8255 * omp-low.c (extract_omp_for_data): For collapsed loops,
8256 if at least one of the loops is known at compile time to
8257 iterate zero times, set count to 0.
8258 (expand_omp_regimplify_p): New function.
8259 (expand_omp_for_generic): For collapsed loops, if at least
8260 one of the loops isn't known to iterate at least once,
8261 add runtime check with setting count to 0.
8262 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8263 For unsigned types if it isn't known at compile time that
8264 the loop will iterate at least once, add runtime check to bypass
8265 the whole loop if initial condition isn't true.
8267 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8269 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8271 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8274 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8275 transformations to avoid an infinite loop.
8277 2013-05-16 Marek Polacek <polacek@redhat.com>
8279 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8281 2013-05-15 Leif Ekblad <leif@rdos.net>
8283 * config/i386/i386.c (ix86_decompose_address): Use
8284 DEFAULT_TLS_SEG_REG to access TLS segment register.
8285 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8286 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8287 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8289 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8292 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8293 sibling calls to functions that would normally be lazily bound,
8294 unless $gp is call-clobbered.
8296 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8298 * config/i386/i386.c (ix86_option_override_internal): Update
8299 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8300 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8301 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8302 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8304 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8306 2013-05-15 Andreas Schwab <schwab@suse.de>
8308 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8309 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8312 2013-05-15 Teresa Johnson <tejohnson@google.com>
8314 * loop-unroll.c (report_unroll_peel): Check decision before
8315 emitting unroll/peel message.
8317 2013-05-15 Teresa Johnson <tejohnson@google.com>
8319 * function.h (has_bb_partition): New rtl_data flag.
8320 (bb_reorder_complete): Ditto.
8321 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8322 instead of flag_reorder_blocks_and_partition.
8323 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8324 with some enhancements.
8325 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8326 * bb-reorder.c (connect_traces): Check for has_bb_partition
8327 instead of flag_reorder_blocks_and_partition.
8328 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8329 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8330 verify_hot_cold_block_grouping.
8331 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8333 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8336 * config/arm/predicates.md (call_insn_operand): New predicate.
8337 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8338 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8339 if insn is not a tail call.
8340 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8342 * config/arm/arm.h (enum reg_class): New caller save register class.
8343 (REG_CLASS_NAMES): Likewise.
8344 (REG_CLASS_CONTENTS): Likewise.
8345 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8348 2013-05-15 Richard Biener <rguenther@suse.de>
8350 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8351 of MSG_OPTIMIZED_LOCATIONS.
8352 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8353 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8355 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8356 of MSG_OPTIMIZED_LOCATIONS.
8357 (execute_vect_slp): Likewise.
8358 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8359 (vect_create_cond_for_alias_checks): Likewise.
8360 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8361 (vect_recog_widen_mult_pattern): Likewise.
8362 (vect_recog_widen_sum_pattern): Likewise.
8363 (vect_recog_over_widening_pattern): Likewise.
8364 (vect_recog_widen_shift_pattern): Likewise.
8365 (vect_recog_vector_vector_shift_pattern): Likewise.
8366 (vect_recog_divmod_pattern): Likewise.
8367 (vect_recog_mixed_size_cond_pattern): Likewise.
8368 (vect_recog_bool_pattern): Likewise.
8369 (vect_pattern_recog_1): Likewise.
8371 2013-05-15 Martin Jambor <mjambor@suse.cz>
8373 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8374 non-functions to builtin_unreachable.
8375 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8376 correct when new direct edges were discovered.
8378 2013-05-15 Martin Jambor <mjambor@suse.cz>
8380 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8381 header, print symbol order instead of node uid, print more information
8382 about indirect edge targets.
8383 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8385 (ipa_make_edge_direct_to_target): Likewise.
8386 (remove_described_reference): Likewise.
8387 (propagate_controlled_uses): Likewise.
8388 (ipa_print_node_params): Also print symbol order.
8389 (ipcp_transform_function): Print symbol order instead of node uids.
8390 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8391 (cgraph_get_create_real_symbol_node): Likewise.
8392 * ipa-cp.c (print_lattice): Likewise.
8393 (print_all_lattices): Likewise.
8394 (determine_versionability): Likewise.
8395 (initialize_node_lattices): Likewise.
8396 (estimate_local_effects): Likewise.
8397 (update_profiling_info): Likewise.
8398 (create_specialized_node): Likewise.
8399 (perhaps_add_new_callers): Likewise.
8400 (decide_about_value): Likewise.
8401 (decide_whether_version_node): Likewise.
8402 (identify_dead_nodes): Likewise.
8403 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8404 (dump_inline_summary): Likewise.
8405 (estimate_node_size_and_time): Likewise.
8406 (inline_analyze_function): Likewise.
8407 * ipa-inline.c (report_inline_failed_reason): Likewise.
8408 (want_early_inline_function_p): Likewise.
8409 (edge_badness): Likewise.
8410 (update_edge_key): Likewise.
8411 (inline_small_functions): Likewise. Add dumping of order to two other
8413 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8414 instead of node uids.
8415 (propagate_pure_const): Likewise.
8416 (propagate_pure_const): Likewise.
8417 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8418 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8419 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8421 * tree-pretty-print.c (dump_function_header): Likewise.
8422 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8423 Print symbol order instead of node uids.
8425 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8427 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8428 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8430 2013-05-15 Richard Biener <rguenther@suse.de>
8432 PR tree-optimization/57275
8433 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8434 return value for fail to do runtime alias checks for gather loads.
8436 2013-05-15 Jan Hubicka <jh@suse.cz>
8440 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8441 weakrefs are not external.
8442 (lto_symtab_merge_decls): Fix thinko when dealing with
8443 non-lto_symtab decls.
8444 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8445 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8446 * varpool.c (dump_varpool_node): Dump more flags.
8448 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8450 * config/i386/i386.c (processor_alias_table): Add instruction
8451 FSGSBASE for AMD bdver3 architecture.
8453 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8455 * tree.c (warn_deprecated_use): Print file:line using locus color.
8456 * diagnostic.c (diagnostic_report_current_module): Print file:line
8457 and file:line:column using locus color.
8459 2013-05-14 Mike Stump <mikestump@comcast.net>
8461 * gdbinit.in: Add __null.
8463 2013-05-14 Mike Stump <mikestump@comcast.net>
8465 * recog.h: Rename struct recog_data to Recog_data.
8466 * recog.c: Likewise.
8467 * reload.c (can_reload_into): Likewise.
8468 * config/picochip/picochip.c: Likewise.
8470 2013-05-14 Mike Stump <mikestump@comcast.net>
8472 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8474 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8476 * resource.h (struct resources): Remove unch_memory member.
8477 (CLEAR_RESOURCE): Don't clear unch_memory.
8478 * resource.c (mark_referenced_resources): Don't set it.
8479 (mark_set_resources): Likewise.
8480 (mark_target_live_regs): Don't clear it.
8481 (init_resource_info): Likewise.
8482 * reorg.c (resource_conflicts_p): Don't compare it.
8483 (redundant_insn): Don't set it.
8485 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8487 * emit-rtl.c (next_label): Remove unused function.
8488 (skip_consecutive_labels, link_cc0_insns): Move to ...
8489 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8490 only place where these functions are used, and make them static.
8492 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8494 * fold-const.c (fold_negate_expr): Handle vectors.
8495 (fold_truth_not_expr): Make it static.
8496 (fold_invert_truthvalue): New static function.
8497 (invert_truthvalue_loc): Handle vectors. Do not call
8498 fold_truth_not_expr directly.
8499 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8500 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8501 (fold_comparison): Handle vector constants.
8502 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8503 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8504 * tree.h (fold_truth_not_expr): Remove declaration.
8506 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8508 * config/aarch64/aarch64-simd.md
8509 (aarch64_vcond_internal<mode>): Rename to...
8510 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8511 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8512 float modes. Clarify all iterator modes.
8513 (vcond<mode><mode>): Use new name for vcond expanders.
8514 (vcond<v_cmp_result><mode>): Likewise.
8515 (vcondu<mode><mode>: Likewise.
8516 * config/aarch64/iterators.md (VDQF_COND): New.
8518 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8521 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8522 variable for the shift amount. Check that we shift by non-negative
8525 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8528 * config/arm/arm.h (EPILOGUE_USES): Only return true
8529 for LR_REGNUM after epilogue_completed.
8531 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8533 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8536 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8539 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8540 and Solaris 11+/x86 with gld.
8541 * configure: Regenerate.
8543 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8545 * expmed.c (expand_shift_1): Canonicalize rotates by
8546 constant bitsize / 2 to bitsize - 1.
8547 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8548 case ROTATERT>: Likewise.
8551 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8553 * config/i386/i386.md (rotateinv): New code attr.
8554 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8555 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8556 roll $31, %eax, etc.
8558 2013-05-14 Richard Biener <rguenther@suse.de>
8561 * tree-eh.c (sink_clobbers): Give up for successors with
8562 multiple predecessors and no virtual uses.
8564 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8566 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8567 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8569 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8572 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8573 the case when both op0 and op1 have VOIDmode.
8575 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8577 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8578 in multiply-accumulate mode.
8580 2013-05-13 Guozhi Wei <carrot@google.com>
8582 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8584 2013-05-13 Kai Tietz <ktietz@redhat.com>
8587 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8588 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8589 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8590 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8591 (ix86_expand_prologue): Likewise.
8592 (ix86_expand_split_stack_prologue): Likewise.
8593 (legitimate_pic_address_disp_p): Likewise.
8594 (legitimize_pic_address): Likewise.
8595 (legitimize_tls_address): Likewise.
8596 (legitimize_pe_coff_symbol): Likewise.
8597 (output_pic_addr_const): Likewise.
8598 (construct_plt_address): Likewise.
8599 (ix86_expand_call): Likewise.
8600 (x86_output_mi_thunk): Likewise.
8601 (x86_function_profiler): Likewise.
8603 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8605 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8606 similar switch cases.
8607 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8608 (aarch64_simd_mov_to_<mode>low): Delete.
8609 (aarch64_simd_mov_to_<mode>high): Delete.
8610 (move_lo_quad_<mode>): Add w<-r alternative.
8611 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8612 (aarch64_simd_mov_from_*): Update type attribute.
8613 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8616 2013-05-13 Jan Hubicka <jh@suse.cz>
8618 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
8619 * config/i386/i386.c (ix86_compute_frame_layout,
8620 ix86_expand_epilogue, emit_i387_cw_initialization,
8621 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
8622 ix86_local_alignment): Fix use of size/speed predicates.
8624 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8626 PR tree-optimization/45216
8627 PR tree-optimization/57157
8628 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
8629 the (-Y) & (B - 1) variant if OP is |.
8630 * expmed.c (expand_shift_1): For rotations by const0_rtx just
8631 return shifted. Use (-op1) & (prec - 1) as other_amount
8632 instead of prec - op1.
8634 2013-05-13 Martin Jambor <mjambor@suse.cz>
8637 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
8638 (ipa_constant_data): New type.
8639 (ipa_jump_func): Use ipa_constant_data to hold information about
8640 constant jump functions.
8641 (ipa_get_jf_constant): Adjust to jump function type changes.
8642 (ipa_get_jf_constant_rdesc): New function.
8643 (ipa_param_descriptor): New field controlled_uses.
8644 (ipa_get_controlled_uses): New function.
8645 (ipa_set_controlled_uses): Likewise.
8646 * ipa-ref.h (ipa_find_reference): Declare.
8647 * ipa-prop.c (ipa_cst_ref_desc): New type.
8648 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
8650 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
8651 New parameter cs. Adjust all callers.
8652 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
8653 (remove_described_reference): New function.
8654 (jfunc_rdesc_usable): Likewise.
8655 (try_make_edge_direct_simple_call): Decrement controlled use count,
8656 attempt to remove reference if it hits zero.
8657 (combine_controlled_uses_counters): New function.
8658 (propagate_controlled_uses): Likewise.
8659 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
8660 (ipa_edge_duplication_hook): Duplicate reference descriptions.
8661 (ipa_print_node_params): Print described use counter.
8662 (ipa_write_jump_function): Adjust to jump function type changes.
8663 (ipa_read_jump_function): New parameter CS, pass it to
8664 ipa_set_jf_constant. Adjust caller.
8665 (ipa_write_node_info): Stream controlled use count
8666 (ipa_read_node_info): Likewise.
8667 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
8669 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
8670 count. Remove cloning-added reference if it reaches zero.
8671 * ipa-ref.c (ipa_find_reference): New function.
8673 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8675 * config/i386/i386.c (processor_target_table): Modified default
8676 alignment values for AMD BD and BT architectures.
8678 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
8680 * tree-vect-generic.c (uniform_vector_p): Move ...
8681 * tree.c (uniform_vector_p): ... here.
8682 * tree.h (uniform_vector_p): Declare it.
8683 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
8686 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8688 PR tree-optimization/57230
8689 * tree-ssa-strlen.c (handle_char_store): Record length for
8690 array store from STRING_CST.
8692 PR tree-optimization/57230
8693 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
8696 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
8698 * config/epiphany/epiphany.c (epiphany_init): Check size of
8699 NUM_MODES_FOR_MODE_SWITCHING.
8700 (epiphany_expand_prologue):
8701 Remove CONFIG_REGNUM initial value handling code.
8702 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
8703 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
8704 (emit_set_fp_mode, epiphany_mode_after): Likewise.
8705 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
8706 Don't return 1 for FP_MODE_NONE.
8707 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
8708 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
8709 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
8710 * config/epiphany/epiphany.md (save_config): New pattern.
8712 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
8714 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
8716 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
8718 * config/i386/i386.md (memory): Handle sseishft1.
8719 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
8720 (*vec_extractv2di_1): Ditto.
8722 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
8724 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
8727 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
8729 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
8730 Add mthumb/march=armv7-a multilib.
8731 Add mthumb/march=armv7-r multilib.
8732 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
8734 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
8736 * config/v850/t-rtems: Add more multilibs.
8738 2013-05-10 Richard Biener <rguenther@suse.de>
8740 PR tree-optimization/57214
8741 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
8742 not propagate from SSA names that occur in abnormal PHI nodes.
8744 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
8746 * stor-layout.c (element_precision): New function.
8747 * machmode.h (element_precision): Declare it.
8748 * tree.c (build_minus_one_cst): New function.
8749 (element_precision): Likewise.
8750 * tree.h (build_minus_one_cst): Declare new function.
8751 (element_precision): Likewise.
8752 * fold-const.c (operand_equal_p): Use element_precision.
8753 (fold_binary_loc): Handle vector types.
8754 * convert.c (convert_to_integer): Use element_precision.
8755 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
8758 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
8760 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
8761 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
8762 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
8763 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8764 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
8765 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
8766 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
8767 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8768 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
8769 (Uuw8): New constraints.
8770 (Usb4): Move into alphabetical order.
8771 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
8772 (sd8_operand, ub8_operand, uw8_operand): New predicates.
8773 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
8774 previously unnamed patterns.
8775 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
8776 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
8777 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
8778 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
8779 of set_attr_alternative/if_then_else. Use extended_mips16 instead
8780 of specific lengths.
8782 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8784 * config/i386/i386.md (rotateinv): New code attr.
8785 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8786 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8787 roll $31, %eax, etc.
8789 PR tree-optimization/45216
8790 PR tree-optimization/57157
8791 * tree-ssa-forwprop.c (simplify_rotate): New function.
8792 (ssa_forward_propagate_and_combine): Call it.
8794 2013-05-10 Richard Biener <rguenther@suse.de>
8796 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8797 disable peeling when we version for aliasing.
8798 (vector_alignment_reachable_p): Honor explicit user alignment.
8799 (vect_supportable_dr_alignment): Likewise.
8800 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8801 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8802 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8803 then peeling to arrange for the cost-model check to come first.
8805 2013-05-10 Alan Modra <amodra@gmail.com>
8807 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8808 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8809 * configure: Regenerate.
8811 2013-05-10 Alan Modra <amodra@gmail.com>
8814 * varasm.c (default_elf_select_section): Move !DECL_P check..
8815 (get_named_section): ..to here before calling get_section_name.
8817 (default_section_type_flags): Add DECL_P check.
8818 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8819 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8821 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8823 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8824 When using gen_stack_adjust_str with a register offset, add a
8825 REG_FRAME_RELATED_EXPR note.
8827 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8829 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8830 (*vec_extractv4si_zext_mem): Ditto.
8831 (*vec_extractv2di): Add 0->x and x->x alternatives.
8832 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8833 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8835 2013-05-09 Jason Merrill <jason@redhat.com>
8837 N3639 C++1y VLA support
8838 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8841 * tree.c (build_constructor_va): New.
8842 * tree.h: Declare it.
8844 2013-05-09 Martin Jambor <mjambor@suse.cz>
8847 * gimple-fold.c (canonicalize_constructor_val): Call
8848 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8850 2013-05-09 Jan Hubicka <jh@suse.cz>
8851 Richard Biener <rguenther@suse.de>
8854 * symtab.c (symtab_make_decl_local): Do not add private names.
8856 2013-05-09 Jan Hubicka <jh@suse.cz>
8859 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8860 (unlink_from_assembler_name_hash): Likewise.
8861 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8862 symtab_unregister_node, symtab_initialize_asm_name_hash,
8863 change_decl_assembler_name): Update.
8865 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8867 * config/aarch64/aarch64.md: New movtf split.
8868 (*movtf_aarch64): Update.
8869 (aarch64_movdi_tilow): Handle TF modes and rename to
8870 aarch64_movdi_<mode>low.
8871 (aarch64_movdi_tihigh): Handle TF modes and rename to
8872 aarch64_movdi_<mode>high
8873 (aarch64_movtihigh_di): Handle TF modes and rename to
8874 aarch64_mov<mode>high_di
8875 (aarch64_movtilow_di): Handle TF modes and rename to
8876 aarch64_mov<mode>low_di
8877 (aarch64_movtilow_tilow): Remove spurious whitespace.
8878 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8880 (aarch64_print_operand): Update.
8882 2013-05-09 Alan Modra <amodra@gmail.com>
8884 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8886 * configure: Regenerate.
8888 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8890 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8891 splitter preparation statements.
8892 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8893 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8896 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8898 * gimple-ssa-strength-reduction.c (count_candidates): Change
8899 return value to int.
8900 (analyze_candidates_and_replace): Change type of length to int.
8902 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8904 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8905 (*vec_extract<mode>): Use VI12_128 mode iterator.
8906 (*vec_extract<mode>_mem): Ditto.
8907 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8910 2013-05-08 Diego Novillo <dnovillo@google.com>
8915 2012-08-17 Diego Novillo <dnovillo@google.com>
8918 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8919 * config.in: Regenerate.
8920 * configure: Regenerate.
8921 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8923 2013-05-08 Jan Hubicka <jh@suse.cz>
8926 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8927 * cgraph.h (symtab_node_base): Add unique_name.
8928 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8929 input_overwrite_node, input_varpool_node): Stream unique_name.
8930 * cgraphclones.c (cgraph_create_virtual_clone,
8931 cgraph_function_versioning): Set unique_name.
8932 * ipa.c (function_and_variable_visibility): Set unique_name.
8934 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8936 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8937 (alloc_cand_and_find_basis): Restrict conditional candidate
8938 processing to CAND_MULTs.
8940 2013-05-08 Jan Hubicka <jh@suse.cz>
8943 lto-symtab.c (lto_symtab_symbol_p): New function.
8944 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8945 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8946 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8947 Skip static symbols.
8949 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8951 PR tree-optimization/57200
8952 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8953 Only call inform if the preceding warning_at returns true.
8955 2013-05-07 Han Shen <shenhan@google.com>
8957 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8958 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8959 * common.opt (fstack-protector-strong): New option.
8960 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8961 * doc/invoke.texi (Optimization Options): Document
8962 "-fstack-protector-strong".
8963 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8965 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8967 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8969 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8971 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8972 it is not lhs of the insn.
8973 (setup_reg_equiv): Remove insn having equiv memory which it is not
8975 * lra-constraints.c (process_address): Try to improve generation
8976 code for address base + disp.
8977 (lra_constraints): Make correct the code for checking insn setting
8978 up backward equivalence. Remove insn only if it is in the init
8980 * lra-eliminations.c (update_reg_eliminate): Change return value.
8981 (lra_eliminate): Use the result.
8983 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8985 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8986 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8987 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8988 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8989 register target operands.
8990 (*vec_extractv8hi_sse2): New pattern.
8991 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8992 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8993 (*vec_extract<mode>_mem): New insn and split pattern.
8995 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8997 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8998 (TARGET_ASAN_SHADOW_OFFSET): Define.
8999 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9000 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9002 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9004 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9005 (incr_vec_index): Return -1 if increment not found.
9006 (create_add_on_incoming_edge): Assert if increment not found.
9007 (record_increment): Limit number of increments recorded.
9008 (all_phi_incrs_profitable): Return false if an increment not found.
9009 (replace_profitable_candidates): Don't process increments that were
9011 (analyze_candidates_and_replace): Limit size of incr_vec.
9013 2013-05-07 Richard Biener <rguenther@suse.de>
9015 * calls.c (special_function_p): setjmp-like functions are leaf.
9016 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9017 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9019 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9021 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9022 (aarch64_simd_mov<mode>): New expander.
9023 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9024 (aarch64_simd_mov_to_<mode>high): Likewise.
9025 (aarch64_simd_mov_from_<mode>low): Likewise.
9026 (aarch64_simd_mov_from_<mode>high): Likewise.
9027 (aarch64_dup_lane<mode>): Update.
9028 (aarch64_dup_lanedi): New instruction pattern.
9029 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9030 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9032 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9034 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9035 (replace_mult_candidate): Remove unnecessary argument; remove
9036 unnecessary parameter from call to introduce_cast_before_cand.
9037 (replace_unconditional_candidate): Remove unnecessary parameter
9038 from call to replace_mult_candidate.
9039 (replace_conditional_candidate): Likewise.
9040 (insert_initializers): Use make_temp_ssa_name.
9041 (introduce_cast_before_cand): Remove unnecessary argument; use
9043 (replace_one_candidate): Remove unnecessary argument; remove
9044 unnecessary parameter from calls to introduce_cast_before_cand.
9045 (replace_profitable_candidates): Remove unnecessary parameters
9046 from calls to replace_one_candidate.
9048 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9050 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9051 phi def as possibly hiding a basis for a CAND_ADD whose operands
9052 have been commuted in the analysis.
9053 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9055 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9057 * config/aarch64/aarch64.md
9058 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9059 shift value between 0-4.
9061 2013-05-07 Richard Biener <rguenther@suse.de>
9063 * double-int.h (rshift): New overload.
9064 * double-int.c (rshift): New function.
9065 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9066 (create_reference_ops_from_ref): Remove.
9067 (vn_reference_insert): Use shared ops for constructing the
9068 reference and copy it.
9070 2013-05-07 Richard Biener <rguenther@suse.de>
9073 * tree-eh.c (sink_clobbers): Properly propagate
9074 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9076 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9078 PR tree-optimization/57149
9079 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9080 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9081 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9082 uninit_undefined_value_p instead of ssa_undefined_value_p.
9085 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9086 for modifier == EXPAND_INITIALIZER.
9088 2013-05-07 Anton Blanchard <anton@samba.org>
9090 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9091 for powerpc64 little endian.
9092 * configure: Regenerate.
9094 2013-05-06 Graham Stott <grahams@btinternet.com>
9096 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9097 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9098 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9099 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9101 2013-05-06 Graham Stott <grahams@btinternet.com>
9103 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9104 codes which allow non-lvalues.
9106 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9108 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9109 components are all 1s.
9110 (integer_minus_onep): New function.
9111 * tree.h (integer_minus_onep): Declare it.
9112 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9113 integer_minus_onep instead of integer_all_onesp.
9115 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9118 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9119 variations of these patterns.
9121 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9123 * config/i386/i386.md (isa): Add x64_sse4 member.
9124 (enabled): Handle x64_sse4.
9125 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9126 instruction for 64bit SSE4_1 targets. Update insn attributes.
9127 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9128 instruction for SSE4_1 targets. Update insn attributes.
9129 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9130 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9131 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9133 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9134 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9135 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9137 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9140 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9142 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9144 * final.c (do_assembler_dialects): Don't handle curly braces and
9145 vertical bar escaped by % as dialect delimiters.
9146 (output_asm_insn): Print curly braces and vertical bar if escaped
9147 by % and ASSEMBLER_DIALECT defined.
9148 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9149 * doc/tm.texi: Regenerated.
9151 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9153 * config/mips/mips.c: Include tree-pass.h.
9154 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9155 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9156 (pass_mips_machine_reorg2): New machine specific pass.
9157 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9158 (mips_option_override): Register the new pass.
9159 * rtl.h (cleanup_barriers): Remove prototype.
9160 (dbr_schedule): Likewise.
9161 * jump.c (cleanup_barriers): Make static.
9162 * reorg.c (dbr_schedule): Likewise.
9164 2013-05-06 Richard Biener <rguenther@suse.de>
9166 PR tree-optimization/57185
9167 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9170 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9173 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9174 between "first_arg" and "insn", not "last" and "insn".
9176 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9178 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9179 (find_candidates_in_block): Re-enable slsr_process_phi.
9180 (create_phi_basis): Fix double counting of candidate adjustment.
9182 2013-05-06 Richard Biener <rguenther@suse.de>
9185 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9186 the edge is also fallthru, preserve it and just clear the
9188 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9189 also complex, preserve that and just clear the fallthru flag.
9190 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9191 update virtual operands.
9193 2013-05-06 Alan Modra <amodra@gmail.com>
9195 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9196 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9197 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9198 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9199 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9200 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9202 2013-05-06 Alan Modra <amodra@gmail.com>
9204 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9205 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9206 (DEFAULT_ASM_ENDIAN): Define.
9207 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9208 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9209 Update -K PIC clause from sysv4.h.
9210 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9211 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9213 2013-05-06 Alan Modra <amodra@gmail.com>
9215 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9216 twice for little-endian.
9217 (ashrdi3_no_power, ashrdi3): Support little-endian.
9219 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9222 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9223 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9225 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9226 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9227 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9230 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9231 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9233 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9234 * config/pa/pa.opt: Make mbig-switch a no-op.
9235 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9236 (CASE_VECTOR_MODE): Always return SImode.
9237 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9238 for the !TARGET_BIG_SWITCH case.
9239 * config/pa/pa-linux.h: Likewise.
9240 * config/pa/pa-openbsd.h: Likewise.
9241 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9242 * config/pa/pa.md (short_jump): Remove define_insn.
9243 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9244 (casesi0): Remove define_insn.
9245 (type): Remove btable_branch.
9246 (pa_combine_type): Likewise.
9247 (in_nullified_branch_delay): Likewise.
9248 (in_call_delay): Likewise.
9249 (define_delay): Likewise.
9250 (define_insn_reservation "Z3"): Likewise.
9251 (define_insn_reservation "Z4"): Likewise.
9252 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9253 (pa_adjust_insn_length): Remove adjustment for btable branches.
9254 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9257 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9259 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9260 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9261 Add m->r,x alternatives.
9262 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9263 splitters using SWI48x mode iterator.
9264 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9265 TARGET_64BIT. Add m->x alternative.
9266 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9267 Add o->x alternative. Enable for TARGET_SSE.
9268 (sse_storeq): Remove expander.
9269 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9270 with memory input operand.
9271 (*vec_extractv2di_1 splitter): New.
9272 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9273 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9275 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9277 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9278 (INT_LOWPART): Delete.
9279 (extract_MB): Adjust.
9280 (extract_ME): Adjust.
9281 (print_operand): Adjust.
9283 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9285 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9286 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9287 (reg_or_logical_cint_operand, easy_fp_constant,
9288 logical_const_operand): Delete "CONST_DOUBLE" case.
9289 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9290 "HOST_BITS_PER_WIDE_INT == 64" test.
9291 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9292 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9293 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9294 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9296 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9297 CONST_DOUBLE DImode/VOIDmode case.
9298 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9299 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9300 CONST_DOUBLE VOIDmode case.
9301 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9302 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9303 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9304 Delete CONST_DOUBLE case.
9305 (splitters for mov FMOVE64 const_double): Delete
9306 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9307 "HOST_BITS_PER_WIDE_INT >= 64" test.
9308 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9310 (mov DI const_double): Delete.
9312 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9314 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9315 on op shows all bits zero in mode of a lowpart subreg, return zero.
9317 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9320 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9321 to save TFmode registers and DImode to save TImode registers for
9322 caller save operations.
9323 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9324 mark being partially clobbered since they only use the first
9327 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9328 and TDmode only use the upper 64-bits of each VSX register.
9330 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9332 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9333 (find_candidates_in_block): Disable slsr_process_phi.
9335 2013-05-03 Guozhi Wei <carrot@google.com>
9337 * coverage.c (coverage_obj_init): Move the construction of gcov
9339 (build_init_ctor): ... here.
9341 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9343 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9344 (slsr_cand_d): Redefine def_phi.
9345 (stride_status, phi_adjust_status, count_phis_status): New enums.
9346 (find_phi_def): New.
9347 (find_basis_for_base_expr): New.
9348 (find_basis_for_candidate): Handle hidden bases.
9349 (alloc_cand_and_find_basis): Handle phi candidates.
9350 (slsr_process_phi): New.
9351 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9352 (create_mul_imm_cand): Likewise.
9353 (create_add_ssa_cand): Exclude phi base candidates.
9354 (create_add_imm_cand): Likewise.
9355 (slsr_process_cast): Likewise.
9356 (slsr_process_copy): Likewise.
9357 (find_candidates_in_block): Handle phi candidates.
9358 (dump_candidate): Likewise.
9359 (unconditional_cands): Delete.
9360 (unconditional_cands_with_known_stride_p): Delete.
9361 (phi_dependent_cand_p): New.
9362 (cand_increment): Handle phi-dependent candidates.
9363 (replace_dependent): Delete.
9364 (replace_mult_candidate): New.
9365 (replace_unconditional_candidate): New.
9366 (incr_vec_index): Move to avoid forward reference.
9367 (create_add_on_incoming_edge): New.
9368 (create_phi_basis): New.
9369 (replace_dependents): Delete.
9370 (replace_conditional_candidate): New.
9371 (phi_add_costs): New.
9372 (replace_uncond_cands_and_profitable_phis): New.
9373 (record_increment): Handle phi adjustments.
9374 (record_phi_increments): New.
9375 (record_increments): Handle phi adjustments.
9376 (phi_incr_cost): New.
9377 (lowest_cost_path): Handle phis.
9378 (total_savings): Likewise.
9379 (analyze_increments): Likewise.
9380 (ncd_with_phi): New.
9381 (ncd_of_cand_and_phis): New.
9382 (nearest_common_dominator_for_cands): Handle phi increments.
9383 (all_phi_incrs_profitable): New.
9384 (replace_profitable_candidates): Handle phi-dependent candidates.
9385 (analyze_candidates_and_replace): Likewise.
9387 2013-05-03 Teresa Johnson <tejohnson@google.com>
9390 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9391 do not exceed REG_BR_PROB_BASE.
9393 2013-05-03 Jeff Law <law@redhat.com>
9395 PR tree-optimization/57144
9396 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9397 operand of the condition will bit into the new type when eliminating
9398 a cast feeding a condition.
9400 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9402 PR rtl-optimization/57130
9403 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9404 of COMPARE as in_code to the recursive call if needed.
9406 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9408 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9409 (enabled): Handle new members.
9410 * config/i386/sse.md (*vec_concatv2si): Merge from
9411 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9412 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9414 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9416 PR tree-optimization/57027
9417 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9418 for fnms opportunity, check we got the prerequisite kind
9419 of tree / gimple before using accessor functions.
9421 2013-05-03 Richard Biener <rguenther@suse.de>
9423 * double-int.h (lshift): New overload without precision
9425 (operator *=, operator +=, operator -=): Move ...
9426 * double-int.c (operator *=, operator +=, operator -=): ... here
9427 and implement more efficiently.
9428 (mul_double_with_sign): Remove.
9429 (lshift_double): Adjust to take unsinged shift argument, push
9430 dispatching code to callers.
9431 (mul_double_wide_with_sign): Add early out for callers that
9432 are not interested in high parts or overflow.
9433 (lshift): New function.
9434 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9436 (lrotate, rrotate): Use logical shifts.
9437 * expr.c (get_inner_reference): Use lshift.
9438 * fixed-value.c (do_fixed_divide): Likewise.
9439 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9440 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9441 (indirect_refs_may_alias_p): Likewise.
9442 (stmt_kills_ref_p_1): Likewise.
9444 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9446 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9448 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9450 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9451 scalar form of FABD instruction.
9453 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9455 * lra-constraints.c (process_alt_operands): Add checking alt
9456 number to choose the best alternative.
9458 2013-05-02 Richard Biener <rguenther@suse.de>
9460 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9461 bitmap and its handling.
9462 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9464 2013-05-02 Richard Biener <rguenther@suse.de>
9467 * tree-inline.c (copy_loops): Properly handle removed loops.
9468 (copy_cfg_body): Mark destination loops for fixup if source
9471 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9474 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9475 generating simple_return for naked functions.
9477 2013-05-02 Martin Jambor <mjambor@suse.cz>
9480 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9481 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9483 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9484 ipa_agg_replacement_value structures.
9485 (known_aggs_to_agg_replacement_list): Likewise.
9486 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9487 (read_agg_replacement_chain): Likewise.
9488 (ipcp_transform_function): Also check that by_ref flags match.
9490 2013-05-02 Richard Biener <rguenther@suse.de>
9492 * graphds.h (struct graph): Add obstack member.
9493 * graphds.c (new_graph): Initialize obstack and allocate
9495 (add_edge): Allocate edge from the obstack.
9496 (free_graph): Free the obstack instead of all edges and vertices.
9498 2013-05-02 Teresa Johnson <tejohnson@google.com>
9500 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9502 * cfg.c (update_bb_profile_for_threading): Ditto.
9503 * tree-inline.c (copy_bb): Ditto.
9504 (copy_edges_for_bb): Ditto.
9505 (initialize_cfun): Ditto.
9506 (copy_cfg_body): Ditto.
9507 (expand_call_inline): Ditto.
9508 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9509 (estimate_node_size_and_time): Ditto.
9510 (inline_merge_summary): Ditto.
9511 * cgraphclones.c (cgraph_clone_edge): Ditto.
9512 (cgraph_clone_node): Ditto.
9513 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9514 (compute_trg_info): Ditto.
9516 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9518 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9519 S reg when fp attribute set.
9520 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9522 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9524 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9526 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9527 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9528 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9530 2013-05-02 Richard Biener <rguenther@suse.de>
9532 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9533 (struct instantiate_cache_entry): New type.
9534 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9535 (struct instantiate_cache_type): New type.
9536 (set_instantiated_value, get_instantiated_value): Remove.
9537 (get_instantiated_value_entry): New function.
9538 (instantiate_scev_name): Use the new cache and adjust.
9539 (instantiate_scev_poly): Adjust.
9540 (instantiate_scev_binary): Likewise.
9541 (instantiate_array_ref): Likewise.
9542 (instantiate_scev_convert): Likewise.
9543 (instantiate_scev_not): Likewise.
9544 (instantiate_scev_3): Likewise.
9545 (instantiate_scev_2): Likewise.
9546 (instantiate_scev_r): Likewise.
9547 (instantiate_scev): Likewise.
9548 (resolve_mixers): Likewise.
9550 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9553 * lra-constraints.c (best_small_class_operands_num): Remove.
9554 (process_alt_operands): Remove small_class_operands_num. Take
9555 small classes operands into losers and only if the operand is not
9556 matched. Modify debugging output.
9557 (curr_insn_transform): Remove best_small_class_operands_num.
9560 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9562 * config/aarch64/aarch64-builtins.c
9563 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9564 * config/aarch64/aarch64-simd-builtins.def
9565 (reduc_splus_): Add new modes.
9566 (reduc_uplus_): New.
9567 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9568 (reduc_uplus_v4sf): Likewise.
9569 (reduc_splus_v4sf): Likewise.
9570 (aarch64_addv<mode>): Likewise.
9571 (reduc_uplus_<mode>): Likewise.
9572 (reduc_splus_<mode>): Likewise.
9573 (aarch64_addvv2di): Likewise.
9574 (reduc_uplus_v2di): Likewise.
9575 (reduc_splus_v2di): Likewise.
9576 (aarch64_addvv2si): Likewise.
9577 (reduc_uplus_v2si): Likewise.
9578 (reduc_splus_v2si): Likewise.
9579 (reduc_<sur>plus_<mode>): New.
9580 (reduc_<sur>plus_v2di): Likewise.
9581 (reduc_<sur>plus_v2si): Likewise.
9582 (reduc_<sur>plus_v4sf): Likewise.
9583 (aarch64_addpv4sf): Likewise.
9584 * config/aarch64/arm_neon.h
9585 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9586 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9587 add UNSPEC_SADDV, UNSPEC_UADDV.
9589 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9591 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9593 * config/aarch64/arm_neon.h
9594 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9596 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9598 * config/aarch64/aarch64-builtins
9599 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9601 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9603 * config/aarch64/aarch64-simd-builtins.def
9605 (reduc_smin_): Likewise.
9606 (reduc_umax_): Likewise.
9607 (reduc_umin_): Likewise.
9608 (reduc_smax_nan_): Likewise.
9609 (reduc_smin_nan_): Likewise.
9612 (smax): Update for V2SF, V4SF and V2DF modes.
9615 (smin_nan): Likewise.
9616 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
9617 (<su><maxmin><mode>3): ...This, refactor.
9618 (s<maxmin><mode>3): New.
9619 (<maxmin_uns><mode>3): Likewise.
9620 (reduc_<maxmin_uns>_<mode>): Refactor.
9621 (reduc_<maxmin_uns>_v4sf): Likewise.
9622 (reduc_<maxmin_uns>_v2si): Likewise.
9623 (aarch64_<fmaxmin><mode>: Remove.
9624 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
9626 (vmin<q>_f<32,64>): Likewise.
9627 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9629 (su): Add mappings for smax, smin, umax, umin.
9631 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9632 (FMAXMIN): Rename as...
9633 (FMAXMIN_UNS): ...This.
9635 (fmaxminv): Likewise.
9636 (fmaxmin): Likewise.
9638 (maxmin_uns_op): Likewise.
9640 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9642 * config/aarch64/arm_neon.h
9643 (vac<ge, gt><sd>_f<32, 64>): Rename to...
9644 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
9645 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
9647 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9649 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
9650 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
9652 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9654 * config/aarch64/aarch64-simd.md
9655 (vcond<mode>_internal): Handle special cases for constant masks.
9656 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
9657 (vcondu<mode><mode>): Likewise.
9658 (vcond<v_cmp_result><mode>): New.
9660 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9662 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
9663 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
9664 * config/aarch64/aarch64-simd-builtins.def
9665 (cmeq): Update to BUILTIN_VALLDI.
9670 * config/aarch64/arm_neon.h
9671 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
9672 to builtins or C as appropriate.
9674 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9676 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
9678 (cmhi): Rename to...
9680 * config/aarch64/aarch64-simd.md
9681 (simd_mode): Add SF.
9682 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
9683 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
9684 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
9685 (cstore<mode>_neg): ...This.
9686 * config/aarch64/iterators.md
9688 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
9690 (UCOMPARISONS): Likewise.
9691 (optab): Add missing comparisons.
9699 (V_cmp_result): Add DF, SF modes.
9700 (v_cmp_result): Likewise.
9703 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
9705 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
9707 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
9708 define_insn to define_insn_and_split.
9709 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
9710 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
9711 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
9712 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
9713 (thumb2_negscc): Likewise.
9715 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9717 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
9719 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9721 * config/arm/thumb2.md: Remove trailing whitespaces.
9723 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9725 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
9726 Use gen_int_mode rather than GEN_INT.
9728 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
9730 * value-prof.c (stream_in_histogram_value): Remove the strayed
9733 2013-04-30 Richard Biener <rguenther@suse.de>
9736 * cfghooks.c (split_edge): Properly check for the loop latch edge.
9738 2013-04-30 Richard Biener <rguenther@suse.de>
9741 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
9743 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
9745 PR rtl-optimization/56957
9746 PR rtl-optimization/57105
9747 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
9748 variable. Use just INSN_UID for determining whether an insn
9749 should be only disconnected from the insn stream.
9750 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
9752 2013-04-30 Jakub Jelinek <jakub@redhat.com>
9754 PR tree-optimization/57104
9755 * tsan.c (instrument_expr): Don't instrument accesses to
9756 DECL_HARD_REGISTER VAR_DECLs.
9758 2013-04-30 Richard Biener <rguenther@suse.de>
9760 * function.h (loops_for_fn): New inline function.
9761 (set_loops_for_fn): Likewise.
9762 * cfgloop.h (place_new_loop): Add struct function parameter.
9763 (get_loop): Likewise.
9764 (get_loops): Likewise.
9765 (number_of_loops): Likewise.
9767 (fel_init): Likewise.
9768 * cfg.c (get_loop_copy): Adjust.
9769 * cfgloop.c (flow_loops_dump): Likewise.
9770 (record_loop_exits): Likewise.
9771 (verify_loop_structure): Likewise.
9772 * cfgloopanal.c (mark_irreducible_loops): Likewise.
9773 (estimate_reg_pressure_cost): Likewise.
9774 (mark_loop_exit_edges): Likewise.
9775 * cfgloopmanip.c (place_new_loop): Likewise.
9776 (add_loop): Likewise.
9777 (duplicate_loop): Likewise.
9778 * graph.c (draw_cfg_nodes): Likewise.
9779 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
9780 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
9781 (extract_affine_chrec): Likewise.
9782 (build_scop_iteration_domain): Likewise.
9783 * graphite.c (graphite_initialize): Likewise.
9784 * ira-build.c (create_loop_tree_nodes): Likewise.
9785 (more_one_region_p): Likewise.
9786 (rebuild_regno_allocno_maps): Likewise.
9787 (mark_loops_for_removal): Likewise.
9788 (mark_all_loops_for_removal): Likewise.
9789 (remove_unnecessary_regions): Likewise.
9790 (ira_build): Likewise.
9791 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9792 * loop-init.c (fix_loop_structure): Likewise.
9793 (gate_rtl_move_loop_invariants): Likewise.
9794 (gate_rtl_unswitch): Likewise.
9795 (gate_rtl_unroll_and_peel_loops): Likewise.
9796 (rtl_doloop): Likewise.
9797 * lto-streamer-in.c (input_cfg): Likewise.
9798 * lto-streamer-out.c (output_cfg): Likewise.
9799 * modulo-sched.c (sms_schedule): Likewise.
9800 * predict.c (tree_estimate_probability): Likewise.
9801 (tree_estimate_probability_driver): Likewise.
9802 (estimate_loops): Likewise.
9803 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9804 (move_sese_region_to_fn): Likewise.
9805 (debug_loop_num): Likewise.
9806 * tree-chrec.c (chrec_evaluate): Likewise.
9807 (hide_evolution_in_other_loops_than_loop): Likewise.
9808 (chrec_component_in_loop_num): Likewise.
9809 (reset_evolution_in_loop): Likewise.
9810 (evolution_function_is_invariant_rec_p): Likewise.
9811 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9812 * tree-inline.c (copy_loops): Likewise.
9813 (copy_cfg_body): Likewise.
9814 (tree_function_versioning): Likewise.
9815 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9816 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9818 (add_to_evolution_1): Likewise.
9819 (scev_const_prop): Likewise.
9820 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9821 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9822 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9823 (tree_ssa_lim_initialize): Likewise.
9824 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9825 (verify_loop_closed_ssa): Likewise.
9826 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9827 (tree_ssa_loop_im): Likewise.
9828 (tree_ssa_loop_unswitch): Likewise.
9829 (tree_vectorize): Likewise.
9830 (check_data_deps): Likewise.
9831 (tree_ssa_loop_ivcanon): Likewise.
9832 (tree_ssa_loop_bounds): Likewise.
9833 (tree_complete_unroll): Likewise.
9834 (tree_complete_unroll_inner): Likewise.
9835 (tree_parallelize_loops): Likewise.
9836 (tree_ssa_loop_prefetch): Likewise.
9837 (tree_ssa_loop_ivopts): Likewise.
9838 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9839 * tree-vectorizer.c (vectorize_loops): Likewise.
9841 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9843 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9844 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9845 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9846 with EABI_LINK_SPEC.
9848 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9851 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9854 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9857 * lra-constraints.c (process_alt_operands): Discourage a bit more
9858 using memory for pseudos. Print cost dump for alternatives.
9859 Modify cost values for conflicts with early clobbers.
9860 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9862 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9865 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9867 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9869 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9871 (movdi_aarch64): Support LDR/STR from/to D register.
9873 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9875 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9876 or higher by default.
9878 2013-04-29 Richard Biener <rguenther@suse.de>
9881 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9882 even if not adding abnormal edges for calls that can make
9885 2013-04-29 Richard Biener <rguenther@suse.de>
9888 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9890 (move_stmt_r): Remove redundant checking.
9892 2013-04-29 Teresa Johnson <tejohnson@google.com>
9895 * basic-block.h (apply_scale): New function.
9896 (apply_probability): Use apply_scale.
9897 * gimple-streamer-in.c (input_bb): Ditto.
9898 * lto-streamer-in.c (input_cfg): Ditto.
9899 * lto-cgraph.c (merge_profile_summaries): Ditto.
9900 * tree-optimize.c (execute_fixup_cfg): Ditto.
9901 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9902 (copy_edges_for_bb): Ditto.
9903 (copy_cfg_body): Ditto.
9905 2013-04-29 Tom de Vries <tom@codesourcery.com>
9907 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9908 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9909 (tail_merge_optimize): Handle current_loops == NULL.
9911 2013-04-26 Jeff Law <law@redhat.com>
9913 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9914 (simplify_cond_using_ranges): Generalize code to simplify
9915 COND_EXPRs where one argument is a constant and the other
9916 is an SSA_NAME created by an integral type conversion.
9918 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9920 * config/arm/arm.md (store_minmaxsi): Use only when
9921 optimize_insn_for_size_p.
9923 2013-04-29 Christian Bruel <christian.bruel@st.com>
9926 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9928 2013-04-29 Richard Biener <rguenther@suse.de>
9931 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9932 loop tree make sure to schedule a fixup for the child as well.
9933 (expand_omp_for_generic): Properly add loops.
9934 (expand_omp_for_static_nochunk): Likewise.
9935 (expand_omp_for_static_chunk): Likewise.
9936 (expand_omp_for): For the degenerate case fixup loops.
9937 (expand_omp_sections): Fix default bb placement in loops.
9938 (expand_omp_atomic_pipeline): Properly add loops.
9940 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9942 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9944 2013-04-29 Tom de Vries <tom@codesourcery.com>
9946 * tree-ssa-tail-merge.c: Update header comment.
9948 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9950 * config/aarch64/arm_neon.h
9951 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9952 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9953 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9954 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9955 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9956 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9957 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9958 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9960 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9962 * config/aarch64/aarch64-simd.md
9963 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9964 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9965 fix_trunc, fixuns_trunc.
9966 (ftrunc<VDQF:mode>2): New.
9967 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9968 (fix_trunc_optab): New.
9970 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9972 * config/aarch64/aarch64-builtins.c
9973 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9974 iceilf, lround, iroundf.
9976 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9979 * config/i386/i386.h (enum ix86_tune_indices)
9980 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9981 New, split from X86_TUNE_INTER_UNIT_MOVES.
9982 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9983 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9984 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9985 (TARGET_INTER_UNIT_MOVES): Remove.
9986 * config/i386/i386.c (initial_ix86_tune_features): Update.
9987 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9988 (ix86_expand_convert_uns_didf_sse): Use
9989 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9990 (ix86_expand_vector_init_one_nonzero): Ditto.
9991 (ix86_expand_vector_init_interleave): Ditto.
9992 (inline_secondary_memory_needed): Return true for moves from SSE class
9993 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9994 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9995 * config/i386/constraints.md (Yi, Ym): Depend on
9996 TARGET_INTER_UNIT_MOVES_TO_VEC.
9997 (Yj, Yn): New constraints.
9998 * config/i386/i386.md (*movdi_internal): Change constraints of
9999 operand 1 from Yi to Yj and from Ym to Yn.
10000 (*movsi_internal): Ditto.
10001 (*movdf_internal): Ditto.
10002 (*movsf_internal): Ditto.
10003 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10004 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10005 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10006 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10007 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10008 * config/i386/sse.md (movdi_to_sse): Ditto.
10009 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10010 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10011 TARGET_INTER_UNIT_MOVES.
10012 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10013 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10014 instead of TARGET_INTER_UNIT_MOVES.
10015 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10016 operand 1 from Yi to Yj and from Ym to Yn.
10018 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10020 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10021 (float_truncate_hi_): Likewise.
10022 (float_extend_lo_): Likewise.
10023 (float_truncate_lo_): Likewise.
10024 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10025 (aarch64_float_extend_lo_v2df): Likewise.
10026 (vec_unpacks_hi_v4sf): Likewise.
10027 (aarch64_float_truncate_lo_v2sf): Likewise.
10028 (aarch64_float_truncate_hi_v4sf): Likewise.
10029 (vec_pack_trunc_v2df): Likewise.
10030 (vec_pack_trunc_df): Likewise.
10032 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10034 * config/aarch64/aarch64-builtins.c
10035 (aarch64_fold_builtin): Fold float conversions.
10036 * config/aarch64/aarch64-simd-builtins.def
10037 (floatv2si, floatv4si, floatv2di): New.
10038 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10039 * config/aarch64/aarch64-simd.md
10040 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10041 * config/aarch64/iterators.md (FLOATUORS): New.
10042 (optab): Add float, floatuns.
10043 (su_optab): Likewise.
10045 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10047 * config/aarch64/aarch64-builtins.c
10048 (aarch64_builtin_vectorized_function): Use new names for
10050 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10051 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10052 (fcvtzu): Split as...
10053 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10054 (fcvtas): Split as...
10055 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10056 (fcvtau): Split as...
10057 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10058 (fcvtps): Split as...
10059 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10060 (fcvtpu): Split as...
10061 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10062 (fcvtms): Split as...
10063 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10064 (fcvtmu): Split as...
10065 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10066 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10067 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10068 (lfrintnusf, lfrintnudf): Likewise.
10069 * config/aarch64/aarch64-simd.md
10070 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10072 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10073 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10074 (fcvt_pattern): Likewise.
10076 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10078 * config/aarch64/aarch64-simd.md
10079 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10080 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10082 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10084 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10085 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10086 (vrnd<a,m,n,p>_f32): Implement using builtins.
10087 (vrnd<i,x><q>_f<32, 64>): New.
10089 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10091 * config/aarch64/aarch64-builtins.c
10092 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10093 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10094 (frintz): Rename to...
10096 (frintp): Rename to...
10098 (frintm): Rename to...
10100 (frinti): Rename to...
10101 (nearbyint): ...this.
10102 (frintx): Rename to...
10104 (frinta): Rename to...
10106 * config/aarch64/aarch64-simd.md
10107 (aarch64_frint<frint_suffix><mode>): Delete.
10108 (<frint_pattern><mode>2): Convert to insn.
10109 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10110 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10111 (frint_pattern): Likewise.
10112 (frint_suffix): Likewise.
10114 2013-04-29 Richard Biener <rguenther@suse.de>
10116 PR tree-optimization/57081
10117 * loop-init.c: Include tree-flow.h.
10118 (loop_optimizer_finalize): Free number of iteration estimates.
10119 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10121 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10123 PR tree-optimization/57083
10124 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10125 non-singleton shift count range, zero extend low_bound for uns case.
10127 * config/i386/predicates.md (general_vector_operand): New predicate.
10128 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10129 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10130 if they aren't nonimmediate operands. If their original values
10131 satisfy const_vector_equal_evenodd_p, don't shift them.
10132 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10133 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10134 if not nonimmediate_operand.
10135 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10136 instead of register_operand.
10137 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10139 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10141 * stor-layout.c (finalize_size_functions): Allocate a structure and
10142 reset cfun before dumping the functions.
10144 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10146 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10149 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10150 use xop_pmacsdqh if uns_p.
10151 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10152 the immediate rotate count.
10154 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10156 * rtl.h (struct rtx_def): Add comment for field jump.
10157 (LRA_SUBREG_P): New macro.
10158 * recog.c (register_operand): Check LRA_SUBREG_P.
10159 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10161 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10162 (lra_final_code_change): Skip subreg change for operators.
10163 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10164 if there are no operand changes.
10165 * lra-constraints.c (curr_insn_set): New.
10166 (match_reload): Set LRA_SUBREG_P.
10167 (emit_spill_move): Ditto.
10168 (check_and_process_move): Use curr_insn_set. Process only single
10169 set insns. Don't initialize sec_mem_p and change_p.
10170 (simplify_operand_subreg): Use LRA_SUBREG_P.
10171 (reg_in_class_p): New function.
10172 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10173 of #ifdef. Add code to remove cycling.
10174 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10175 non-null disp. Reload inner instead of disp when base and index
10176 are null. Try to put lo_sum into register.
10177 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10178 (check_and_process_move): Move code for move cost check to
10179 simple_move_p. Remove equiv_substitution.
10180 (simple_move_p): New function.
10181 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10182 curr_insn_set. Call check_and_process_move only for single set
10183 insns. Use the new function. Move call of check_and_process_move
10184 after operand equiv substitution and address process.
10186 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10189 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10190 with nonlocal goto receivers or returns twice calls, ignore
10191 unininitialized values from abnormal edges to nl goto receiver
10192 or returns twice call.
10194 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10196 PR tree-optimization/57051
10197 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10198 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10201 2013-04-26 Richard Biener <rguenther@suse.de>
10203 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10204 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10205 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10206 (fixup_loop_arrays_after_move): New function.
10207 (move_sese_region_to_fn): Properly outline the loop tree parts
10208 of the SESE region.
10210 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10212 * config/i386/i386.md (type, unit): Fix long lines.
10214 2013-04-26 Richard Biener <rguenther@suse.de>
10216 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10217 (lto-streamer-out.o): Likewise.
10218 * cfgloop.c (init_loops_structure): Export, add struct function
10219 argument and adjust.
10220 (flow_loops_find): Adjust.
10221 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10222 (init_loops_structure): Declare.
10223 * lto-streamer-in.c: Include cfgloop.h.
10224 (input_cfg): Input the loop tree.
10225 * lto-streamer-out.c: Include cfgloop.h.
10226 (output_cfg): Output the loop tree.
10227 (output_struct_function_base): Do not drop PROP_loops.
10229 2013-03-26 Richard Biener <rguenther@suse.de>
10231 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10232 (pass_build_cfg): Provide PROP_loops.
10233 (move_sese_region_to_fn): Remove loops that are outlined into fn
10235 * tree-inline.c: Include cfgloop.h.
10236 (initialize_cfun): Do not drop PROP_loops.
10237 (copy_loops): New function.
10238 (copy_cfg_body): Copy loop structure.
10239 (tree_function_versioning): Initialize destination loop tree.
10240 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10241 (pass_parallelize_loops): Do IL verification.
10242 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10243 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10244 the CFG make sure we fixup loops as well.
10245 * tree-ssa-tail-merge.c: Include cfgloop.h.
10246 (replace_block_by): When merging loop latches mark loops for fixup.
10247 * lto-streamer-out.c (output_struct_function_base): Drop
10248 PROP_loops for now.
10249 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10250 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10251 * ipa-split.c: Include cfgloop.h.
10252 (split_function): Add the new return block to the loop tree root.
10253 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10254 whether we have removed the forwarder block.
10255 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10256 * cfgloop.h (place_new_loop): Declare.
10257 * cfgloopmanip.c (place_new_loop): Export.
10258 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10259 (tree-switch-conversion.o): Likewise.
10260 (tree-complex.o): Likewise.
10261 (tree-inline.o): Likewise.
10262 (tree-ssa-tailmerge.o): Likewise.
10263 (ipa-split.o): Likewise.
10264 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10265 (tree-ssa-copy.o): Likewise.
10266 * tree-switch-conversion.c: Include cfgloop.h
10267 (process_switch): If we emit a bit-test cascade, schedule loops
10269 * tree-complex.c: Include cfgloop.h.
10270 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10271 * asan.c: Include cfgloop.h.
10272 (create_cond_insert_point): Properly add new basic-blocks to
10273 loops, schedule loop fixup.
10274 * cfgloop.c (verify_loop_structure): Check that looks are not
10276 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10278 (expand_omp_for_generic): Likewise.
10279 (expand_omp_sections): Likewise.
10280 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10281 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10282 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10283 is initialized, not when loops are present.
10284 * tree-parloops.c (parallelize_loops): Remove checking here.
10285 * passes.c (init_optimization_passes): Schedule a copy-propagation
10286 pass before complete unrolling of inner loops.
10288 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10290 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10291 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10292 (colorize_init): Add argument to _WIN32 version.
10293 * toplev.c: Include diagnostic-color.h.
10294 (process_options): Default to -fdiagnostics-color=auto if
10295 GCC_COLORS env var is in the environment.
10296 * common.opt (fdiagnostics-color=): Add Var and Init.
10297 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10298 env var is in the environment, the default is auto rather than never.
10300 * diagnostic.h (file_name_as_prefix): Add context argument.
10301 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10302 the string as locus.
10303 * langhooks.c (lhd_print_error_function): Adjust caller.
10305 2013-04-25 Lawrence Crowl <crowl@google.com>
10307 * var-tracking.c (shared_hash_def::htab):
10308 Change type to hash_table. Update dependent calls and types.
10310 2013-04-25 Lawrence Crowl <crowl@google.com>
10312 * Makefile.in: Update as needed below.
10314 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10315 Move declaration to after the type's method definitons.
10317 * attribs.c (htab_t scoped_attributes::attribute_hash):
10318 Change type to hash_table. Update dependent calls and types.
10320 * bitmap.c (htab_t bitmap_desc_hash):
10321 Change type to hash_table. Update dependent calls and types.
10323 * cselib.c (htab_t cselib_hash_table):
10324 Change type to hash_table. Update dependent calls and types.
10326 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10327 (hash_string_slot_node): Move implementation into lto-streamer.h
10328 struct string_slot_hasher.
10329 (eq_string_slot_node): Likewise.
10331 * data-streamer-out.c: Update output_block::string_hash_table
10332 dependent calls and types.
10334 * dwarf2cfi.c (htab_t trace_index):
10335 Change type to hash_table. Update dependent calls and types.
10337 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10338 Change type to hash_table. Update dependent calls and types.
10339 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10340 (htab_t optimize_external_refs::map): Likewise.
10341 (htab_t output_comp_unit::extern_map): Likewise.
10342 (htab_t output_comdat_type_unit::extern_map): Likewise.
10343 (htab_t output_macinfo::macinfo_htab): Likewise.
10344 (htab_t optimize_location_lists::htab): Likewise.
10345 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10347 * except.c (htab_t ehspec_hash_type):
10348 Change type to hash_table. Update dependent calls and types.
10349 (assign_filter_values::ttypes): Likewise.
10350 (assign_filter_values::ehspec): Likewise.
10351 (sjlj_assign_call_site_values::ar_hash): Likewise.
10352 (convert_to_eh_region_ranges::ar_hash): Likewise.
10354 * gcse.c (htab_t pre_ldst_table):
10355 Change type to hash_table. Update dependent calls and types.
10357 * ggc-common.c (htab_t saving_htab):
10358 Change type to hash_table. Update dependent calls and types.
10359 (htab_t loc_hash): Likewise.
10360 (htab_t ptr_hash): Likewise.
10361 (call_count): Rename ggc_call_count.
10362 (call_alloc): Rename ggc_call_alloc.
10363 (loc_descriptor): Rename make_loc_descriptor.
10364 (add_statistics): Rename ggc_add_statistics.
10366 * ggc-common.c (saving_htab):
10367 Change type to hash_table. Update dependent calls and types.
10369 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10370 (push_gimplify_context): Likewise.
10371 (pop_gimplify_context): Likewise.
10372 (struct gimple_temp_hash_elt): Added.
10373 (struct gimplify_hasher): Likewise.
10374 (struct gimplify_ctx.temp_htab):
10375 Change type to hash_table. Update dependent calls and types.
10377 * gimple-fold.c: Include gimplify-ctx.h.
10379 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10380 Change type to hash_table. Update dependent calls and types.
10381 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10382 avoid potential global name collision.
10384 * gimplify.c: Include gimplify-ctx.h.
10385 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10386 (htab_t gimplify_ctx::temp_htab):
10387 Update dependent calls and types for new type hash_table.
10388 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10389 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10391 * gimplify-ctx.h: New.
10392 (struct gimple_temp_hash_elt): Move from gimplify.c.
10393 (class gimplify_hasher): New.
10394 (struct gimplify_ctx): Move from gimple.h.
10395 (htab_t gimplify_ctx::temp_htab):
10396 Change type to hash_table. Update dependent calls and types.
10398 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10399 (htab_t ivs_params::newivs_index):
10400 Change type to hash_table. Update dependent calls and types.
10401 (htab_t ivs_params::params_index): Likewise.
10402 (htab_t print_generated_program::params_index): Likewise.
10403 (htab_t gloog::newivs_index): Likewise.
10404 (htab_t gloog::params_index): Likewise.
10406 * graphite.c: Include graphite-htab.h.
10407 4htab_t graphite_transform_loops::bb_pbb_mapping):
10408 Change type to hash_table. Update dependent calls and types.
10410 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10411 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10412 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10414 * graphite-dependences.c: Include graphite-htab.h.
10415 (loop_is_parallel_p): Change hash table type of parameter.
10417 * graphite-htab.h: New.
10418 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10419 (extern find_pbb_via_hash): Move from graphite-poly.h.
10420 (extern loop_is_parallel_p): Move from graphite-poly.h.
10421 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10423 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10424 (extern loop_is_parallel_p): Move to graphite-htab.h.
10425 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10427 * haifa-sched.c (htab_t delay_htab):
10428 Change type to hash_table. Update dependent calls and types.
10429 (htab_t delay_htab_i2): Likewise.
10431 * ira-color.c (htab_t allocno_hard_regs_htab):
10432 Change type to hash_table. Update dependent calls and types.
10434 * ira-costs.c (htab_t cost_classes_htab):
10435 Change type to hash_table. Update dependent calls and types.
10437 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10438 Change type to hash_table. Update dependent calls and types.
10440 * loop-iv.c (htab_t bivs):
10441 Change type to hash_table. Update dependent calls and types.
10443 * loop-unroll.c (htab_t opt_info::insns_to_split):
10444 Change type to hash_table. Update dependent calls and types.
10445 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10447 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10448 (struct string_slot_hasher): New.
10449 (htab_t output_block::string_hash_table):
10450 Change type to hash_table. Update dependent calls and types.
10452 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10453 (htab_t file_name_hash_table):
10454 Change type to hash_table. Update dependent calls and types.
10456 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10459 * lto-streamer.c (htab_t tree_htab):
10460 Change type to hash_table. Update dependent calls and types.
10462 * omp-low.c: Include gimplify-ctx.h.
10464 * passes.c (htab_t name_to_pass_map):
10465 Change type to hash_table. Update dependent calls and types.
10466 (pass_traverse): Rename to passes_pass_traverse.
10468 * plugin.c (htab_t event_tab):
10469 Change type to hash_table. Update dependent calls and types.
10471 * postreload-gcse.c (htab_t expr_table):
10472 Change type to hash_table. Update dependent calls and types.
10473 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10475 * sese.c (debug_rename_map_1): Make extern.
10476 (htab_t copy_bb_and_scalar_dependences::rename_map):
10477 Change type to hash_table. Update dependent calls and types.
10479 * sese.h (extern debug_rename_map): Move to .c file.
10481 * store-motion.c (htab_t store_motion_mems_table):
10482 Change type to hash_table. Update dependent calls and types.
10484 * trans-mem.c (htab_t tm_new_mem_hash):
10485 Change type to hash_table. Update dependent calls and types.
10487 * tree-browser.c (htab_t TB_up_ht):
10488 Change type to hash_table. Update dependent calls and types.
10490 * tree-cfg.c (htab_t discriminator_per_locus):
10491 Change type to hash_table. Update dependent calls and types.
10493 * tree-complex.c: Include tree-hasher.h
10494 (htab_t complex_variable_components):
10495 Change type to hash_table. Update dependent calls and types.
10497 * tree-eh.c (htab_t finally_tree):
10498 Change type to hash_table. Update dependent calls and types.
10500 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10501 struct int_tree_hasher.
10502 (extern int_tree_map_eq): Likewise.
10503 (uid_decl_map_hash): Removed.
10504 (extern decl_tree_map_eq): Likewise.
10506 * tree-hasher.h: New.
10507 (struct int_tree_hasher): New.
10508 (typedef int_tree_htab_type): New.
10510 * tree-inline.c: Include gimplify-ctx.h.
10512 * tree-mudflap.c: Include gimplify-ctx.h.
10514 * tree-parloops.c: Include tree-hasher.h.
10515 (htab_t eliminate_local_variables_stmt::decl_address):
10516 Change type to hash_table. Update dependent calls and types.
10517 (htab_t separate_decls_in_region::decl_copies): Likewise.
10519 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10520 Change type to hash_table. Update dependent calls and types.
10522 * tree-sra.c (candidates):
10523 Change type to hash_table. Update dependent calls and types.
10525 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10527 (int_tree_map_hash): Likewise.
10529 * tree-ssa-dom.c (htab_t avail_exprs):
10530 Change type to hash_table. Update dependent calls and types.
10532 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10533 Change type to hash_table. Update dependent calls and types.
10535 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10536 Change type to hash_table. Update dependent calls and types.
10538 * tree-ssa-phiopt.c (seen_ssa_names):
10539 Change type to hash_table. Update dependent calls and types.
10541 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10542 Change type to hash_table. Update dependent calls and types.
10544 * tree-ssa-uncprop.c (equiv):
10545 Change type to hash_table. Update dependent calls and types.
10547 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10549 PR rtl-optimization/57003
10550 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10551 call note_stores with kill_clobbered_value callback again after
10552 killing regs_invalidated_by_call.
10554 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10556 * config/aarch64/aarch64-simd.md
10557 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10558 (aarch64_simd_bsl<mode>): Likewise.
10559 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10561 2013-04-25 Marek Polacek <polacek@redhat.com>
10563 PR tree-optimization/57066
10564 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10566 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10568 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10570 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10572 * config/aarch64/aarch64-builtins.c
10573 (aarch64_fold_builtin): New.
10574 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10575 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10576 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10577 * config/aarch64/arm_neon.h
10578 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10580 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10581 Tejas Belagod <tejas.belagod@arm.com>
10583 * config/aarch64/aarch64-builtins.c
10584 (aarch64_gimple_fold_builtin): New.
10585 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10586 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10587 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10588 (addvv4sf): Update.
10589 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10591 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10593 * config/aarch64/aarch64.md
10594 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10596 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10598 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10599 (*ngcsi_uxtw): New pattern.
10601 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10602 Julian Brown <julian@codesourcery.com>
10604 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10605 (TB_DREG): Add T_V4HF.
10606 (v4hf_UP): New macro.
10607 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10608 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10609 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10611 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10612 (arm_vector_mode_supported_p): Handle V4HF.
10613 (arm_mangle_map): Handle V4HFmode.
10614 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
10615 * config/arm/arm_neon_builtins.def: Add entries for
10616 vcvtv4hfv4sf, vcvtv4sfv4hf.
10617 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
10618 (neon_vcvtv4hfv4sf): Likewise.
10619 * config/arm/neon-gen.ml: Handle half-precision floating point
10621 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
10622 * config/arm/arm_neon.h: Regenerate.
10623 * config/arm/neon.ml (type elts): Add F16.
10624 (type vectype): Add T_float16x4, T_floatHF.
10625 (type vecmode): Add V4HF.
10626 (type features): Add Requires_FP_bit feature.
10627 (elt_width): Handle F16.
10628 (elt_class): Likewise.
10629 (elt_of_class_width): Likewise.
10630 (mode_of_elt): Refactor.
10631 (type_for_elt): Handle F16, fix error messages.
10632 (vectype_size): Handle T_float16x4.
10633 (vcvt_sh): New function.
10634 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
10635 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
10636 (string_of_mode): Handle V4HF.
10637 * doc/arm-neon-intrinsics.texi: Regenerate.
10639 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10641 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
10642 format specifier in 'X' case.
10644 2013-04-25 Alan Modra <amodra@gmail.com>
10647 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
10648 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
10649 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
10650 Repeat for many other rotate/shift and mask patterns using subregs.
10651 Name lshiftrt insns.
10652 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
10653 on WORDS_BIG_ENDIAN.
10655 2013-04-25 Alan Modra <amodra@gmail.com>
10657 * config.gcc: Support little-endian powerpc-linux targets.
10658 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
10659 (LINK_OS_LINUX_SPEC): Define.
10660 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
10661 Preserve MASK_LITTLE_ENDIAN.
10662 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
10663 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
10664 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
10665 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
10666 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
10667 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
10668 Correct fp word order for little-endian. Don't shift toc entries
10669 smaller than a word for little-endian.
10670 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
10671 (bswapdi2 splits): Correct low-part subreg for little-endian.
10672 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
10673 low/high where such is correct only for be.
10674 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
10675 little-endian for -mcall-aixdesc.
10677 2013-04-25 Alan Modra <amodra@gmail.com>
10679 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10680 replace_equiv_address_nv.
10682 2013-04-25 Alan Modra <amodra@gmail.com>
10684 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
10686 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10689 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10690 * rtl.h (struct rtx_def): ...
10692 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10694 PR rtl-optimizations/57046
10695 * lra-constraints (split_reg): Set up lra_risky_transformations_p
10696 for multi-reg splits.
10698 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
10700 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
10702 2013-04-24 Sterling Augustine <saugustine@google.com>
10704 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
10705 (comp_dir_string, debug_str_dwo_section): New.
10706 (DEBUG_STR_DWO_SECTION): Rename to ...
10707 (DEBUG_DWO_STR_SECTION): ... this.
10708 (DEBUG_NORM_STR_SECTION): Delete.
10709 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
10710 (DEBUG_STR_DWO_SECTION_FLAGS): New.
10711 (find_AT_string): Move most logic to ...
10712 (find_AT_string_in_table): ... here. New.
10713 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
10714 add_skeleton_AT_string. Delete logic.
10715 (output_skeleton_debug_sections): Remove call to
10716 add_top_level_skeleton_die_attrs.
10717 (add_comp_dir_attribute): Move logic to comp_dir_string.
10718 (dwarf2out_init): Initialize debug_str_dwo_section.
10719 (output_indirect_string): Call find_string_form.
10720 (output_indirect_strings): Rewrite.
10721 (prune_unused_types): Empty skeleton_debug_str_hash.
10722 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
10723 (dwarf2out_finish): Call output_indirect_strings.
10725 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10727 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
10729 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10731 * rtl.h (struct rtx_def): Add comment for field jump.
10732 (LRA_SUBREG_P): New macro.
10733 * recog.c (register_operand): Check LRA_SUBREG_P.
10734 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10736 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10737 (lra_final_code_change): Skip subreg change for operators.
10738 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10739 if there are no operand changes.
10740 * lra-constraints.c (curr_insn_set): New.
10741 (match_reload): Set LRA_SUBREG_P.
10742 (emit_spill_move): Ditto.
10743 (check_and_process_move): Use curr_insn_set. Process only single
10744 set insns. Don't initialize sec_mem_p and change_p.
10745 (simplify_operand_subreg): Use LRA_SUBREG_P.
10746 (reg_in_class_p): New function.
10747 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10748 of #ifdef. Add code to remove cycling.
10749 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10750 non-null disp. Reload inner instead of disp when base and index
10751 are null. Try to put lo_sum into register.
10752 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10753 (check_and_process_move): Move code for move cost check to
10754 simple_move_p. Remove equiv_substitution.
10755 (simple_move_p): New function.
10756 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10757 curr_insn_set. Call check_and_process_move only for single set
10758 insns. Use the new function. Move call of check_and_process_move
10759 after operand equiv substitution and address process.
10761 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
10763 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
10764 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
10765 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
10767 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10769 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
10771 2013-04-24 Marek Polacek <polacek@redhat.com>
10773 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
10774 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
10775 (select_loops_exit_conditions): Likewise.
10776 (number_of_iterations_for_all_loops): Likewise.
10777 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
10778 (scev_analysis): Likewise.
10780 2013-04-02 Catherine Moore <clm@codesourcery.com>
10781 Chao-ying Fu <fu@mips.com>
10783 * config/mips/micromips.md (jraddiusp): New pattern.
10784 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
10785 instruction if possible.
10787 2013-04-24 Alan Modra <amodra@gmail.com>
10789 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10791 2013-04-24 Julian Brown <julian@codesourcery.com>
10792 Chung-Lin Tang <cltang@codesourcery.com>
10794 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10795 dependency behavior in enumeration type DIE generation. Add TODO note
10796 to comments about future DW_FORM_sdata/udata re-work of related code.
10798 2013-04-23 Lawrence Crowl <crowl@google.com>
10800 * Makefile.in: Update as needed below.
10802 * hash-table.h (class hash_table):
10803 Correct many methods with parameter types compare_type to the correct
10804 value_type. (Correct code was unlikely to notice the change.)
10805 (hash_table::elements_with_deleted) New.
10806 (class hashtable::iterator): New.
10807 (hashtable::begin()): New.
10808 (hashtable::end()): New.
10809 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10811 * statistics.c (statistics_hashes):
10812 Change type to hash_table. Update dependent calls and types.
10814 * tree-into-ssa.c (var_infos):
10815 Change type to hash_table. Update dependent calls and types.
10817 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10818 Change type to hash_table. Update dependent calls and types.
10820 * tree-ssa-loop-im.c (struct mem_ref.refs):
10821 Change type to hash_table. Update dependent calls and types.
10823 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10824 Change type to hash_table. Update dependent calls and types.
10826 * tree-ssa-sccvn.c (vn_tables_s::nary):
10827 Change type to hash_table. Update dependent calls and types.
10828 (vn_tables_s::phis): Likewise.
10829 (vn_tables_s::references): Likewise.
10831 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10832 (vn_reference_eq): Update parameter and return types.
10834 * tree-ssa-structalias.c (pointer_equiv_class_table):
10835 Change type to hash_table. Update dependent calls and types.
10836 (location_equiv_class_table): Likewise.
10838 * tree-vect-data-refs.c: Consequential changes for making
10839 peeling a hash_table.
10841 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10842 (destroy_loop_vec_info): Dependent hash_table update.
10844 * tree-vectorizer.h (peeling_htab):
10845 Change type to hash_table. Update dependent calls and types.
10847 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10849 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10850 to check the register content is equal or not.
10851 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10852 to assign register content record.
10853 * lra-eliminations.c (update_reg_eliminate): Use
10854 lra_update_reg_val_offset to update register content offset.
10855 * lra-int.h (struct lra_reg): Add offset member.
10856 (lra_reg_val_equal_p): New static inline function.
10857 (lra_update_reg_val_offset): New static inline function.
10858 (lra_assign_reg_val): New static inline function.
10859 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10860 to assign register content record.
10861 (initialize_lra_reg_info_element): Initial offset to zero.
10863 2013-04-23 Catherine Moore <clm@codesourcery.com>
10865 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10866 operands. Record compression.
10868 2013-04-23 Xinliang David Li <davidxl@google.com>
10870 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10872 2013-04-23 Richard Biener <rguenther@suse.de>
10874 PR middle-end/57036
10875 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10876 parameter, only add abnormal goto edges from the copied body
10877 if the call could perform abnormal gotos.
10878 (copy_cfg_body): Adjust.
10880 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10882 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10884 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10886 * coretypes.h (gimple_stmt_iterator): Add struct to make
10889 2013-04-23 Richard Biener <rguenther@suse.de>
10891 PR tree-optimization/57026
10892 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10893 from SSA names occuring in abnormal PHI nodes.
10895 2013-04-22 Andi Kleen <ak@linux.intel.com>
10897 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10899 2013-04-22 Andi Kleen <ak@linux.intel.com>
10901 * lto/lto.c (print_lto_report_1): Declare early.
10902 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10904 2013-04-22 Andi Kleen <ak@linux.intel.com>
10906 * common.opt (-flto-report-wpa): Add.
10907 * doc/invoke.texi (-flto-report-wpa): Add.
10908 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10911 2013-04-22 Xinliang David Li <davidxl@google.com>
10913 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10914 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10915 * Makefile.in: New dependency
10917 David Daney <ddaney.cavm@gmail.com>
10919 * configure.ac (gcc_cv_as_micromips_support): Use the
10920 --fatal-warnings option.
10921 * configure: Regenerate.
10923 2013-04-22 Marek Polacek <polacek@redhat.com>
10926 * tsan.c (instrument_expr): Don't instrument expression
10927 in case its size is zero.
10929 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10933 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10935 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10937 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10939 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10940 (gimple_stmt_iterator): New typedef.
10941 * gimple.h (gimple_stmt_iterator): Rename to...
10942 (gimple_stmt_iterator_d): ... This.
10943 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10944 trees be valid for GIMPLE and GENERIC.
10945 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10946 * gimple-fold.c (gimple_fold_call): Call target hook
10947 gimple_fold_builtin.
10948 * hooks.c (hook_bool_gsiptr_false): New.
10949 * hooks.h (hook_bool_gsiptr_false): New.
10950 * target.def (fold_stmt): New.
10951 * doc/tm.texi: Regenerate.
10953 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10956 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10957 a set sp if no stack realignment.
10959 2013-04-22 Nick Clifton <nickc@redhat.com>
10961 * config.gcc (tilegx-linux): Extend extra_objs rather than
10963 (tilepro-linux): Likewise.
10965 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10967 * config/aarch64/aarch64-builtins.c
10969 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10970 (VAR<1-12>): Add MAP parameter.
10971 (BUILTIN_*): Likewise.
10972 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10973 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10974 (aarch64_ushl_n<mode>): Likewise.
10975 (aarch64_sshr_n<mode>): Likewise.
10976 (aarch64_ushr_n<mode>): Likewise.
10977 (aarch64_<maxmin><mode>): Likewise.
10978 (aarch64_sqrt<mode>): Likewise.
10979 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10980 (vshr<q>_n_*): Likewise.
10982 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10984 * config/aarch64/aarch64-builtins.c
10985 (aarch64_simd_builtin_type_mode): Handle SF types.
10987 (BUILTIN_GPF): Define.
10988 (aarch64_init_simd_builtins): Handle SF types.
10989 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10990 (frecps): Likewise.
10991 (frecpx): Likewise.
10992 * config/aarch64/aarch64-simd.md
10993 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10994 (aarch64_frecpe<mode>): New.
10995 (aarch64_frecps<mode>): Likewise.
10996 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10997 (v8type): Add frecp<esx>.
10998 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10999 (aarch64_frecps<mode>): Likewise.
11000 * config/aarch64/iterators.md (FRECP): New.
11001 (frecp_suffix): Likewise.
11002 * config/aarch64/arm_neon.h
11003 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11005 2013-04-22 Christian Bruel <christian.bruel@st.com>
11008 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11009 (REG_CLASS_NAMES): Idem.
11010 (REG_CLASS_CONTENTS): Idem.
11011 (REGCLASS_HAS_FP_REG): Idem.
11012 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11013 (sh_conditional_register_usage): Idem.
11015 2013-04-21 Jeff Law <law@redhat.com>
11017 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11018 (ssa_forward_propagate_and_combine): Use it.
11020 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11022 * lra.c: Update the flow chart diagram.
11024 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11026 PR rtl-optimization/56847
11027 * lra-constraints.c (process_alt_operands): Discourage alternative
11028 with non-matche doffsettable memory constraint fro memory with
11031 2013-04-19 Richard Biener <rguenther@suse.de>
11033 PR tree-optimization/56982
11034 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11036 * gimplify.c (gimplify_call_expr): Notice special calls.
11037 (gimplify_modify_expr): Likewise.
11038 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11039 abnormal control flow receivers.
11040 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11041 in the same way as cfun->has_nonlocal_labels.
11042 (gimple_purge_dead_abnormal_call_edges): Likewise.
11043 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11044 receivers start a basic-block.
11046 2013-04-19 Richard Biener <rguenther@suse.de>
11048 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11050 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11051 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11052 (SLP_TREE_LOAD_PERMUTATION): Add.
11053 (vect_transform_slp_perm_load): Adjust prototype.
11054 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11055 (vect_free_slp_instance): Likewise.
11056 (vect_create_new_slp_node): Likewise.
11057 (vect_supported_slp_permutation_p): Remove.
11058 (vect_slp_rearrange_stmts): Adjust.
11059 (vect_supported_load_permutation_p): Likewise. Inline
11060 vect_supported_slp_permutation_p here.
11061 (vect_analyze_slp_instance): Compute load permutations per
11062 slp node instead of per instance.
11063 (vect_get_slp_defs): Adjust.
11064 (vect_transform_slp_perm_load): Likewise.
11065 (vect_schedule_slp_instance): Remove redundant code.
11066 (vect_schedule_slp): Remove hack for PR56270, add it ...
11067 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11068 CSE loads for SLP. Adjust.
11070 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11072 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11073 spelling in two comments.
11075 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11078 * config/arm/arm.c (load_multiple_sequence): Require SP
11079 as base register for loads if SP is in the register list.
11081 2013-04-19 Martin Jambor <mjambor@suse.cz>
11083 PR tree-optimization/56718
11084 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11085 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11086 and made public. Adjusted all callers.
11087 (ipa_intraprocedural_devirtualization): New function.
11088 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11089 (ipa_intraprocedural_devirtualization): Likewise.
11090 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11092 2013-04-19 Richard Biener <rguenther@suse.de>
11094 PR tree-optimization/57000
11095 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11097 2013-04-19 Terry Guo <terry.guo@arm.com>
11099 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11101 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11102 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11103 (cortex_m4_fmacs): Use new reservations.
11104 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11106 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11108 PR rtl-optimization/56999
11109 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11111 (lra_coalesce): Remove split_origin_bitmap and related code.
11112 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11113 ranges if necessary.
11115 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11117 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11119 (ix86_expand_call): Remove clobbered_registers array and use
11120 x86_64_ms_sysv_extra_clobbered_registers instead.
11121 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11123 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11125 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11126 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11127 (*call_value_rex64_ms_sysv): Ditto.
11129 2013-04-18 Cary Coutant <ccoutant@google.com>
11131 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11132 parent instead of die_mark.
11134 2013-04-18 Diego Novillo <dnovillo@google.com>
11136 * gimple.c (create_gimple_tmp): New.
11137 (get_expr_type): New.
11138 (build_assign): New.
11139 (build_type_cast): New.
11140 * gimple.h (enum ssa_mode): Define.
11141 (gimple_seq_set_location): New.
11142 * asan.c (build_check_stmt): Change some gimple_build_* calls
11143 to use build_assign and build_type_cast.
11145 2013-04-18 Richard Biener <rguenther@suse.de>
11147 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11148 handle negative step. Remove redundant checks.
11149 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11150 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11151 for negative step and grouped loads fail to vectorize.
11153 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11155 * emit-rtl.c (reset_insn_used_flags): New function.
11156 (reset_all_used_flags): Use it.
11157 (verify_insn_sharing): New function.
11158 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11160 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11162 PR tree-optimization/56984
11163 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11164 and (x >> M) >= N don't register any assertion if N << M is the
11167 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11169 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11170 do not delete_insn insns that have not yet been emitted, only
11171 unlink them with remove_insn.
11172 * df-scan.c (df_insn_delete): Revert r197492.
11174 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11176 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11177 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11179 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11181 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11182 define_insn_and_split.
11183 (and_scc,ior_scc,negscc): Likewise.
11184 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11186 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11188 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11189 can benefit from using a sequence of LDRD instructions in epilogue
11190 instead of a single LDM instruction.
11192 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11195 * doc/extend.texi: Fix typo.
11197 2013-04-17 Richard Biener <rguenther@suse.de>
11199 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11200 (vect_build_slp_tree): ... here.
11201 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11202 match. Remove special-casing of mismatched complex loads.
11203 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11204 re-try the match with swapped commutative operands.
11205 (vect_supported_load_permutation_p): Remove special-casing of
11206 mismatched complex loads.
11207 (vect_analyze_slp_instance): Adjust.
11209 2013-04-17 Richard Biener <rguenther@suse.de>
11211 PR rtl-optimization/56921
11212 * cfgloop.h (struct loop): Add simple_loop_desc member.
11213 (struct niter_desc): Mark with GTY(()).
11214 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11215 * loop-iv.c (get_simple_loop_desc): Likewise.
11216 (free_simple_loop_desc): Likewise.
11219 2013-04-16 Richard Biener <rguenther@suse.de>
11221 PR rtl-optimization/56921
11222 * loop-init.c (pass_rtl_move_loop_invariants): Add
11223 TODO_do_not_ggc_collect to todo_flags_finish.
11224 (pass_rtl_unswitch): Same.
11225 (pass_rtl_unroll_and_peel_loops): Same.
11226 (pass_rtl_doloop): Same.
11228 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11230 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11231 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11232 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11234 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11235 * tree-streamer.c (record_common_node): Adjust reference in comment.
11237 2013-04-17 Terry Guo <terry.guo@arm.com>
11239 * config/arm/cortex-m4.md: Add a new bypass.
11241 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11243 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11245 (*subs_<optab><mode>_multp2): New pattern.
11246 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11247 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11249 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11251 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11252 (*subs_mul_imm_<mode>): New pattern.
11254 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11257 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11258 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11259 (vsx_movti_32bit): Change j->wa to O->wa.
11261 2013-04-16 Richard Biener <rguenther@suse.de>
11263 PR rtl-optimization/56921
11264 * loop-init.c (pass_rtl_move_loop_invariants): Add
11265 TODO_do_not_ggc_collect to todo_flags_finish.
11266 (pass_rtl_unswitch): Same.
11267 (pass_rtl_unroll_and_peel_loops): Same.
11268 (pass_rtl_doloop): Same.
11270 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11272 * config/arm/arm.c (emit_multi_reg_push): New declaration
11273 for an existing function.
11274 (arm_emit_strd_push): New function.
11275 (arm_expand_prologue): Used here.
11276 (arm_emit_ldrd_pop): New function.
11277 (arm_expand_epilogue): Used here.
11278 (arm_get_frame_offsets): Update condition.
11279 (arm_emit_multi_reg_pop): Add a special case for load of a single
11280 register with writeback.
11282 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11284 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11287 2013-04-16 Richard Biener <rguenther@suse.de>
11289 PR tree-optimization/56756
11290 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11291 (first_mem_ref_loc): New.
11292 (execute_sm): Place the load temporarily before a previous
11293 access instead of in the latch edge to ensure its SSA dependencies
11294 are defined at points dominating the load.
11296 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11298 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11299 correct fix by moving header and footer insn to the footer of
11300 the merged basic block. Clear BB_END of the merged-away block.
11302 PR middle-end/43631
11303 * emit-rtl.c (make_note_raw): New function.
11304 (link_insn_into_chain): New static inline function.
11305 (add_insn): Use it.
11306 (add_insn_before, add_insn_after): Factor insn chain linking code...
11307 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11308 using link_insn_into_chain.
11309 (note_outside_basic_block_p): New helper function for emit_note_after
11310 and emit_note_before.
11311 (emit_note_after): Use nobb variant of add_insn_after if the note
11312 should not be contained in a basic block.
11313 (emit_note_before): Use nobb variant of add_insn_before if the note
11314 should not be contained in a basic block.
11315 (emit_note_copy): Use make_note_raw.
11316 (emit_note): Likewise.
11317 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11318 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11319 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11320 the moved barrier the tail of the basic block it follows.
11321 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11323 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11325 PR tree-optimization/56962
11326 * gimple-ssa-strength-reduction.c (record_increment): Only set
11327 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11328 either rhs1 or rhs2 is equal to c->base_expr.
11330 2013-04-15 Richard Biener <rguenther@suse.de>
11332 PR tree-optimization/56933
11333 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11335 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11336 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11337 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11338 dependence check ...
11339 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11341 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11342 GROUP_READ_WRITE_DEPENDENCE.
11344 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11346 * emit-rtl.c (reset_all_used_flags): New function.
11347 (verify_rtl_sharing): Call reset_all_used_flags before and after
11348 performing the checks.
11350 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11352 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11353 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11354 * config/arm/constraints.md (De): New constraint.
11355 * config/arm/neon.md (anddi3_neon): Delete.
11356 (neon_vand<mode>): Expand to standard anddi3 pattern.
11357 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11358 Move earlier in the file.
11359 (neon_inv_logic_op2): Likewise.
11360 (arm_anddi_operand_neon): New predicate.
11362 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11364 * configure.ac (gcc_cv_ld_as_needed): Set
11365 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11366 Use -z ignore, -z record on *-*-solaris2*.
11367 (HAVE_LD_AS_NEEDED): Update comment.
11368 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11369 * configure: Regenerate.
11370 * config.in: Regenerate.
11371 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11372 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11373 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11374 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11375 equivalents. Fix markup.
11376 * doc/tm.texi: Regenerate.
11378 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11380 * config/i386/i386.opt: New option mstack-protector-guard=.
11381 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11382 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11383 TARGET_SSP_TLS_GUARD.
11384 * config/i386/i386.c (ix86_option_override_internal): Set
11385 ix86_stack_protector_guard.
11386 * config/i386/i386.md (stack_protect_set): Enable for
11387 TARGET_SSP_TLS_GUARD only.
11388 (stack_protect_set_<mode>): Ditto.
11389 (stack_protect_test): Ditto.
11390 (stack_protect_test_<mode>): Ditto.
11391 * doc/invoke.texi (i386 Option): Document.
11393 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11396 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11397 (S_MODES): Set H_MODE bit.
11398 (SF_MODES): Set only S_MODE and SF_MODE bits.
11399 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11400 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11401 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11402 <MODE_FLOAT>: Likewise.
11404 2013-04-15 Joey Ye <joey.ye@arm.com>
11406 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11408 2013-04-15 Joey Ye <joey.ye@arm.com>
11410 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11412 (thumb_far_jump_used_p): Count instruction size and set
11415 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11417 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11418 * resource.c (mark_target_live_regs): Reformat conditional block.
11420 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11422 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11423 notes, they are emitted only just before final.
11424 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11426 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11428 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11429 * cfgrtl.c (delete_insn): Call it here instead.
11430 * lra-spills.c (lra_final_code_change): Use delete_insn.
11431 * haifa-sched.c (sched_remove_insn): Likewise.
11432 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11433 returning to the nop pool.
11434 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11435 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11437 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11439 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11440 * doc/tm.texi: Regenerated.
11442 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11444 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11447 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11449 * df-core.c (df_find_def): Compare register numbers.
11450 (df_find_use): Likewise.
11452 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11455 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11456 lra_in_progress for return.
11458 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11460 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11461 define_insn into define_insn_and_split and emit movsicc patterns.
11463 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11465 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11467 2013-04-12 Richard Biener <rguenther@suse.de>
11469 * tree-pass.h (TODO_do_not_ggc_collect): New.
11470 * passes.c (execute_one_ipa_transform_pass): Honor
11471 TODO_do_not_ggc_collect.
11472 (execute_one_pass): Likewise.
11475 2013-04-10 Richard Biener <rguenther@suse.de>
11477 * passes.c (init_optimization_passes): Remove reload pass.
11478 * ira.c (do_reload): Merge into ...
11480 (rest_of_handle_reload): Remove.
11481 (pass_reload): Likewise.
11482 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11483 of reload for vzeroupper pass placement.
11485 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11487 PR tree-optimization/56918
11488 PR tree-optimization/56920
11489 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11490 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11491 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11492 use wide_mul_with_sign method.
11494 2013-04-12 Richard Biener <rguenther@suse.de>
11496 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11497 not be considered a gimple constant.
11499 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11501 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11502 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11505 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11506 Jakub Jelinek <jakub@redhat.com>
11508 * opts.c: Include diagnostic-color.h.
11509 (common_handle_option): Handle OPT_fdiagnostics_color_.
11510 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11511 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11512 (diagnostic-color.o): New.
11513 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11514 (diagnostic_color_rule): New enum.
11515 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11516 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11517 the location string.
11518 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11519 either NULL, or color kind.
11520 * diagnostic-color.c: New file.
11521 * diagnostic-color.h: New file.
11522 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11524 * doc/invoke.texi (-fdiagnostics-color): Document.
11525 * pretty-print.h (pp_show_color): Define.
11526 (struct pretty_print_info): Add show_color field.
11527 * diagnostic.c: Include diagnostic-color.h.
11528 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11529 macros. Colorize error:, warning: etc. strings and also the location
11531 (diagnostic_show_locus): Colorize the caret line.
11532 * pretty-print.c: Include diagnostic-color.h.
11533 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11534 inside of %< %> quotes or quoted through q format modifier.
11536 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11538 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11540 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11542 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11543 code in CC_NZ mode.
11544 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11547 2013-04-11 Marek Polacek <polacek@redhat.com>
11549 PR tree-optimization/48184
11550 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11552 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11554 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11555 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11556 (skip_simple_arithmetic): Tidy up.
11557 * tree.h (skip_simple_constant_arithmetic): Declare.
11559 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11561 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11563 2013-04-11 Richard Biener <rguenther@suse.de>
11565 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11566 generate vector constants.
11568 2013-04-11 Richard Biener <rguenther@suse.de>
11570 PR tree-optimization/56878
11571 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11572 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11574 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11575 Prefer to align the DR with the most invariant base address.
11577 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11579 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11582 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11584 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11585 floating-point vector comparisons against 0.
11587 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11589 PR tree-optimization/56899
11590 * fold-const.c (extract_muldiv_1): Apply distributive law
11591 only if TYPE_OVERFLOW_WRAPS (ctype).
11593 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11596 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11597 loading has side effect.
11599 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11601 * config/sparc/sparc.c: Include tree-pass.h.
11602 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11603 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11604 head of file. Change return type. Split off gate function.
11605 (sparc_gate_work_around_errata): New function.
11606 (pass_work_around_errata): New pass definition.
11607 (insert_pass_work_around_errata) New pass insert definition to
11608 insert pass_work_around_errata just after delayed-branch scheduling.
11609 (sparc_option_override): Insert the pass.
11610 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11612 2013-04-10 David S. Miller <davem@davemloft.net>
11614 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
11615 or -mcpu=hypersparc.
11617 * target.def (cstore_mode): New hook.
11618 * target.h: Include insn-codes.h
11619 * targhooks.c: Likewise.
11620 (default_cstore_mode): New function.
11621 * targhooks.h: Declare it.
11622 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
11623 * doc/tm.texi: Rebuild.
11624 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
11625 target hook, rather than inspecting the insn_data.
11626 * config/sparc/sparc.c (sparc_cstore_mode): New function.
11627 (TARGET_CSTORE_MODE): Redefine.
11628 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
11630 * config/sparc/predicates.md (cstore_result_operand): New special
11632 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
11633 Use it for operand 0.
11634 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
11635 (*snesi_special): Likewise.
11636 (*snesi_zero): Likewise.
11637 (*seqsi_zero): Likewise.
11638 (*sltu_insn): Likewise.
11639 (*sgeu_insn): Likewise.
11640 (*seqdi_special): Make operand 0 and comparison operation be of
11642 (*snedi_special): Likewise.
11643 (*snedi_special_vis3): Likewise.
11644 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
11645 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
11646 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
11647 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
11648 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
11649 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
11650 (*sltu_extend_sp64): Likewise.
11651 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
11652 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
11653 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
11654 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
11655 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
11657 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
11659 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
11660 (aarch64_start_file): Use the new function.
11662 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11663 Jason Merrill <jason@redhat.com>
11665 * common.opt: Add -gdwarf.
11666 * opts.c (common_handle_option): Handle it.
11667 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
11669 2013-04-10 Richard Biener <rguenther@suse.de>
11671 * passes.c (execute_todo): Do not call ggc_collect conditional here.
11672 (execute_one_ipa_transform_pass): But unconditionally here.
11673 (execute_one_pass): And here.
11674 (init_optimization_passes): Remove reload pass.
11675 * tree-pass.h (TODO_ggc_collect): Remove.
11676 (pass_reload): Likewise.
11677 * ira.c (do_reload): Merge into ...
11679 (rest_of_handle_reload): Remove.
11680 (pass_reload): Likewise.
11681 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11682 of reload for vzeroupper pass placement.
11683 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
11684 and todo_flags_finish of all passes.
11686 2013-04-10 Richard Biener <rguenther@suse.de>
11688 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
11689 first_const_oprnd field, rename first_def_type to first_op_type.
11690 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
11691 (vect_get_and_check_slp_defs): Always use the type of the
11692 operand. Allow mixed vect_external_def, vect_constant_def types.
11693 (vect_get_constant_vectors): Handle mixed vect_external_def,
11694 vect_constant_def types.
11696 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
11698 PR tree-optimization/55524
11699 * tree-ssa-math-opts.c
11700 (convert_mult_to_fma): Don't use an fms construct
11701 when we don't have an fms operation, but fnma, and it looks
11702 likely that we'll be able to use the latter.
11704 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11706 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
11708 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
11709 inline fail caused by overwritable functions.
11711 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
11713 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
11714 unnecessary bits in the constant power of two case.
11716 2013-04-10 Richard Biener <rguenther@suse.de>
11718 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
11719 broken code swapping operands.
11720 (vect_build_slp_tree): Do not compute load permutations here.
11721 (vect_analyze_slp_instance): Compute load permutations here,
11722 after building the SLP tree.
11724 2013-04-09 Christian Bruel <christian.bruel@st.com>
11726 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
11727 of next/prev_real_insn.
11729 2013-04-09 Jan Hubicka <jh@suse.cz>
11731 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11732 Drop aliased parameter.
11733 (function_and_variable_visibility): Do not handle alias pairs.
11734 * cgraph.c (varpool_externally_visible_p): Update prototype.
11735 * varpool.c (varpool_add_new_variable): Update.
11737 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11739 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
11741 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
11743 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
11745 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
11747 2013-04-09 Marek Polacek <polacek@redhat.com>
11749 PR tree-optimization/48762
11750 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
11752 2013-04-09 Richard Biener <rguenther@suse.de>
11754 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
11756 (vect_build_slp_tree): Likewise.
11757 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
11758 calculating the cost of a SLP instance.
11759 (vect_analyze_slp_instance): Use it from here, after building
11762 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11764 PR middle-end/56883
11765 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
11766 expand_omp_for_static_chunk): Use simple_p = true in
11767 force_gimple_operand_gsi calls when assigning to addressable decls.
11769 2013-04-09 Jeff Law <law@redhat.com>
11771 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
11772 when the boolean was created by converting a wider object which
11773 had a boolean range.
11775 2013-04-09 Richard Biener <rguenther@suse.de>
11777 * tree-vectorizer.h (slp_void_p): Remove.
11778 (slp_tree): Typedef before _slp_tree declaration.
11779 (struct _slp_tree): Use a vector of slp_tree as children.
11780 (vect_get_place_in_interleaving_chain): Remove.
11781 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
11783 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
11785 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
11786 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
11787 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
11788 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
11789 Use slp_node instead of slp_void_p and adjust.
11791 2013-04-09 Richard Biener <rguenther@suse.de>
11793 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11794 work that is not necessary.
11796 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11798 PR tree-optimization/56854
11799 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11800 forward into clobber stmts if it would change MEM_REF lhs into
11803 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11805 * tree.c (type_hash_lookup, type_hash_add): Make static.
11806 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11808 2013-04-09 Richard Biener <rguenther@suse.de>
11810 * tree.h (unsave_expr_now): Remove.
11811 * tree-inline.c (mark_local_for_remap_r): Remove.
11812 (unsave_expr_1): Likewise.
11813 (unsave_r): Likewise.
11814 (unsave_expr_now): Likewise.
11815 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11816 (propagate_tree_value): Likewise.
11818 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11820 * doc/rtl.texi (sequence): Rewrite documentation to match the
11821 current use of SEQUENCE rtl objects.
11822 * rtl.def (SEQUENCE): Likewise.
11824 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11825 Update documentation.
11826 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11827 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11829 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11831 2013-04-08 Teresa Johnson <tejohnson@google.com>
11833 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11834 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11836 (estimate_edge_size_and_time): Add comment to suggest using rounding
11838 (estimate_node_size_and_time): Ditto.
11839 (remap_edge_change_prob): Use helper rounding divide methods.
11840 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11841 (gimple_mod_pow2_value_transform): Ditto.
11842 (gimple_mod_subtract_transform): Ditto.
11843 (gimple_ic_transform): Ditto.
11844 (gimple_stringops_transform): Ditto.
11845 * stmt.c (conditional_probability): Ditto.
11846 (emit_case_dispatch_table): Ditto.
11847 * lto-cgraph.c (merge_profile_summaries): Ditto.
11848 * tree-optimize.c (execute_fixup_cfg): Ditto.
11849 * cfgcleanup.c (try_forward_edges): Ditto.
11850 * cfgloopmanip.c (scale_loop_profile): Ditto.
11852 (duplicate_loop_to_header_edge): Ditto.
11853 (lv_adjust_loop_entry_edge): Ditto.
11854 * tree-vect-loop.c (vect_transform_loop): Ditto.
11855 * profile.c (compute_branch_probabilities): Ditto.
11856 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11857 * lto-streamer-in.c (input_cfg): Ditto.
11858 * gimple-streamer-in.c (input_bb): Ditto.
11859 * ipa-cp.c (update_profiling_info): Ditto.
11860 (update_specialized_profile): Ditto.
11861 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11862 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11864 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11865 (compute_trg_info): Ditto.
11866 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11867 (purge_dead_edges): Ditto.
11868 * loop-unswitch.c (unswitch_loop): Ditto.
11869 * cgraphclones.c (cgraph_clone_edge): Ditto.
11870 (cgraph_clone_node): Ditto.
11871 * tree-inline.c (copy_bb): Ditto.
11872 (copy_edges_for_bb): Ditto.
11873 (initialize_cfun): Ditto.
11874 (copy_cfg_body): Ditto.
11875 (expand_call_inline): Ditto.
11877 2013-04-08 Kai Tietz <ktietz@redhat.com>
11879 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11880 TARGET_CYGWIN64 by TARGET_64BIT.
11882 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11884 * config/epiphany/epiphany.md (GPR_1): New constant.
11885 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11886 * config/epiphany/epiphany.c (gen_compare_reg):
11887 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11888 is already in place.
11889 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11890 Don't require being called during rtl expansion; If y operlaps r0,
11892 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11893 (epiphany_expand_epilogue): Likewise.
11895 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11896 Don't use CC_FPmode for ORDERED / UNORDERED.
11897 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11899 * config/epiphany/constraints.md (CnL): New constraint.
11900 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11901 * config/epiphany/predicates.md (add_operand): Allow 1024.
11903 * config/epiphany/epiphany.md (logical_op): New code iterator.
11904 (op_mnc): New code attribute.
11905 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11906 (mov_f+1, mov_f+2): New peephole2 patterns.
11908 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11909 (cstoresi4): Also allow re-use of zero result when doing a NE
11910 comparison to a non-zero operand.
11911 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11913 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11914 Use gen_addsi3_i / gen_subsi3_i.
11916 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11920 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11921 contain anything but clobbers, at most one __builtin_stack_restore,
11922 optionally debug stmts and final resx, and if it has at least one
11923 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11924 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11925 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11926 which isn't defaut definition, remove them.
11927 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11928 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11929 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11930 with MEM_REF LHS with SSA_NAME address.
11932 2013-04-08 Jeff Law <law@redhat.com>
11934 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11936 2013-04-08 Richard Biener <rguenther@suse.de>
11938 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11940 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11941 determined vector type.
11942 (vect_analyze_data_refs): Likewise.
11943 (vect_get_new_vect_var): Adjust.
11944 (vect_create_destination_var): Preserve SSA name versions.
11945 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11946 not dump anything here.
11948 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11950 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11951 Add member lr_slot_known.
11952 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11954 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11955 Remove code that sets lr_slot_offset according to what a previous
11956 version of epiphany_emit_save_restore used to do.
11957 (epiphany_emit_save_restore): When doing an lr save or restore,
11958 set/verify lr_slot_known and lr_slot_offset.
11960 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11963 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11966 2013-04-08 Richard Biener <rguenther@suse.de>
11968 * alias.c (find_base_term): Fix thinko in previous change.
11970 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11972 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11973 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11974 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11975 if possible to compute val.
11976 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11977 For QImode integers don't require anything about precision. Use
11978 const_with_all_bytes_same to find out if the constant doesn't have
11979 repeated bytes in it.
11981 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11983 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11986 2013-04-08 Marek Polacek <polacek@redhat.com>
11988 PR rtl-optimization/48182
11989 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11992 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11995 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11996 nuses, make sure we have a label.
11998 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12001 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12002 (rs6000_emit_swdiv_low_precision): Remove.
12003 (rs6000_emit_swdiv): Rewrite to handle between one and four
12004 iterations of Newton-Raphson generally; modify required number of
12005 iterations for some cases.
12006 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12008 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12010 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12011 set-but-unused variable.
12013 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12014 basic blocks of released function bodies garbage-collectable.
12016 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12017 (struct rtl_opt_pass): Add TODO_df_finish.
12019 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12021 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12023 * config/arm/constraints.md (q): New constraint.
12024 * config/arm/ldrdstrd.md: New file.
12025 * config/arm/arm.md (ldrdstrd.md) New include.
12026 (arm_movdi): Use "q" instead of "r" constraint
12027 for double-word memory access.
12028 (movdf_soft_insn): Likewise.
12029 * config/arm/vfp.md (movdi_vfp): Likewise.
12030 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12031 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12032 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12033 (mem_ok_for_ldrd_strd): Likewise.
12034 (output_move_double): Update assertion.
12036 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12038 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12040 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12042 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12043 define_insn_and_split.
12044 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12046 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12048 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12049 define_insn_and_split.
12050 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12051 (shiftsi3_compare): New pattern.
12052 (rrx): New pattern.
12053 * config/arm/unspecs.md (UNSPEC_RRX): New.
12055 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12057 * config/arm/arm.md (negdi_extendsidi): New pattern.
12058 (negdi_zero_extendsidi): Likewise.
12060 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12062 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12063 define_insn_and_split.
12064 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12065 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12067 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12069 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12070 define_insn_and_split.
12071 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12072 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12074 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12076 * config/arm/arm.md (subsi3_carryin): New pattern.
12077 (subsi3_carryin_const): Likewise.
12078 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12079 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12081 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12083 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12085 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12087 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12088 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12090 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12092 * config/arm/arm.c (arm_expand_builtin): Change fcode
12093 type to unsigned int.
12095 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12097 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12099 2013-04-04 Ian Lance Taylor <iant@google.com>
12101 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12104 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12106 PR middle-end/56729
12107 * df-scan.c (df_insn_delete): Disable failing assert.
12109 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12111 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12112 New function prototype.
12113 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12114 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12115 (arm_builtin_vectorized_function): New function.
12117 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12119 * config/arm/arm_neon_builtins.def: New file.
12120 * config/arm/arm.c (neon_builtin_data): Move contents to
12121 arm_neon_builtins.def.
12122 (enum arm_builtins): Include neon builtin definitions.
12123 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12124 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12126 2013-04-04 Marek Polacek <polacek@redhat.com>
12128 PR tree-optimization/48186
12129 * predict.c (maybe_hot_frequency_p): Return false if
12130 HOT_BB_FREQUENCY_FRACTION is 0.
12131 (cgraph_maybe_hot_edge_p): Likewise.
12133 2013-04-04 Richard Biener <rguenther@suse.de>
12135 PR tree-optimization/56826
12136 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12139 2013-04-04 Richard Biener <rguenther@suse.de>
12141 PR tree-optimization/56213
12142 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12143 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12145 2013-04-04 Richard Biener <rguenther@suse.de>
12147 PR tree-optimization/56837
12148 * tree-loop-distribution.c (classify_partition): For non-zero
12149 values require that the value has the same precision as its
12150 mode to be useful as memset value.
12152 2013-04-03 Nick Clifton <nickc@redhat.com>
12154 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12155 (fmssf4): Use fmsf.s on E3V5 architectures.
12156 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12157 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12159 2013-04-03 Jeff Law <law@redhat.com>
12161 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12162 (lra-eliminations.o): Likewise.
12164 2013-04-03 Teresa Johnson <tejohnson@google.com>
12166 * gcov-io.c (compute_working_sets): Moved most of body of old
12167 compute_working_sets here from profile.c.
12168 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12169 (gcov_working_set_t): Moved typedef here from basic-block.h
12170 (compute_working_set): Declare.
12171 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12172 (get_working_sets): Renamed from compute_working_set,
12173 replace most of body with call to new compute_working_sets.
12174 (get_exec_counts): Replace call to compute_working_sets
12175 to get_working_sets.
12176 * profile.h (get_working_sets): Renamed from compute_working_set.
12177 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12178 to get_working_sets.
12179 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12180 * gcov-dump.c (dump_working_sets): New function.
12182 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12184 * hwint.c (sext_hwi, zext_hwi): New functions.
12185 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12186 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12187 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12188 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12189 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12190 (sext_hwi, zext_hwi): New functions.
12192 2013-04-03 Jeff Law <law@redhat.com>
12194 PR tree-optimization/56799
12195 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12196 back test for widening conversion erroneously dropped in prior change.
12198 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12201 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12202 instead of next_real_insn.
12204 2013-04-03 Marek Polacek <polacek@redhat.com>
12207 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12209 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12212 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12214 (thumb1_output_casesi): Likewise.
12215 (thumb2_output_casesi): Likewise.
12217 2013-04-03 Richard Biener <rguenther@suse.de>
12219 PR tree-optimization/56817
12220 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12222 (tree_unroll_loops_completely_1): ... new function to manually
12223 walk the loop tree, properly defering outer loops of unrolled
12224 loops to later iterations.
12226 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12228 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12229 (vectorizable_load): Likewise.
12230 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12231 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12233 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12235 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12238 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12240 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12242 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12244 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12246 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12248 PR tree-optimization/56790
12249 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12252 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12254 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12256 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12257 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12260 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12263 * tree.h (force_folding_builtin_constant_p): New decl.
12264 * builtins.c (force_folding_builtin_constant_p): New variable.
12265 (fold_builtin_constant_p): Fold immediately also if
12266 force_folding_builtin_constant_p.
12268 2013-04-03 Richard Biener <rguenther@suse.de>
12270 PR tree-optimization/56812
12271 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12272 DRs of the same interleaving chain are independent.
12274 2013-04-02 Jason Merrill <jason@redhat.com>
12276 * gdbinit.in (pbb): Use debug fn.
12278 2013-04-02 Lawrence Crowl <crowl@google.com>
12280 * sese.h (struct ivtype_map_elt_s): Remove unused.
12281 (extern debug_ivtype_map): Remove unused.
12282 (extern eq_ivtype_map_elts): Remove unused.
12283 * sese.c (debug_ivtype_map): Removed unused.
12284 (debug_ivtype_map_1): Removed unused.
12285 (debug_ivtype_elt): Remove unused.
12286 (eq_ivtype_map_elts): Remove unused.
12289 2013-04-02 Kai Tietz <ktietz@redhat.com>
12292 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12293 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12294 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12296 (legitimize_pe_coff_symbol): Likewise.
12297 (is_imported_p): New helper-function.
12298 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12299 for Windows x64 targets.
12300 (ix86_expand_prologue): Optimize for pe-coff targets.
12301 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12302 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12303 medium/large code-model.
12304 (legitimize_pic_address): Likewise.
12305 (legitimize_tls_address): Likewise.
12306 (ix86_expand_call): Likewise.
12307 (x86_output_mi_thunk): Likewise.
12308 (get_dllimport_decl): Add new beimport argument.
12309 (construct_plt_address): Don't assert for x64 pe-coff targets.
12310 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12312 (SYMBOL_FLAG_STUBVAR): New macro.
12313 (SYMBOL_REF_STUBVAR_P): Likewise.
12314 * config/i386/winnt.c (stub_list): New structure.
12315 (stub_head): New local variable.
12316 (i386_pe_record_stub): New function.
12317 (i386_pe_file_end): Emit refptr-stubs.
12319 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12321 PR rtl-optimization/56745
12322 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12323 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12326 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12327 and both of them are MEM_REFs, just compare first argument for
12328 equality and attempt to deal even with differing offsets.
12331 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12332 of gimple_clobber_p to be MEM_REF.
12333 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12334 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12335 after gimplification.
12336 * asan.c (get_mem_ref_of_assignment): Don't instrument
12337 gimple_clobber_p stmts.
12338 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12339 gimple_clobber_p stmt if they have MEM_REF lhs and
12340 are dead because of another gimple_clobber_p stmt.
12341 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12342 gimple_clobber_p stmts like debug stmts.
12343 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12344 that refer to unused VAR_DECLs or uninitialized values.
12345 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12346 gimple_clobber_p stmts if they refer to removed parameters.
12347 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12350 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12352 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12353 using SWI48 mode attribute.
12355 2013-04-02 Wei Mi <wmi@google.com>
12357 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12358 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12359 *<rotate_insn><mode>3_mask in i386.md.
12361 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12363 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12365 2013-04-02 Richard Biener <rguenther@suse.de>
12367 PR tree-optimization/56778
12368 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12369 Runtime alias tests are not supported for gather loads.
12370 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12371 stmts referenced from SSA operands before updating SSA form.
12373 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12374 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12376 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12377 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12378 * config/arm/cortex-a53.md: New file.
12379 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12380 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12381 * config/arm/arm.c (arm_issue_rate): Likewise.
12382 * config/arm/arm-tune.md: Regenerate
12383 * config/arm/arm-tables.opt: Regenerate.
12384 * config/arm/arm-cores.def: Add cortex-a53.
12386 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12388 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12391 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12393 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12394 scalar load/store operations using B/H registers.
12395 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12397 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12399 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12401 * config/aarch64/aarch64.c
12402 (aarch64_simd_scalar_immediate_valid_for_move): New.
12403 * config/aarch64/aarch64-protos.h
12404 (aarch64_simd_scalar_immediate_valid_for_move): New.
12405 * config/aarch64/constraints.md (Dh, Dq): New.
12406 * config/aarch64/iterators.md (hq): New.
12408 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12410 * reorg.c (get_branch_condition): Deal with conditional returns.
12411 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12413 2013-04-01 Wei Mi <wmi@google.com>
12415 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12416 Truncate operand 2 using %b asm operand modifier.
12417 (*<shift_insn><mode>3_mask): Ditto.
12418 (*<rotate_insn><mode>3_mask): Ditto.
12420 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12422 PR middle-end/56798
12423 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12425 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12427 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12429 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12431 2013-03-30 Lawrence Crowl <crowl@google.com>
12433 * dse.c (clear_alias_sets): Remove never set.
12434 (disqualified_clear_alias_sets): Remove never set.
12435 (clear_alias_mode_pool): Remove never set.
12436 (dse_step0): Remove condition that is never true.
12437 (canon_address): Remove condition that is never true.
12438 (dse_step7): Remove condition that is never true.
12439 (rest_of_handle_dse): Remove condition that is never true.
12440 (rest_of_handle_dse::did_global): Remove never read from above.
12441 (dse_step2_spill): Remove never called from above.
12442 (dse_step5_spill): Remove never called from above.
12444 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12446 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12447 JUMP_TABLE_DATA changes.
12448 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12449 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12450 (Insns) <jump_table_data>: New entry.
12451 * doc/tm.texi: Regenerate.
12453 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12455 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12456 for table jump at the end of a basic block using tablejump_p.
12457 * targhooks.c (default_invalid_within_doloop): Likewise.
12458 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12459 target hook implementation that is identical to the default hook.
12460 (rs6000_invalid_within_doloop): Remove.
12462 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12463 unused variable from tablejump_p call.
12465 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12466 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12467 (INSN_DELETED_P): Likewise.
12468 (emit_jump_table_data): New prototype.
12469 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12470 after 4th as unused.
12471 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12472 * sched-vis.c (print_insn): Likewise.
12473 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12474 insn for compatibility with back ends that use next_active_insn to
12475 identify jump table data.
12476 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12477 (remove_insn): Likewise.
12478 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12480 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12481 (emit_jump_table_data): New function.
12483 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12484 basic block, a JUMP_TABLE_DATA never is.
12485 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12486 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12487 off from code handling real insns.
12488 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12489 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12490 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12491 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12492 is not a NONDEBUG_INSN_P.
12493 * ira-costs.c (scan_one_insn): Likewise.
12494 * jump.c (mark_all_labels): Likewise.
12495 (mark_jump_label_1): Likewise.
12496 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12497 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12499 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12500 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12501 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12502 JUMP_TABLE_DATA_P insns.
12503 (calculate_elim_costs_all_insns): Likewise.
12504 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12505 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12506 (delete_output_reload): Code style fixups.
12507 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12508 insn flags on this non-insn.
12509 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12510 as scheduling barriers, for pre-change compatibility.
12511 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12512 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12514 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12515 redundant JUMP_TABLE_DATA_P test.
12516 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12517 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12518 (frv_for_each_packet): Likewise.
12519 * config/i386/i386.c (min_insn_size): Likewise.
12520 (ix86_avoid_jump_mispredicts): Likewise.
12521 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12522 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12523 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12524 (mips16_insn_length): Robustify.
12525 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12526 (mips16_split_long_branches): Likewise.
12527 * config/pa/pa.c (pa_combine_instructions): Likewise.
12528 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12529 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12530 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12531 as contributing to pool range lengths.
12532 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12533 Remove redundant JUMP_TABLE_DATA_P test.
12534 (sh_loop_align): Likewise.
12535 (split_branches): Likewise.
12536 (sh_insn_length_adjustment): Likewise.
12537 * config/spu/spu.c (get_branch_target): Likewise.
12539 2013-03-29 Jan Hubicka <jh@suse.cz>
12541 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12542 gcov streaming; stream hot bb threshold to ltrans.
12543 * predict.c (get_hot_bb_threshold): Break out from ....
12544 (maybe_hot_count_p): ... here.
12545 (set_hot_bb_threshold): New function.
12546 * lto-section-in.c (lto_section_name): Add profile.
12547 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12548 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12549 and data-streamer.h
12550 (histogram_entry): New structure.
12551 (histogram, histogram_pool): New global vars.
12552 (histogram_hash): New structure.
12553 (histogram_hash::hash): New method.
12554 (histogram_hash::equal): Likewise.
12555 (account_time_size): New function.
12556 (cmp_counts): New function.
12557 (dump_histogram): New function.
12558 (ipa_profile_generate_summary): New function.
12559 (ipa_profile_write_summary): New function.
12560 (ipa_profile_read_summary): New function.
12561 (ipa_profile): Decide on threshold.
12562 (pass_ipa_profile): Add ipa_profile_write_summary and
12563 ipa_profile_read_summary.
12564 * Makefile.in (ipa.o): Update dependencies.
12565 * lto-streamer.h (LTO_section_ipa_profile): New section.
12567 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12569 * tree.h (VAR_P): New.
12571 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12574 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12576 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12578 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12579 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12580 (control_flow_insn_p): Likewise.
12581 * cfgrtl.c (duplicate_insn_chain): Likewise.
12582 * final.c (get_attr_length_1): Likewise.
12583 (shorten_branches): Likewise.
12584 (final_scan_insn): Likewise.
12585 * function.c (instantiate_virtual_regs): Likewise.
12586 * gcse.c (insert_insn_end_basic_block): Likewise.
12587 * ira-costs.c (scan_one_insn): Likewise.
12588 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12589 * lra.c (check_rtl): Likewise.
12590 * reload1.c (elimination_costs_in_insn): Likewise.
12591 * reorg.c (follow_jumps): Likewise.
12593 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12594 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12595 (thumb_far_jump_used_p): Likewise.
12596 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12597 (workaround_speculation): Likewise.
12598 (add_sched_insns_for_speculation): Likewise.
12599 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12600 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12601 (frv_for_each_packet): Likewise.
12602 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12603 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12604 (final_emit_insn_group_barriers): Likewise.
12605 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12606 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12607 (mips16_insn_length): Likewise.
12608 * config/pa/pa.c (pa_reorg): Likewise.
12609 (pa_combine_instructions): Likewise.
12610 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12611 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12612 (sh_reorg): Likewise.
12613 (split_branches): Likewise.
12614 * config/spu/spu.c (get_branch_target): Likewise.
12616 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
12619 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
12621 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
12622 Fix declaration name.
12624 2013-03-28 Lawrence Crowl <crowl@google.com>
12626 * graphds.h (struct graph.indicies): Remove unused.
12627 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
12628 (SCOP_ORIGINAL_PDDRS): Remove unused.
12629 * sese.h (extern insert_loop_close_phis): Removed unused.
12630 (extern insert_guard_phis): Removed unused.
12631 (extern ivtype_map_elt_info): Removed unused.
12632 (new_ivtype_map_elt): Removed unused.
12633 * sese.c (ivtype_map_elt_info): Removed unused.
12635 2013-03-28 Lawrence Crowl <crowl@google.com>
12637 * Makefile.in: Add several missing include dependences.
12639 (test-dump.o): New. This object is not added to any executable,
12640 but is present for ad-hoc testing.
12642 (debug (const bitmap_head_def &)): New.
12643 (debug (const bitmap_head_def *)): New.
12645 (extern debug (const bitmap_head_def &)): New.
12646 (extern debug (const bitmap_head_def *)): New.
12648 (debug (edge_def &)): New.
12649 (debug (edge_def *)): New.
12651 (debug (basic_block_def &)): New.
12652 (debug (basic_block_def *)): New.
12654 (dump_node (const_tree, int, FILE *)): Correct source file.
12656 (debug (die_struct &)): New.
12657 (debug (die_struct *)): New.
12659 (extern debug (die_struct &)): New.
12660 (extern debug (die_struct *)): New.
12661 * gimple-pretty-print.c
12662 (debug (gimple_statement_d &)): New.
12663 (debug (gimple_statement_d *)): New.
12664 * gimple-pretty-print.h
12665 (extern debug (gimple_statement_d &)): New.
12666 (extern debug (gimple_statement_d *)): New.
12668 (debug (ira_allocno_copy &)): New.
12669 (debug (ira_allocno_copy *)): New.
12670 (debug (ira_allocno &)): New.
12671 (debug (ira_allocno *)): New.
12673 (extern debug (ira_allocno_copy &)): New.
12674 (extern debug (ira_allocno_copy *)): New.
12675 (extern debug (ira_allocno &)): New.
12676 (extern debug (ira_allocno *)): New.
12678 (debug (live_range &)): New.
12679 (debug (live_range *)): New.
12681 (debug (lra_live_range &)): New.
12682 (debug (lra_live_range *)): New.
12684 (debug (lra_live_range &)): New.
12685 (debug (lra_live_range *)): New.
12687 (debug (omega_pb_d &)): New.
12688 (debug (omega_pb_d *)): New.
12690 (extern debug (omega_pb_d &)): New.
12691 (extern debug (omega_pb_d *)): New.
12693 (debug (const rtx_def &)): New.
12694 (debug (const rtx_def *)): New.
12696 (debug_tree (tree): Move within file.
12697 (debug_raw (const tree_node &)): New.
12698 (debug_raw (const tree_node *)): New.
12699 (dump_tree_via_hooks (const tree_node *, int)): New.
12700 (debug (const tree_node &)): New.
12701 (debug (const tree_node *)): New.
12702 (debug_verbose (const tree_node &)): New.
12703 (debug_verbose (const tree_node *)): New.
12704 (debug_head (const tree_node &)): New.
12705 (debug_head (const tree_node *)): New.
12706 (debug_body (const tree_node &)): New.
12707 (debug_body (const tree_node *)): New.
12708 (debug_vec_tree (tree): Move and reimplement in terms of dump.
12709 (debug (vec<tree, va_gc> &)): New.
12710 (debug (vec<tree, va_gc> *)): New.
12712 (extern debug (const rtx_def &)): New.
12713 (extern debug (const rtx_def *)): New.
12715 (debug_raw (simple_bitmap_def &)): New.
12716 (debug_raw (simple_bitmap_def *)): New.
12717 (debug (simple_bitmap_def &)): New.
12718 (debug (simple_bitmap_def *)): New.
12720 (extern debug (simple_bitmap_def &)): New.
12721 (extern debug (simple_bitmap_def *)): New.
12722 (extern debug_raw (simple_bitmap_def &)): New.
12723 (extern debug_raw (simple_bitmap_def *)): New.
12725 (debug (vinsn_def &)): New.
12726 (debug (vinsn_def *)): New.
12727 (debug_verbose (vinsn_def &)): New.
12728 (debug_verbose (vinsn_def *)): New.
12729 (debug (expr_def &)): New.
12730 (debug (expr_def *)): New.
12731 (debug_verbose (expr_def &)): New.
12732 (debug_verbose (expr_def *)): New.
12733 (debug (vec<rtx> &)): New.
12734 (debug (vec<rtx> *)): New.
12736 (extern debug (vinsn_def &)): New.
12737 (extern debug (vinsn_def *)): New.
12738 (extern debug_verbose (vinsn_def &)): New.
12739 (extern debug_verbose (vinsn_def *)): New.
12740 (extern debug (expr_def &)): New.
12741 (extern debug (expr_def *)): New.
12742 (extern debug_verbose (expr_def &)): New.
12743 (extern debug_verbose (expr_def *)): New.
12744 (extern debug (vec<rtx> &)): New.
12745 (extern debug (vec<rtx> *)): New.
12747 (_list_iter_cond_expr): Make inline instead of static.
12749 (debug (sreal &)): New.
12750 (debug (sreal *)): New.
12752 (extern debug (sreal &)): New.
12753 (extern debug (sreal *)): New.
12755 (extern debug_raw (const tree_node &)): New.
12756 (extern debug_raw (const tree_node *)): New.
12757 (extern debug (const tree_node &)): New.
12758 (extern debug (const tree_node *)): New.
12759 (extern debug_verbose (const tree_node &)): New.
12760 (extern debug_verbose (const tree_node *)): New.
12761 (extern debug_head (const tree_node &)): New.
12762 (extern debug_head (const tree_node *)): New.
12763 (extern debug_body (const tree_node &)): New.
12764 (extern debug_body (const tree_node *)): New.
12765 (extern debug (vec<tree, va_gc> &)): New.
12766 (extern debug (vec<tree, va_gc> *)): New.
12768 (debug (struct loop &)): New.
12769 (debug (struct loop *)): New.
12770 (debug_verbose (struct loop &)): New.
12771 (debug_verbose (struct loop *)): New.
12772 * tree-dump.c: Add header dependence.
12774 (extern debug (struct loop &)): New.
12775 (extern debug (struct loop *)): New.
12776 (extern debug_verbose (struct loop &)): New.
12777 (extern debug_verbose (struct loop *)): New.
12779 (debug (data_reference &)): New.
12780 (debug (data_reference *)): New.
12781 (debug (vec<data_reference_p> &)): New.
12782 (debug (vec<data_reference_p> *)): New.
12783 (debug (vec<ddr_p> &)): New.
12784 (debug (vec<ddr_p> *)): New.
12786 (extern debug (data_reference &)): New.
12787 (extern debug (data_reference *)): New.
12788 (extern debug (vec<data_reference_p> &)): New.
12789 (extern debug (vec<data_reference_p> *)): New.
12790 (extern debug (vec<ddr_p> &)): New.
12791 (extern debug (vec<ddr_p> *)): New.
12793 (debug (pt_solution &)): New.
12794 (debug (pt_solution *)): New.
12796 (extern debug (pt_solution &)): New.
12797 (extern debug (pt_solution *)): New.
12799 (debug (_var_map &)): New.
12800 (debug (_var_map *)): New.
12801 (debug (tree_live_info_d &)): New.
12802 (debug (tree_live_info_d *)): New.
12804 (extern debug (_var_map &)): New.
12805 (extern debug (_var_map *)): New.
12806 (extern debug (tree_live_info_d &)): New.
12807 (extern debug (tree_live_info_d *)): New.
12809 2013-03-28 Jan Hubicka <jh@suse.cz>
12811 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12813 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12815 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12816 record only when desired or required.
12818 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12820 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12821 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12823 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12825 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12826 (*andsi3_compare0_uxtw): New pattern.
12827 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12828 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12830 2013-03-28 Jan Hubicka <jh@suse.cz>
12832 * data-streamer-in.c (streamer_read_gcov_count): New function.
12833 * gimple-streamer-out.c: Include value-prof.h.
12834 (output_gimple_stmt): Output histogram.
12835 (output_bb): Use streamer_write_gcov_count.
12836 * value-prof.c: Include data-streamer.h
12837 (dump_histogram_value): Add HIST_TYPE_MAX.
12838 (stream_out_histogram_value): New function.
12839 (stream_in_histogram_value): New function.
12840 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12841 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12842 * data-streamer-out.c (streamer_write_gcov_count): New function.
12843 (streamer_write_gcov_count_stream): New function.
12844 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12845 (lto_output_node): Likewise.
12846 (input_node, input_edge): Likewise.
12847 * lto-streamer-out.c (output_cfg): Update streaming.
12848 * lto-streamer-in.c (input_cfg): Likewise.
12849 * data-streamer.h (streamer_write_gcov_count,
12850 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12851 * gimple-streamer-in.c: Include value-prof.h
12852 (input_gimple_stmt): Input histograms.
12853 (input_bb): Update profile streaming.
12855 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12857 * genmodes.c (emit_max_int): New function.
12858 (emit_insn_modes_h): Added call to emit_max_function.
12859 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12861 * machmode.def: Fixed comment.
12863 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12865 * combine.c (try_combine): Removed useless assert.
12866 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12868 2013-03-28 Marek Polacek <polacek@redhat.com>
12869 Richard Biener <rguenther@suse.de>
12871 PR tree-optimization/56695
12872 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12873 build signed result of a vector comparison.
12874 * tree-cfg.c (verify_gimple_comparison): Check that a result
12875 of a vector comparison has signed type.
12877 2013-03-28 Richard Biener <rguenther@suse.de>
12879 PR tree-optimization/37021
12880 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12881 do not restrict gaps between groups.
12882 * tree-vect-stmts.c (vectorizable_load): Properly account for
12883 a gap between groups.
12885 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12887 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12888 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12891 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12893 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12894 * doc/extend.texi (Named Address Spaces): Ditto.
12895 (Variable Attributes): Ditto.
12897 2013-03-27 Kai Tietz <ktietz@redhat.com>
12899 * config.build: Add support for cygwin x64 target.
12900 * config.gcc: Likewise.
12901 * config.host: Likewise.
12902 * configure.ac: Likewise
12903 * configure: Regenerated.
12905 2013-03-27 Kai Tietz <ktietz@redhat.com>
12907 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12908 * config/i386/t-cygwin-w64: New file.
12909 * config/i386/cygwin-w64.h: New file.
12910 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12911 and add support for x64-cygwin target.
12912 (CPP_SPEC): Likewise.
12913 (CXX_WRAP_SPEC_LIST): Undefine before define.
12914 (LIBGCJ_SONAME): Use 15 as version.
12916 2013-03-27 Richard Biener <rguenther@suse.de>
12918 PR tree-optimization/56716
12919 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12920 dumping for ref nodes.
12922 2013-03-27 Martin Jambor <mjambor@suse.cz>
12924 PR tree-optimization/55334
12925 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12926 restricted pointers to arrays.
12928 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12930 * Makefile.in (.SUFFIXES): Add .cc.
12931 (.c.o): Apply same recipe for implicit rule .cc.o.
12933 2013-03-27 Richard Biener <rguenther@suse.de>
12935 PR tree-optimization/37021
12936 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12937 REALPART/IMAGPART_EXPRs around the supported refs.
12938 * tree-ssa-structalias.c (find_func_aliases): Assume that
12939 floating-point values are not used to transfer pointers.
12941 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12943 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
12944 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12945 * gcc/doc/tm.texi: Regenerate.
12946 * gcc/targhooks.h (default_has_ifunc_p): New.
12947 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
12948 * gcc/config/linux-protos.h: New file.
12949 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12950 this hook for linux which disables support of indirect functions in
12952 * gcc/config/linux-android.c: New file.
12953 * gcc/config/t-linux-android.c: Ditto.
12954 * gcc/config.gcc: Added new object file linux-android.o.
12955 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
12956 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12957 * gcc/varasm.c (do_assemble_alias): Likewise.
12958 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12959 doesn't support indirect functions.
12960 * configure: Regenerate.
12962 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12965 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12966 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12969 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12971 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12973 2013-03-27 Terry Guo <terry.guo@arm.com>
12975 * config/arm/arm-cores.def: Added core cortex-r7.
12976 * config/arm/arm-tune.md: Regenerated.
12977 * config/arm/arm-tables.opt: Regenerated.
12978 * doc/invoke.texi: Added entry for core cortex-r7.
12980 2013-03-27 Walter Lee <walt@tilera.com>
12982 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12983 double-decrement of next_scratch_regno.
12985 2013-03-27 Walter Lee <walt@tilera.com>
12987 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12989 (insn_v1mulus): Ditto.
12990 (insn_v2muls): Ditto.
12992 2013-03-27 Walter Lee <walt@tilera.com>
12994 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12995 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12997 2013-03-27 Walter Lee <walt@tilera.com>
12999 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13000 (*sibcall_value): Ditto.
13002 2013-03-27 Walter Lee <walt@tilera.com>
13004 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13005 (insn_mnz_v8qi): ... this ...
13006 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13008 (insn_v<n>mnz): Replaced by ...
13009 (insn_v1mnz): ... this ...
13010 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13012 (insn_mz_<mode>): Replaced by ...
13013 (insn_mz_v8qi): ... this ...
13014 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13016 (insn_v<n>mz): Replaced by ...
13017 (insn_v1mz): ... this ...
13018 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13021 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13023 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13025 2013-03-26 Roland McGrath <mcgrathr@google.com>
13027 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13028 than fprintf with a non-constant, non-format string.
13030 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13032 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13033 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13034 operand 0 predicate.
13035 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13036 attribute. Use general_x64nomem_operand as operand 1 predicate.
13037 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13038 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13039 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13040 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13041 (mov<mode>_insv_1): Remove expander. Merge insn with
13042 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13043 Use general_x64nomem_operand as operand 1 predicate.
13044 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13045 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13046 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13047 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13048 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13049 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13050 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13051 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13052 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13053 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13054 (general_x64nomem_operand): Ditto.
13056 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13058 * config/rtems.opt: Add -pthread option.
13060 2013-03-26 Richard Biener <rguenther@suse.de>
13062 * alias.c (find_base_term): Avoid redundant and not used recursion.
13063 (base_alias_check): Get the initial base term from the caller.
13064 (true_dependence_1): Compute and pass base terms to base_alias_check.
13065 (write_dependence_p): Likewise.
13066 (may_alias_p): Likewise.
13068 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13070 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13071 PC-relative load in SI modes and above only.
13073 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13075 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13076 * config/arm/iwmmxt.md (WCGR0): Update.
13077 (WCGR1, WCGR2, WCGR3): Likewise.
13079 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13081 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13082 Use x64 and nox64 isa attributes.
13084 2013-03-26 Richard Biener <rguenther@suse.de>
13086 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13087 alignment computations and rely on get_object_alignment_1
13088 for the !TYPE_P case.
13089 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13091 2013-03-26 Walter Lee <walt@tilera.com>
13093 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13094 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13096 2013-03-25 Jeff Law <law@redhat.com>
13098 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13099 check for INTEGRAL_TYPE_P that was missing due to checking in
13100 wrong version of prior patch.
13102 2013-03-25 Walter Lee <walt@tilera.com>
13104 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13105 TILEGX_INSN_SHUFFLEBYTES1.
13106 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13108 (tilegx_builtins): Ditto.
13109 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13111 2013-03-25 Walter Lee <walt@tilera.com>
13113 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13114 (floatunssisf2): New pattern.
13115 (floatsidf2): New pattern.
13116 (floatunssidf2): New pattern.
13118 2013-03-25 Walter Lee <walt@tilera.com>
13120 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13121 tests for constraint J, K, N, P.
13123 2013-03-25 Walter Lee <walt@tilera.com>
13125 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13126 Use indirect/pcrel encoding.
13127 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13130 2013-03-25 Steve Ellcey <sellcey@mips.com>
13132 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13133 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13134 * config/mips/mips.c (mips_option_override): Set IMADD default.
13135 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13136 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13137 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13138 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13139 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13141 2013-03-25 Jeff Law <law@redhat.com>
13143 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13144 slightly to avoid creating and folding useless trees. Simplify
13145 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13147 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13149 * config/i386/i386.md (*zero_extendsidi2): Merge with
13150 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13151 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13152 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13153 targets. Clarify comment.
13155 2013-03-25 Martin Jambor <mjambor@suse.cz>
13157 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13158 pass-through jump functions differently.
13159 (ipa_read_jump_function): Likewise. Also use setter functions to set
13162 2013-03-25 Martin Jambor <mjambor@suse.cz>
13164 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13165 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13167 (ipa_get_indirect_edge_target): New function.
13168 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13169 ipa_get_indirect_edge_target. Update all callers.
13170 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13171 ipa_get_indirect_edge_target_1 instead of calling
13172 ipa_get_indirect_edge_target.
13173 (create_specialized_node): Pass aggvlas to
13174 ipcp_discover_new_direct_edges.
13176 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13178 * config/arm/arm.md (f_sels, f_seld): New types.
13179 (*cmov<mode>): New pattern.
13180 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13183 2013-03-25 Kai Tietz <ktietz@redhat.com>
13185 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13186 POSIX-printf for mingw-hosted builds.
13188 2013-03-25 Richard Biener <rguenther@suse.de>
13190 PR middle-end/56694
13191 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13192 must-not-throw stmt location.
13194 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13196 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13197 Emit load-acquire versions when acq is true.
13198 (arm_emit_store_exclusive): Add rel parameter.
13199 Emit store-release versions when rel is true.
13200 (arm_split_compare_and_swap): Use acquire-release instructions
13202 of barriers when appropriate.
13203 (arm_split_atomic_op): Likewise.
13204 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13205 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13206 (VUNSPEC_SLX): Likewise.
13207 (VUNSPEC_LDA): Likewise.
13208 (VUNSPEC_STL): Likewise.
13209 * config/arm/sync.md (atomic_load<mode>): New pattern.
13210 (atomic_store<mode>): Likewise.
13211 (arm_load_acquire_exclusive<mode>): Likewise.
13212 (arm_load_acquire_exclusivesi): Likewise.
13213 (arm_load_acquire_exclusivedi): Likewise.
13214 (arm_store_release_exclusive<mode>): Likewise.
13216 2013-03-25 Catherine Moore <clm@codesourcery.com>
13218 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13219 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13220 * config/mip/predicates.md (lwsp_swsp_operand,
13221 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13222 sb16_operand, db4_operand, db7_operand, ib3_operand,
13223 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13224 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13225 andi16_operand): New predicates.
13226 * config/mips/mips.md (compression): New attribute.
13227 (enabled): New attribute.
13228 (length): Consider compression in computing length.
13229 (shift_compression): New code attribute.
13230 (*add<mode>3): New operands. Record compression.
13231 (sub<mode>3): Likewise.
13232 (one_cmpl<mode>2): Likewise.
13233 (*and<mode>3): Likewise.
13234 (*ior<mode>3): Likewise.
13235 (unnamed pattern for xor): Likewise.
13236 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13237 (*<optab><mode>3): Likewise.
13238 (*mov<mode>_internal: Likewise.
13239 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13240 (mips_unsigned_immediate_p): New.
13241 (umips_lwsp_swsp_address_p): New.
13242 (m16_based_address_p): New.
13243 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13244 (mips_unsigned_immediate_p): New prototype.
13245 (lwsp_swsp_address_p): New prototype.
13246 (m16_based_address_p): New prototype.
13247 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13248 (mips_signed_immediate_p): New function.
13249 (m16_based_address_p): New function.
13250 (lwsp_swsp_address_p): New function.
13251 (mips_print_operand_punctuation): Recognize short delay slot insns
13252 for microMIPS.add<mode>3"
13254 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13257 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13258 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13260 2013-03-25 Richard Biener <rguenther@suse.de>
13262 PR tree-optimization/56689
13263 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13266 2013-03-25 Richard Biener <rguenther@suse.de>
13268 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13270 (memory_references): Likewise.
13271 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13272 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13273 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13274 (gather_mem_refs_in_loops): Fold into ...
13275 (analyze_memory_references): ... this. Move initialization
13276 to tree_ssa_lim_initialize.
13277 (fill_always_executed_in): Rename to ...
13278 (fill_always_executed_in_1): ... this.
13279 (fill_always_executed_in): Move contains_call computation to
13280 this new function from ...
13281 (tree_ssa_lim_initialize): ... here.
13282 (tree_ssa_lim): Call fill_always_executed_in.
13284 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13286 * postreload.c (reload_combine): Fix code detecting returns.
13288 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13290 * function.c (emit_use_return_register_into_block): On cc0 targets,
13291 do not emit the sequence between cc0 setter and user.
13293 2013-03-25 Kai Tietz <ktietz@redhat.com>
13295 * config/i386/predicates.md (local_symbolic_operand): Interpret
13296 dll-imported symbols as none-local.
13298 2013-03-25 Richard Biener <rguenther@suse.de>
13300 * tree-ssa-loop-im.c (struct depend): Remove.
13301 (struct lim_aux_data): Make depends a vec of gimples.
13302 (free_lim_aux_data): Adjust.
13303 (add_dependency): Likewise.
13304 (set_level): Likewise.
13306 2013-03-25 Richard Biener <rguenther@suse.de>
13308 PR middle-end/56434
13309 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13310 the pointer returned by calls with ECF_MALLOC set.
13312 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13314 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13316 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13318 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13319 using MMXMODE mode iterator.
13320 (*move<mode>_internal): Merge with *movv2sf_internal and
13321 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13323 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13325 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13326 (record_last_mem_set_info): Likewise.
13328 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13329 of XNEWVEC followed by memset.
13330 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13332 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13334 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13335 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13336 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13337 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13338 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13339 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13340 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13341 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13342 BARRIER_P instead of GET_CODE.
13344 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13346 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13347 inaccuracy in the probing code.
13349 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13350 (ctrapdi4): Likewise.
13352 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13354 * calls.c (expand_call): Add missing guard to code handling return
13355 of non-BLKmode structures in MSB.
13356 * function.c (expand_function_end): Likewise.
13358 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13360 * combine.c (try_combine): Adjust comment. Do not add the set of
13361 insn #0 if the destination indirectly is set or dies in insn #2.
13362 Tidy up code to distribute a new note.
13364 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13366 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13367 also for alternatives 16 and 17.
13369 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13371 * config/i386/sse.md (*mov<mode>_internal): Merge with
13372 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13373 Emit insn template depending on type attribute. Use
13374 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13375 movd instead of movq mnemonic for interunit moves. Rewrite mode
13376 attribute calculation. Remove unit attribute calculation.
13377 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13378 Set prefix_data16 attribute for DImode ssemov types.
13379 Use Ym instead of y for SSE-MMX conversion alternatives.
13380 Reorder operand constraints.
13382 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13384 * df.h (df_insn_delete): Adjust prototype.
13385 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13386 and let it decide whether mark the basic block dirty.
13387 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13388 * df-scan.c (df_insn_info_delete): New helper function, split
13389 off from df_insn_delete.
13390 (df_scan_free_bb_info): Use it.
13391 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13393 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13394 that the insn is actually an insn and it has a non-NULL basic block.
13395 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13397 2013-03-22 Richard Biener <rguenther@suse.de>
13399 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13401 (mem_ref_alloc): Do not allocate them.
13402 (refs_independent_p): Do not query or maintain a cache.
13404 2013-03-22 Richard Biener <rguenther@suse.de>
13406 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13407 (gather_mem_refs_in_loops): Do not compute it.
13408 (analyze_memory_references): Do not allocate it.
13409 (tree_ssa_lim_finalize): Do not free it.
13410 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13412 2013-03-22 Richard Biener <rguenther@suse.de>
13414 * is-a.h (as_a): Use gcc_checking_assert.
13416 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13418 * config/aarch64/aarch64.c (aarch64_print_operand): New
13419 format specifier for printing a constant in hex.
13420 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13421 format specifier for printing second operand.
13423 2013-03-22 Richard Biener <rguenther@suse.de>
13425 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13427 (gather_mem_refs_in_loops): Perform store accumulation here.
13428 (create_vop_ref_mapping_loop): Remove.
13429 (create_vop_ref_mapping): Likewise.
13430 (analyze_memory_references): Initialize refs_stored_in_loop.
13431 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13432 (record_indep_loop): Remove.
13433 (record_dep_loop): New function.
13434 (ref_indep_loop_p_1): Adjust to only walk over references
13435 in the loop, not its subloops.
13436 (ref_indep_loop_p): Rename to ...
13437 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13438 maintaining a more fine-grained cache.
13439 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13440 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13442 2013-03-22 Richard Biener <rguenther@suse.de>
13444 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13445 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13446 aggregate mem_ref_loc.
13447 (free_mem_ref_locs): Inline into ...
13448 (memref_free): ... this and adjust.
13449 (mem_ref_alloc): Adjust.
13450 (mem_ref_locs_alloc): Remove.
13451 (record_mem_ref_loc): Adjust.
13452 (get_all_locs_in_loop): Rewrite into ...
13453 (for_all_locs_in_loop): ... this iterator.
13454 (rewrite_mem_ref_loc): New functor.
13455 (rewrite_mem_refs): Use for_all_locs_in_loop.
13456 (sm_set_flag_if_changed): New functor.
13457 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13458 (ref_always_accessed): New functor.
13459 (ref_always_accessed_p): Use for_all_locs_in_loop.
13461 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13463 * tree-pass.h (PROP_gimple_lvec): New.
13464 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13465 (init_optimization_passes): Move pass_lower_vector.
13466 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13468 (pass_lower_vector): Provide PROP_gimple_lvec.
13469 (pass_lower_vector_ssa): Likewise.
13470 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13472 2013-03-21 Mark Wielaard <mjw@redhat.com>
13474 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13476 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13478 * config/i386/i386.md (*movdi_internal): Disparage slightly
13479 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13480 conversion alternatives.
13482 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13484 PR middle-end/48087
13485 * diagnostic.def (DK_WERROR): New kind.
13486 * diagnostic.h (werrorcount): Define.
13487 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13488 promoted to DK_ERROR, increment DK_WERROR counter instead of
13490 * toplev.c (toplev_main): Call print_ignored_options even if
13491 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13492 even if just werrorcount is non-zero.
13495 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13497 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13498 (string_cst_pool_decl): New function.
13499 (optimize_one_addr_into_implicit_ptr): New function.
13500 (resolve_addr_in_expr): Optimize DWARF location expression
13501 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13502 which doesn't live in memory, but has DW_AT_location or
13503 DW_AT_const_value, or refers to a string literal, into
13504 DW_OP_GNU_implicit_pointer.
13505 (optimize_location_into_implicit_ptr): New function.
13506 (resolve_addr): If removing DW_AT_location of a variable because
13507 it was DW_OP_addr of address of the variable, but the variable doesn't
13508 live in memory, try to emit const value attribute for the initializer.
13510 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13512 * tree.h (VECTOR_TYPE_P): New macro.
13513 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13514 TYPE_MODE): Use it.
13515 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13516 VEC_COND_EXPR cannot be lvalues.
13517 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13519 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13521 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13522 Restrict the transformation to equal modes.
13524 2013-03-21 Richard Biener <rguenther@suse.de>
13526 PR tree-optimization/39326
13527 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13528 (MEM_ANALYZABLE): Adjust.
13529 (record_mem_ref_loc): Move bitmap ops ...
13530 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13531 unanalyzable refs, do not record locations for it.
13532 (analyze_memory_references): Allocate ref zero as shared
13534 (refs_independent_p): Do not test for unanalyzed mems here.
13535 (ref_indep_loop_p_1): Special-case disambiguation against
13536 the unanalyzed ref.
13537 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13539 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13541 * config/arm/arm-protos.h (tune_params): Add
13542 prefer_neon_for_64bits field.
13543 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13544 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13545 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13546 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13547 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13548 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13549 (arm_option_override): Handle -mneon-for-64bits new option.
13550 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13551 (prefer_neon_for_64bits): Declare new variable.
13552 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13553 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13554 (arch_enabled): Handle new arch types. Remove support for onlya8
13556 (one_cmpldi2): Use new arch names.
13557 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13558 * config/arm/arm.opt (mneon-for-64bits): Add option.
13559 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13560 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13561 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13563 * doc/invoke.texi (-mneon-for-64bits): Document.
13565 2013-03-21 Richard Biener <rguenther@suse.de>
13567 PR tree-optimization/39326
13568 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13569 (sort_bbs_in_loop_postorder_cmp): New function.
13570 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13572 2013-03-21 Richard Biener <rguenther@suse.de>
13574 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13575 (vect_insert_into_interleaving_chain): Likewise.
13576 (vect_drs_dependent_in_basic_block): Inline ...
13577 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13579 (vect_analyze_data_ref_dependence): ... here. Simplify.
13580 (vect_check_interleaving): Simplify.
13581 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13582 (vect_slp_analyze_data_ref_dependences): ... this new function.
13583 (dr_group_sort_cmp): New function.
13584 (vect_analyze_data_ref_accesses): Compute data-reference groups
13585 here instead of in vect_analyze_data_ref_dependence. Use
13586 a more efficient algorithm.
13587 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13588 vect_slp_analyze_data_ref_dependences. Call
13589 vect_analyze_data_ref_accesses earlier.
13590 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13591 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13592 (vect_slp_analyze_data_ref_dependences): New prototype.
13594 2013-03-21 Richard Biener <rguenther@suse.de>
13596 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13597 ref is stored in the loop.
13598 (find_refs_for_sm): Walk only over all stores.
13599 (store_motion_loop): Allocate from lim_bitmap_obstack.
13600 (store_motion): Likewise.
13602 2013-03-21 Richard Biener <rguenther@suse.de>
13604 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13605 Update virtual SSA form.
13607 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13609 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13610 * configure: Regenerate.
13611 * config.in: Regenerate.
13612 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13613 if !HAVE_LD_EH_FRAME_CIEV3.
13615 2013-03-21 Richard Biener <rguenther@suse.de>
13617 * tree-cfg.c (verify_expr_no_block): New function.
13618 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
13619 nor DECL_VALUE_EXPR have locations with associated blocks.
13620 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
13621 (clear_unused_block_pointer): Remove code dealing with
13622 blocks in DECL_DEBUG_EXPR locations.
13624 2013-03-21 Richard Biener <rguenther@suse.de>
13626 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
13627 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
13628 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
13629 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
13630 instead of DECL_DEBUG_EXPR_IS_FROM.
13631 * gimplify.c (gimplify_modify_expr): Likewise.
13632 * tree-cfg.c (verify_expr_location_1): Likewise.
13633 * tree-complex.c (create_one_component_var): Likewise.
13634 * tree-sra.c (create_access_replacement): Likewise.
13635 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
13636 (clear_unused_block_pointer): Likewise.
13637 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
13638 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
13639 * var-tracking.c (var_debug_decl): Likewise.
13640 (track_expr_p): Likewise.
13641 * tree-inline.c (add_local_variables): Likewise. Set
13642 DECL_HAS_DEBUG_EXPR_P after copying it.
13643 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
13644 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
13646 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13649 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
13650 * configure: Regenerate.
13651 * config.in: Regenerate.
13652 * config/i386/i386.md (*movdf_internal): Use
13653 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13654 movd instead of movq mnemonic for interunit moves.
13655 (*movdi_internal): Ditto.
13657 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13659 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
13660 (abd<mode>_3): New pattern.
13661 (aba<mode>_3): New pattern.
13662 (fabd<mode>_3): New pattern.
13664 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13666 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
13667 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
13668 occurrence of REGISTER_PREFIX as its empty string.
13670 2013-03-20 Jeff Law <law@redhat.com>
13672 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
13673 addititional equivalences for equality comparisons between an SSA_NAME
13674 and a constant where the SSA_NAME was set from a widening conversion.
13676 2013-03-20 Walter Lee <walt@tilera.com>
13678 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
13680 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13682 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
13683 depending on type attribute.
13684 (*movti_internal): Ditto.
13685 (*movtf_internal): Ditto.
13686 (*movxf_internal): Ditto.
13687 (*movdf_internal): Ditto.
13688 (*movsf_internal): Ditto.
13690 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13692 * config/i386/i386.md (*movti_internal): Set prefix attribute to
13693 maybe_vex for sselog1 and ssemov types.
13694 (*movdi_internal): Reorder operand constraints.
13695 (*movsi_internal): Ditto. Set prefix attribute to
13696 maybe_vex for sselog1 and ssemov types.
13697 (*movtf_internal): Set prefix attribute to maybe_vex
13698 for sselog1 and ssemov types.
13699 (*movdf_internal): Ditto. Set prefix_data16 attribute for
13700 DImode ssemov types. Reorder operand constraints.
13701 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
13702 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
13703 attribute for SImode ssemov types. Reorder operand constraints.
13705 2013-03-20 Martin Jambor <mjambor@suse.cz>
13707 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
13708 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
13710 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
13712 * config/rs6000/predicates.md (indexed_address, update_address_mem
13713 update_indexed_address_mem): New predicates.
13714 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
13715 attribute for load/store instructions.
13716 * config/rs6000/dfp.md (movsd_store): Likewise.
13717 (movsd_load): Likewise.
13718 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
13719 (unnamed HI->DI extend define_insn): Likewise.
13720 (unnamed SI->DI extend define_insn): Likewise.
13721 (unnamed QI->SI extend define_insn): Likewise.
13722 (unnamed QI->HI extend define_insn): Likewise.
13723 (unnamed HI->SI extend define_insn): Likewise.
13724 (unnamed HI->SI extend define_insn): Likewise.
13725 (extendsfdf2_fpr): Likewise.
13726 (movsi_internal1): Likewise.
13727 (movsi_internal1_single): Likewise.
13728 (movhi_internal): Likewise.
13729 (movqi_internal): Likewise.
13730 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
13731 attribute for load/store instructions.
13732 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
13734 (mov<mode>_softfloat): Likewise.
13735 (mov<mode>_hardfloat32): Likewise.
13736 (mov<mode>_hardfloat64): Likewise.
13737 (mov<mode>_softfloat64): Likewise.
13738 (movdi_internal32): Likewise.
13739 (movdi_internal64): Likewise.
13740 (probe_stack_<mode>): Likewise.
13742 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
13744 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
13745 floating point, and decimal floating point to reload iterator.
13747 * config/rs6000/constraints.md (wl constraint): New constraints to
13748 return FLOAT_REGS if certain options are used to reduce the number
13749 of separate patterns that exist in the file.
13750 (wx constraint): Likewise.
13751 (wz constraint): Likewise.
13753 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13754 -mdebug=reg, print wg, wl, wx, and wz constraints.
13755 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13756 Initialize the reload functions for 64-bit binary/decimal floating
13758 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13759 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13760 create the buffer on the stack to overcome not having a 32-bit
13762 (rs6000_emit_move): Likewise.
13763 (rs6000_secondary_memory_needed_rtx): Likewise.
13764 (rs6000_alloc_sdmode_stack_slot): Likewise.
13765 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13766 via xxlxor, just like DFmode 0.0.
13768 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
13769 define as 1 if we are running on a power7 or newer.
13770 (enum r6000_reg_class_enum): Add new constraints.
13772 * config/rs6000/dfp.md (movsd): Delete, combine with binary
13773 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13774 with other moves by using conditional constraits (wg). Use LFIWZX
13775 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
13776 (movsd splitter): Likewise.
13777 (movsd_hardfloat): Likewise.
13778 (movsd_softfloat): Likewise.
13780 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
13781 binary and decimal floating point moves.
13782 (fmove_ok): New attributes to combine binary and decimal floating
13783 point moves, and to combine power6x (mfpgpr) moves along normal
13785 (real_value_to_target): Likewise.
13786 (f32_lr): Likewise.
13787 (f32_lm): Likewise.
13788 (f32_li): Likewise.
13789 (f32_sr): Likewise.
13790 (f32_sm): Likewise.
13791 (f32_si): Likewise.
13792 (movsf): Combine binary and decimal floating point moves. Combine
13793 power6x (mfpgpr) moves with other moves by using conditional
13794 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13795 (mov<mode> for SFmode/SDmode); Likewise.
13796 (SFmode/SDmode splitters): Likewise.
13797 (movsf_hardfloat): Likewise.
13798 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13799 (movsf_softfloat): Likewise.
13800 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13802 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13803 wx and wz constraints.
13805 * config/rs6000/constraints.md (wg constraint): New constraint to
13806 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13808 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13811 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13812 -mdebug=reg, print wg, wl, wx, and wz constraints.
13813 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13814 Initialize the reload functions for 64-bit binary/decimal floating
13816 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13817 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13818 create the buffer on the stack to overcome not having a 32-bit
13820 (rs6000_emit_move): Likewise.
13821 (rs6000_secondary_memory_needed_rtx): Likewise.
13822 (rs6000_alloc_sdmode_stack_slot): Likewise.
13823 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13824 via xxlxor, just like DFmode 0.0.
13826 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13827 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13828 with other moves by using conditional constraits (wg). Use LFIWZX
13829 and STFIWX for loading SDmode on power7.
13830 (movdd splitters): Likewise.
13831 (movdd_hardfloat32): Likewise.
13832 (movdd_softfloat32): Likewise.
13833 (movdd_hardfloat64_mfpgpr): Likewise.
13834 (movdd_hardfloat64): Likewise.
13835 (movdd_softfloat64): Likewise.
13837 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13838 64-bit binary and decimal floating point moves.
13839 (FMOVE64X): Likewise.
13840 (movdf): Combine 64-bit binary and decimal floating point moves.
13841 Combine power6x (mfpgpr) moves with other moves by using
13842 conditional constraits (wg).
13843 (mov<mode> for DFmode/DDmode): Likewise.
13844 (DFmode/DDmode splitters): Likewise.
13845 (movdf_hardfloat32): Likewise.
13846 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13847 (movdf_softfloat32): Likewise.
13848 (movdf_hardfloat64_mfpgpr): Likewise.
13849 (movdf_hardfloat64): Likewise.
13850 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13851 (movdf_softfloat64): Likewise.
13852 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13853 (reload_<mode>_load): Move to later in the file so they aren't in
13854 the middle of the floating point move insns.
13855 (reload_<mode>_store): Likewise.
13857 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13860 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13861 constraint if -mdebug=reg.
13862 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13863 Enable using dd reload support if needed.
13865 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13866 binary and decimal floating point moves in rs6000.md.
13867 (movtd_internal): Likewise.
13869 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13870 decimal floating point moves.
13872 (movtf_internal): Likewise.
13873 (mov<mode>_internal, TDmode/TFmode): Likewise.
13874 (movtf_softfloat): Likewise.
13875 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13877 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13878 movdi_internal64, using wg constraint for move direct operations.
13879 (movdi_internal64): Likewise.
13881 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13882 MODES_TIEABLE_P for selected modes. Print the numerical value of
13883 the various virtual registers. Use GPR/FPR first/last values,
13884 instead of hard coding the register numbers. Print which modes
13885 have reload functions registered.
13886 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13887 settings before/after setting cpu, target and subtarget settings.
13888 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13889 and for secondary reload failures in rs6000_secondary_reload_inner.
13890 (rs6000_secondary_reload_fail): Likewise.
13891 (rs6000_secondary_reload_inner): Likewise.
13893 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13894 macros for first/last GPR and FPR registers.
13895 (LAST_GPR_REGNO): Likewise.
13896 (FIRST_FPR_REGNO): Likewise.
13897 (LAST_FPR_REGNO): Likewise.
13899 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13900 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13901 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13902 (vcond<mode><mode>): Likewise.
13903 (vcondu<mode><mode>): Likewise.
13904 (vector_gtu<mode>): Likewise.
13905 (vector_gte<mode>): Likewise.
13906 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13907 to prevent the compiler from converting DImode operations to TImode.
13908 (ior<mode>3): Likewise.
13909 (and<mode>3): Likewise.
13910 (one_cmpl<mode>2): Likewise.
13911 (nor<mode>3): Likewise.
13912 (andc<mode>3): Likewise.
13914 * config/rs6000/constraints.md (wt constraint): New constraint
13915 that returns VSX_REGS if TImode is allowed in VSX registers.
13917 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13918 constant under VSX.
13920 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13921 similar to TImode, but it is restricted to being in the GPRs.
13923 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13924 TImode to occupy a single VSX register.
13926 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13927 -mvsx-timode for power7/power8.
13928 (power7 cpu): Likewise.
13929 (power8 cpu): Likewise.
13931 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13932 sure that TFmode/TDmode take up two registers if they are ever
13933 allowed in the upper VSX registers.
13934 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13936 (rs6000_init_hard_regno_mode_ok): Likewise.
13937 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13938 constraint. Print if LRA is turned on.
13939 (rs6000_option_override_internal): Give an error if -mvsx-timode
13940 and VSX is not enabled.
13941 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13942 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13943 to reg+offset addressing. Use PTImode when checking offset
13944 addresses for validity.
13945 (reg_offset_addressing_ok_p): Likewise.
13946 (rs6000_legitimate_offset_address_p): Likewise.
13947 (rs6000_legitimize_address): Likewise.
13948 (rs6000_legitimize_reload_address): Likewise.
13949 (rs6000_legitimate_address_p): Likewise.
13950 (rs6000_eliminate_indexed_memrefs): Likewise.
13951 (rs6000_emit_move): Likewise.
13952 (rs6000_secondary_reload): Likewise.
13953 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13954 reloads to fpr registers to continue to use reg+offset addressing,
13955 but 64-bit reloads to altivec registers need reg+reg addressing.
13956 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13957 it. Treat LO_SUM like a PLUS operation.
13958 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13959 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13960 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13961 registers to share a register with a smaller sized type, since VSX
13962 puts scalars in the upper 64-bits.
13963 (print_operand): Add support for PTImode.
13964 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13965 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13966 registers, but don't have arithmetic support.
13967 (rs6000_memory_move_cost): Add test for VSX.
13968 (rs6000_opt_masks): Add -mvsx-timode.
13970 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13973 (VSr): Use wt constraint for TImode.
13974 (VSv): Drop TImode support.
13975 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13976 (vsx_movti_64bit): Likewise.
13977 (vsx_movti_32bit): Likewise.
13978 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13979 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13980 one '?' on the appropriate output constraint. Do not allow TImode
13981 logical operations on 32-bit systems.
13982 (vsx_ior<mode>3): Likewise.
13983 (vsx_xor<mode>3): Likewise.
13984 (vsx_one_cmpl<mode>2): Likewise.
13985 (vsx_nor<mode>3): Likewise.
13986 (vsx_andc<mode>3): Likewise.
13987 (vsx_concat_<mode>): Likewise.
13988 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13990 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13991 OPTION_MASK_VSX_TIMODE.
13992 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13993 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13995 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13996 (TI2 iterator): New iterator for TImode, PTImode.
13997 (wd mode attribute): Add values for vector types.
13998 (movti_string): Replace TI move operations with operations for TImode
13999 and PTImode. Add support for TImode being allowed in VSX registers.
14000 (mov<mode>_string, TImode/PTImode): Likewise.
14001 (movti_ppc64): Likewise.
14002 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14003 (TI mode splitters): Likewise.
14005 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14008 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14010 PR tree-optimization/56355
14011 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14012 Also handle integers with undefined overflow.
14014 2013-03-20 Catherine Moore <clm@codesourcery.com>
14015 Maciej W. Rozycki <macro@codesourcery.com>
14016 Tom de Vries <tom@codesourcery.com>
14017 Nathan Sidwell <nathan@codesourcery.com>
14018 Iain Sandoe <iain@codesourcery.com>
14019 Nathan Froyd <froydnj@codesourcery.com>
14020 Chao-ying Fu <fu@mips.com>
14022 * doc/extend.texi: (micromips, nomicromips, nocompression):
14023 Document new function attributes.
14024 * doc/invoke.texi (minterlink-compressed, mmicromips,
14025 m14k, m14ke, m14kec): Document new options.
14026 (minterlink-mips16): Update documentation.
14027 * doc/md.texi (ZC, ZD): Document new constraints.
14028 * configure.ac (gcc_cv_as_micromips): Check if linker
14029 supports the .set micromips directive.
14030 * configure: Regenerate.
14031 * config.in: Regenerate.
14032 * config/mips/mips-tables.opt: Regenerate.
14033 * config/mips/micromips.md: New file.
14034 * constraints.md (ZC, ZD): New constraints.
14035 * config/mips/predicates.md (movep_src_register): New predicate.
14036 (movep_src_operand): New predicate.
14037 (non_volatile_mem_operand): New predicate.
14038 * config/mips/mips.md (multimem): New type.
14039 (length): Differentiate between 17-bit and 18-bit branch offsets.
14040 (MOVEP1, MOVEP2): New mode iterator.
14041 (mov_<load>l): Use ZC constraint.
14042 (mov_<load>r): Likewise.
14043 (mov_<store>l): Likewise.
14044 (mov_<store>r): Likewise.
14045 (*branch_equality<mode>_inverted): Add microMIPS support.
14046 (*branch_equality<mode>): Likewise.
14047 (*jump_absolute): Likewise.
14048 (indirect_jump_<mode>): Likewise.
14049 (tablejump_<mode>): Likewise.
14050 (<optab>_internal): Likewise.
14051 (sibcall_internal): Likewise.
14052 (sibcall_value_internal): Likewise.
14053 (prefetch): Use constraint ZD.
14054 * config/mips/mips.opt (minterlink-compressed): New option.
14055 (minterlink-mips16): Now an alias for minterlink-compressed.
14056 (mmicromips): New option.
14057 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14058 (compare_and_swap_12): Likewise.
14059 (sync_add<mode>): Likewise.
14060 (sync_<optab>_12): Likewise.
14061 (sync_old_<optab>_12): Likewise.
14062 (sync_new_<optab>_12): Likewise.
14063 (sync_nand_12): Likewise.
14064 (sync_old_nand_12): Likewise.
14065 (sync_new_nand_12): Likewise.
14066 (sync_sub<mode>): Likewise.
14067 (sync_old_add<mode>): Likewise.
14068 (sync_old_sub<mode>): Likewise.
14069 (sync_new_add<mode>): Likewise.
14070 (sync_new_sub<mode>): Likewise.
14071 (sync_<optab><mode>): Likewise.
14072 (sync_old_<optab><mode>): Likewise.
14073 (sync_new_<optab><mode>): Likewise.
14074 (sync_nand<mode>): Likewise.
14075 (sync_old_nand<mode>): Likewise.
14076 (sync_new_nand<mode>): Likewise.
14077 (sync_lock_test_and_set<mode>): Likewise.
14078 (test_and_set_12): Likewise.
14079 (atomic_compare_and_swap<mode>): Likewise.
14080 (atomic_exchange<mode>_llsc): Likewise.
14081 (atomic_fetch_add<mode>_llsc): Likewise.
14082 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14083 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14084 (umips_save_restore_pattern_p): Likewise.
14085 (umips_load_store_pair_p): Likewise.
14086 (umips_output_load_store_pair): Likewise.
14087 (umips_movep_target_p): Likewise.
14088 (umips_12bit_offset_address_p): Likewise.
14089 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14090 (mips_base_mips16): Rename this...
14091 (mips_base_compression_flags): ...to this. Update all uses.
14092 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14093 (mips_mips16_decl_p): Delete.
14094 (mips_nomips16_decl_p): Delete.
14095 (mips_get_compress_on_flags): New function.
14096 (mips_get_compress_off_flags): New function.
14097 (mips_get_compress_mode): New function.
14098 (mips_get_compress_on_name): New function.
14099 (mips_get_compress_off_name): New function.
14100 (mips_insert_attributes): Support multiple compression types.
14101 (mips_merge_decl_attributes): Likewise.
14102 (umips_12bit_offset_address_p): New function.
14103 (mips_start_function_definition): Emit .set micromips directive.
14104 (mips_call_may_need_jalx_p): New function.
14105 (mips_function_ok_for_sibcall): Add microMIPS support.
14106 (mips_print_operand_punctuation): Support short delay slots and
14108 (umips_swm_mask, umips_swm_encoding): New.
14109 (umips_build_save_restore): New function.
14110 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14111 (was_mips16_p): Remove.
14112 (old_compression_mode): New.
14113 (mips_set_compression_mode): New function.
14114 (mips_set_current_function): Add microMIPS support.
14115 (mips_option_override): Likewise.
14116 (umips_save_restore_pattern_p): New function.
14117 (umips_output_save_restore): New function.
14118 (umips_load_store_pair_p_1): New function.
14119 (umips_load_store_pair_p): New function.
14120 (umips_output_load_store_pair_1): New function.
14121 (umips_output_load_store_pair): New function.
14122 (umips_movep_target_p) New function.
14123 (mips_prepare_pch_save): Add microMIPS support.
14124 * config/mips/mips.h (TARGET_COMPRESSION): New.
14125 (TARGET_CPU_CPP_BUILTINS): Update macro
14126 to use new compression flags and to support microMIPS.
14127 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14128 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14129 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14130 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14131 (ASM_SPEC): Support mmicromips and mno-micromips.
14132 (M16STORE_REG_P): New macro.
14133 (MIPS_CALL): Support TARGET_MICROMIPS.
14134 (MICROMIPS_J): New macro.
14135 (mips_base_mips16): Rename this...
14136 (mips_base_compression_flags): ...to this.
14137 (UMIPS_12BIT_OFFSET_P): New macro.
14138 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14139 (MULTILIB_DIRNAMES): Likewise.
14140 2013-03-20 Richard Biener <rguenther@suse.de>
14142 PR tree-optimization/56661
14143 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14144 the result does not have to be distinct.
14146 2013-03-20 Richard Biener <rguenther@suse.de>
14148 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14151 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14152 Steven Bosscher <steven@gcc.gnu.org>
14154 PR rtl-optimization/56605
14155 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14157 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14160 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14161 that require movd instead of movq.
14163 2013-03-20 Richard Biener <rguenther@suse.de>
14165 * tree-ssa-structalias.c (struct variable_info): Add pointer
14166 to the first field of an aggregate with sub-vars. Make
14167 this and the pointer to the next subfield its ID.
14168 (vi_next): New function.
14169 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14170 storedanything_id, integer_id): Increment by one.
14171 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14172 get_call_clobber_vi): Adjust.
14173 (solution_set_expand): Simplify and speedup.
14174 (solution_set_add): Inline into ...
14175 (set_union_with_increment): ... this. Adjust accordingly.
14176 (do_sd_constraint): Likewise.
14177 (do_ds_constraint): Likewise.
14178 (do_complex_constraint): Simplify.
14179 (build_pred_graph): Adjust.
14180 (solve_graph): Likewise. Simplify and speedup.
14181 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14182 get_constraint_for_component_ref, get_constraint_for_1,
14183 first_vi_for_offset, first_or_preceding_vi_for_offset,
14184 create_function_info_for, create_variable_info_for_1,
14185 create_variable_info_for, intra_create_variable_infos): Adjust.
14186 (init_base_vars): Push NULL for ID zero.
14187 (compute_points_to_sets): Adjust.
14189 2013-03-20 Richard Biener <rguenther@suse.de>
14191 * cfgloop.c (verify_loop_structure): Streamline and avoid
14192 ICEing on corrupt loop tree.
14193 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14196 2013-03-20 Richard Biener <rguenther@suse.de>
14198 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14199 check whether an SSA update is needed.
14201 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14203 * config/mips/constraints.md (T): Rename to...
14207 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14208 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14210 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14212 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14213 (*subsi3_carryin_uxtw): Likewise.
14215 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14217 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14218 (*rorsi3_insn_uxtw): Likewise.
14220 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14222 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14223 (*extrsi5_insn_uxtw): Likewise.
14225 2013-03-19 Richard Biener <rguenther@suse.de>
14227 PR tree-optimization/56273
14228 * passes.c (init_optimization_passes): Move second VRP after DOM.
14230 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14232 * config/i386/i386.md (*movti_internal): Merge from
14233 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14234 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14235 nox64 isa attributes.
14237 2013-03-18 Richard Biener <rguenther@suse.de>
14239 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14241 (merge_node_constraints): Likewise.
14242 (build_succ_graph): Likewise.
14243 (valid_graph_edge): Inline into single caller.
14244 (unify_nodes): Likewise. Use bitmap_set_bit return value
14246 (scc_visit): Fix formatting and variable use.
14247 (do_sd_constraint): Use gcc_checking_assert.
14248 (do_ds_constraint): Likewise.
14249 (do_complex_constraint): Likewise.
14250 (condense_visit): Likewise. Cleanup.
14251 (dump_pred_graph): New function.
14252 (perform_var_substitution): Dump the pred-graph before
14253 variable substitution.
14254 (find_equivalent_node): Use gcc_checking_assert.
14255 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14257 2013-03-18 Richard Biener <rguenther@suse.de>
14259 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14260 Remove cond_expr_stmt_list argument and do not gimplify the
14262 (vect_loop_versioning): Adjust.
14263 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14264 Cleanup to use less temporaries.
14265 (vect_create_data_ref_ptr): Cleanup.
14267 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14269 PR tree-optimization/56635
14270 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14271 require types_compatible_p types.
14273 2013-03-18 Nick Clifton <nickc@redhat.com>
14275 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14276 spurious backslash.
14278 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14279 Add missing line to comment describing function.
14281 2013-03-18 Richard Biener <rguenther@suse.de>
14283 PR tree-optimization/56210
14284 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14285 Handle string / character search functions.
14286 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14288 2013-03-18 Richard Biener <rguenther@suse.de>
14290 PR middle-end/56483
14291 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14292 and implement properly.
14293 * gimple.h (gimple_cond_single_var_p): Remove.
14295 2013-03-18 Richard Biener <rguenther@suse.de>
14297 * tree-data-ref.h (find_data_references_in_loop): Declare.
14298 * tree-data-ref.c (get_references_in_stmt): Use a stack
14299 vector pre-allocated in the callers.
14300 (find_data_references_in_stmt): Adjust.
14301 (graphite_find_data_references_in_stmt): Likewise.
14302 (create_rdg_vertices): Likewise.
14303 (find_data_references_in_loop): Export.
14304 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14305 Compute dependences here...
14306 (vect_analyze_data_refs): ...not here. When we encounter
14307 a non-vectorizable data reference in basic-block vectorization
14308 truncate the data reference vector. Do not bother to
14309 fixup data-dependence information for gather loads.
14310 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14311 of data references, as reported.
14313 2013-03-18 Richard Biener <rguenther@suse.de>
14315 PR tree-optimization/3713
14316 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14317 has_constants and expr.
14318 (stmt_has_constants): Properly valueize SSA names when deciding
14319 whether the stmt has constants.
14321 2013-03-18 Richard Biener <rguenther@suse.de>
14323 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14324 whole function when there is nothing to do.
14325 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14326 * tree-vectorizer.c (vectorize_loops): Update virtual and
14327 loop-closed SSA once.
14328 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14330 2013-03-18 Richard Biener <rguenther@suse.de>
14332 PR middle-end/56113
14333 * domwalk.c (bb_postorder): New global static.
14334 (cmp_bb_postorder): New function.
14335 (walk_dominator_tree): Replace scheme imposing an order for
14336 visiting dominator sons by one sorting them at the time they
14337 are pushed on the stack.
14339 2013-03-18 Richard Biener <rguenther@suse.de>
14341 PR tree-optimization/39326
14342 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14343 (struct mem_ref): Replace mem member with ao_ref typed member.
14344 (MEM_ANALYZABLE): Adjust.
14345 (memref_eq): Likewise.
14346 (mem_ref_alloc): Likewise.
14347 (gather_mem_refs_stmt): Likewise.
14348 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14349 (execute_sm_if_changed_flag_set): Adjust.
14350 (execute_sm): Likewise.
14351 (ref_always_accessed_p): Likewise.
14352 (refs_independent_p): Likewise.
14353 (can_sm_ref_p): Likewise.
14355 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14358 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14359 return 1 even for !unsignedp.
14361 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14363 * config/i386/i386.md (isa): Add x64 and nox64.
14364 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14365 (*pushtf): Enable *roF alternative for x64 isa only.
14366 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14367 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14368 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14369 (*movtf_internal): Merge from *movtf_internal_rex64 and
14370 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14371 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14372 nox64 isa attributes.
14373 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14374 nox64 isa attributes.
14375 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14377 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14379 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14381 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14384 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14385 class names. Remove trailing comma after "ALL_REGS".
14387 2013-03-16 Jan Hubicka <jh@suse.cz>
14389 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14390 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14391 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14392 of cgraph_get_create_node.
14393 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14395 2013-03-16 Jason Merrill <jason@redhat.com>
14398 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14399 with DW_AT_default_value.
14401 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14403 * BASE-VER: Set to 4.9.0.
14405 2013-03-14 Andi Kleen <ak@linux.intel.com>
14408 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14409 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14410 Document _x* TSX intrinsics.
14412 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14413 David Holsgrove <david.holsgrove@xilinx.com>
14415 * configure.ac: Add MicroBlaze TLS support detection.
14416 * configure: Regenerate.
14417 * config/microblaze/microblaze-protos.h
14418 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14419 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14420 * config/microblaze/microblaze.c (microblaze_address_type): Add
14421 ADDRESS_TLS and tls_reloc address types.
14422 (microblaze_address_info): Add tls_reloc.
14423 (TARGET_HAVE_TLS): Define.
14424 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14425 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14426 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14427 load_tls_operand, microblaze_call_tls_get_addr,
14428 microblaze_legitimize_tls_address): New functions.
14429 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14430 (get_base_reg): Use microblaze_tls_symbol_p.
14431 (microblaze_classify_address): Handle TLS.
14432 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14433 label_mentioned_p and microblaze_tls_referenced_p.
14434 (microblaze_legitimize_address): Handle TLS.
14435 (microblaze_address_insns): Handle ADDRESS_TLS.
14436 (pic_address_needs_scratch): Handle TLS.
14437 (print_operand_address): Handle TLS.
14438 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14439 (microblaze_expand_move): Handle TLS.
14440 (microblaze_legitimate_constant_p): Check
14441 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14442 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14443 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14444 (PIC_OFFSET_TABLE_REGNUM): Set.
14445 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14446 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14447 (addsi3, movsi_internal2, movdf_internal): Update constraints
14448 * config/microblaze/predicates.md (arith_plus_operand): Define
14449 (move_operand): Redefine as move_src_operand,
14450 check microblaze_tls_referenced_p.
14452 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14454 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14455 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14457 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14459 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14462 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14464 PR tree-optimization/53265
14465 * common.opt (Waggressive-loop-optimizations): New option.
14466 * tree-ssa-loop-niter.c: Include tree-pass.h.
14467 (do_warn_aggressive_loop_optimizations): New function.
14468 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14469 if number_of_latch_executions returned constant.
14470 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14471 early. If number_of_latch_executions returned constant, set
14472 nb_iterations_upper_bound back to it.
14473 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14475 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14476 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14478 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14479 (MULTILIB_OSDIRNAMES): Set.
14480 * genmultilib: If defaultosdirname doesn't start with :: , set
14481 defaultosdirname2 instead, clear it and emit two . multilib_raw
14482 entries instead of just one.
14484 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14486 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14487 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14488 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14489 (SUBTARGET_OVERRIDE_OPTIONS): New.
14491 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14494 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14495 (musermode): Convert to Var(TARGET_USERMODE).
14496 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14497 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14498 * config/sh/sh.c (sh_option_override): Use
14499 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14500 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14502 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14504 (udivsi3_i4_single, divsi3_i4_single): Use
14505 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14507 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14509 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14512 2013-03-13 Richard Biener <rguenther@suse.de>
14514 PR tree-optimization/56608
14515 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14516 calls when vectorizing basic-blocks.
14518 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14521 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14524 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14526 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14528 2013-03-11 Jan Hubicka <jh@suse.cz>
14531 * lto-streamer-out.c (output_symbol_p): Skip references from
14532 constructors of external variables.
14534 2013-03-11 Jan Hubicka <jh@suse.cz>
14536 PR middle-end/56571
14537 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14539 * emit-rtl.c (verify_rtx_sharing): Likewise.
14540 (copy_insn_1): Likewise.
14541 * rtl.c (copy_rtx): Likewise.
14543 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14546 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14547 output_operand_lossage message.
14549 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14552 * arm.c (shift_op): Validate RTL pattern on the fly.
14553 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14556 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14559 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14560 new call patterns that clobber %fr12.
14561 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14562 split and postreload patterns.
14563 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14564 registers %fr12 and %fr12R as call used.
14566 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14568 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14569 (canon_address, record_store, replace_read, check_mem_read_rtx,
14570 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14571 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14572 rest_of_handle_dse): Likewise.
14574 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14576 PR middle-end/56524
14577 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14579 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14580 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14581 (save_optabs_if_changed): Replace with...
14582 (init_tree_optimization_optabs): ...this.
14583 * optabs.c (save_optabs_if_changed): Rename to...
14584 (init_tree_optimization_optabs): ...this. Take the optimization node
14585 as argument. Do nothing if the base optabs are already correct.
14586 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14587 to recompute optabs.
14588 * function.h (function): Remove optabs field.
14589 * function.c (invoke_set_current_function_hook): Call
14590 init_tree_optimization_optabs. Use the result to initialize
14593 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14595 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14596 if GTMA_HAS_NO_INSTRUMENTATION.
14597 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14598 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14599 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14600 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14601 GTMA_HAS_NO_INSTRUMENTATION.
14603 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14605 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14608 2013-03-08 Marek Polacek <polacek@redhat.com>
14609 Jakub Jelinek <jakub@redhat.com>
14611 PR tree-optimization/56478
14612 * predict.c (is_comparison_with_loop_invariant_p): Change the
14613 type of loop_step to tree.
14614 (predict_loops): Adjust.
14615 (predict_iv_comparison): Perform the computations on double_ints.
14617 2013-03-08 Richard Biener <rguenther@suse.de>
14619 PR tree-optimization/56570
14620 * tree-cfg.c (verify_expr_location_1): Verify locations for
14622 * tree-sra.c (create_access_replacement): Strip locations
14623 from DECL_DEBUG_EXPRs.
14625 2013-03-08 Richard Biener <rguenther@suse.de>
14627 * tree-inline.c (expand_call_inline): Do not associate
14628 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
14629 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
14631 2013-03-08 Richard Biener <rguenther@suse.de>
14633 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
14634 or block changes with -Og. Fix for location / block encoding
14635 changes and PHI arguments with locations.
14637 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
14639 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
14641 (struct output_info): Likewise.
14642 (register_overhead): Remove bad gcc_assert.
14643 (bitmap_find_bit): If there is only a single bitmap element, do not
14644 count a miss as a search.
14645 (print_statistics): Update for counter type changes.
14646 (dump_bitmap_statistics): Likewise. Print headers such that they
14647 are properly lined up with the printed counters.
14649 2013-03-07 Jakub Jelinek <jakub@redhat.com>
14651 PR tree-optimization/56559
14652 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
14653 check that it has only a single use.
14655 2013-03-07 Richard Biener <rguenther@suse.de>
14657 * doc/invoke.texi (fwhole-program): Discourage use in combination
14660 2013-03-06 Jakub Jelinek <jakub@redhat.com>
14662 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
14664 PR tree-optimization/56539
14665 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
14666 instead of GSI_CONTINUE_LINKING as last argument to
14667 force_gimple_operand_gsi. Adjust function comment.
14669 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
14672 PR middle-end/56548
14673 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
14674 promoted mode, convert the result back to the original mode.
14676 2013-03-06 Richard Biener <rguenther@suse.de>
14678 PR middle-end/56294
14679 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
14680 (insert_updated_phi_nodes_compare_uids): New function.
14681 (update_ssa): Sort symbols_to_rename after UID before
14682 traversing it to insert PHI nodes.
14684 2013-03-06 Richard Biener <rguenther@suse.de>
14686 PR middle-end/50494
14687 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
14688 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
14691 2013-02-13 Richard Biener <rguenther@suse.de>
14694 * varasm.c (output_constant_def_1): Get the decl representing
14695 the constant as argument.
14696 (output_constant_def): Wrap output_constant_def_1.
14697 (make_decl_rtl): Use output_constant_def_1 with the decl
14698 representing the constant.
14699 (build_constant_desc): Optionally re-use a decl already
14700 representing the constant.
14701 (tree_output_constant_def): Adjust.
14703 2013-03-06 Joey Ye <joey.ye@arm.com>
14706 * gcc.c (convert_white_space): New function.
14707 (main): Handles white space in function name.
14709 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
14712 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
14713 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
14714 to SH_DIV_CALL_TABLE for TARGET_SH2.
14715 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
14717 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
14718 call-table options.
14720 2013-03-05 Sterling Augustine <saugustine@google.com>
14721 Cary Coutant <ccoutant@google.com>
14724 * dwarf2out.c (resolve_addr): Don't call
14725 remove_loc_list_addr_table_entries a second time for the same
14728 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14731 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
14732 (avoid_complex_debug_insns): New function.
14733 (expand_debug_locations): Call it.
14735 PR rtl-optimization/56484
14736 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
14737 lifetimes of hard registers on small register class machines.
14739 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14741 * config/microblaze/microblaze-protos.h: Rename
14742 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
14743 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
14745 (microblaze_fast_interrupt_function_p): New function.
14746 (microblaze_is_interrupt_handler): Rename to
14747 microblaze_is_interrupt_variant and add fast_interrupt check.
14748 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
14749 (save_restore_insns): Likewise.
14750 (compute_frame_size): Likewise.
14751 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
14752 (microblaze_globalize_label): Likewise.
14753 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
14754 * config/microblaze/microblaze.md: Use wrapper
14755 microblaze_is_interrupt_variant.
14757 2013-03-05 Kai Tietz <ktietz@redhat.com>
14759 * sdbout.c (sdbout_one_type): Switch to current function's section
14760 supporting cold/hot.
14762 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14764 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
14767 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14769 PR middle-end/56461
14770 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
14771 if VALGRIND_GET_VBITS is defined, temporarily make object
14772 memory all defined, and restore previous valgrind addressability
14773 and definability afterwards. Free this_object at the end.
14775 PR middle-end/56461
14776 * lra.c (lra): Call lra_clear_live_ranges if live_p,
14777 right before calling lra_create_live_ranges, also call it
14778 when clearing live_p. Only call lra_clear_live_ranges
14779 at the end if live_p.
14781 PR middle-end/56461
14782 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
14784 2013-03-05 Richard Biener <rguenther@suse.de>
14786 PR tree-optimization/56521
14787 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14790 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14793 * except.h (remove_unreachable_eh_regions): New prototype.
14794 * except.c (remove_eh_handler_splicer): New function, split out
14795 of remove_eh_handler.
14796 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14797 warning about running it on many EH regions one at a time.
14798 (remove_unreachable_eh_regions_worker): New function, walk the
14799 EH tree in depth-first order and remove non-marked regions.
14800 (remove_unreachable_eh_regions): New function.
14801 * tree-eh.c (mark_reachable_handlers): New function, split out
14802 from remove_unreachable_handlers.
14803 (remove_unreachable_handlers): Use mark_reachable_handlers and
14804 remove_unreachable_eh_regions.
14805 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14806 and remove_unreachable_eh_regions.
14808 2013-03-05 Richard Biener <rguenther@suse.de>
14810 PR middle-end/56525
14811 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14812 not freeing them until the end.
14814 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14816 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14818 2013-03-05 Richard Biener <rguenther@suse.de>
14820 PR tree-optimization/56270
14821 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14822 of loads after scheduling an SLP instance.
14824 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14826 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14828 (check_gcc_parallelize): Run guality.exp as a separate job from
14829 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14830 struct-layout-1.exp with stackalign.exp.
14832 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14834 PR middle-end/56461
14835 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14836 load_index sbitmap even if some bit in it isn't set.
14838 PR middle-end/56461
14839 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14840 (discover_iteration_bound_by_body_walk): Change queues to
14841 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14842 spelling in comment. Call safe_push on queues[bound_index] directly.
14843 Release queues[queue_index] in every iteration unconditionally.
14844 Release bounds vector.
14846 PR middle-end/56461
14847 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14848 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14849 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14852 2013-03-05 Richard Biener <rguenther@suse.de>
14855 * tree-inline.c (remap_blocks_to_null): New function.
14856 (expand_call_inline): When expanding a call stmt without
14857 an associated block inline remap all callee blocks to NULL.
14859 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14861 PR rtl-optimization/56494
14862 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14863 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14864 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14866 PR middle-end/56461
14867 * sel-sched-ir.c (free_sched_pools): Release
14868 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14869 if succs_info_pool.max_top isn't -1.
14872 * opts.c (opts_obstack, opts_concat): Moved to...
14873 * opts-common.c (opts_obstack, opts_concat): ... here.
14875 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14877 PR middle-end/56461
14878 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14880 2013-03-04 Martin Jambor <mjambor@suse.cz>
14882 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14883 all appropriate places.
14885 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14887 PR tree-optimization/56424
14888 * ipa-split.c (split_function): Do not set the RSO flag if result is
14889 not by reference and its type is a register type.
14891 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14893 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14894 (microblaze_legitimate_pic_operand): Likewise
14895 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14896 new function microblaze_legitimate_pic_operand
14897 * config/microblaze/microblaze-protos.h
14898 (microblaze_legitimate_pic_operand): Declare.
14900 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14902 * config/microblaze/predicates.md (call_insn_simple_operand):
14903 New predicate for supported rtx code types.
14904 * config/microblaze/microblaze.md (call_internal1): Use
14905 call_insn_simple_operand predicate.
14907 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14909 PR middle-end/56461
14910 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14911 partitions.ordered_remove.
14913 PR middle-end/56461
14914 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14915 vec_oprnds0.create (1) for modifier == NONE.
14917 PR middle-end/56461
14918 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14919 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14920 vec_oprnds1 right before pushing anything to it for
14923 PR middle-end/56461
14924 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14925 set nbbs to 0 instead of having separate code path.
14926 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14927 instead of false as last argument if returning NULL.
14929 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14931 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14932 the attribute is now called "target" instead of "option".
14933 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14934 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14935 attribute/pragma name for TARGET_OPTION_VALID_P and
14936 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14937 * doc/tm.texi: Regenerated.
14939 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14941 * config/microblaze/microblaze.c:
14942 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14943 * config/microblaze/microblaze.h: Add -mxl-reorder to
14945 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14946 instructions emitted if TARGET_REORDER.
14947 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14948 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14951 2013-03-01 Xinliang David Li <davidxl@google.com>
14953 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14956 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14958 PR middle-end/56461
14959 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14960 vector even when returning true. Fix up function comment formatting.
14962 PR middle-end/56461
14963 * ira-build.c (ira_loop_nodes_count): New variable.
14964 (create_loop_tree_nodes): Initialize it.
14965 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14967 PR middle-end/56461
14968 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14969 method on dr_chain and result_chain.
14970 * tree-vect-stmts.c (vectorizable_store): Only call
14971 result_chain.create if j == 0.
14973 PR middle-end/56461
14974 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14975 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14976 before overwriting it.
14978 2013-03-01 Tobias Burnus <burnus@net-b.de>
14980 * doc/extended.texi (C Extensions): Change order in @menu
14982 (Other MIPS Built-in Functions): Move last MIPS entry before
14983 "picoChip Built-in Functions".
14984 (SH Built-in Functions): Move after RX Built-in Functions.
14985 * doc/gcc.texi (Introduction): Change order in @menu
14987 * doc/md.texi (Constraints): Ditto.
14988 * gty.texi (Type Information): Ditto.
14989 (User-provided marking routines for template types): Make
14991 * doc/invoke.texi (AArch64 Options): Move before
14992 "Adapteva Epiphany Options".
14994 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14995 Jakub Jelinek <jakub@redhat.com>
14998 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14999 no_address_safety_analysis attribute.
15000 * doc/extend.texi (no_address_safety_attribute): Rename to
15001 no_sanitize_address attribute, mention no_address_safety_analysis
15002 attribute as deprecated alias.
15004 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15006 PR middle-end/56461
15007 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15008 type to vec<vec<tree> > *.
15009 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15010 to be vec<tree> instead of vec<tree> *, set vec_defs
15011 to vNULL and call vec_defs.create (number_of_vects), adjust other
15013 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15014 vectorizable_condition): Adjust vect_get_slp_defs callers.
15016 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15018 * config/aarch64/aarch64.c
15019 (aarch64_float_const_representable): Remove unused variable.
15021 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15023 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15025 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15027 * config/aarch64/aarch64-builtins.c
15028 (aarch64_init_simd_builtins): Make static.
15030 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15032 * config/aarch64/aarch64.c
15033 (aarch64_simd_make_constant): Make static.
15035 2013-02-28 Martin Jambor <mjambor@suse.cz>
15037 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15038 with no initialization to the RHS of debug statements.
15040 2013-02-28 Martin Jambor <mjambor@suse.cz>
15042 PR tree-optimization/56294
15043 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15045 (get_access_replacement): Do not call create_access_replacement.
15046 Assert a replacement exists.
15047 (get_repl_default_def_ssa_name): Create the replacement declaration
15050 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15052 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15053 final_end_function.
15055 2013-02-28 Marek Polacek <polacek@redhat.com>
15057 PR rtl-optimization/56466
15058 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15059 if we're changing a loop.
15060 (peel_loops_completely): Likewise.
15062 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15065 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15067 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15070 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15071 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15072 INTX_FTYPE_FX, FX_FTYPE_INTX.
15073 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15075 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15077 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15078 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15079 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15080 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15081 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15082 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15083 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15084 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15085 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15086 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15087 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15088 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15089 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15090 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15091 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15092 (avrxmega6): Increase max flash segments from 5 to 6.
15093 * config/avr/t-multilib: Regenerate.
15094 * config/avr/avr-tables.opt: Regenerate.
15095 * doc/avr-mmcu.texi: Regenerate.
15097 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15099 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15100 (avr_device_to_arch): Rename to avr_device_to_ld.
15101 (avr_device_to_as): New prototype.
15102 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15103 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15104 * config/avr/driver-avr.c (avr_device_to_as): New.
15105 (avr_device_to_arch): Rename to avr_device_to_ld.
15107 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15109 PR middle-end/56461
15110 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15111 method on dr_chain and result_chain.
15113 PR middle-end/56461
15114 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15115 pointer_set_destroy on not_executed_last_iteration.
15117 PR middle-end/56461
15118 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15120 PR middle-end/56461
15121 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15122 FOR_EACH_DEFINED_FUNCTION when freeing state.
15124 PR middle-end/56461
15125 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15127 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15130 PR middle-end/56461
15131 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15132 known_aggs[i].items and release known_aggs vector.
15134 PR middle-end/56461
15135 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15137 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15139 * config/microblaze/microblaze.c (microblaze_emit_compare):
15140 Use xor for EQ/NE comparisions.
15141 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15142 (cbranchsf4): Adjust operator to comparison_operator.
15144 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15146 PR middle-end/56461
15147 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15149 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15150 vec_safe_push, always update *slot.
15151 (redirect_edge_var_map_clear): Use vec_free.
15152 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15153 (free_var_map_entry): Use vec_free.
15154 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15155 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15157 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15159 PR middle-end/45472
15160 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15161 when the may_trap_p bit of the exprs being merged differs.
15162 Reorder tests for speculativeness in the logical and operator.
15164 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15166 * incpath.c (add_standard_paths): Use reconcat instead of concat
15167 where appropriate and avoid leaking memory.
15169 * opts.h: Include obstack.h.
15170 (opts_concat): New prototype.
15171 (opts_obstack): New declaration.
15172 * opts.c (opts_concat): New function.
15173 (opts_obstack): New variable.
15174 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15175 (finish_options): Use opts_concat instead of concat
15176 and XOBNEWVEC instead of XNEWVEC.
15177 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15178 generate_option): Likewise.
15179 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15180 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15183 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15184 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15186 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15188 PR middle-end/56461
15189 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15191 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15193 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15194 (arm_block_move_unaligned_straight): Likewise.
15195 (arm_adjust_block_mem): Likewise.
15197 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15200 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15201 temp, cond and label.
15202 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15205 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15206 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15209 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15210 of prologue/epilogue functions.
15213 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15214 Remove unused variable cfa_offset.
15215 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15218 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15219 type promotion to unsigned.
15222 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15223 for HOST_WIDE_INT of 32 bit / same size as int.
15224 (arm_block_move_unaligned_straight): Likewise.
15225 (arm_adjust_block_mem): Likewise.
15228 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15231 2013-02-26 Marek Polacek <polacek@redhat.com>
15233 PR tree-optimization/56426
15234 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15236 2013-02-26 Richard Biener <rguenther@suse.de>
15239 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15240 unused variable loops.
15242 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15244 PR tree-optimization/56448
15245 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15246 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15247 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15248 later operands of the references, or even first operand for
15249 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15251 PR tree-optimization/56443
15252 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15253 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15254 to type_for_mode langhook.
15256 2013-02-25 Matt Turner <mattst88@gmail.com>
15258 * doc/invoke.texi: Document r4700.
15260 2013-02-25 Richard Biener <rguenther@suse.de>
15262 PR tree-optimization/56175
15263 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15265 (simplify_bitwise_binary): ... here. Also guard the conversion
15266 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15268 2013-02-25 Catherine Moore <clm@codesourcery.com>
15271 2013-02-24 Catherine Moore <clm@codesourcery.com>
15272 Maciej W. Rozycki <macro@codesourcery.com>
15273 Tom de Vries <tom@codesourcery.com>
15274 Nathan Sidwell <nathan@codesourcery.com>
15275 Iain Sandoe <iain@codesourcery.com>
15276 Nathan Froyd <froydnj@codesourcery.com>
15277 Chao-ying Fu <fu@mips.com>
15279 * doc/extend.texi: (micromips, nomicromips, nocompression):
15280 Document new function attributes.
15281 * doc/invoke.texi (minterlink-compressed, mmicromips,
15282 m14k, m14ke, m14kec): Document new options.
15283 (minterlink-mips16): Update documentation.
15284 * doc/md.texi (ZC, ZD): Document new constraints.
15285 * configure.ac (gcc_cv_as_micromips): Check if linker
15286 supports the .set micromips directive.
15287 * configure: Regenerate.
15288 * config.in: Regenerate.
15289 * config/mips/mips-tables.opt: Regenerate.
15290 * config/mips/micromips.md: New file.
15291 * constraints.md (ZC, AD): New constraints.
15292 * config/mips/predicates.md (movep_src_register): New predicate.
15293 (movep_src_operand): New predicate.
15294 (non_volatile_mem_operand): New predicate.
15295 * config/mips/mips.md (multimem): New type.
15296 (length): Differentiate between 17-bit and 18-bit branch offsets.
15297 (MOVEP1, MOVEP2): New mode iterator.
15298 (mov_<load>l): Use ZC constraint.
15299 (mov_<load>r): Likewise.
15300 (mov_<store>l): Likewise.
15301 (mov_<store>r): Likewise.
15302 (*branch_equality<mode>_inverted): Add microMIPS support.
15303 (*branch_equality<mode>): Likewise.
15304 (*jump_absolute): Likewise.
15305 (indirect_jump_<mode>): Likewise.
15306 (tablejump_<mode>): Likewise.
15307 (<optab>_internal): Likewise.
15308 (sibcall_internal): Likewise.
15309 (sibcall_value_internal): Likewise.
15310 (prefetch): Use constraint ZD.
15311 * config/mips/mips.opt (minterlink-compressed): New option.
15312 (minterlink-mips16): Now an alias for minterlink-compressed.
15313 (mmicromips): New option.
15314 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15315 (compare_and_swap_12): Likewise.
15316 (sync_add<mode>): Likewise.
15317 (sync_<optab>_12): Likewise.
15318 (sync_old_<optab>_12): Likewise.
15319 (sync_new_<optab>_12): Likewise.
15320 (sync_nand_12): Likewise.
15321 (sync_old_nand_12): Likewise.
15322 (sync_new_nand_12): Likewise.
15323 (sync_sub<mode>): Likewise.
15324 (sync_old_add<mode>): Likewise.
15325 (sync_old_sub<mode>): Likewise.
15326 (sync_new_add<mode>): Likewise.
15327 (sync_new_sub<mode>): Likewise.
15328 (sync_<optab><mode>): Likewise.
15329 (sync_old_<optab><mode>): Likewise.
15330 (sync_new_<optab><mode>): Likewise.
15331 (sync_nand<mode>): Likewise.
15332 (sync_old_nand<mode>): Likewise.
15333 (sync_new_nand<mode>): Likewise.
15334 (sync_lock_test_and_set<mode>): Likewise.
15335 (test_and_set_12): Likewise.
15336 (atomic_compare_and_swap<mode>): Likewise.
15337 (atomic_exchange<mode>_llsc): Likewise.
15338 (atomic_fetch_add<mode>_llsc): Likewise.
15339 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15340 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15341 (umips_save_restore_pattern_p): Likewise.
15342 (umips_load_store_pair_p): Likewise.
15343 (umips_output_load_store_pair): Likewise.
15344 (umips_movep_target_p): Likewise.
15345 (umips_12bit_offset_address_p): Likewise.
15346 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15347 (mips_base_mips16): Rename this...
15348 (mips_base_compression_flags): ...to this. Update all uses.
15349 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15350 (mips_mips16_decl_p): Delete.
15351 (mips_nomips16_decl_p): Delete.
15352 (mips_get_compress_on_flags): New function.
15353 (mips_get_compress_off_flags): New function.
15354 (mips_get_compress_mode): New function.
15355 (mips_get_compress_on_name): New function.
15356 (mips_get_compress_off_name): New function.
15357 (mips_insert_attributes): Support multiple compression types.
15358 (mips_merge_decl_attributes): Likewise.
15359 (umips_12bit_offset_address_p): New function.
15360 (mips_start_function_definition): Emit .set micromips directive.
15361 (mips_call_may_need_jalx_p): New function.
15362 (mips_function_ok_for_sibcall): Add microMIPS support.
15363 (mips_print_operand_punctuation): Support short delay slots and
15365 (umips_swm_mask, umips_swm_encoding): New.
15366 (umips_build_save_restore): New function.
15367 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15368 (was_mips16_p): Remove.
15369 (old_compression_mode): New.
15370 (mips_set_compression_mode): New function.
15371 (mips_set_current_function): Add microMIPS support.
15372 (mips_option_override): Likewise.
15373 (umips_save_restore_pattern_p): New function.
15374 (umips_output_save_restore): New function.
15375 (umips_load_store_pair_p_1): New function.
15376 (umips_load_store_pair_p): New function.
15377 (umips_output_load_store_pair_1): New function.
15378 (umips_output_load_store_pair): New function.
15379 (umips_movep_target_p) New function.
15380 (mips_prepare_pch_save): Add microMIPS support.
15381 * config/mips/mips.h (TARGET_COMPRESSION): New.
15382 (TARGET_CPU_CPP_BUILTINS): Update macro
15383 to use new compression flags and to support microMIPS.
15384 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15385 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15386 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15387 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15388 (ASM_SPEC): Support mmicromips and mno-micromips.
15389 (M16STORE_REG_P): New macro.
15390 (MIPS_CALL): Support TARGET_MICROMIPS.
15391 (MICROMIPS_J): New macro.
15392 (mips_base_mips16): Rename this...
15393 (mips_base_compression_flags): ...to this.
15394 (UMIPS_12BIT_OFFSET_P): New macro.
15395 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15396 (MULTILIB_DIRNAMES): Likewise.
15398 2013-02-25 Tom de Vries <tom@codesourcery.com>
15400 PR rtl-optimization/56131
15401 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15402 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15403 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15405 2013-02-25 Tobias Burnus <burnus@net-b.de>
15407 * doc/invoke.texi (-fsanitize=): Move from optimization
15408 to debugging options.
15410 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15412 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15414 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15415 Alexander Monakov <amonakov@ispras.ru>
15417 PR middle-end/56077
15418 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15419 flush pending lists also on non-jumps. Adjust comment.
15421 2013-02-24 Catherine Moore <clm@codesourcery.com>
15422 Maciej W. Rozycki <macro@codesourcery.com>
15423 Tom de Vries <tom@codesourcery.com>
15424 Nathan Sidwell <nathan@codesourcery.com>
15425 Iain Sandoe <iain@codesourcery.com>
15426 Nathan Froyd <froydnj@codesourcery.com>
15427 Chao-ying Fu <fu@mips.com>
15429 * doc/extend.texi: (micromips, nomicromips, nocompression):
15430 Document new function attributes.
15431 * doc/invoke.texi (minterlink-compressed, mmicromips,
15432 m14k, m14ke, m14kec): Document new options.
15433 (minterlink-mips16): Update documentation.
15434 * doc/md.texi (ZC, ZD): Document new constraints.
15435 * configure.ac (gcc_cv_as_micromips): Check if linker
15436 supports the .set micromips directive.
15437 * configure: Regenerate.
15438 * config.in: Regenerate.
15439 * config/mips/mips-tables.opt: Regenerate.
15440 * config/mips/micromips.md: New file.
15441 * constraints.md (ZC, AD): New constraints.
15442 * config/mips/predicates.md (movep_src_register): New predicate.
15443 (movep_src_operand): New predicate.
15444 (non_volatile_mem_operand): New predicate.
15445 * config/mips/mips.md (multimem): New type.
15446 (length): Differentiate between 17-bit and 18-bit branch offsets.
15447 (MOVEP1, MOVEP2): New mode iterator.
15448 (mov_<load>l): Use ZC constraint.
15449 (mov_<load>r): Likewise.
15450 (mov_<store>l): Likewise.
15451 (mov_<store>r): Likewise.
15452 (*branch_equality<mode>_inverted): Add microMIPS support.
15453 (*branch_equality<mode>): Likewise.
15454 (*jump_absolute): Likewise.
15455 (indirect_jump_<mode>): Likewise.
15456 (tablejump_<mode>): Likewise.
15457 (<optab>_internal): Likewise.
15458 (sibcall_internal): Likewise.
15459 (sibcall_value_internal): Likewise.
15460 (prefetch): Use constraint ZD.
15461 * config/mips/mips.opt (minterlink-compressed): New option.
15462 (minterlink-mips16): Now an alias for minterlink-compressed.
15463 (mmicromips): New option.
15464 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15465 (compare_and_swap_12): Likewise.
15466 (sync_add<mode>): Likewise.
15467 (sync_<optab>_12): Likewise.
15468 (sync_old_<optab>_12): Likewise.
15469 (sync_new_<optab>_12): Likewise.
15470 (sync_nand_12): Likewise.
15471 (sync_old_nand_12): Likewise.
15472 (sync_new_nand_12): Likewise.
15473 (sync_sub<mode>): Likewise.
15474 (sync_old_add<mode>): Likewise.
15475 (sync_old_sub<mode>): Likewise.
15476 (sync_new_add<mode>): Likewise.
15477 (sync_new_sub<mode>): Likewise.
15478 (sync_<optab><mode>): Likewise.
15479 (sync_old_<optab><mode>): Likewise.
15480 (sync_new_<optab><mode>): Likewise.
15481 (sync_nand<mode>): Likewise.
15482 (sync_old_nand<mode>): Likewise.
15483 (sync_new_nand<mode>): Likewise.
15484 (sync_lock_test_and_set<mode>): Likewise.
15485 (test_and_set_12): Likewise.
15486 (atomic_compare_and_swap<mode>): Likewise.
15487 (atomic_exchange<mode>_llsc): Likewise.
15488 (atomic_fetch_add<mode>_llsc): Likewise.
15489 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15490 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15491 (umips_save_restore_pattern_p): Likewise.
15492 (umips_load_store_pair_p): Likewise.
15493 (umips_output_load_store_pair): Likewise.
15494 (umips_movep_target_p): Likewise.
15495 (umips_12bit_offset_address_p): Likewise.
15496 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15497 (mips_base_mips16): Rename this...
15498 (mips_base_compression_flags): ...to this. Update all uses.
15499 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15500 (mips_mips16_decl_p): Delete.
15501 (mips_nomips16_decl_p): Delete.
15502 (mips_get_compress_on_flags): New function.
15503 (mips_get_compress_off_flags): New function.
15504 (mips_get_compress_mode): New function.
15505 (mips_get_compress_on_name): New function.
15506 (mips_get_compress_off_name): New function.
15507 (mips_insert_attributes): Support multiple compression types.
15508 (mips_merge_decl_attributes): Likewise.
15509 (umips_12bit_offset_address_p): New function.
15510 (mips_start_function_definition): Emit .set micromips directive.
15511 (mips_call_may_need_jalx_p): New function.
15512 (mips_function_ok_for_sibcall): Add microMIPS support.
15513 (mips_print_operand_punctuation): Support short delay slots and
15515 (umips_swm_mask, umips_swm_encoding): New.
15516 (umips_build_save_restore): New function.
15517 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15518 (was_mips16_p): Remove.
15519 (old_compression_mode): New.
15520 (mips_set_compression_mode): New function.
15521 (mips_set_current_function): Add microMIPS support.
15522 (mips_option_override): Likewise.
15523 (umips_save_restore_pattern_p): New function.
15524 (umips_output_save_restore): New function.
15525 (umips_load_store_pair_p_1): New function.
15526 (umips_load_store_pair_p): New function.
15527 (umips_output_load_store_pair_1): New function.
15528 (umips_output_load_store_pair): New function.
15529 (umips_movep_target_p) New function.
15530 (mips_prepare_pch_save): Add microMIPS support.
15531 * config/mips/mips.h (TARGET_COMPRESSION): New.
15532 (TARGET_CPU_CPP_BUILTINS): Update macro
15533 to use new compression flags and to support microMIPS.
15534 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15535 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15536 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15537 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15538 (ASM_SPEC): Support mmicromips and mno-micromips.
15539 (M16STORE_REG_P): New macro.
15540 (MIPS_CALL): Support TARGET_MICROMIPS.
15541 (MICROMIPS_J): New macro.
15542 (mips_base_mips16): Rename this...
15543 (mips_base_compression_flags): ...to this.
15544 (UMIPS_12BIT_OFFSET_P): New macro.
15545 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15546 (MULTILIB_DIRNAMES): Likewise.
15548 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15551 * target-globals.c (save_target_globals): For init_reg_sets and
15552 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15554 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15556 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15557 * config/aarch64/t-aarch64
15558 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15560 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15562 PR inline-asm/56148
15563 * lra-constraints.c (process_alt_operands): Reload operand
15564 conflicting with earlier clobber only if no more other conflicting
15567 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15570 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15571 if not linking a shared library.
15573 2013-02-22 Seth LaForge <sethml@google.com>
15575 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15577 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15579 * config/arm/arm.md (split for extendsidi): Update condition.
15580 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15581 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15582 (qhs_zextenddi_cstr): Likewise.
15584 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15586 PR middle-end/56420
15587 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15588 avoid signed wrapping.
15589 (expand_mult): Handle properly multiplication by
15590 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15591 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15592 in the compiler if coeff is HOST_WIDE_INT_MIN.
15593 (expand_divmod): Don't make ext_op1 static, change it's type to
15594 uhwi. Avoid undefined behavior in -INTVAL (op1).
15596 PR rtl-optimization/50339
15597 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15599 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15600 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15601 into splitting_ashiftrt field.
15602 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15604 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15607 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15609 PR middle-end/56108
15610 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15611 are sure to go irrevocable.
15613 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15615 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
15616 scalars are valid operands.
15618 2013-02-21 Martin Jambor <mjambor@suse.cz>
15620 PR tree-optimization/56310
15621 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
15622 only matching indices and non-negative final offsets.
15623 (intersect_aggregates_with_edge): Pass src_idx to
15624 agg_replacements_to_vector. Pass src_idx insstead of index to
15625 intersect_with_agg_replacements.
15627 2013-02-21 Martin Jambor <mjambor@suse.cz>
15629 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
15630 instead of hard-wired defaults.
15632 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
15634 * doc/invoke.texi (MIPS Options): Update documentation of the
15635 floating-point multiply-accumulate instruction restrictions.
15637 2013-02-21 Kostya Serebryany <kcc@google.com>
15639 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
15640 asan_shadow_offset on x86_64 linux.
15642 2013-02-21 Richard Biener <rguenther@suse.de>
15644 PR tree-optimization/56415
15646 2013-02-11 Richard Biener <rguenther@suse.de>
15648 PR tree-optimization/56273
15649 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15652 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15655 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
15658 PR inline-asm/56405
15659 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
15660 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
15662 2013-02-20 Jan Hubicka <jh@suse.cz>
15664 PR tree-optimization/56265
15665 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
15666 when target is referenced for first time.
15668 2013-02-20 Richard Biener <rguenther@suse.de>
15670 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
15671 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
15672 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
15673 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
15674 not return anything.
15675 (rename_ssa_copies): Do not remove unused locals.
15676 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
15677 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
15678 * passes.c (execute_function_todo): Do not schedule unused locals
15679 removal if cleanup_tree_cfg did something.
15680 * tree-ssa-live.c (remove_unused_locals): Dump statistics
15681 about the number of removed locals.
15683 2013-02-20 Richard Biener <rguenther@suse.de>
15685 PR tree-optimization/56398
15686 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
15688 2013-02-20 Martin Jambor <mjambor@suse.cz>
15690 PR tree-optimization/55334
15691 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
15692 restricted pointers to arrays.
15694 2013-02-20 Richard Biener <rguenther@suse.de>
15695 Jakub Jelinek <jakub@redhat.com>
15697 PR tree-optimization/56396
15698 * tree-ssa-ccp.c (n_const_val): New static variable.
15699 (get_value): Return NULL for SSA names we don't have a lattice
15701 (ccp_initialize): Initialize n_const_val.
15702 * tree-ssa-copy.c (n_copy_of): New static variable.
15703 (init_copy_prop): Initialize n_copy_of.
15704 (get_value): Return NULL_TREE for SSA names we don't have a
15707 2013-02-20 Martin Jambor <mjambor@suse.cz>
15709 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
15711 2013-02-20 Richard Biener <rguenther@suse.de>
15713 * genpreds.c (write_lookup_constraint): Do not compare first
15714 letter of the constraint again.
15716 2013-02-20 Richard Biener <rguenther@suse.de>
15718 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
15720 (get_use_iv_cost): Terminate hashtable walk when coming across
15723 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
15725 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
15726 reassociation for avx2 targets.
15728 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15730 * config/microblaze/microblaze.c: microblaze_has_clz = 0
15731 Add version check for v8.10.a to enable microblaze_has_clz
15732 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
15733 version and TARGET_PATTERN_COMPARE check
15734 * config/microblaze/microblaze.md: New clzsi2 instruction
15736 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15738 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
15739 function before branching.
15741 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15743 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
15745 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
15747 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15749 PR middle-end/55889
15750 * sel-sched.c: Include ira.h.
15751 (implicit_clobber_conflict_p): New function.
15752 (moveup_expr): Use it.
15753 * Makefile.in (sel-sched.o): Depend on ira.h.
15755 2013-02-19 Richard Biener <rguenther@suse.de>
15757 PR tree-optimization/56384
15758 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
15759 (vn_hash_type): Split out from ...
15760 (vn_hash_constant_with_type): ... here.
15761 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
15762 (vn_phi_eq): Compare types from vn_phi_s structure.
15763 (vn_phi_lookup): Populate vn_phi_s type.
15764 (vn_phi_insert): Likewise.
15766 2013-02-19 Jakub Jelinek <jakub@redhat.com>
15768 PR tree-optimization/56350
15769 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
15770 if haven't found reduction or nested cycle operand, rather than
15771 asserting we must find it.
15773 PR tree-optimization/56381
15774 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
15777 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
15778 Jakub Jelinek <jakub@redhat.com>
15781 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
15782 (swap_optab_enable): Same.
15783 (init_all_optabs): Use argument instead of global.
15784 * tree.h (struct tree_optimization_option): New field target_optabs.
15785 * expr.h (init_all_optabs): Add argument to prototype.
15786 (TREE_OPTIMIZATION_OPTABS): New.
15787 (save_optabs_if_changed): Protoize.
15788 * optabs.h: Declare this_fn_optabs.
15789 * optabs.c (save_optabs_if_changed): New.
15790 Declare this_fn_optabs.
15791 (init_optabs): Add argument to init_all_optabs() call.
15792 * function.c (invoke_set_current_function_hook): Handle per
15794 * function.h (struct function): New field optabs.
15795 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15796 optimization_current_node has changed.
15797 * target-globals.h (save_target_globals_default_opts): Protoize.
15798 * target-globals.c (save_target_globals_default_opts): New.
15800 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15803 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15804 registers %fr12 and %fr12R as call used.
15807 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15808 and HImode, require all displacements to be an integer multiple of
15810 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15811 only allow QImode and HImode when reload is in progress and strict is
15812 true. Likewise for symbolic addresses. Use base14_operand to check
15813 displacements in REG+BASE addresses.
15815 2013-02-18 Richard Biener <rguenther@suse.de>
15817 PR tree-optimization/56366
15818 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15819 handle sign-conversion of outer-loop initial induction value.
15821 2013-02-18 Richard Biener <rguenther@suse.de>
15823 PR middle-end/56349
15824 * cfghooks.c (merge_blocks): If we merge a latch into another
15825 block adjust references to it.
15826 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15827 (verify_loop_structure): Verify that a recorded latch is in fact
15830 2013-02-18 Richard Biener <rguenther@suse.de>
15832 PR tree-optimization/56321
15833 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15834 order SSA name release and virtual operand unlinking.
15836 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15838 * config/microblaze/microblaze.md (save_stack_block): Define.
15839 (restore_stack_block): Likewise.
15841 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15843 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15844 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15845 * config/microblaze/microblaze.c (microblaze_option_override):
15846 Bail out early for PIC modes when target does not support PIC.
15848 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15850 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15851 Replace with a microblaze version.
15852 (microblaze_trampoline_init): Adapt for microblaze.
15853 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15856 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15857 Dodji Seketeli <dodji@redhat.com>
15860 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15861 (instrument_mem_region_access): Do not forget to always put
15862 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15863 0) statement, even for cases where either 'base' or 'base + len'
15864 are not instrumented -- because they have been previously
15865 instrumented. Simplify the logic by putting all the statements
15866 instrument 'base + len' inside a sequence, and then insert that
15867 sequence right before the current insertion point. Then, to
15868 instrument 'base + len', just get an iterator on that statement.
15869 And do not forget to update the pointer to iterator the function
15870 received as argument.
15872 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15874 PR rtl-optimization/56348
15875 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15877 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15879 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15880 (clean_graph_dump_file): Pass base to start_graph_dump.
15882 2013-02-14 Richard Henderson <rth@redhat.com>
15885 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15887 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15889 * collect2-aix.h: Define F_LOADONLY.
15891 2013-02-14 Richard Biener <rguenther@suse.de>
15894 * varasm.c (output_constant_def_1): Get the decl representing
15895 the constant as argument.
15896 (output_constant_def): Wrap output_constant_def_1.
15897 (make_decl_rtl): Use output_constant_def_1 with the decl
15898 representing the constant.
15899 (build_constant_desc): Optionally re-use a decl already
15900 representing the constant.
15901 (tree_output_constant_def): Adjust.
15903 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15906 * asan.c (instrument_builtin_call): Really put the length of the
15907 second source argument into src1_len.
15909 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15911 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15912 argument. If it is false, don't create edge from then_bb to
15914 (insert_if_then_before_iter): Pass true to it.
15915 (build_check_stmt): Pass false to it.
15916 (transform_statements): Flush hash table only on extended basic
15917 block boundaries, rather than at the beginning of every bb.
15918 Don't flush hash table on nonfreeing_call_p calls.
15919 * tree-flow.h (nonfreeing_call_p): New prototype.
15920 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15922 2013-02-13 David S. Miller <davem@davemloft.net>
15924 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15926 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15929 * ira.c (max_regno_before_ira): Move from ...
15931 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15934 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15936 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15938 2013-02-13 Richard Biener <rguenther@suse.de>
15941 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15942 globals in MEM_REFs.
15944 2013-02-13 Richard Biener <rguenther@suse.de>
15946 * loop-init.c (loop_optimizer_init): Clear loop state when
15947 re-initializing preserved loops.
15948 * loop-unswitch.c (unswitch_single_loop): Return whether
15949 we unswitched the loop. Do not verify loop state here.
15950 (unswitch_loops): When we unswitched a loop discover new loops.
15952 2013-02-13 Kostya Serebryany <kcc@google.com>
15954 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15956 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15958 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15960 Avoid instrumenting duplicated memory access in the same basic block
15961 * Makefile.in (asan.o): Add new dependency on hash-table.h
15962 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15963 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15964 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15965 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15966 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15967 (get_mem_ref_of_assignment): New functions.
15968 (get_mem_refs_of_builtin_call): Extract from
15969 instrument_builtin_call and tweak a little bit to make it fit with
15971 (instrument_builtin_call): Use the new
15972 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15973 of is_gimple_builtin_call.
15974 (instrument_derefs, instrument_mem_region_access): Insert the
15975 instrumented memory reference into the hash table.
15976 (maybe_instrument_assignment): Renamed instrument_assignment into
15977 this, and change it to advance the iterator when instrumentation
15978 actually happened and return true in that case. This makes it
15979 homogeneous with maybe_instrument_assignment, and thus give a
15980 chance to callers to be more 'regular'.
15981 (transform_statements): Clear the memory reference hash table
15982 whenever we enter a new BB, when we cross a function call, or when
15983 we are done transforming statements. Use
15984 maybe_instrument_assignment instead of instrumentation. No more
15985 need to special case maybe_instrument_assignment and advance the
15986 iterator after calling it; it's now handled just like
15987 maybe_instrument_call. Update comment.
15989 2013-02-13 Richard Biener <rguenther@suse.de>
15991 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15992 Fix loop discovery code.
15994 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15996 PR inline-asm/56148
15997 * lra-constraints.c (process_alt_operands): Match early clobber
15998 operand with itself. Check conflicts with earlyclobber only if
15999 the operand is not reloaded. Prefer to reload conflicting operand
16000 if earlyclobber and matching operands are the same.
16002 2013-02-12 Richard Biener <rguenther@suse.de>
16005 * lto-streamer-out.c (write_symbol): Do not output symbols
16006 for hard register variables.
16008 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16011 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16012 (umulsidi3_insn, mulsidi3_insn): New insns.
16014 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16016 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16017 (struct tune_params): Add vec_costs field.
16018 * config/arm/arm.c (arm_builtin_vectorization_cost)
16019 (arm_add_stmt_cost): New functions.
16020 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16021 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16022 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16023 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16024 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16025 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16026 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16028 2013-02-12 Richard Biener <rguenther@suse.de>
16031 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16032 decls again if possible.
16034 2013-02-12 Richard Biener <rguenther@suse.de>
16036 PR middle-end/56288
16037 * tree-ssa.c (verify_ssa_name): Fix check, move
16038 SSA_NAME_IN_FREE_LIST check up.
16040 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16041 Steven Bosscher <steven@gcc.gnu.org>
16043 PR rtl-optimization/56151
16044 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16045 equal to op0 or op1, and last_insn pattern is CODE operation
16046 with MEM dest and one of the operands matches that MEM.
16048 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16050 * doc/extend.texi: Document Function Multiversioning and "default"
16051 parameter string to target attribute.
16052 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16053 target attribute parameter is "default".
16054 (ix86_compare_version_priority): Remove checks for target attribute.
16055 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16056 Remove check for target attribute equal to NULL. Add assert.
16057 (ix86_generate_version_dispatcher_body): Change error to sorry.
16059 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16060 Jack Howarth <howarth@bromo.med.uc.edu>
16061 Patrick Marlier <patrick.marlier@gmail.com>
16064 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16065 define ENDFILE_SPEC as TM_DESTRUCTOR.
16066 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16068 2013-02-11 Alexander Potapenko <glider@google.com>
16069 Jack Howarth <howarth@bromo.med.uc.edu>
16070 Jakub Jelinek <jakub@redhat.com>
16073 * config/darwin.c (cdtor_record): Rename ctor_record.
16074 (sort_cdtor_records): Rename sort_ctor_records.
16075 (finalize_dtors): New routine to sort destructors by
16076 priority before use in assemble_integer.
16077 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16079 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16081 PR rtl-optimization/56275
16082 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16083 offset is non-negative and less than cmode size before
16084 calling simplify_subreg.
16086 2013-02-11 Richard Biener <rguenther@suse.de>
16088 PR tree-optimization/56264
16089 * cfgloop.h (fix_loop_structure): Adjust prototype.
16090 * loop-init.c (fix_loop_structure): Return the number of
16091 newly discovered loops.
16092 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16093 are discovered, do a full loop-closed SSA rewrite.
16095 2013-02-11 Richard Biener <rguenther@suse.de>
16097 PR tree-optimization/56273
16098 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16100 (check_array_ref): Fix missing newline in dumps.
16101 (search_for_addr_array): Likewise.
16103 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16105 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16107 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16110 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16112 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16114 PR rtl-optimization/56246
16115 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16117 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16118 constraints are satisfied.
16120 2013-02-08 Jeff Law <law@redhat.com>
16123 * emit-rtl.c (reg_is_parm_p): New function.
16124 * regs.h (reg_is_parm_p): New prototype.
16125 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16126 callee-clobbered registers.
16128 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16131 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16132 If there is no implicit builtin declaration, just return NULL.
16134 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16136 * config/i386/sse.md (FMAMODEM): New mode iterator.
16137 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16138 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16140 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16142 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16143 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16144 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16146 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16148 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16149 (microblaze*-*-elf): Likewise.
16150 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16152 * config/microblaze/microblaze-c.c: Add builtin defines for
16153 _LITTLE_ENDIAN and _BIG_ENDIAN.
16154 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16155 add to TARGET_DEFAULT flags.
16156 Expand ASM_SPEC and LINK_SPEC.
16157 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16158 * config/microblaze/microblaze.md: Update extendsidi2 and
16159 movdi_internal instructions to use low-order / high-order reg
16161 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16162 options and inversemask / mask of LITTLE_ENDIAN.
16163 * config/microblaze/t-microblaze: Expand multilib options to
16164 include mlittle-endian (le) and update exceptions patterns.
16166 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16168 PR rtl-optimization/56195
16169 * lra-constraints.c (get_reload_reg): Don't reuse regs
16170 if they have smaller mode than requested, if they have
16171 wider mode than requested, try to return a SUBREG.
16173 PR tree-optimization/56250
16174 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16175 if type is unsigned and code isn't MULT_EXPR.
16177 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16179 PR tree-optimization/56064
16180 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16181 bits according to mode.
16182 * fixed-value.h (fixed_from_double_int)
16183 (const_fixed_from_double_int): Adjust comments.
16185 2013-02-08 Richard Biener <rguenther@suse.de>
16188 * lto-streamer.h (struct data_in): Remove current_file, current_line
16189 and current_col members.
16190 * lto-streamer-out.c (lto_output_location): Stream changed bits
16191 en-block for efficiency.
16192 * lto-streamer-in.c (clear_line_info): Remove.
16193 (lto_input_location): Cache current file, line and column
16194 globally via local statics. Read changed bits en-block.
16195 (input_function): Do not call clear_line_info.
16196 (lto_read_body): Likewise.
16197 (lto_input_toplevel_asms): Likewise.
16199 2013-02-08 Michael Matz <matz@suse.de>
16201 PR tree-optimization/52448
16202 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16203 (nt_call_phase): New static.
16204 (add_or_mark_expr): Only mark accesses with newer phase than any
16206 (nonfreeing_call_p): New.
16207 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16208 (nt_fini_block): Keep blocks marked as visited.
16209 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16211 2013-02-08 Richard Biener <rguenther@suse.de>
16213 * ira.c (ira): Free broken dominator information.
16215 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16217 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16219 2013-02-08 Marek Polacek <polacek@redhat.com>
16221 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16223 2013-02-08 Richard Biener <rguenther@suse.de>
16225 PR middle-end/56181
16226 * cfgloop.h (flow_loops_find): Adjust.
16227 (bb_loop_header_p): Declare.
16228 * cfgloop.c (bb_loop_header_p): New function split out from ...
16229 (flow_loops_find): ... here. Adjust function signature,
16230 support incremental loop structure update.
16231 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16232 * cfgloopmanip.c (fix_loop_structure): Move ...
16233 * loop-init.c (fix_loop_structure): ... here.
16234 (apply_loop_flags): Split out from ...
16235 (loop_optimizer_init): ... here.
16236 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16237 in incremental mode, only remove dead loops here.
16239 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16242 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16243 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16244 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16245 (*round<mode>3.libgcc): New insns for fixed-modes.
16246 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16247 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16248 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16249 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16250 implementations. Define to __builtin_avr_absFX,
16251 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16252 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16253 __builtin_avr_countlsFX, respectively.
16254 * config/avr/avr-c.c (target.h): Include it.
16255 (enum avr_builtin_id): New enum.
16256 (avr_resolve_overloaded_builtin): New static function.
16257 (avr_register_target_pragmas): Use it to set
16258 targetm.resolve_overloaded_builtin.
16259 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16260 tree nodes used by DEF_BUILTIN.
16261 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16262 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16263 <AVR_BUILTIN_xxBITS>: Same.
16265 2013-02-08 Richard Biener <rguenther@suse.de>
16267 * cfgloop.c (verify_loop_structure): Properly handle
16268 a loop exiting to another loop header.
16269 * ira-int.h (ira_loops): Remove.
16270 * ira.c (ira_loops): Remove.
16271 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16272 (do_reload): Use loop_optimizer_finalize.
16273 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16274 number_of_loops to access the loop tree.
16275 (more_one_region_p): Likewise.
16276 (finish_loop_tree_nodes): Likewise.
16277 (rebuild_regno_allocno_maps): Likewise.
16278 (mark_loops_for_removal): Likewise.
16279 (mark_all_loops_for_removal): Likewise.
16280 (remove_unnecessary_regions): Likewise.
16281 (ira_build): Likewise.
16282 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16284 2013-02-08 Richard Biener <rguenther@suse.de>
16286 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16287 * ipa-pure-const.c (analyze_function): Avoid calling
16288 mark_irreducible_loops twice.
16289 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16291 2013-02-07 David S. Miller <davem@davemloft.net>
16293 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16295 * var-tracking.c (vt_add_function_parameter): Test the presence of
16296 HAVE_window_save properly and do not remap argument registers when
16297 we have a leaf function.
16299 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16302 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16304 * config/i386/i386.c (ix86_print_operand): Ditto.
16306 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16308 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16310 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16312 PR rtl-optimization/56225
16313 * lra-constraints.c (process_alt_operands): Check that reload hard
16314 reg can hold value for strict_low_part.
16316 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16319 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16320 dwarf2out_end_function.
16321 (in_first_function_p, maybe_at_text_label_p,
16322 first_loclabel_num_not_at_text_label): New variables.
16323 (dwarf2out_var_location): In the first function find out
16324 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16325 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16328 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16330 PR rtl-optimization/56178
16331 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16332 SUBREG of a register. Tidy up related block of code.
16333 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16334 note if the source is a register or a SUBREG of a register.
16336 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16339 * config/rs6000/rs6000.md (ptrm): New mode attr.
16340 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16341 call_value_indirect_aix<pttrsize>,
16342 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16345 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16347 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16348 if -bnortl. Convert to strcmp and strncmp.
16350 2013-02-07 Alan Modra <amodra@gmail.com>
16353 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16354 addresses won't wrap when offsetting.
16355 (rs6000_secondary_reload): Provide secondary reloads needed for
16356 wrapping LO_SUM addresses.
16358 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16360 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16361 MACH, just __MACH__.
16363 2013-02-06 Richard Biener <rguenther@suse.de>
16365 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16366 instead of calling fix_loop_structure.
16368 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16370 PR middle-end/56217
16371 * omp-low.c (use_pointer_for_field): Return false if
16372 lower_send_shared_vars doesn't generate any copy-out code.
16374 2013-02-06 Tom de Vries <tom@codesourcery.com>
16376 PR rtl-optimization/56131
16377 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16378 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16379 of the label is NULL. Add comment.
16381 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16383 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16386 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16387 (STATIC_LIBTSAN_LIBS): Likewise.
16388 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16389 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16390 is defined, don't add anything else beyond that.
16391 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16392 (LINK_COMMAND_SPEC): Use them.
16394 PR tree-optimization/56205
16395 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16396 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16397 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16399 2013-02-05 Richard Biener <rguenther@suse.de>
16401 PR tree-optimization/53342
16402 PR tree-optimization/53185
16403 * tree-vectorizer.h (vect_check_strided_load): Remove.
16404 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16405 not disallow peeling for vectorized strided loads.
16406 (vect_check_strided_load): Make static and simplify.
16407 (vect_analyze_data_refs): Adjust.
16408 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16409 correctly when vectorizing strided loads.
16411 2013-02-05 Richard Biener <rguenther@suse.de>
16413 * doc/install.texi: Refer to ISL, not PPL.
16415 2013-02-05 Jan Hubicka <jh@suse.cz>
16417 PR tree-optimization/55789
16418 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16420 2013-02-05 Jan Hubicka <jh@suse.cz>
16422 PR tree-optimization/55789
16423 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16424 the dead call anyway.
16426 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16429 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16431 2013-02-04 Alexander Potapenko <glider@google.com>
16432 Jack Howarth <howarth@bromo.med.uc.edu>
16433 Jakub Jelinek <jakub@redhat.com>
16436 * config/darwin.c (sort_ctor_records): Stabilized qsort
16437 on constructor priority by using original position.
16438 (finalize_ctors): New routine to sort constructors by
16439 priority before use in assemble_integer.
16440 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16442 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16445 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16446 about visibility on artificial decls.
16447 * config/sol2.c (solaris_assemble_visibility): Likewise.
16449 2013-02-04 Kai Tietz <ktietz@redhat.com>
16452 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16453 argument and improve checking of return-argument types for 16-byte
16455 (ix86_function_value_1): Add additional valtype argument on call
16456 of function_value_64.
16457 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16458 handling infunction_value_64 function.
16460 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16462 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16464 2013-02-04 Richard Biener <rguenther@suse.de>
16466 PR tree-optimization/56188
16467 * tree-ssa-structalias.c (label_visit): Consider case with
16468 initially non-empty points-to set.
16469 (perform_var_substitution): Dump node mapping and clean up.
16471 2013-02-04 Richard Guenther <rguenther@suse.de>
16474 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16475 node prevail as last resort.
16476 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16477 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16479 2013-02-04 Richard Biener <rguenther@suse.de>
16481 PR tree-optimization/56113
16482 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16484 (equiv_class_lookup_or_add): ... this.
16485 (label_visit): Adjust and fix error in previous patch.
16486 (perform_var_substitution): Adjust.
16488 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16490 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16491 * config/sh/sh4-300.md: Likewise.
16492 * config/sh/sh4a.md: Likewise.
16493 * config/sh/constraints.md: Likewise.
16494 * config/sh/sh.md: Likewise.
16495 * config/sh/netbsd-elf.h: Likewise.
16496 * config/sh/predicates.md: Likewise.
16497 * config/sh/sh-protos.h: Likewise.
16498 * config/sh/ushmedia.h: Likewise.
16499 * config/sh/linux.h: Likewise.
16500 * config/sh/sh.c: Likewise.
16501 * config/sh/superh.h: Likewise.
16502 * config/sh/elf.h: Likewise.
16503 * config/sh/sh4.md: Likewise.
16504 * config/sh/sh.h: Likewise.
16506 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16508 * config/pa/constraints.md: Adjust unused letters. Change "T"
16509 constraint to match_test floating_point_store_memory_operand().
16510 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16511 (base14_operand): New.
16512 (floating_point_store_memory_operand): New.
16513 (integer_store_memory_operand): Revise to use base14_operand and
16514 reg_plus_base_memory_operand.
16515 (move_dest_operand): Allow symbolic_memory_operands.
16516 (symbolic_memory_operand): Check for LO_SOM.
16517 (symbolic_operand): Change default case to break.
16518 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16519 CONST_DOUBLE values to be reloaded by putting them into memory when
16520 the destination is a floating point register.
16521 (movdf): Remove code to handle CONST_DOUBLE.
16523 (reload_indf_r1): New.
16524 (reload_insf_r1): New.
16525 Consistently use "Q" and "T" constraints with integer and floating
16526 point move instructions, respectively.
16527 (movdi): Remove FAIL.
16528 Change predicate for source operand unamed DImode move from
16529 general_operand to move_src_operand.
16530 (umulsidi3): Change predicate for destination operand to
16532 Likewise for similar unamed patterns.
16533 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16534 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16535 (hppa_legitimize_address): Simplify mask calculation.
16536 (pa_emit_move_sequence): Revised handling of secondary reloads from
16537 REG+D addresses for floating point loads and stores. Directly handle
16538 loading CONST0_RTX (mode) to a floating point register.
16539 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16540 to floating point registers. Don't restrict secondary reloads to
16541 floating point registers to integer modes. Revise some comments and
16543 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16544 (pa_legitimate_address_p): New.
16545 (pa_legitimize_reload_address): New.
16546 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16547 (STRICT_REG_OK_FOR_BASE_P): New.
16548 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16549 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16551 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16552 Andrew Dixie <andrewd@gentrack.com>
16554 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16557 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16559 * expmed.c (extract_bit_field_1): Pass the full width of the
16560 structure to get_best_reg_extraction_insn.
16562 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16565 * configure.ac (use_cxa_atexit): Add AIX.
16566 * configure: Regenerate.
16568 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16570 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16573 * final.c (need_profile_function): New variable.
16574 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16575 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16576 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16577 notes, targetm.asm_out.function_prologue doesn't emit anything,
16578 HAVE_prologue and profiler should be emitted before prologue,
16579 set need_profile_function instead of emitting it.
16580 (final_scan_insn): If need_profile_function, emit
16581 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16582 NOTE_INSN_FUNCTION_BEG note.
16584 2013-02-01 Richard Henderson <rth@redhat.com>
16586 * config/rs6000/rs6000.md (smulditi3): New.
16589 * config/alpha/alpha.md (umulditi3): New.
16591 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16593 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16594 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16596 2013-02-01 Richard Biener <rguenther@suse.de>
16598 PR tree-optimization/56113
16599 * tree-ssa-structalias.c (label_visit): Reduce work for
16600 single-predecessor nodes.
16602 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16604 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16605 range isn't testing for zero.
16607 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16609 PR middle-end/56113
16610 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16612 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16613 Nick Clifton <nickc@redhat.com>
16615 * config/v850/constraints.md (Q): Define as a memory constraint.
16616 * config/v850/predicates.md (label_ref_operand): New predicate.
16617 (e3v5_shift_operand): New predicate.
16618 (ior_operator): New predicate.
16619 * config/v850/t-v850: Add e3v5 multilib.
16620 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
16621 (v850_gen_movdi): Prototype.
16622 * config/v850/v850.c: Add support for e3v5 architecture.
16623 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
16625 (construct_save_jarl): Add e3v5 long JARL support.
16626 (v850_adjust_insn_length): New function. Adjust length of call
16627 insns when using e3v5 instructions.
16628 (v850_gen_movdi): New function: Generate instructions to move a
16630 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
16631 (CPP_SPEC): Define __v850e3v5__ as appropriate.
16632 (TARGET_USE_FPU): Enable for e3v5.
16633 (CONST_OK_FOR_W): New macro.
16634 (ADJUST_INSN_LENGTH): Define.
16635 * config/v850/v850.md (UNSPEC_LOOP): Define.
16636 (attr cpu): Add v850e3v5.
16637 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
16638 (movdi): New pattern.
16639 (movdi_internal): New pattern.
16640 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
16641 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
16642 (cstoresf4): Likewise.
16643 (cstoredf4): Likewise.
16644 (insv): New pattern.
16645 (rotlso3_a): New pattern.
16646 (rotlsi3_b): New pattern
16647 (rotlsi3_v850e3v5): New pattern.
16648 (doloop_begin): New pattern.
16649 (fix_loop_counter): New pattern.
16650 (doloop_end): New pattern.
16651 (branch_normal): Add e3v5 long branch support.
16652 (branch_invert): Likewise.
16653 (branch_z_normal): Likewise.
16654 (branch_z_invert): Likewise.
16655 (branch_nz_normal): Likewise.
16656 (branch_nz_invert): Likewise.
16657 (call_internal_short): Add e3v5 register-indirect JARL support.
16658 (call_internal_long): Likewise.
16659 (call_value_internal_short): Likewise.
16660 (call_value_internal_long): Likewise.
16661 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
16662 (mloop): New option.
16663 * config.gcc: Add support for configuring v840e3v5 target.
16664 * doc/invoke.texi: Document new v850 specific command line options.
16666 2013-01-31 Paul Koning <ni1d@arrl.net>
16670 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
16671 children if parent is a class.
16672 (prune_unused_types_prune): Don't add DW_AT_declaration.
16674 2013-01-31 Richard Biener <rguenther@suse.de>
16676 PR tree-optimization/56157
16677 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
16678 match up operand with SLP child.
16680 2013-01-31 Jason Merrill <jason@redhat.com>
16683 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
16684 parameters the first time.
16685 (gen_scheduled_generic_parms_dies): Check completeness here.
16687 2013-01-31 Richard Biener <rguenther@suse.de>
16689 PR middle-end/53073
16690 * common.opt (faggressive-loop-optimizations): New flag,
16691 enabled by default.
16692 * doc/invoke.texi (faggressive-loop-optimizations): Document.
16693 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
16694 infer_loop_bounds_from_undefined by it.
16696 2013-01-31 Richard Biener <rguenther@suse.de>
16698 PR tree-optimization/56150
16699 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
16700 visit virtual operands.
16701 (find_uses_to_rename_bb): Likewise.
16703 2013-01-31 Richard Biener <rguenther@suse.de>
16705 PR tree-optimization/56150
16706 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
16707 mixed store non-store stmts.
16709 2013-01-30 Jakub Jelinek <jakub@redhat.com>
16712 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
16713 LIBASAN_EARLY_SPEC is defined.
16714 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
16715 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
16717 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
16720 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
16721 invalid args instead of ICEing on it.
16722 (ix86_valid_target_attribute_tree): Return error_mark_node if
16723 ix86_valid_target_attribute_inner_p failed.
16724 (ix86_valid_target_attribute_p): Return false only if
16725 ix86_valid_target_attribute_tree returned error_mark_node. Allow
16726 target("default") attribute.
16727 (sorted_attr_string): Change argument from const char * to tree,
16728 merge in all target attribute arguments rather than just one.
16729 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
16730 instead of free. Avoid using strcat.
16731 (ix86_mangle_function_version_assembler_name): Mangle
16732 target("default") as if no target attribute is present. Adjust
16733 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
16734 instead of xmalloc and XDELETEVEC instead of free.
16735 (ix86_function_versions): Don't return true if one of the decls
16736 doesn't have target attribute. If they don't and one of the decls
16737 is DECL_FUNCTION_VERSIONED, report an error. Adjust
16738 sorted_attr_string caller. Use XDELETEVEC instead of free.
16739 (ix86_supports_function_versions): Remove.
16740 (make_name): Fix up formatting.
16741 (make_dispatcher_decl): Remove resolver_name and its initialization.
16742 Avoid leaking memory.
16743 (is_function_default_version): Return true if there is
16744 target("default") attribute rather than no target attribute at all.
16745 (make_resolver_func): Avoid leaking memory.
16746 (ix86_generate_version_dispatcher_body): Likewise.
16747 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
16748 * target.def (supports_function_versions): Remove.
16749 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
16750 * doc/tm.texi: Regenerated.
16752 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
16754 PR rtl-optimization/56144
16755 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
16756 for values with side effects.
16758 2013-01-30 Richard Biener <rguenther@suse.de>
16760 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
16761 (sparseset_pop): Likewise.
16762 * cfganal.c (compute_idf): Likewise. Increase work-stack size
16763 to be able to use quick_push in the worker loop.
16765 2013-01-30 Marek Polacek <polacek@redhat.com>
16767 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
16769 2013-01-30 Richard Biener <rguenther@suse.de>
16772 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
16774 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
16776 PR tree-optimization/56064
16777 * fixed-value.c (fixed_from_double_int): New function.
16778 * fixed-value.h (fixed_from_double_int): New prototype.
16779 (const_fixed_from_double_int): New static inline function.
16780 * fold-const.c (native_interpret_fixed): New static function.
16781 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
16782 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
16783 (native_encode_fixed): New static function.
16784 (native_encode_expr) <FIXED_CST>: Use it.
16785 (native_interpret_int): Move double_int worker code to...
16786 * double-int.c (double_int::from_buffer): ...this new static method.
16787 * double-int.h (double_int::from_buffer): Prototype it.
16789 2013-01-30 Richard Biener <rguenther@suse.de>
16791 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16792 New pointer-map and obstack.
16793 (init_alias_vars): Allocate pointer-map and obstack.
16794 (delete_points_to_sets): Free them.
16795 (find_what_var_points_to): Cache result.
16796 (find_what_p_points_to): Adjust for changed interface of
16797 find_what_var_points_to.
16798 (compute_points_to_sets): Likewise.
16799 (ipa_pta_execute): Likewise.
16801 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16803 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16804 * configure: Regenerate.
16805 * config.in: Regenerate.
16806 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16807 #nobits/#progbits if supported.
16809 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16812 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16813 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16814 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16816 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16818 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16819 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16821 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16823 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16824 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16826 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16828 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16830 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16831 * config/arm/cortex-a7.md: New bypasses using
16832 arm_mac_accumulator_is_result.
16834 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16836 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16837 (cortex_a7_neon_mla): Likewise.
16838 (cortex_a7_fpfmad): New reservation.
16839 (cortex_a7_fpmacs): Use ffmas and update required units.
16840 (cortex_a7_fpmuld): Update required units and latency.
16841 (cortex_a7_fpmacd): Likewise.
16842 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16843 (cortex_a7_neon). Likewise.
16844 (bypass) Update participating units.
16846 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16848 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16849 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16851 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16852 (vfp_fmul): Use ffmad.
16853 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16854 (cortex_r4_fmacd): Use ffmad.
16855 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16856 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16857 (cortex_a9_fmacd): Use ffmad.
16858 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16859 (cortex_a8_vfp_macd): Use ffmad.
16860 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16861 (cortex_a5_fpmacd): Use ffmad.
16862 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16863 (cortex_a15_vfp_macd): Use ffmad.
16864 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16866 2013-01-29 Jason Merrill <jason@redhat.com>
16869 * varasm.c (default_assemble_visibility): Don't warn about
16870 visibility on artificial decls.
16872 2013-01-29 Richard Biener <rguenther@suse.de>
16874 PR tree-optimization/56113
16875 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16877 (label_visit): Free duplicate bitmaps and record the leader instead.
16878 (perform_var_substitution): Adjust.
16880 2013-01-29 Richard Biener <rguenther@suse.de>
16882 PR tree-optimization/55270
16883 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16884 the CFG, schedule loops for fixup.
16886 2013-01-29 Nick Clifton <nickc@redhat.com>
16888 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16891 2013-01-28 Leif Ekblad <leif@rdos.net>
16893 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16894 * config/i386/i386.h (TARGET_RDOS): New macro.
16895 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16896 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16897 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16898 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16899 DEFAULT_LARGE_SECTION_THRESHOLD.
16900 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16901 * config/i386/rdos.h: New file.
16902 * config/i386/rdos64.h: New file.
16904 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16907 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16910 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16912 PR rtl-optimization/56117
16913 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16914 call cselib_lookup_from_insn on the MEM before calling
16915 add_insn_mem_dependence.
16917 2013-01-28 Richard Biener <rguenther@suse.de>
16919 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16920 to a stmt that didn't have one.
16921 (copy_phis_for_bb): Likewise for PHI arguments.
16922 (copy_debug_stmt): Likewise for debug stmts.
16924 2013-01-28 Richard Biener <rguenther@suse.de>
16926 PR tree-optimization/56034
16927 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16928 (partition_builtin_p): Adjust.
16929 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16930 it is the last partition.
16931 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16932 up the vertex for the definition.
16933 (classify_partition): Classify whether a partition is a
16934 PKIND_REDUCTION, thus has uses outside of the loop.
16935 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16936 Merge all PKIND_REDUCTION partitions into the last partition.
16937 (tree_loop_distribution): Seed partitions from reductions as well.
16939 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16941 PR tree-optimization/56125
16942 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16943 pow(x,c) into sqrt(x) * powi(x, n/2) or
16944 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16945 optimizing for size.
16946 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16947 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16950 PR tree-optimization/56094
16951 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16952 to UNKNOWN_LOCATION while gimplifying expr.
16954 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16957 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16958 operand 0 in movabs insn template for -masm=intel asm alternative.
16959 (*movabs<mode>_2): Ditto for operand 1.
16961 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16964 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16967 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16969 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16972 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16974 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16975 Undef to avoid warning.
16977 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16979 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16980 * configure: Regenerate.
16982 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16984 PR tree-optimization/56098
16985 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16986 for stmts with volatile ops.
16987 (cond_store_replacement): Don't optimize if assign has volatile ops.
16988 (cond_if_else_store_replacement_1): Don't optimize if either
16989 then_assign or else_assign have volatile ops.
16990 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16993 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16995 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16997 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16999 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17000 missing ':' in asm example.
17002 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17004 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17005 entries into lane and laneq entries.
17006 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17007 Remove AdvSIMD scalar modes.
17008 (aarch64_sq<r>dmulh_laneq<mode>): New.
17009 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17011 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17012 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17013 * config/aarch64/iterators.md (VCOND): New.
17016 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17019 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17020 Add NULL LIBNAME argument to existing definitions.
17021 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17022 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17023 * config/avr/avr.c (DEF_BUILTIN): Same.
17024 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17025 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17026 implementation is available (DECL_ASSEMBLER_NAME is set).
17027 (avr_fold_absfx): New static function.
17028 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17029 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17030 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17031 AVR_BUILTIN_ABSLLK.
17032 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17033 (abshk, absk, abslk, absllk): Provide as static inline functions.
17035 2013-01-25 Marek Polacek <polacek@redhat.com>
17037 PR tree-optimization/56035
17038 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17040 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17042 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17043 (*movtf_internal_rex64): Add (!o,C) alternative
17044 (*movxf_internal_rex64): Ditto.
17045 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17047 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17049 * doc/invoke.texi: fix typo.
17050 * doc/objc.texi: fix typo.
17052 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17054 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17055 for the first two alternatives.
17057 2013-01-24 Diego Novillo <dnovillo@google.com>
17059 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17060 (ggc-zone.o): Remove.
17061 * configure.ac: Remove option --with-gc.
17062 * configure: Re-generate.
17063 * doc/install.texi: Remove documentation for --with-gc.
17064 * gengtype.c (write_enum_defn): Remove. Update all users.
17065 (write_Types_process_field): Remove generation of gt_e_* argument.
17066 (output_type_enum): Remove. Update all users.
17067 (write_enum_defn): Remove. Update all users.
17068 (enum alloc_zone): Remove. Update all users.
17069 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17070 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17071 Update all callers.
17072 (struct ptr_data): Remove field TYPE. Update all users.
17073 (gt_pch_note_object): Remove argument TYPE. Update all users.
17074 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17076 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17077 (struct alloc_zone): Remove.
17078 (ggc_internal_alloc_zone_stat): Remove.
17079 (ggc_internal_cleared_alloc_zone_stat): Remove.
17080 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17081 (ggc_pch_count_object): Remove last argument. Update all users.
17082 (ggc_pch_alloc_object): Remove last argument. Update all users.
17083 (struct alloc_zone): Remove.
17084 * ggc-zone.c: Remove.
17085 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17086 (struct alloc_zone): Remove.
17087 (ggc_alloc_typed_stat): Remove.
17088 (ggc_alloc_typed): Remove.
17089 (ggc_splay_alloc): Remove first argument.
17090 (rtl_zone): Remove. Update all users.
17091 (tree_zone): Remove. Update all users.
17092 (tree_id_zone): Remove. Update all users.
17093 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17094 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17095 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17096 * tree-ssanames.c: Remove references to zone allocator in comments.
17098 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17100 * config/avr/avr.c (avr_out_fract): Make register numbers that
17101 might be outside of source operand signed.
17103 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17105 * config/i386/constraints.md (Yf): New constraint.
17106 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17107 of f constraint to conditionaly disable x87 register preferences.
17108 (*movdf_internal): Ditto.
17109 (*movsf_internal): Ditto.
17111 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17113 PR inline-asm/55934
17114 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17115 that have operands with impossible constraints.
17116 Add a FIXME for a speed-up opportunity.
17117 * lra-constraints.c (process_alt_operands): Verify that a class
17118 selected from constraints on asms is valid for the operand mode.
17119 (curr_insn_transform): Remove incorrect comment.
17121 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17123 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17124 TOC operand is a valid symbol ref in the constant pool.
17126 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17128 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17130 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17133 * config/avr/stdfix.h: New file.
17134 * t-avr (stdfix-gcc.h): New rule to build it.
17135 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17137 2013-01-23 Kostya Serebryany <kcc@google.com>
17139 * config/darwin.h: remove dependency on
17140 CoreFoundation (asan on Mac OS).
17142 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17145 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17146 instead of cmpdi_operand for first comparison operand.
17147 Don't assert that comparison operands aren't both constants.
17149 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17151 * doc/install.texi (Downloading the Source): Update references to
17152 downloading separate components.
17154 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17156 * doc/extend.texi (__int128): Improve grammar.
17158 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17161 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17162 alternative to (o,r).
17163 (*movdi_internal_rex64): Remove (!o,n) alternative.
17164 (DImode immediate->memory splitter): Remove.
17165 (DImode immediate->memory peephole2): Remove.
17166 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17167 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17168 alternative to (!o,*r).
17169 (*movtf_internal_sse): New pattern.
17170 (*movxf_internal_rex64): New pattern.
17171 (*movxf_internal): Disable for TARGET_64BIT.
17172 (*movdf_internal_rex64): Remove (!o,F) alternative.
17174 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17176 PR middle-end/56074
17177 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17178 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17179 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17180 stmt locations where LOCATION_LOCUS of the stmt location is
17181 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17184 * config/i386/i386.md (UNSPEC_STOS): New.
17185 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17186 *strsetqi_1): Add UNSPEC_STOS.
17188 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17191 * doc/invoke.texi: Remove left over -Wsynth example.
17193 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17195 PR tree-optimization/56051
17196 * fold-const.c (fold_binary_loc): Don't fold
17197 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17198 a narrowing conversion, or widening conversion from signed
17201 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17203 PR rtl-optimization/56023
17204 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17205 dependent on debug instruction.
17207 2013-01-21 Martin Jambor <mjambor@suse.cz>
17209 PR middle-end/56022
17210 * function.c (allocate_struct_function): Call
17211 invoke_set_current_function_hook earlier.
17213 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17215 * reload1.c (init_reload): Only initialize reload_obstack
17216 during the first call.
17218 2013-01-21 Marek Polacek <polacek@redhat.com>
17220 * cfgloop.c (verify_loop_structure): Fix up grammar.
17222 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17224 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17225 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17227 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17230 * config/arm/marvell-pj4.md: Update copyright year.
17231 Fix up use of alu to alu_reg and simple_alu_imm.
17233 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17235 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17237 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17240 * lra-constraints.c (curr_insn_transform): Don't reuse original
17241 insn for secondary memory move when memory mode should be different.
17243 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17245 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17246 atomic_storedi_1): New patterns.
17248 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17250 btver2 pipeline descriptions.
17251 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17253 * config/i386/i386.md (btver2_decode): New type attributes.
17254 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17256 * config/i386/btver2.md: New file describing btver2 pipelines.
17258 2013-01-19 Andrew Pinski <apinski@cavium.com>
17260 PR tree-optimization/52631
17261 * tree-ssa-sccvn (visit_use): Before looking up the original
17262 statement, try looking up the simplified expression.
17264 2013-01-19 Anthony Green <green@moxielogic.com>
17266 * config/moxie/moxie.c (moxie_expand_prologue): Set
17267 current_function_static_stack_size.
17269 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17271 PR tree-optimization/56029
17272 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17273 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17275 2013-01-18 Sharad Singhai <singhai@google.com>
17277 PR tree-optimization/55995
17278 * dumpfile.c (dump_loc): Print location only if available.
17279 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17281 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17284 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17285 secondary memory move.
17286 (inherit_reload_reg): Use rclass instead of cl for
17287 check_secondary_memory_needed_p.
17289 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17291 PR middle-end/56015
17292 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17293 the case where writing real complex part of target modifies op1.
17295 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17297 * config/aarch64/aarch64-simd.md
17298 (aarch64_vcond_internal<mode>): Handle unordered cases.
17299 * config/aarch64/iterators.md (v_cmp_result): New.
17301 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17302 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17304 * config/arm/marvell-pj4.md: New file.
17305 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17306 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17307 (generic_vfp): Likewise.
17308 * config/arm/arm-cores.def: Add marvell-pj4.
17309 * config/arm/arm-tune.md: Regenerate.
17310 * config/arm/arm-tables.opt: Regenerate.
17311 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17312 * doc/invoke.texi: Document marvell-pj4.
17314 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17316 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17318 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17323 * var-tracking.c (negative_power_of_two_p): New.
17324 (global_get_addr_cache, local_get_addr_cache): New.
17325 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17326 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17328 (vt_stack_offset_p): Remove.
17329 (vt_canon_true_dep): Always canonicalize loc's address.
17330 (clobber_overlapping_mems): Make sure we have a MEM.
17331 (local_get_addr_clear_given_value): New.
17332 (val_reset): Clear local cached entries.
17333 (compute_bb_dataflow): Create and release the local cache.
17334 Disable duplicate MEMs clobbering.
17335 (emit_notes_in_bb): Clobber MEMs likewise.
17336 (vt_emit_notes): Create and release the local cache.
17337 (vt_initialize, vt_finalize): Create and release the global
17338 cache, respectively.
17339 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17341 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17343 PR libmudflap/53359
17344 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17345 not found in the symtab.
17347 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17350 PR rtl-optimization/55547
17351 PR rtl-optimization/53827
17354 * alias.c (offset_overlap_p): New, factored out of...
17355 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17356 the conservative special case for symbolic constants. Don't
17357 adjust zero sizes on alignment.
17359 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17361 PR rtl-optimization/52573
17362 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17363 REG_UNUSED for the same register.
17365 2013-01-17 Richard Biener <rguenther@suse.de>
17366 Marek Polacek <polacek@redhat.com>
17368 PR rtl-optimization/55833
17369 * loop-unswitch.c (unswitch_loops): Move loop verification...
17370 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17371 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17372 Set it to true when we're removing a loop from hierarchy tree in
17373 an irreducible region.
17374 (fix_bb_placements): Adjust caller.
17375 (fix_loop_placements): Likewise.
17377 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17379 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17380 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17381 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17382 Remove ID. Adjust comments.
17383 * config/avr/avr-c.c (avr_builtin_name): Remove.
17384 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17385 * config/avr/avr.c (avr_tolower): New static function.
17386 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17387 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17388 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17391 2013-01-17 Jan Hubicka <jh@suse.cz>
17393 PR tree-optimization/55273
17394 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17396 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17399 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17400 store through atomic_store<mode>_1.
17401 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17403 2013-01-17 Martin Jambor <mjambor@suse.cz>
17405 PR tree-optimizations/55264
17406 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17407 for virtual methods.
17408 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17409 virtual methods before inlining is over.
17410 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17412 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17415 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17417 PR rtl-optimization/56005
17418 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17419 pending reads for prefetch.
17421 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17423 * config/aarch64/aarch64.md
17424 (*cstoresi_neg_uxtw): New pattern.
17425 (*cmovsi_insn_uxtw): New pattern.
17426 (*<optab>si3_uxtw): New pattern.
17427 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17428 (*<optab>si3_insn_uxtw): New pattern.
17429 (*bswapsi2_uxtw): New pattern.
17431 2013-01-16 Richard Biener <rguenther@suse.de>
17433 * tree-inline.c (tree_function_versioning): Remove set but
17434 never used variable.
17436 2013-01-16 Richard Biener <rguenther@suse.de>
17438 PR tree-optimization/55964
17439 * tree-flow.h (rename_variables_in_loop): Remove.
17440 (rename_variables_in_bb): Likewise.
17441 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17442 (copy_loop_before): Adjust and delete update-ssa status.
17443 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17444 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17445 (rename_variables_in_loop): Remove.
17446 (slpeel_update_phis_for_duplicate_loop): Likewise.
17447 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17448 use available cfg machinery instead of duplicating it.
17449 Update PHI nodes and perform poor-mans SSA update here.
17450 (slpeel_tree_peel_loop_to_edge): Adjust.
17452 2013-01-16 Richard Biener <rguenther@suse.de>
17454 PR tree-optimization/54767
17455 PR tree-optimization/53465
17456 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17457 (vrp_visit_phi_node): For PHI arguments coming via backedges
17458 drop all symbolical range information.
17459 (execute_vrp): Compute backedges.
17461 2013-01-16 Richard Biener <rguenther@suse.de>
17463 * doc/install.texi: Update CLooG and ISL requirements to
17466 2013-01-16 Christian Bruel <christian.bruel@st.com>
17469 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17470 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17471 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17473 2013-01-16 DJ Delorie <dj@redhat.com>
17475 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17476 (UNSPECV_SP_SWITCH_E): New.
17477 (sp_switch_1): Change to an unspec.
17478 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17481 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17483 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17484 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17485 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17486 (expand_mem_thread_fence): Ditto.
17487 (expand_mem_signal_fence): Ditto.
17488 (expand_atomic_load): Ditto.
17489 (expand_atomic_store): Ditto.
17491 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17493 PR rtl-optimization/55547
17494 PR rtl-optimization/53827
17497 * alias.c (memrefs_conflict_p): Set sizes to negative after
17500 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17503 * function.c (thread_prologue_and_epilogue_insns): Always
17504 add crtl->drap_reg to set_up_by_prologue.set, even if
17505 stack_realign_drap is false.
17507 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17509 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17510 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17511 *call): Fix indention.
17513 2013-01-15 Tom de Vries <tom@codesourcery.com>
17516 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17519 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17521 PR rtl-optimization/55153
17522 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17524 2013-01-15 Martin Jambor <mjambor@suse.cz>
17526 PR tree-optimization/55920
17527 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17528 accesses as grp_to_be_debug_replaced.
17530 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17532 PR tree-optimization/55920
17533 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17534 there is non-useless type conversion needed from debug rhs to lhs,
17535 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17537 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17538 Mikael Pettersson <mikpe@it.uu.se>
17541 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17543 (ASM_OUTPUT_CASE_LABEL): Remove.
17544 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17545 * final.c (shorten_branches): Update alignment of labels before
17546 jump tables if CASE_VECTOR_SHORTEN_MODE.
17548 2013-01-15 Richard Biener <rguenther@suse.de>
17551 * system.h: Do not include gmp.h for building host tools.
17553 2013-01-15 Richard Biener <rguenther@suse.de>
17555 PR middle-end/55882
17556 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17557 account for bitpos when computing alignment.
17559 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17561 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17562 (ix86_target_macros_internal): Likewise.
17564 * config/i386/i386.c (m_CORE2I7): Removed.
17565 (m_CORE_HASWELL): New macro.
17566 (m_CORE_ALL): Likewise.
17567 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17568 (initial_ix86_arch_features): Likewise.
17569 (processor_target_table): Initializations for Core avx2.
17570 (cpu_names): New names "core-avx2".
17571 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17572 PROCESSOR_CORE_HASWELL.
17573 (ix86_issue_rate): New case.
17574 (ia32_multipass_dfa_lookahead): Likewise.
17575 (ix86_sched_init_global): Likewise.
17577 * config/i386/i386.h (TARGET_HASWELL): New macro.
17578 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17579 (processor_type): New PROCESSOR_HASWELL.
17581 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17583 PR tree-optimization/55955
17584 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17585 *SHIFT_EXPR and *ROTATE_EXPR codes.
17587 PR tree-optimization/48766
17588 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17589 -ftrapv disable -fwrapv.
17591 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17594 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17595 etc. to 1 and not to __flash.
17596 Use LL suffix for __INT24_MAX__ with -mint8.
17597 Use ULL suffix for __UINT24_MAX__ with -mint8.
17599 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17601 * config/avr/avr-arch.h
17602 (struct base_arch_s): Use typedef avr_arch_t instead.
17603 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17604 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17605 * config/avr/avr.c: Same.
17606 * config/avr/avr-devices.c: Same.
17607 * config/avr/driver-avr.c: Same.
17608 * config/avr/gen-avr-mmcu-texi.c: Same.
17609 * config/avr/avr-mcus.def: Adjust comment.
17611 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17613 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17614 * config/aarch64/iterators.md (VALLDI): New.
17616 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
17617 Andi Kleen <ak@linux.intel.com>
17620 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
17621 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
17624 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17626 * config/avr/avr-stdint.h: Remove trailing blanks.
17627 * config/avr/avr-log.h: Same.
17628 * config/avr/avr-arch.h: Same.
17629 * config/avr/avr-devices.c: Same.
17630 * config/avr/avr-dimode.md: Same.
17631 * config/avr/predicates.md: Same.
17632 * config/avr/avr-c.c: Same. And fix typo.
17634 * config/avr/avr-protos.h: Same. And:
17635 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
17636 (init_cumulative_args): Rename to avr_init_cumulative_args.
17637 (expand_prologue): Rename to avr_expand_prologue.
17638 (expand_epilogue): Rename to avr_expand_epilogue.
17639 (adjust_insn_length): Rename to avr_adjust_insn_length.
17640 (notice_update_cc): Rename to avr_notice_update_cc.
17641 (final_prescan_insn): Rename to avr_final_prescan_insn.
17642 * config/avr/avr.c: Same.
17643 * config/avr/avr.h: Same.
17644 * config/avr/avr.md: Remove trailing blanks.
17645 (prologue): Use avr_expand_prologue.
17646 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
17648 2013-01-14 Richard Biener <rguenther@suse.de>
17650 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
17651 verify_location, collect_subblocks): New functions.
17652 (verify_gimple_in_cfg): Verify that locations only reference
17653 BLOCKs in the functions BLOCK tree.
17655 2013-01-14 Richard Biener <rguenther@suse.de>
17657 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
17659 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
17661 (insert_out_of_ssa_copy_on_edge): Likewise.
17662 (rewrite_close_phi_out_of_ssa): Likewise.
17663 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
17665 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
17666 propagated constants.
17667 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
17670 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17672 * config/avr/avr-modes.def: Add GPL copyright notice.
17674 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
17676 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
17677 MEMMODEL_MASK to determine memory model.
17678 (atomic_store<mode>): Ditto from operands[2].
17679 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
17681 2013-01-13 Jakub Jelinek <jakub@redhat.com>
17684 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
17685 (fold_gimple_assign): Don't call unshare_expr here.
17686 (fold_ctor_reference): Call unshare_expr.
17688 2013-01-13 Terry Guo <terry.guo@arm.com>
17690 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
17691 * doc/fragments.texi: Document MULTILIB_REUSE.
17692 * gcc.c (multilib_reuse): New internal spec.
17693 (set_multilib_dir): Also search multilib from multilib_reuse.
17694 * genmultilib (tmpmultilib3): Refactor code.
17695 (tmpmultilib4): Ditto.
17696 (multilib_reuse): New multilib argument.
17698 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
17700 * Makefile.in: Update copyright.
17702 2013-01-12 Tom de Vries <tom@codesourcery.com>
17704 PR middle-end/55890
17705 * calls.c (expand_call): Check if arg_nr is valid.
17707 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17709 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
17710 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
17711 documentation. Add missing '__' in front of
17712 __builtin_ia32_packssdw256.
17714 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17717 * config/s390/s390.c (s390_preferred_reload_class): Do not return
17718 NO_REGS for larl operands.
17719 (s390_reload_larl_operand): Use s390_load_address instead of
17722 2013-01-11 Richard Biener <rguenther@suse.de>
17724 * tree-cfg.c (verify_node_sharing_1): Split out from ...
17725 (verify_node_sharing): ... here.
17726 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
17728 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
17730 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
17731 Substitute TREECHECKING.
17732 * configure: Regenerate.
17733 * Makefile.in (TREECHECKING): New.
17735 2013-01-11 Richard Guenther <rguenther@suse.de>
17737 PR tree-optimization/44061
17738 * tree-vrp.c (extract_range_basic): Compute zero as
17739 value-range for __builtin_constant_p of function parameters.
17741 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
17743 Update copyright years.
17745 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
17747 PR rtl-optimization/55672
17748 * lra-eliminations.c (mark_not_eliminable): Permit addition with
17749 const to be eliminable.
17751 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
17753 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
17754 * configure: Regenerate.
17756 2013-01-10 Richard Biener <rguenther@suse.de>
17758 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
17760 2013-01-10 Richard Biener <rguenther@suse.de>
17763 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
17764 locations for virtual PHI arguments.
17765 (rewrite_update_phi_arguments): Likewise.
17767 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
17769 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
17772 2013-01-10 Jakub Jelinek <jakub@redhat.com>
17774 PR tree-optimization/55921
17775 * tree-complex.c (expand_complex_asm): New function.
17776 (expand_complex_operations_1): Call it for GIMPLE_ASM.
17778 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17781 * config/s390/s390.c (s390_symref_operand_p)
17782 (s390_loadrelative_operand_p): Merge the two functions.
17783 (s390_check_qrst_address, print_operand_address): Add parameters
17784 to s390_loadrelative_operand_p invokation.
17785 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
17786 (s390_reload_larl_operand, s390_secondary_reload): Use
17787 s390_loadrelative_operand_p instead of s390_symref_operand_p.
17788 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17790 2013-01-09 Mike Stump <mikestump@comcast.net>
17792 * dse.c (record_store): Remove unnecessary assert.
17794 2013-01-09 Jan Hubicka <jh@suse.cz>
17796 PR tree-optimization/55569
17797 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17798 * cfgloop.h (scale_loop_profile): Likewise.
17800 2013-01-09 Jan Hubicka <jh@suse.cz>
17803 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17805 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17807 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17809 PR middle-end/55114
17810 * expr.h (maybe_emit_group_store): Declare.
17811 * expr.c (maybe_emit_group_store): New function.
17812 * builtins.c (expand_builtin_int_roundingfn): Call it.
17813 (expand_builtin_int_roundingfn_2): Likewise.
17815 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17817 PR rtl-optimization/55829
17818 * lra-constraints.c (match_reload): Add code for absent output.
17819 (curr_insn_transform): Add code for reloads of matched inputs
17822 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17824 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17825 attribute of movddup insn to DF.
17826 (*vec_interleave_lowv2df): Ditto.
17827 (vec_dupv2df): Ditto.
17829 2013-01-09 Jan Hubicka <jh@suse.cz>
17831 PR tree-optimiation/55875
17832 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17833 EVERY_ITERATION parameter.
17834 (number_of_iterations_exit): Check if exit is executed every iteration.
17835 (idx_infer_loop_bounds): Similarly here.
17836 (n_of_executions_at_most): Simplify
17837 to only test for cases where statement is dominated by the
17838 particular bound; handle correctly the "postdominance" test.
17839 (scev_probably_wraps_p): Use max loop iterations info
17840 as a global bound first.
17842 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17843 Nick Clifton <nickc@redhat.com>
17845 * config/v850/v850.md (cbranchsf4): New pattern.
17846 (cstoresf4): New pattern.
17847 (cbranchdf4): New pattern.
17848 (cstoredf4): New pattern.
17849 (movsicc): Disallow floating point comparisons.
17850 (cmpsf_le_insn): Fix order of operators.
17851 (cmpsf_lt_insn): Likewise.
17852 (cmpsf_eq_insn): Likewise.
17853 (cmpdf_le_insn): Likewise.
17854 (cmpdf_lt_insn): Likewise.
17855 (cmpdf_eq_insn): Likewise.
17856 (cmpsf_ge_insn): Use LE comparison.
17857 (cmpdf_ge_insn): Likewise.
17858 (cmpsf_gt_insn): Use LT comparison.
17859 (cmpdf_gt_insn): Likewise.
17860 (cmpsf_ne_insn): Delete pattern.
17861 (cmpdf_ne_insn): Delete pattern.
17862 * config/v850/v850.c (v850_gen_float_compare): Use
17863 gen_cmpdf_eq_insn for NE comparison.
17864 (v850_float_z_comparison_operator)
17865 (v850_float_nz_comparison_operator): Move from here ...
17866 * config/v850/predicates.md: ... to here. Move GT and GE
17867 comparisons into v850_float_z_comparison_operator.
17868 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17870 (v850_float_nz_comparison_operator): Likewise.
17872 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17874 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17875 with calls to gen_insvsi/gen_insvdi.
17877 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17879 * config/i386/i386.c (initial_ix86_tune_features): Set up
17880 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17882 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17883 Jakub Jelinek <jakub@redhat.com>
17885 PR tree-optimization/48189
17886 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17887 If nitercst is 0, don't predict the exit edge.
17889 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17891 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17892 in asm_fprintf with reg_names.
17893 (aarch64_print_operand_address): Likewise.
17894 (aarch64_return_addr): Likewise.
17895 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17897 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17899 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17900 (INT_U6_BITS): Likewise.
17901 * config/pa/predicates.md (uint6_operand): New predicate.
17902 (shift5_operand, shift6_operand): Likewise.
17903 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17905 (lshrdi3): Use shift6_operand.
17906 (shrpsi4, shrpdi4): New insn patterns.
17907 (extzv): Delete expander.
17908 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17909 predicates in unamed zero extract patterns. Tighten common constraint.
17910 (extv): Delete expander.
17911 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17912 predicates in unamed sign extract patterns. Tighten common constraint.
17913 (insv): Delete expander.
17914 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17915 predicates in unamed insert patterns. Tighten common constraint.
17916 Change uint32_operand predicate to uint6_operand predicate in unamed
17917 DImode pattern to insert constant values of type 1...1xxxx.
17919 2013-01-04 Jan Hubicka <jh@suse.cz>
17921 PR tree-optimization/55823
17922 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17925 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17926 Uros Bizjak <ubizjak@gmail.com>
17928 PR rtl-optimization/55845
17929 * df-problems.c (can_move_insns_across): Stop scanning at
17930 volatile_insn_p source instruction or give up if
17931 across_from .. across_to range contains any volatile_insn_p
17934 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17936 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17937 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17939 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17940 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17942 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17945 * asan.c (asan_clear_shadow): New function.
17946 (asan_emit_stack_protection): Use it.
17948 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17950 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17951 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17952 with tab instead of space.
17954 2013-01-08 Nick Clifton <nickc@redhat.com>
17956 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17957 register bank 0 at the start of an interrupt handler.
17958 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17961 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17963 * config/aarch64/aarch64-simd.md
17964 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17965 (aarch64_simd_bsl): Likewise.
17966 (aarch64_vcond_internal<mode>): Likewise.
17967 (vcond<mode><mode>): Likewise.
17968 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17969 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17971 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17973 * config/aarch64/aarch64-builtins.c
17974 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17976 2013-01-08 Martin Jambor <mjambor@suse.cz>
17979 * tree-sra.c (analyze_access_subtree): Return true also after
17980 potentially creating a debug-only replacement.
17982 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17984 PR middle-end/55890
17985 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17987 PR tree-optimization/54120
17988 * tree-vrp.c (range_fits_type_p): Don't allow
17989 src_precision < precision from signed vr to unsigned_p
17990 if vr->min or vr->max is negative.
17991 (simplify_float_conversion_using_ranges): Test can_float_p
17992 against CODE_FOR_nothing.
17994 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17995 Richard Biener <rguenther@suse.de>
17997 PR middle-end/55851
17998 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17999 types instead of just INTEGER_TYPE types.
18001 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18003 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18006 2013-01-07 Steve Ellcey <sellcey@mips.com>
18009 * config/mips/mips.opt: Change mad to mmad to match documentation.
18011 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18014 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18015 .progmemx.data now.
18017 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18020 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18021 (avr_addrspace_t): Add .section_name field.
18022 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18024 (avr_addrspace): Same. Initialize .section_name. Remove last
18025 NULL entry. Put __memx into .progmemx.data.
18026 (progmem_section_prefix): Remove.
18027 (avr_asm_init_sections): No need to initialize progmem_section.
18028 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18029 section name prefix.
18030 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18031 retrieve the progmem section.
18032 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18033 boundary to run over avr_addrspace[].
18034 (avr_register_target_pragmas): Ditto.
18036 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18038 * varasm.c (output_constant_def_contents): For asan_protect_global
18039 protected strings, adjust DECL_ALIGN if needed, before testing for
18041 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18042 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18044 (output_object_block): For asan protected decls, emit asan padding
18045 after their contents.
18046 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18047 (asan_finish_file): Test it here instead.
18049 2013-01-07 Nick Clifton <nickc@redhat.com>
18050 Matthias Klose <doko@debian.org>
18051 Doug Kwan <dougkwan@google.com>
18052 H.J. Lu <hongjiu.lu@intel.com>
18055 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18057 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18059 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18061 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18063 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18065 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18068 * doc/install.texi (Cross-Compiler-Specific Options): Document
18071 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18073 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18074 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18075 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18076 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18077 vqmovun_high_s64): Fix source operand number and update copyright.
18079 2013-01-07 Richard Biener <rguenther@suse.de>
18081 PR middle-end/55890
18082 * gimple.h (gimple_call_builtin_p): New overload.
18083 * gimple.c (validate_call): New function.
18084 (gimple_call_builtin_p): Likewise.
18085 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18086 Use gimple_call_builtin_p.
18087 (find_func_clobbers): Likewise.
18088 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18089 (strlen_optimize_stmt): Likewise.
18091 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18093 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18094 (vld1q_dup_*): Likewise.
18095 (vld1_*): Likewise.
18096 (vld1q_*): Likewise.
18097 (vld1_lane_*): Likewise.
18098 (vld1q_lane_*): Likewise.
18100 2013-01-07 Richard Biener <rguenther@suse.de>
18102 * lto-streamer.h (LTO_minor_version): Bump to 2.
18104 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18106 * config/aarch64/aarch64-protos.h
18107 (aarch64_const_double_zero_rtx_p): Rename to...
18108 (aarch64_float_const_zero_rtx_p): ...this.
18109 (aarch64_float_const_representable_p): New.
18110 (aarch64_output_simd_mov_immediate): Likewise.
18111 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18112 move immediate case.
18113 * config/aarch64/aarch64.c
18114 (aarch64_const_double_zero_rtx_p): Rename to...
18115 (aarch64_float_const_zero_rtx_p): ...this.
18116 (aarch64_print_operand): Allow printing of new constants.
18117 (aarch64_valid_floating_const): New.
18118 (aarch64_legitimate_constant_p): Check for valid floating-point
18120 (aarch64_simd_valid_immediate): Likewise.
18121 (aarch64_vect_float_const_representable_p): New.
18122 (aarch64_float_const_representable_p): Likewise.
18123 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18124 (aarch64_output_simd_mov_immediate): New.
18125 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18126 (*movdf_aarch64): Likewise.
18127 * config/aarch64/constraints.md (Ufc): New.
18128 (Y): call aarch64_float_const_zero_rtx.
18129 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18131 2013-01-07 Richard Biener <rguenther@suse.de>
18133 PR tree-optimization/55888
18134 PR tree-optimization/55862
18135 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18136 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18137 not if it is contained therein.
18139 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18141 * config/avr/t-avr: Typo.
18143 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18146 * config/avr/t-avr: Don't automatically rebuild
18147 $(srcdir)/config/avr/t-multilib
18148 $(srcdir)/config/avr/avr-tables.opt
18149 $(srcdir)/doc/avr-mmcu.texi
18150 (avr-mcus): New phony target to build them on request.
18151 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18152 * avr/avr-mcus.def: Adjust comments.
18154 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18156 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18158 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18160 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18162 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18164 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18166 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18168 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18169 to generate profiling.
18170 * config/rs6000/aix64.h (LIB_SPEC): Same.
18172 2013-01-04 Andrew Pinski <apinski@cavium.com>
18174 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18176 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18178 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18180 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18182 (ix86_expand_move): Ditto.
18183 (ix86_zero_extend_to_Pmode): Ditto.
18184 (ix86_expand_call): Ditto.
18185 (ix86_expand_special_args_builtin): Ditto.
18186 (ix86_expand_builtin): Ditto.
18188 2013-01-04 Richard Biener <rguenther@suse.de>
18190 PR tree-optimization/55862
18191 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18192 translating them through PHI nodes.
18194 2013-01-04 Martin Jambor <mjambor@suse.cz>
18196 PR tree-optimization/55755
18197 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18198 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18200 2013-01-04 Marek Polacek <polacek@redhat.com>
18202 PR middle-end/55859
18203 * opts.c (default_options_optimization): Clarify error message.
18205 2013-01-04 Richard Biener <rguenther@suse.de>
18207 PR middle-end/55863
18208 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18211 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18214 * config/pa/pa.md (movsi): Revert previous change.
18215 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18218 2013-01-03 Richard Henderson <rth@redhat.com>
18220 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18221 after eliminating TLS symbols.
18223 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18226 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18227 * graphite-poly.c (debug_gmp_value): Likewise.
18229 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18232 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18233 selected code model, define __code_mode_small__, __code_model_medium__,
18234 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18235 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18236 xchg temporary register with %k. Declare temporary register as
18238 [__x86_64__]: For medium and large code models, preserve %rbx register.
18240 2013-01-03 Richard Biener <rguenther@suse.de>
18242 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18243 (dump_subscript): Adjust.
18244 (finalize_ddr_dependent): Do not dump redundant info.
18245 (analyze_siv_subscript): Adjust.
18246 (subscript_dependence_tester): Likewise.
18247 (compute_affine_dependence): Likewise.
18249 2013-01-03 Richard Biener <rguenther@suse.de>
18252 2013-01-03 Richard Biener <rguenther@suse.de>
18254 PR tree-optimization/55857
18255 * tree-vect-stmts.c (vectorizable_load): Do not setup
18256 re-alignment for invariant loads.
18258 2013-01-02 Richard Biener <rguenther@suse.de>
18260 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18261 invariant load do not generate a vector load from the scalar location.
18263 2013-01-03 Richard Biener <rguenther@suse.de>
18265 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18266 for not vectorizing.
18267 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18268 not build INDIRECT_REFs, call get_name once only.
18269 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18270 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18272 2013-01-03 Richard Biener <rguenther@suse.de>
18274 PR tree-optimization/55857
18275 * tree-vect-stmts.c (vectorizable_load): Do not setup
18276 re-alignment for invariant loads.
18278 2013-01-03 Richard Biener <rguenther@suse.de>
18281 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18282 prefer a built-in decl.
18284 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18286 * gcc.c (process_command): Update copyright notice dates.
18287 * gcov.c (print_version): Likewise.
18288 * gcov-dump.c (print_version): Likewise.
18290 PR rtl-optimization/55838
18291 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18292 iv0.step, iv1.step and step.
18294 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18295 Marc Glisse <marc.glisse@inria.fr>
18297 PR tree-optimization/55832
18298 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18299 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18300 integer_{one,zero}_node.
18302 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18305 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18306 * var-tracking.c (reverse_op): Don't add reverse ops to
18307 VALUEs that have already
18308 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18310 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18312 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18314 2013-01-02 Teresa Johnson <tejohnson@google.com>
18316 * dumpfile.c (dump_loc): Print filename with location.
18317 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18318 new location_t parameter to emit complete unroll message with
18319 new dump framework.
18320 (canonicalize_loop_induction_variables): Compute loops location
18321 and pass to try_unroll_loop_completely.
18322 * loop-unroll.c (report_unroll_peel): New function.
18323 (peel_loops_completely): Use new dump format with location
18324 for main dumpfile message, and invoke report_unroll_peel on success.
18325 (decide_unrolling_and_peeling): Ditto.
18326 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18327 by report_unroll_peel.
18328 (decide_peel_completely): Ditto.
18329 (decide_unroll_constant_iterations): Ditto.
18330 (decide_unroll_runtime_iterations): Ditto.
18331 (decide_peel_simple): Ditto.
18332 (decide_unroll_stupid): Ditto.
18333 * cfgloop.c (get_loop_location): New function.
18334 * cfgloop.h (get_loop_location): Declare.
18336 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18338 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18341 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18343 PR middle-end/55198
18344 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18345 BLKmode objects when EXPAND_MEMORY is specified.
18347 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18349 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18351 (fold_builtin_cpu): Do not share cpu model decls across statements.
18353 2013-01-02 Jason Merrill <jason@redhat.com>
18356 * tree.c (build_array_type_1): Revert earlier change.
18358 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18360 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18362 * config/aarch64/aarch64-tune.md: Re-generate.
18364 2013-01-02 Richard Biener <rguenther@suse.de>
18366 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18367 invariant load do not generate a vector load from the scalar location.
18369 2013-01-02 Richard Biener <rguenther@suse.de>
18372 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18373 * configure: Regenerate.
18375 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18377 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18378 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18379 (expand_builtin_int_roundingfn_2): Keep the original target around
18380 for the fallback case.
18382 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18384 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18385 to be clear for sign changes.
18387 2013-01-01 Jan Hubicka <jh@suse.cz>
18389 * ipa-inline-analysis.c: Fix formatting.
18391 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18393 PR tree-optimization/55831
18394 * tree-vect-loop.c (get_initial_def_for_induction): Use
18395 gsi_after_labels instead of gsi_start_bb.
18397 Copyright (C) 2013 Free Software Foundation, Inc.
18399 Copying and distribution of this file, with or without modification,
18400 are permitted in any medium without royalty provided the copyright
18401 notice and this notice are preserved.