1 2013-09-25 Tom Tromey <tromey@redhat.com>
3 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
5 2013-09-25 Tom Tromey <tromey@redhat.com>
7 * config/i386/t-i386 (i386.o): Remove.
8 (i386-c.o): Use COMPILE and POSTCOMPILE.
10 2013-09-25 Tom Tromey <tromey@redhat.com>
12 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
14 2013-09-25 Tom Tromey <tromey@redhat.com>
16 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
17 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
18 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
19 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
20 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
21 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
22 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
23 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
24 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
25 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
26 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
27 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
28 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
29 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
30 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
31 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
32 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
33 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
34 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
35 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
36 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
37 (tree-ssa-pre.o, tree-ssa-sccvn.o)
38 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
39 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
40 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
41 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
42 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
43 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
44 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
45 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
46 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
47 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
48 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
49 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
50 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
51 (graphite-clast-to-gimple.o, graphite-dependences.o)
52 (graphite-interchange.o, graphite-poly.o)
53 (graphite-scop-detection.o, graphite-sese-to-poly.o)
54 (graphite-optimize-isl.o, tree-vect-loop.o)
55 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
56 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
57 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
58 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
59 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
60 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
61 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
62 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
63 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
64 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
65 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
66 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
67 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
68 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
69 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
70 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
71 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
72 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
73 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
74 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
75 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
76 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
77 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
78 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
79 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
80 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
81 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
82 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
83 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
84 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
85 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
86 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
87 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
88 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
89 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
90 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
91 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
92 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
93 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
94 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
95 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
96 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
97 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
98 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
99 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
100 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
101 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
102 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
103 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
104 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
105 (gcov-dump.o): Remove.
106 (default-c.o): Use COMPILE and POSTCOMPILE.
107 (CFLAGS-gcc.o): New variable.
108 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
110 2013-09-25 Tom Tromey <tromey@redhat.com>
112 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
113 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
114 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
115 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
116 (c-family/c-pragma.o, c-family/c-pretty-print.o)
117 (c-family/c-semantics.o, c-family/c-ada-spec.o)
118 (c-family/array-notation-common.o, c-family/stub-objc.o)
119 (c-family/c-ubsan.o): Remove.
121 2013-09-25 Tom Tromey <tromey@redhat.com>
123 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
125 2013-09-25 Tom Tromey <tromey@redhat.com>
127 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
128 to add -DENABLE_SHARED_LIBGCC.
129 (gcc.o): Don't use subshell.
131 2013-09-25 Tom Tromey <tromey@redhat.com>
133 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
134 * configure.ac: Don't invoke AM_PROG_CC_C_O.
135 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
136 * configure, config.in: Rebuild.
138 2013-09-25 Tom Tromey <tromey@redhat.com>
140 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
141 (COMPILE, POSTCOMPILE): New variables.
142 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
143 (DEPFILES): New variable.
145 * configure.ac: Add checks for dependency checking.
146 * configure, aclocal.m4: Regenerate.
148 2013-09-25 Tom Tromey <tromey@redhat.com>
150 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
151 ($(ALL_HOST_OBJS)): Move order-only dependency to end
154 2013-09-25 Tom Tromey <tromey@redhat.com>
156 * Makefile.in (generated_files): Add options.h,
157 target-hooks-def.h, insn-opinit.h,
158 common/common-target-hooks-def.h, pass-instances.def,
159 c-family/c-target-hooks-def.h.
161 2013-09-25 Jeff Law <law@redhat.com>
163 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
164 than foo[foo.length () - 1] to access last member in a vec.
165 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
167 2013-09-25 Richard Biener <rguenther@suse.de>
170 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
172 2013-09-25 Jan Hubicka <jh@suse.cz>
174 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
177 2013-09-25 Marek Polacek <polacek@redhat.com>
180 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
181 when determining the type name.
183 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
185 * config/sh/sh.md: Fix formatting.
187 2013-09-24 Xinliang David Li <davidxl@google.com>
189 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
190 Check max peel iterations parameter.
191 * param.def: New parameter.
192 * doc/invoke.texi: Document New parameter.
194 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
196 * gimple-pretty-print.c: Various whitespace tweaks.
197 * tree-core.h: Likewise.
198 * tree-pretty-print.c: Likewise.
199 * tree-ssa-alias.c: Likewise.
200 * tree-ssa-copy.c: Likewise.
201 * tree-ssanames.c: Likewise.
202 * tree-ssanames.h: Likewise.
203 * tree-vrp.c: Likewise.
205 2013-09-24 Alan Modra <amodra@gmail.com>
209 * stmt.c (expand_asm_operands): Call expand_expr with
210 EXPAND_MEMORY for output operands that disallow regs. Don't
211 use EXPAND_WRITE on inout operands.
213 2013-09-24 Richard Biener <rguenther@suse.de>
216 * tree.c (reference_alias_ptr_type): Move ...
217 * alias.c (reference_alias_ptr_type): ... here and implement
218 in terms of the new reference_alias_ptr_type_1.
219 (ref_all_alias_ptr_type_p): New helper.
220 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
221 use ref_all_alias_ptr_type_p.
222 (get_deref_alias_set): Add flag_strict_aliasing check here.
223 (reference_alias_ptr_type_1): New function, split out from ...
224 (get_alias_set): ... here.
225 (alias_ptr_types_compatible_p): New function.
226 * alias.h (reference_alias_ptr_type): Declare.
227 (alias_ptr_types_compatible_p): Likewise.
228 * tree.h (reference_alias_ptr_type): Remove.
229 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
230 to compare MEM_REF alias types.
232 2013-09-24 Richard Biener <rguenther@suse.de>
234 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
236 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
238 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
239 reload helper function arrays into a single array reg_addr.
240 (reload_fpr_gpr): Likewise.
241 (reload_gpr_vsx): Likewise.
242 (reload_vsx_gpr): Likewise.
243 (struct rs6000_reg_addr): Likewise.
244 (reg_addr): Likewise.
245 (rs6000_debug_reg_global): Change rs6000_vector_reload,
246 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
247 (rs6000_init_hard_regno_mode_ok): Likewise.
248 (rs6000_secondary_reload_direct_move): Likewise.
249 (rs6000_secondary_reload): Likewise.
251 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
252 constraints: wu, ww, and wy. Repurpose wv constraint added during
253 power8 changes. Put wg constraint in alphabetical order.
255 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
256 for future work to add ISA 2.07 VSX single precision support.
257 (-mvsx-scalar-double): Change default from -1 to 1, update
258 documentation comment.
259 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
260 (-mupper-regs-df): New debug switch to control whether DF values
261 can go in the traditional Altivec registers.
262 (-mupper-regs-sf): New debug switch to control whether SF values
263 can go in the traditional Altivec registers.
265 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
267 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
268 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
269 Add new constraints, wu/ww/wy. Repurpose wv constraint.
270 (rs6000_debug_legitimate_address_p): Print if we are running
271 before, during, or after reload.
272 (rs6000_secondary_reload): Add a comment.
273 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
275 * config/rs6000/constraints.md (wa constraint): Sort w<x>
276 constraints. Update documentation string.
277 (wd constraint): Likewise.
278 (wf constraint): Likewise.
279 (wg constraint): Likewise.
280 (wn constraint): Likewise.
281 (ws constraint): Likewise.
282 (wt constraint): Likewise.
283 (wx constraint): Likewise.
284 (wz constraint): Likewise.
285 (wu constraint): New constraint for ISA 2.07 SFmode scalar
287 (ww constraint): Likewise.
288 (wy constraint): Likewise.
289 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
290 the previous submissions.
291 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
293 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
295 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
296 (REG_BR_PROB): Say that the probability is stored in an int_list.
297 * reg-notes.def: Update commentary to mention INT_LIST.
298 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
300 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
301 * rtlanal.c (int_reg_note_p): New function.
302 (alloc_reg_note): Assert that the note does not have an int argument.
303 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
304 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
305 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
306 rather than an INSN_LIST. Handle INT_LIST.
307 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
308 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
309 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
310 Manipulate them as ints rather than rtxes.
311 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
312 * regmove.c (copy_src_to_dest): Likewise.
313 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
315 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
316 into the cases that need it.
317 * config/arm/arm.c (arm_unwind_emit): Likewise.
319 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
320 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
321 * loop-doloop.c (add_test, doloop_modify): Likewise.
322 * loop-unswitch.c (compare_and_jump_seq): Likewise.
323 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
324 * predict.c (combine_predictions_for_insn): Likewise.
325 * print-rtl.c (print_rtx): Handle INT_LIST.
326 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
327 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
328 * config/arm/arm.c (emit_unlikely_jump): Likewise.
329 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
330 (ix86_split_fp_branch, predict_jump): Likewise.
331 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
332 * config/sh/sh.c (expand_cbranchsi4): Likewise.
333 * config/spu/spu.c (ea_load_store_inline): Likewise.
335 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
336 value of a REG_BR_PROB note.
337 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
338 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
339 * emit-rtl.c (try_split): Likewise.
340 * predict.c (br_prob_note_reliable_p): Likewise.
341 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
342 * reorg.c (mostly_true_jump): Likewise.
343 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
344 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
345 * config/i386/i386.c (ix86_print_operand): Likewise.
346 * config/ia64/ia64.c (ia64_print_operand): Likewise.
347 * config/mmix/mmix.c (mmix_print_operand): Likewise.
348 * config/rs6000/rs6000.c (output_cbranch): Likewise.
349 * config/s390/s390.c (s390_expand_tbegin): Likewise.
350 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
351 * config/sparc/sparc.c (output_cbranch): Likewise.
352 * config/spu/spu.c (get_branch_target): Likewise.
353 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
354 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
356 2013-09-23 Jan Hubicka <jh@suse.cz>
358 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
360 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
361 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
363 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
364 (gimple_fold_call): Dump inconsistent devirtualizations; add
365 sanity check for type based devirtualizations.
366 * ipa-prop.c: Include ipa-utils.h
367 (ipa_intraprocedural_devirtualization): Add sanity check.
368 (try_make_edge_direct_virtual_call): Likewise.
370 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
372 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
373 assignment statements.
375 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
377 * gimple-pretty-print.c (dump_ssaname_info): New function.
378 (dump_gimple_phi): Call it.
379 (pp_gimple_stmt_1): Likewise.
380 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
381 (range_info_def): Declare.
382 * tree-pretty-print.c (pp_double_int): New function.
383 (dump_generic_node): Call it.
384 * tree-pretty-print.h (pp_double_int): Declare.
385 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
386 * tree-ssanames.h (range_info_def): New structure.
387 (value_range_type): Move definition here.
388 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
390 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
392 (set_range_info): New function.
393 (get_range_info): Likewise.
394 (duplicate_ssa_name_range_info): Likewise.
395 (duplicate_ssa_name_fn): Check pointer type and call
396 duplicate_ssa_name_range_info.
397 * tree-ssa-copy.c (fini_copy_prop): Likewise.
398 * tree-vrp.c (value_range_type): Remove definition, now in
400 (vrp_finalize): Call set_range_info to update value range of
402 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
403 (SSA_NAME_RANGE_INFO): New macro.
405 2013-09-23 Richard Biener <rguenther@suse.de>
407 PR tree-optimization/58464
408 * tree-ssa-pre.c (phi_trans_lookup): Remove.
409 (phi_trans_add): Change to add conditionally on being not
411 (phi_translate_1): Remove recursion detection here.
412 (phi_translate): Pre-seed the cache with NULL to catch
413 recursion here in a more generic way.
414 (bitmap_find_leader): Adjust comment.
415 (get_representative_for): Dump value-numbers.
416 (create_expression_by_pieces): Likewise.
417 (insert_into_preds_of_block): Likewise.
419 2013-09-23 Christian Bruel <christian.bruel@st.com>
422 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
423 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
425 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
428 2013-09-20 Renlin Li <renlin.li@arm.com>
430 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
431 (aarch64_expand_epilogue): Likewise.
432 (aarch64_legitimize_reload_address): Likewise.
434 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
436 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
437 NULL_TREE before pushing them onto the vector. Likewise for labels.
439 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
441 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
442 comparison operators when -fno-trapping-math is in effect.
443 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
444 comparison operators in TFmode and assert that unsupported operators
446 (ia64_print_operand): Likewise.
448 2013-09-21 Jan Hubicka <jh@suse.cz>
450 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
451 (sse_typeless_stores): Enable for core
452 (sse_load0_by_pxor): Likewise.
453 (four_jump_limit): Disable for core.
454 (pad_returns): Likewise.
455 (avoid_vector_decode): Likewise.
456 (fuse_cmp_and_branch): Enable for cores.
457 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
459 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
462 * config/pa/pa.c (pa_option_override): Disable auto increment and
463 decrement instructions until reload is completed.
465 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
466 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
467 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
469 2013-09-20 DJ Delorie <dj@redhat.com>
470 Nick Clifton <nickc@redhat.com>
472 * config/rl78/rl78.c: Various whitespace and comment tweaks.
473 (need_to_save): Save bank 0 on interrupts.
474 (characterize_address): Strip far address wrappers.
475 (rl78_as_legitimate_address): Likewise.
476 (transcode_memory_rtx): Likewise.
477 (rl78_peep_movhi_p): Disable this peephole after devirt.
478 (rl78_propogate_register_origins): Forget all origins when a
480 * config/rl78/rl78-virt.md: Various whitespace tweaks.
481 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
483 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
484 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
485 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
487 * config/rl78/constraints.md: Various whitespace and paren tweaks.
489 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
491 * config/pa/pa.md: In "scc" insn patterns, change output template to
492 handle const0_rtx in reg_or_0_operand operands.
494 2013-09-20 Martin Husemann <martin@NetBSD.org>
497 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
498 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
500 2013-09-20 Richard Biener <rguenther@suse.de>
503 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
504 remembering SSA name version and block index.
505 (new_scev_info_str): Adjust.
506 (hash_scev_info): Likewise. Also hash the block index.
507 (eq_scev_info): Adjust.
508 (find_var_scev_info): Likewise.
509 (struct instantiate_cache_entry): Remove.
510 (struct instantiate_cache_type): Use a htab to map name, block
512 (instantiate_cache_type::~instantiate_cache_type): Adjust.
513 (get_instantiated_value_entry): Likewise.
514 (hash_idx_scev_info, eq_idx_scev_info): New functions.
515 (instantiate_scev_name): Adjust.
517 2013-09-20 Jeff Law <law@redhat.com>
519 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
521 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
523 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
524 Call aarch64_simd_expand_args to update op[argc].
526 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
528 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
531 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
533 * gengtype.c (file_rules): Added rule for *.cc files.
534 (get_output_file_with_visibility): Give fatal message when no
537 2013-09-20 Renlin Li <renlin.li@arm.com>
539 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
540 (aarch64_expand_epilogue): Likewise.
541 (aarch64_legitimize_reload_address): Likewise.
543 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
546 * expr.c (expand_assignment): Remove misalignp code path.
548 2013-09-20 Marek Polacek <polacek@redhat.com>
551 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
552 TYPE_PRECISION. Add asserts.
554 2013-09-20 Richard Biener <rguenther@suse.de>
556 PR tree-optimization/58453
557 * tree-loop-distribution.c (distribute_loop): Apply the cost
558 model for -ftree-loop-distribute-patterns, too.
560 2013-09-20 Richard Biener <rguenther@suse.de>
563 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
564 make type comparison less strict.
566 2013-09-20 Alan Modra <amodra@gmail.com>
568 * configure: Regenerate.
569 * aclocal.m4: Regenerate.
571 2013-09-20 Marek Polacek <polacek@redhat.com>
574 * doc/extend.texi: Document that attribute used is meant to be used
575 on variables with static storage duration.
577 2013-09-19 Jakub Jelinek <jakub@redhat.com>
579 PR tree-optimization/58472
580 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
581 simd_lane_access set inv_p = false.
582 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
583 the simduid magic VAR_DECL.
585 2013-09-19 Jan Hubicka <jh@suse.cz>
587 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
589 2013-09-17 Jeff Law <law@redhat.com>
591 * tree-ssa-dom.c (record_temporary_equivalences): New function
592 split out of dom_opt_dom_walker::after_dom_children.
593 (dom_opt_dom_walker::thread_across_edge): Move common code
594 in here from dom_opt_dom_walker::after_dom_children.
595 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
597 2013-09-19 Jan Hubicka <jh@suse.cz>
599 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
600 (TARGET_GENERIC): Use PROCESOR_GENERIC
601 (enum processor_type): Unify generic32 and 64.
602 * i386.md (cpu): Likewise.
603 * x86-tune.def (use_leave): Enable for generic32.
604 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
605 * athlon.md: Change generic64 to generic in all occurences.
606 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
607 (ix86_target_macros_internal): Likewise.
608 * driver-i386.c (host_detect_local_cpu): Likewise.
609 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
610 (generic_memcpy, generic_memset, generic_cost): This one.
611 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
612 (m_GENERIC32, m_GENERIC64): Remove.
613 (m_GENERIC): Turn into one flag.
614 (processor_target): Unify generic tunnings.
615 (ix86_option_override_internal): Replace generic32/64 by generic.
616 (ix86_issue_rate): Likewise.
617 (ix86_adjust_cost): Likewise.
619 2013-09-19 Jan Hubicka <jh@suse.cz>
621 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
624 2013-09-19 Jakub Jelinek <jakub@redhat.com>
626 * omp-low.c (expand_omp_sections): Always pass len - 1 to
627 GOMP_sections_start, even if !exit_reachable.
629 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
631 * lra-constraints.c (need_for_all_save_p): Use macro
632 HARD_REGNO_CALL_PART_CLOBBERED.
633 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
634 set up pseudo conflict hard regs.
636 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
639 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
642 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
644 PR rtl-optimization/58438
645 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
646 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
647 from previous subpasses.
649 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
651 * arm.c (arm_get_frame_offsets): Validate architecture supports
652 LDRD/STRD before accepting the tuning preference.
653 (arm_expand_prologue): Likewise.
654 (arm_expand_epilogue): Likewise.
656 2013-09-18 Richard Biener <rguenther@suse.de>
658 PR tree-optimization/58417
659 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
660 have chrecs with symbols defined in the loop as operands.
661 (chrec_fold_multiply): Likewise.
662 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
663 parameters before folding binary operations.
664 (struct instantiate_cache_entry_hasher): Remove.
665 (struct instantiate_cache_type): Use a pointer-map.
666 (instantiate_cache_type::instantiate_cache_type): New function.
667 (instantiate_cache_type::get): Likewise.
668 (instantiate_cache_type::set): Likewise.
669 (instantiate_cache_type::~instantiate_cache_type): Adjust.
670 (get_instantiated_value_entry): Likewise.
671 (global_cache): New global.
672 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
673 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
674 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
675 (instantiate_scev_name): Adjust.
676 (instantiate_scev): Construct global instead of local cache.
677 (resolve_mixers): Likewise.
679 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
680 Paolo Carlini <paolo.carlini@oracle.com>
683 * doc/implement-cxx.texi: Fix references to the C++ standards.
685 2013-09-18 Jakub Jelinek <jakub@redhat.com>
687 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
688 * tree-vect-data-refs.c (vect_analyze_data_refs): For
689 simd_lane_access drs, update also DR_ALIGNED_TO.
691 2013-09-18 Marek Polacek <polacek@redhat.com>
694 * doc/extend.texi: Document no_sanitize_undefined attribute.
695 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
696 no_sanitize_undefined attribute.
698 2013-09-18 Nick Clifton <nickc@redhat.com>
700 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
701 (ASM_DECLARE_FUNCTION_NAME): Define.
703 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
705 * compare-elim.c (find_comparison_dom_walker): New class
706 (find_comparisons_in_bb): Rename to
707 find_comparison_dom_walker::before_dom_children
708 (find_comparisons): Adjust
709 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
711 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
712 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
713 (init_walk_dominator_tree): Remove declaration.
714 (fini_walk_dominator_tree): Remove declaration.
715 * fwprop.c (single_def_use_dom_walker): New class
716 (single_def_use_enter_block): Convert to
717 single_def_use_dom_walker::before_dom_children.
718 (single_def_use_leave_block): Convert to
719 single_def_use_dom_walker::after_dom_children.
720 (build_single_def_use_links): Adjust.
721 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
723 (find_candidates_in_block): Convert to
724 find_candidates_dom_walker::before_dom_children.
725 (execute_strength_reduction): Adjust.
726 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
727 (sese_dom_walker): New class.
728 (sese_dom_walker::sese_dom_walker): New constructor.
729 (sese_dom_walker::~sese_dom_walker): New destructor.
730 (build_sese_conditions_before): Convert to
731 sese_dom_walker::before_dom_children.
732 (build_sese_conditions_after): Convert to
733 sese_dom_walker::after_dom_children.
734 (build_poly_scop): Adjust
735 * tree-into-ssa.c (rewrite_dom_walker): New class
736 (rewrite_enter_block): Convert to
737 rewrite_dom_walker::before_dom_children.
738 (rewrite_leave_block): Convert to
739 rewrite_dom_walker::after_dom_children.
740 (rewrite_update_dom_walker): New class.
741 (rewrite_update_enter_block): Convert to
742 rewrite_update_dom_walker::before_dom_children.
743 (rewrite_update_leave_block): Convert to
744 rewrite_update_dom_walker::after_dom_children.
745 (rewrite_blocks, rewrite_into_ssa): Adjust.
746 (mark_def_dom_walker): New class.
747 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
748 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
749 (mark_def_sites_blocks): Convert to
750 mark_def_dom_walker::before_dom_children.
751 (mark_def_site_blocks): Remove.
752 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
753 (tree_ssa_dominator_optimize): Adjust.
754 (dom_thread_across_edge): Convert to method
755 dom_opt_dom_walker::thread_across_edge.
756 (dom_opt_enter_block): Convert to member function
757 dom_opt_dom_walker::before_dom_children.
758 (dom_opt_leave_block): Convert to member function
759 dom_opt_dom_walker::after_dom_children.
760 * tree-ssa-dse.c (dse_dom_walker): New class.
761 (dse_enter_block): Convert to member function
762 dse_dom_walker::before_dom_children.
763 (tree_ssa_dse): Adjust.
764 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
765 (determine_invariantness_stmt): Convert to method
766 invariantness_dom_walker::before_dom_children.
767 (determine_invariantness): Remove
768 (move_computations_dom_walker): New class.
769 (move_computations_stmt): Convert to method
770 move_computations_dom_walker::before_dom_children.
771 (move_computations, tree_ssa_lim): Adjust.
772 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
773 (nt_init_block): Make method
774 notrappping_dom_walker::before_dom_children.
775 (nt_fini_block): Make
776 method nontrapping_dom_walker::after_dom_children.
777 (get_non_trapping): Adjust.
778 * tree-ssa-pre.c (eliminate_dom_walker): New class.
779 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
780 (eliminate_leave_block): Make method.
781 eliminate_dom_walker::after_dom_children.
783 * tree-ssa-strlen.c (strlen_dom_walker): New class.
784 (strlen_enter_block): Make method
785 strlen_dom_walker::before_dom_children.
786 (strlen_leave_block): Make
787 method strlen_dom_walker::after_dom_children.
788 (tree_ssa_strlen): Adjust.
789 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
790 (tree_ssa_uncprop): Adjust.
791 (uncprop_leave_block): Make method
792 uncprop_dom_walker::after_dom_children.
793 (uncprop_leave_block): Make method
794 uncprop_dom_walker::before_dom_children.
796 2013-09-18 Bin Cheng <bin.cheng@arm.com>
798 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
799 before branch in same basic block. Check both src and dest of
802 2013-09-17 Nick Clifton <nickc@redhat.com>
804 * config/rl78/rl78-real.md (bf): New pattern.
806 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
807 (rl78_print_operand): Do not put a # before a %B.
808 * config/rl78/rl78.opt: Tweak doc strings.
810 2013-09-17 DJ Delorie <dj@redhat.com>
812 * config/rl78/constraints.md (Wcv): Allow up to $r31.
813 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
814 (rl78_option_override): Likewise, if -mallregs.
815 (is_virtual_register): Likewise.
816 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
817 (REGNO_OK_FOR_BASE_P): Likewise.
818 * config/rl78/rl78.opt (-mallregs): New.
820 2013-09-17 Nick Clifton <nickc@redhat.com>
822 * config/rl78/rl78.c (need_to_save): Change return type to bool.
823 For interrupt functions: save all call clobbered registers if the
824 interrupt handler is not a leaf function.
825 (rl78_expand_prologue): Always recompute the frame information.
826 For interrupt functions: only select bank 0 if one of the bank 0
827 registers is going to be psuhed.
829 2013-09-17 DJ Delorie <dj@redhat.com>
831 * config/rl78/constraints.md: For each W* constraint, rename to C*
832 and create a W* constraint that checks for an optional ES: prefix
834 * config/rl78/rl78.md (UNS_ES_ADDR): New.
835 (es_addr): New. Used to wrap far addresses.
836 * config/rl78/rl78-protos.h (rl78_es_addr): New.
838 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
839 wrapped far addresses.
840 (rl78_print_operand_1): Unwrap far addresses before processing.
841 (rl78_lo16): Wrap far addresses in unspecs.
844 (insn_ok_now): Check for not-yet-wrapped far addresses.
845 (transcode_memory_rtx): Properly re-wrap far addresses.
847 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
849 * config/sparc/t-rtems: Add leon3 multilibs.
851 2013-09-17 Cong Hou <congh@google.com>
853 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
854 when checking the dot production pattern. The type of rhs operand
855 of multiply is now checked correctly.
857 2013-09-17 Jeff Law <law@redhat.com>
859 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
860 edge implied equivalences into successor phis.
861 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
862 here from tree-ssa-threadedge.c.
863 (mark_threaded_blocks): When threading through a joiner, if both
864 successors of the joiner's clone reach the same block, verify the
865 PHI arguments are equal. If not, cancel the jump threading request.
866 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
867 tree-ssa-threadupdate.c
868 (thread_across_edge): Don't check PHI argument equality when
869 threading through joiner block here.
871 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
873 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
874 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
875 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
876 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
877 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
878 (has_undefined_value_p): New. Pass dependant parts of
879 ssa_undefined_value_p.
880 (uninit_undefined_value_p): Use has_undefined_value_p.
881 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
882 make_pass_early_warn_uninitialized): Move from tree-ssa.c
883 * tree-ssa.h: Adjust prototypes
885 2013-09-17 Jan Hubicka <jh@suse.cz>
888 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
889 * ipa-inline.c (can_inline_edge_p): Do not downgrade
890 FUNCTION_NOT_OPTIMIZED.
891 * ipa-inline-analysis.c (compute_inline_parameters): Function
892 not optimized is not inlinable unless it is alwaysinline.
893 (inline_analyze_function): Force calls in not optimized
894 function not inlinable.
896 2013-09-17 Jan Hubicka <jh@suse.cz>
899 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
901 * ipa.c (function_and_variable_visibility): Likewise.
902 * ipa-profile.c (ipa_profile): Likewise.
904 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
907 * cgraph.c (cgraph_function_body_availability): Check for ifunc
908 attribute, and don't inline the resolver in this case.
910 2013-09-17 Teresa Johnson <tejohnson@google.com>
912 * coverage.c (get_coverage_counts): Add missing newline.
914 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
916 PR tree-optimization/58088
917 * fold-const.c (mask_with_trailing_zeros): New function.
918 (fold_binary_loc): Make sure we don't recurse infinitely
919 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
920 Use mask_with_trailing_zeros where appropriate.
922 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
924 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
925 of 'prev' var to get better distance estimation.
927 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
929 * tree-inline.h (struct copy_body_data): Add transform_parameter.
930 * tree-inline.c (is_parameter_of): New predicate.
931 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
932 a parameter has been remapped.
933 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
934 (optimize_inline_calls): Initialize transform_parameter.
935 (copy_gimple_seq_and_replace_locals): Likewise.
936 (tree_function_versioning): Likewise.
937 (maybe_inline_call_in_expr): Likewise.
939 2013-09-17 Nick Clifton <nickc@redhat.com>
941 * config/msp430/msp430-protos.h: Add prototypes for new functions.
942 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
944 (is_attr_func): New function.
945 (msp430_is_interrupt_func): New function.
946 (is_naked_func): New function.
947 (is_reentrant_func): New function.
948 (is_critical_func): New function.
949 (msp430_start_function): Add annotations for function attributes.
950 (msp430_attr): New function.
951 (msp430_attribute_table): New.
952 (msp430_function_section): New function.
953 (TARGET_ASM_FUNCTION_SECTION): Define.
954 (msp430_builtin): New enum.
955 (msp430_init_builtins): New function.
956 (msp430_builtin_devl): New function.
957 (msp430_expand_builtin): New function.
958 (TARGET_INIT_BUILTINS): Define.
959 (TARGET_EXPAND_BUILTINS): Define.
960 (TARGET_BUILTIN_DECL): Define.
961 (msp430_expand_prologue): Add support for naked, interrupt,
962 critical and reentrant functions.
963 (msp430_expand_epilogue): Likewise.
964 (msp430_print_operand): Handle 'O' character.
965 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
967 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
968 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
969 (pushm): Use a 'n' rather than an 'i' constraint.
970 (msp_return): Add generation of the interrupt return instruction.
971 (disable_interrupts): New pattern.
972 (enable_interrupts): New pattern.
973 (push_intr_state): New pattern.
974 (pop_intr_state): New pattern.
975 (bic_SR): New pattern.
976 (bis_SR): New pattern.
977 * doc/extend.texi: Document MSP430 function attributes and builtin
980 2013-09-17 Richard Biener <rguenther@suse.de>
982 PR tree-optimization/58432
983 * tree-loop-distribution.c (tree_loop_distribution): Also
984 scan PHIs for outside loop uses and seed a partition from them.
986 2013-09-17 Bin Cheng <bin.cheng@arm.com>
988 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
989 (restructure_reference): Call backtrace_base_for_ref.
991 2013-09-17 Alan Modra <amodra@gmail.com>
994 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
996 2013-09-16 DJ Delorie <dj@redhat.com>
998 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
999 vregs location for RL78/G10.
1000 (rl78_expand_prologue): Avoid SEL on G10.
1001 (rl78_expand_epilogue): Likewise.
1002 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1003 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1004 __RL78_G10__ when appropriate.
1005 (ASM_SPEC): Pass -mg10 along to the assembler.
1006 * config/rl78/rl78.md (sel_rb): Disable for G10.
1007 * config/rl78/rl78.opt: Add -mg10 option.
1008 * config/rl78/t-rl78: Add -mg10 multilib.
1010 2013-09-16 Xinliang David Li <davidxl@google.com>
1012 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1013 * omp-low.c (omp_max_vf): Ditto.
1014 (expand_omp_simd): Ditto.
1015 * tree-vectorizer.c (vectorize_loops): Ditto.
1016 (gate_vect_slp): Ditto.
1017 (gate_increase_alignment): Ditto.
1018 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1019 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1020 (gate_tree_vectorize): Name change.
1021 (tree_vectorize): Ditto.
1022 (pass_vectorize::gate): Call new function.
1023 (pass_vectorize::execute): Ditto.
1024 opts.c: O3 default setting change.
1025 (finish_options): Check new flag.
1026 * doc/invoke.texi: Document new flags.
1027 * common.opt: New flags.
1029 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1031 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1032 (Condition Code): Don't reference it.
1033 * doc/tm.texi: Regenerate.
1035 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1038 * lra-constraints.c (undo_optional_reloads): Consider all optional
1039 reload even if it did not get a hard reg.
1041 2013-09-16 Teresa Johnson <tejohnson@google.com>
1043 * dumpfile.c (dump_loc): Remove newline emission.
1044 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1045 emission to dump_printf_loc calls where missing.
1046 (vect_mark_for_runtime_alias_test): Ditto.
1047 (vect_analyze_data_ref_dependence): Ditto.
1048 (vect_analyze_data_ref_dependences): Ditto.
1049 (vect_slp_analyze_data_ref_dependence): Ditto.
1050 (vect_slp_analyze_data_ref_dependences): Ditto.
1051 (vect_compute_data_ref_alignment): Ditto.
1052 (vect_update_misalignment_for_peel): Ditto.
1053 (vect_verify_datarefs_alignment): Ditto.
1054 (vector_alignment_reachable_p): Ditto.
1055 (vect_get_data_access_cost): Ditto.
1056 (vect_enhance_data_refs_alignment): Ditto.
1057 (vect_find_same_alignment_drs): Ditto.
1058 (vect_analyze_data_refs_alignment): Ditto.
1059 (vect_analyze_group_access): Ditto.
1060 (vect_analyze_data_ref_access): Ditto.
1061 (vect_analyze_data_ref_accesses): Ditto.
1062 (vect_prune_runtime_alias_test_list): Ditto.
1063 (vect_analyze_data_refs): Ditto.
1064 (vect_create_addr_base_for_vector_ref): Ditto.
1065 (vect_create_data_ref_ptr): Ditto.
1066 (vect_grouped_store_supported): Ditto.
1067 (vect_grouped_load_supported): Ditto.
1068 * value-prof.c (check_counter): Ditto.
1069 (check_ic_target): Ditto.
1070 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1071 (vect_recog_widen_mult_pattern): Ditto.
1072 (vect_recog_widen_sum_pattern): Ditto.
1073 (vect_recog_over_widening_pattern): Ditto.
1074 (vect_recog_widen_shift_pattern): Ditto.
1075 (vect_recog_rotate_pattern): Ditto.
1076 (vect_recog_vector_vector_shift_pattern): Ditto.
1077 (vect_recog_divmod_pattern): Ditto.
1078 (vect_recog_mixed_size_cond_pattern): Ditto.
1079 (vect_recog_bool_pattern): Ditto.
1080 (vect_pattern_recog_1): Ditto.
1081 (vect_pattern_recog): Ditto.
1082 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1083 (vect_is_simple_iv_evolution): Ditto.
1084 (vect_analyze_scalar_cycles_1): Ditto.
1085 (vect_get_loop_niters): Ditto.
1086 (vect_analyze_loop_1): Ditto.
1087 (vect_analyze_loop_form): Ditto.
1088 (vect_analyze_loop_operations): Ditto.
1089 (vect_analyze_loop_2): Ditto.
1090 (vect_analyze_loop): Ditto.
1091 (report_vect_op): Ditto.
1092 (vect_is_slp_reduction): Ditto.
1093 (vect_is_simple_reduction_1): Ditto.
1094 (vect_get_known_peeling_cost): Ditto.
1095 (vect_estimate_min_profitable_iters): Ditto.
1096 (vect_model_reduction_cost): Ditto.
1097 (vect_model_induction_cost): Ditto.
1098 (get_initial_def_for_induction): Ditto.
1099 (vect_create_epilog_for_reduction): Ditto.
1100 (vectorizable_reduction): Ditto.
1101 (vectorizable_induction): Ditto.
1102 (vectorizable_live_operation): Ditto.
1103 (vect_loop_kill_debug_uses): Ditto.
1104 (vect_transform_loop): Ditto.
1105 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1106 (vect_stmt_relevant_p): Ditto.
1107 (process_use): Ditto.
1108 (vect_mark_stmts_to_be_vectorized): Ditto.
1109 (vect_model_simple_cost): Ditto.
1110 (vect_model_promotion_demotion_cost): Ditto.
1111 (vect_model_store_cost): Ditto.
1112 (vect_get_store_cost): Ditto.
1113 (vect_model_load_cost): Ditto.
1114 (vect_get_load_cost): Ditto.
1115 (vect_init_vector_1): Ditto.
1116 (vect_get_vec_def_for_operand): Ditto.
1117 (vect_finish_stmt_generation): Ditto.
1118 (vectorizable_call): Ditto.
1119 (vectorizable_conversion): Ditto.
1120 (vectorizable_assignment): Ditto.
1121 (vectorizable_shift): Ditto.
1122 (vectorizable_operation): Ditto.
1123 (vectorizable_store): Ditto.
1124 (vectorizable_load): Ditto.
1125 (vectorizable_condition): Ditto.
1126 (vect_analyze_stmt): Ditto.
1127 (vect_transform_stmt): Ditto.
1128 (vect_is_simple_use): Ditto.
1129 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1130 (vect_can_advance_ivs_p): Ditto.
1131 (vect_update_ivs_after_vectorizer): Ditto.
1132 (vect_do_peeling_for_loop_bound): Ditto.
1133 (vect_gen_niters_for_prolog_loop): Ditto.
1134 (vect_update_inits_of_drs): Ditto.
1135 (vect_create_cond_for_alias_checks): Ditto.
1136 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1137 (vect_build_slp_tree_1): Ditto.
1138 (vect_supported_load_permutation_p): Ditto.
1139 (vect_analyze_slp_instance): Ditto.
1140 (vect_analyze_slp): Ditto.
1141 (vect_make_slp_decision): Ditto.
1142 (vect_detect_hybrid_slp): Ditto.
1143 (vect_bb_vectorization_profitable_p): Ditto.
1144 (vect_slp_analyze_bb_1): Ditto.
1145 (vect_update_slp_costs_according_to_vf): Ditto.
1146 (vect_get_mask_element): Ditto.
1147 (vect_transform_slp_perm_load): Ditto.
1148 (vect_schedule_slp_instance): Ditto.
1149 (vect_schedule_slp): Ditto.
1150 (vect_slp_transform_bb): Ditto.
1151 * profile.c (read_profile_edge_counts): Ditto.
1152 (compute_branch_probabilities): Ditto.
1153 * coverage.c (get_coverage_counts): Ditto.
1155 2013-09-16 Diego Novillo <dnovillo@google.com>
1157 * tree-core.h: Add missing comment lines from refactoring
1160 2013-09-16 Jan Hubicka <jh@suse.cz>
1162 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1163 abstract functions; for static functions check the presence
1166 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1168 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1169 * config/aarch64/aarch64-simd.md
1170 (aarch64_mla_elt<mode>): New.
1171 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1172 (aarch64_mls_elt<mode>): Likewise.
1173 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1174 (aarch64_fma4_elt<mode>): Likewise.
1175 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1176 (aarch64_fma4_elt_to_128v2df): Likewise.
1177 (aarch64_fma4_elt_to_64df): Likewise.
1178 (fnma<mode>4): Likewise.
1179 (aarch64_fnma4_elt<mode>): Likewise.
1180 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1181 (aarch64_fnma4_elt_to_128v2df): Likewise.
1182 (aarch64_fnma4_elt_to_64df): Likewise.
1183 * config/aarch64/iterators.md (VDQSF): New.
1184 * config/aarch64/arm_neon.h
1185 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1186 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1188 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1190 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1191 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1192 (aarch64_mul3_elt_to_128df): Likewise.
1193 (aarch64_mul3_elt_to_64v2df): Likewise.
1194 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1196 (VMUL_CHANGE_NLANES) Likewise.
1199 * config/aarch64/arm_neon.h
1200 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1202 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1204 * config/aarch64/arm_neon.h
1205 (vcvtx_high_f32_f64): Fix parameters.
1207 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1208 Uros Bizjak <ubizjak@gmail.com>
1210 * config/alpha.c: Include tree-ssanames.h.
1212 2013-09-16 Richard Biener <rguenther@suse.de>
1214 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1215 (dot_rdg_1): Handle control_dd.
1216 (create_edge_for_control_dependence): New function.
1217 (create_rdg_edges): Add control dependences if asked for.
1218 (build_rdg): Likewise.
1219 (generate_loops_for_partition): If there are not necessary
1220 control stmts remove all their dependencies.
1221 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1222 (distribute_loop): Pass on control dependences.
1223 (tree_loop_distribution): Compute control dependences and remove
1224 restriction on number of loop nodes.
1226 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1228 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1231 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1233 * cse.c (try_const_anchors): Punt on CC modes.
1235 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1237 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1239 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1242 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1245 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1248 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1250 (darwin_objc1_section): Likewise.
1251 (darwin_file_end): Emit Image Info section when required.
1253 2013-09-14 Jan Hubicka <jh@suse.cz>
1255 * tree-into-ssa.c (gate_into_ssa): New.
1256 (pass_data_build_ssa): Use it.
1257 * cgraph.h (expand_thunk): Update prototype.
1258 * cgraphunit.c (analyze_function): Expand thunks early.
1259 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1260 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1261 set lowered flag; do not add new function.
1262 (assemble_thunks_and_aliases): Update.
1263 * tree-ssa.c (gate_init_datastructures): New gate.
1264 (pass_data_init_datastructures): Use it.
1266 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1269 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1270 xmm register set described in the psABI.
1272 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1274 * dwarf2out.c (should_emit_struct_debug): Add check
1275 for type_decl variable is not NULL.
1277 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1279 * config.gcc: Use new winnt-c.c target hooks
1280 * config/t-winnt: New file
1281 * config/winnt-c.c: New file
1282 * doc/tm.texi.in: Document new hook
1283 * doc/tm.texi: Regenerated
1285 2013-09-13 Jan Hubicka <jh@suse.cz>
1288 * ipa-inline.c (check_callers): New function.
1289 (check_caller_edge): Remove.
1290 (want_inline_function_to_all_callers_p): Also permit alises that are
1292 (inline_to_all_callers): Terminate the walk when devirtualization
1293 introduce new calls.
1295 2013-09-13 Jan Hubicka <jh@suse.cz>
1297 * ipa-inline-analysis.c (struct growth_data): Add node.
1298 (do_estimate_growth_1): Fix detection of recursion.
1300 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1302 PR tree-optimization/58392
1303 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1304 to avoid shadowing of outer loop variable. If
1305 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1306 replace_by_duplicate_decl simduid of loops that have it set and
1307 set dest_cfun->has_simduid_loops and/or
1308 dest_cfun->has_force_vect_loops.
1309 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1310 instead of maybe_lookup_decl.
1311 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1312 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1313 call. Copy over force_vect and copy and remap simduid. Set
1314 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1315 (copy_cfg_body): Remove blocks_to_copy argument. Use
1316 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1317 caller. Don't set cfun->has_simduid_loops and/or
1318 cfun->has_force_vect_loops here.
1319 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1321 (expand_call_inline, tree_function_versioning): Adjust copy_body
1324 2013-09-13 Martin Jambor <mjambor@suse.cz>
1327 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1328 the assert if the edge was a speculative one.
1330 2013-09-13 Richard Biener <rguenther@suse.de>
1332 * tree-data-ref.h (known_dependences_p): Move here ...
1333 * tree-loop-distribution.c (known_dependences_p): ... from here.
1334 (dump_rdg_component, debug_rdg_component): Remove.
1336 (generate_loops_for_partition): Use gimple_uid instead of
1337 relying on matching stmt visit order.
1338 (rdg_build_partitions): Take starting stmt vector.
1339 (ldist_gen): Merge into ...
1340 (distribute_loop): ... this function. Do not compute starting
1342 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1344 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1346 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1347 Set type attribute correctly. Set predicable_short_it attribute.
1348 (cmpsi_shiftsi): Remove %? from output template.
1350 2013-09-13 Richard Biener <rguenther@suse.de>
1352 * tree-loop-distribution.c (struct rdg_component,
1353 rdg_defs_used_in_other_loops_p, free_rdg_components,
1354 rdg_build_components): Remove.
1355 (stmts_from_loop): Do not record virtual PHIs.
1356 (generate_loops_for_partition): Skip virtual PHIs.
1357 (build_rdg_partition_for_component): Rename to ...
1358 (build_rdg_partition_for_vertex): ... this and adjust.
1359 (rdg_build_partitions): Take a vector of starting vertices
1360 instead of components. Remove unnecessary leftover handling.
1361 (ldist_gen): Do not build components or record other stores.
1362 (distribute_loop): Do not distribute loops containing stmts
1365 2013-09-13 Christian Bruel <christian.bruel@st.com>
1368 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1370 2013-09-13 Kai Tietz <ktietz@redhat.com>
1372 * config.gcc: Separate cases for mingw and cygwin targets,
1373 and add 64-bit cygwin target case.
1375 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1376 dll-export inline-functions.
1377 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1380 2013-09-13 Jeff Law <law@redhat.com>
1384 2013-09-06 Jeff Law <law@redhat.com>
1386 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1387 edge implied equivalences into successor phis.
1389 2013-09-12 DJ Delorie <dj@redhat.com>
1391 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1394 2013-09-12 Brooks Moses <bmoses@google.com>
1397 * Makefile.in: Do not install driver binaries in $(target)/bin.
1399 2013-09-12 DJ Delorie <dj@redhat.com>
1401 * config/rl78/rl78.opt (mrelax): New.
1402 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1403 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1405 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1407 (rl78_expand_epilogue): Likewise.
1409 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1412 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1413 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1414 pointer elimination with using remove_reg_equal_offset_note.
1416 2013-09-12 DJ Delorie <dj@redhat.com>
1418 * config/msp430/: New port.
1419 * config.gcc (msp430): Added.
1420 * doc/invoke.texi: Document MSP430 options.
1421 * doc/install.texi: Document msp430-elf
1422 * doc/md.texi: Document msp430-elf
1423 * doc/contrib.texi: Document msp430-elf
1425 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1428 2013-09-12 Martin Jambor <mjambor@suse.cz>
1431 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1432 reference descriptor is NULL.
1433 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1436 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1438 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1439 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1440 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1441 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1442 make_temp_ssa_name): move to tree-ssanames.h
1443 * tree-ssa-alias.h: Move prototype.
1444 * tree-ssa.h: Include tree-ssanames.h.
1445 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1446 * tree-ssanames.h: New. Move items from tree-flow*.h
1447 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1449 2013-09-12 Richard Biener <rguenther@suse.de>
1451 PR tree-optimization/58404
1452 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1453 propagate non-invariant addresses into dereferences wrapped
1454 in component references.
1456 2013-09-12 Richard Biener <rguenther@suse.de>
1458 PR tree-optimization/58402
1459 * passes.def: Move pass_late_warn_uninitialized later.
1461 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1463 * tree-ssa.h: New. Move content from tree-flow.h and
1465 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1466 Move prototypes belonging to tree-ssa.c.
1467 * tree-flow-inline.h (redirect_edge_var_map_def,
1468 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1470 * gimple.h: Adjust prototypes.
1471 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1473 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1474 * tree.h: Move prototype to tree-ssa.h.
1475 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1476 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1477 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1478 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1479 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1480 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1481 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1482 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1483 graphite-blocking.c, graphite-clast-to-gimple.c,
1484 graphite-dependences.c, graphite-interchange.c,
1485 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1486 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1487 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1488 ipa-reference.c, ipa-split.c, ipa-utils.c,
1489 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1490 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1491 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1492 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1493 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1494 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1495 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1496 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1497 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1498 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1499 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1500 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1501 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1502 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1503 value-prof.c, var-tracking.c,
1504 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1506 2013-09-12 Richard Biener <rguenther@suse.de>
1508 PR tree-optimization/58396
1509 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1510 (build_rdg): Take a loop-nest parameter, fix memleaks.
1511 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1513 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1515 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1518 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1520 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1521 The FMA instruction names should have a 'v' prefix.
1523 2013-09-12 Richard Biener <rguenther@suse.de>
1525 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1526 (dot_rdg): Use popen instead of system in optional code.
1527 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1528 (already_processed_vertex_p): Adjust.
1529 (has_anti_or_output_dependence, predecessor_has_mem_write,
1530 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1531 rdg_flag_uses): Remove.
1532 (rdg_flag_vertex): Simplify.
1533 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1535 (build_rdg_partition_for_component): Process the first vertex
1536 of a component only.
1537 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1539 2013-09-12 Alan Modra <amodra@gmail.com>
1541 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1543 2013-09-11 DJ Delorie <dj@redhat.com>
1544 Nick Clifton <nickc@redhat.com>
1546 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1547 (rl78_stack_based_mem): New.
1548 * config/rl78/constraints.md (Iv08): New.
1557 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1559 (movsi): Change from expand to insn-and-split.
1560 (ashrsi3): Clobber AX.
1564 * config/rl78/rl78.md (CC_REG): Fix.
1565 (addsi3): Allow memory and immediate operands.
1566 (addsi3_internal): Split into...
1567 (addsi3_internal_virt): ...new, and ...
1568 (addsi3_internal_real): ...new.
1570 (subsi3_internal_virt): New.
1571 (subsi3_internal_real): New.
1572 (mulsi3): Add memory operand.
1573 (mulsi3_rl78): Likewise.
1574 (mulsi3_g13): Likewise.
1575 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1576 (cbranchqi4_real): Add more constraint options.
1577 (cbranchhi4_real): Expand pattern.
1578 (cbranchhi4_real_signed): New.
1579 (cbranchhi4_real_inverted): New.
1580 (cbranchsi4_real_lt): New.
1581 (cbranchsi4_real_ge): New.
1582 (cbranchsi4_real_signed): New.
1583 (cbranchsi4_real): New.
1585 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1587 (lshrsi3_virt): Likewise.
1588 (ashlsi3_virt): Likewise.
1589 (cbranchqi4_virt_signed): New.
1590 (cbranchhi4_virt_signed): New.
1591 (cbranchsi4_virt): New.
1592 * config/rl78/rl78.c: Whitespace fixes throughout.
1593 (move_elim_pass): New.
1594 (pass_data_rl78_move_elim): New.
1595 (pass_rl78_move_elim): New.
1596 (make_pass_rl78_move_elim): New.
1597 (rl78_devirt_info): Run devirt earlier.
1598 (rl78_move_elim_info): New.
1599 (rl78_asm_file_start): Register it.
1600 (rl78_split_movsi): New.
1601 (rl78_as_legitimate_address): Allow virtual base registers when
1603 (rl78_addr_space_convert): Remove spurious debug stuff.
1604 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1605 (rl78_print_operand): More cases for not printing '#'.
1606 (rl78_expand_compare): Remove most of the logic.
1607 (content_memory): New.
1608 (clear_content_memory): New.
1609 (get_content_index): New.
1610 (get_content_name): New.
1611 (display_content_memory): New.
1612 (update_content): New.
1613 (record_content): New.
1614 (already_contains): New.
1615 (insn_ok_now): Re-recog insns with virtual registers.
1616 (add_postponed_content_update): New.
1617 (process_postponed_content_update): New.
1618 (gen_and_emit_move): New.
1619 (transcode_memory_rtx): Record new location content.
1620 Use gen_and_emit_move.
1621 (force_into_acc): New.
1622 (move_to_acc): Use gen_and_emit_move.
1623 (move_from_acc): Likewise.
1624 (move_acc_to_reg): Likewise.
1625 (move_to_x): Likewise.
1626 (move_to_hl): Likewise.
1627 (move_to_de): Likewise.
1628 (rl78_alloc_physical_registers_op1): Record location content.
1629 (has_constraint): New.
1630 (rl78_alloc_physical_registers_op2): Record location content.
1632 (rl78_alloc_physical_registers_ro1): Likewise.
1633 (rl78_alloc_physical_registers_cmp): Likewise.
1634 (rl78_alloc_physical_registers_umul): Likewise.
1635 (rl78_alloc_address_registers_macax): New.
1636 (rl78_alloc_physical_registers): Initialize and set location
1637 content memory as needed.
1638 (rl78_reorg): Make sure split2 is called.
1639 (rl78_rtx_costs): New.
1641 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1643 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1644 for (not (neg ...)) and (neg (not ...)) cases.
1646 2013-09-11 Richard Biener <rguenther@suse.de>
1649 * passes.def: Split critical edges before late uninit warning passes.
1650 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1652 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1654 PR tree-optimization/58385
1655 * fold-const.c (build_range_check): If both low and high are NULL,
1656 use omit_one_operand_loc to preserve exp side-effects.
1658 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1660 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1662 2013-09-11 Richard Biener <rguenther@suse.de>
1664 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1665 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1666 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1667 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1668 stmts_from_loop, known_dependences_p, build_empty_rdg,
1669 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1670 * tree-loop-distribution.c: ... here.
1671 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1672 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1673 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1674 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1675 * tree-loop-distribution.c: ... here.
1676 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1677 (struct partition_s): Add loops member.
1678 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1679 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1680 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1682 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1683 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1684 Sergey Lega <sergey.s.lega@intel.com>
1685 Anna Tikhonova <anna.tikhonova@intel.com>
1686 Ilya Tocar <ilya.tocar@intel.com>
1687 Andrey Turetskiy <andrey.turetskiy@intel.com>
1688 Ilya Verbin <ilya.verbin@intel.com>
1689 Kirill Yukhin <kirill.yukhin@intel.com>
1690 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1692 * config/i386/constraints.md (k): New.
1694 * config/i386/i386.c (const regclass_map): Add new mask registers.
1695 (dbx_register_map): Ditto.
1696 (dbx64_register_map): Ditto.
1697 (svr4_dbx_register_map): Ditto.
1698 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1700 (ix86_preferred_reload_class): Disable constants for mask registers.
1701 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1702 (ix86_hard_regno_mode_ok): Support new mask registers.
1703 (x86_order_regs_for_local_alloc): Ditto.
1704 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1705 (FIXED_REGISTERS): Add new mask registers.
1706 (CALL_USED_REGISTERS): Ditto.
1707 (REG_ALLOC_ORDER): Ditto.
1708 (VALID_MASK_REG_MODE): New.
1709 (FIRST_MASK_REG): Ditto.
1710 (LAST_MASK_REG): Ditto.
1711 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1712 (MAYBE_MASK_CLASS_P): New.
1713 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1714 (REG_CLASS_CONTENTS): Ditto.
1715 (MASK_REGNO_P): New.
1716 (ANY_MASK_REG_P): Ditto.
1717 (HI_REGISTER_NAMES): Add new mask registers.
1718 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1719 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1721 (attribute "type"): Add mskmov, msklog.
1722 (attribute "length_immediate"): Support them.
1723 (attribute "memory"): Ditto.
1724 (attribute "prefix_0f"): Ditto.
1725 (*movhi_internal): Support new mask registers.
1726 (*movqi_internal): Ditto.
1727 (define_split): Split out clobber pattern is a logic
1728 insn on mask registers.
1729 (*k<logic><mode>): New.
1730 (*andhi_1): Extend to support mask regs.
1731 (*andqi_1): Extend to support mask regs.
1733 (define_split): Split and-not to and and not if operands
1735 (*<code><mode>_1): Separate HI mode to new pattern...
1736 (*<code>hi_1): This.
1737 (*<code>qi_1): Extend to support mask regs.
1739 (kortestzhi): Ditto.
1740 (kortestchi): Ditto.
1742 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1743 (*one_cmplhi2_1): ...Here, now with mask registers support.
1744 (*one_cmplqi2_1): Support new mask registers.
1745 (HI/QImode arithmetics splitter): Don't split if mask registers
1747 (HI/QImode not splitter): Ditto.
1748 * config/i386/predicated.md (mask_reg_operand): New.
1749 (general_reg_operand): Ditto.
1751 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1753 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1754 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1756 2013-09-10 Jeff Law <law@redhat.com>
1758 PR tree-optimization/58380
1759 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1760 of threading through a buried loop header.
1762 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1763 return value for single successor case.
1765 2013-09-10 Jan Hubicka <jh@suse.cz>
1767 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1769 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1772 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1773 support conditional execution.
1774 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1776 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1778 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1780 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1781 Check destination too to check move insn.
1782 (undo_optional_reloads): Add check that the original peudo did not
1783 changed its allocation and the optional reload was inherited on last
1784 inheritance pass. Break loop after deciding to keep optional reload.
1785 (lra_undo_inheritance): Add check that inherited pseudo still in
1788 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1790 * config/aarch64/aarch64.md (generic_sched): New.
1791 * config/aarch64/aarch64-generic.md (load): Make conditional
1792 on generic_sched attribute.
1793 (nonload): Likewise.
1795 2013-09-10 Jan Hubicka <jh@suse.cz>
1797 * lto-cgraph.c: Include ipa-utils.h.
1798 (compute_ltrans_boundary): Also add possible targets into the boundary.
1800 2013-09-10 Jan Hubicka <jh@suse.cz>
1802 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1803 VAR_DECL of vtable rather than full expression.
1805 2013-09-10 Jan Hubicka <jh@suse.cz>
1806 Paolo Carlini <paolo.carlini@oracle.com>
1808 * cgraphunit.c (analyze_functions): Save input_location, set it
1809 to UNKNOWN_LOCATION and restore it at the end.
1811 2013-09-10 Martin Jambor <mjambor@suse.cz>
1813 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1814 represented by a thunk.
1816 2013-09-10 Jeff Law <law@redhat.com>
1818 PR tree-optimization/58343
1819 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1820 jump threading requests through loop headers buried in the middle
1821 of a jump threading path.
1823 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1824 in return value/type.
1826 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1828 PR rtl-optimization/58365
1829 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1830 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1833 2013-09-10 Richard Biener <rguenther@suse.de>
1835 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1836 * tree-data-ref.c (create_rdg_vertices): Collect all data
1837 references, signal failure to the caller, use data-ref API.
1838 (build_rdg): Compute data references only once. Maintain lifetime
1839 of data references and data dependences from within RDG.
1840 (free_rdg): Free dependence relations.
1841 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1842 inventing extra dependences.
1843 (distribute_loop): Update for RDG API changes.
1845 2013-09-10 Kai Tietz <ktietz@redhat.com>
1847 * doc/invoke.texi (fms-extensions): Document changed
1848 behavior for ms-abi targets.
1849 * config/i386/i386.c (ix86_option_override_internal):
1850 Set default value of option -fms-extension for ms-abi targets.
1852 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1854 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
1856 2013-09-10 Alan Modra <amodra@gmail.com>
1859 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
1861 2013-09-10 Alan Modra <amodra@gmail.com>
1863 * config/rs6000/predicates.md (add_cint_operand): New.
1864 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
1865 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
1866 using add_cint_operand.
1867 (largetoc_high_plus_aix): Likewise.
1869 2013-09-09 Jakub Jelinek <jakub@redhat.com>
1871 PR tree-optimization/58364
1872 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
1873 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
1874 the current range can't be an unconditional true or false.
1876 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
1878 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
1880 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
1882 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
1884 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
1887 * doc/invoke.texi (-Wdelete-incomplete): Document it.
1889 2013-09-09 Ian Bolton <ian.bolton@arm.com>
1891 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
1892 NO_REGS for immediate that can't be moved directly into FP_REGS.
1894 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1896 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
1897 comparison with negated operand.
1898 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
1901 2013-09-09 Richard Biener <rguenther@suse.de>
1904 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
1905 of a subloop record all its block as affecting loop-closed SSA form.
1907 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1909 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
1910 of an rtx/bitpos pair.
1911 (store_fixed_bit_field): Update accordingly.
1913 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1915 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
1917 * builtins.c (expand_errno_check): Likewise.
1918 * dwarf2cfi.c (init_return_column_size): Likewise.
1919 * except.c (sjlj_mark_call_sites): Likewise.
1920 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
1921 * lra-constraints.c (emit_inc): Likewise.
1922 * ree.c (combine_set_extension): Likewise.
1923 * regmove.c (fixup_match_2): Likewise.
1924 * reload1.c (inc_for_reload): Likewise.
1926 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1928 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
1929 (force_to_mode, simplify_shift_const_1, simplify_comparison):
1930 Use gen_int_mode with the mode of the associated simplify_* call.
1931 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
1932 * expmed.c (expand_shift_1): Likewise.
1933 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1934 * loop-iv.c (iv_number_of_iterations): Likewise.
1935 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1936 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
1938 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1940 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
1941 of the associated expand_* call.
1942 (asan_emit_stack_protection): Likewise.
1943 * builtins.c (round_trampoline_addr): Likewise.
1944 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
1945 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
1946 (emit_store_flag): Likewise.
1947 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
1948 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
1950 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1951 * ifcvt.c (noce_try_store_flag_constants): Likewise.
1952 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1953 * modulo-sched.c (generate_prolog_epilog): Likewise.
1954 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
1955 (expand_ctz, expand_ffs, expand_unop): Likewise.
1957 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1959 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
1960 of the associated gen_rtx_* call.
1961 * caller-save.c (init_caller_save): Likewise.
1962 * combine.c (find_split_point, make_extraction): Likewise.
1963 (make_compound_operation): Likewise.
1964 * dwarf2out.c (mem_loc_descriptor): Likewise.
1965 * explow.c (plus_constant, probe_stack_range): Likewise.
1966 * expmed.c (expand_mult_const): Likewise.
1967 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
1968 * reload1.c (init_reload): Likewise.
1969 * valtrack.c (cleanup_auto_inc_dec): Likewise.
1970 * var-tracking.c (adjust_mems): Likewise.
1971 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
1972 rather than gen_rtx_fmt_ee.
1974 2013-09-09 Jan Hubicka <jh@suse.cz>
1977 * value-prof.c (gimple_ic): Copy also abnormal edges.
1979 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1981 * asan.c (asan_shadow_cst): Use gen_int_mode.
1983 2013-09-08 Jan Hubicka <jh@suse.cz>
1985 * ipa-profile.c: Add toplevel comment.
1986 (ipa_propagate_frequency_1): Be more conservative when profile is read.
1987 (contains_hot_call_p): New function.
1988 (ipa_propagate_frequency): Set frequencies based on counts when
1990 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
1991 profile; do not tamper with profile after inlining if it is read.
1993 2013-09-08 Jan Hubicka <jh@suse.cz>
1995 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
1998 2013-09-08 Jan Hubicka <jh@suse.cz>
2000 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2003 2013-09-08 Jeff Law <law@redhat.com>
2006 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2009 2013-09-08 Andi Kleen <ak@linux.intel.com>
2011 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2013 2013-09-08 Jan Hubicka <jh@suse.cz>
2015 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2017 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2019 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2020 for non-debug insns.
2021 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2022 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2023 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2026 2013-09-08 Jan Hubicka <jh@suse.cz>
2028 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2029 targets and devirtualize to BUILT_IN_UNREACHABLE.
2030 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2031 * ipa.c (walk_polymorphic_call_targets): New function.
2032 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2033 functions; use the new timevar.
2034 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2035 was removed from the program.
2036 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2037 not consider it in the walk when its vtable is dead.
2038 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2040 (devirt_variable_node_removal_hook): New function.
2041 (possible_polymorphic_call_targets): Also register
2042 devirt_variable_node_removal_hook.
2043 (ipa_devirt): Do not do non-speculative devirtualization.
2044 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2046 2013-09-08 Jan Hubicka <jh@suse.cz>
2048 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2049 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2050 varpool_remove_variable_insertion_hook): Declare.
2051 * varpool.c (varpool_node_hook_list): New structure.
2052 (first_varpool_node_removal_hook,
2053 first_varpool_variable_insertion_hook): New variables.
2054 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2055 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2056 varpool_remove_variable_insertion_hook,
2057 varpool_call_variable_insertion_hooks): New functions.
2058 (varpool_remove_node): Use it.
2060 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2063 * diagnostic.c (diagnostic_build_prefix): When s.file is
2064 "<built-in>" don't output line and column numbers.
2066 2013-09-06 Jan Hubicka <jh@suse.cz>
2068 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2069 * lto-streamer-out.c: Stream thunks, too.
2070 * lto-streamer-in.c (input_function): Pop cfun here
2071 (lto_read_body): Instead of here.
2073 2013-09-06 Caroline Tice <cmtice@google.com>
2075 * doc/install.texi: Add documentation for the --enable-vtable-verify
2076 and the --disable-libvtv configure options.
2078 2013-09-06 Jeff Law <law@redhat.com>
2080 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2081 edge implied equivalences into successor phis.
2083 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2085 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2087 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2089 * resource.c (mark_target_live_regs): Compute resources taking
2090 into account if a call is predicated or not.
2092 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2094 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2095 the compiler in the function names.
2097 2013-09-06 Jan Hubicka <jh@suse.cz>
2100 * ipa-inline.c (has_caller_p): New function.
2101 (want_inline_function_to_all_callers_p): Use it.
2102 (sum_callers, inline_to_all_callers): Break out from ...
2103 (ipa_inline): ... here.
2105 2013-09-06 Jan Hubicka <jh@suse.cz>
2107 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2108 only when AVX is enabled.
2110 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2112 * config/aarch64/aarch64.md
2113 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2114 is fpsimd_<load/store>2.
2115 (load_pair<mode>): Likewise.
2116 (store_pair<mode>): Likewise.
2118 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2120 * config/arm/types.md (type): Add "mrs" type.
2121 * config/aarch64/aarch64.md
2122 (aarch64_load_tp_hard): Make type "mrs".
2124 (load_tp_hard): Make type "mrs".
2125 * config/arm/cortex-a15.md: Update with new attributes.
2126 * config/arm/cortex-a5.md: Update with new attributes.
2127 * config/arm/cortex-a53.md: Update with new attributes.
2128 * config/arm/cortex-a7.md: Update with new attributes.
2129 * config/arm/cortex-a8.md: Update with new attributes.
2130 * config/arm/cortex-a9.md: Update with new attributes.
2131 * config/arm/cortex-m4.md: Update with new attributes.
2132 * config/arm/cortex-r4.md: Update with new attributes.
2133 * config/arm/fa526.md: Update with new attributes.
2134 * config/arm/fa606te.md: Update with new attributes.
2135 * config/arm/fa626te.md: Update with new attributes.
2136 * config/arm/fa726te.md: Update with new attributes.
2138 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2140 * config/aarch64/aarch64.md
2141 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2142 (*movtf_aarch64): Likewise.
2144 (thumb1_movdi_insn): Use "multiple" for type where more than one
2145 instruction is used for a move.
2146 (*arm32_movhf): Likewise.
2147 (*thumb_movdf_insn): Likewise.
2149 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2151 * config/arm/types.md (type): Rename fcpys to fmov.
2153 (*arm_movsi_vfp): Rename type fcpys as fmov.
2154 (*thumb2_movsi_vfp): Likewise
2155 (*movhf_vfp_neon): Likewise
2156 (*movhf_vfp): Likewise
2157 (*movsf_vfp): Likewise
2158 (*thumb2_movsf_vfp): Likewise
2159 (*movsfcc_vfp): Likewise
2160 (*thumb2_movsfcc_vfp): Likewise
2161 * config/aarch64/aarch64-simd.md
2162 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2163 * config/aarch64/aarch64.md
2164 (*movsi_aarch64): Replace type mov_reg with fmovs.
2165 (*movdi_aarch64): Likewise
2166 (*movsf_aarch64): Likewise
2167 (*movdf_aarch64): Likewise
2169 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2170 * config/arm/iwmmxt.md
2171 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2172 * config/arm/arm1020e.md: Update with new attributes.
2173 * config/arm/cortex-a15-neon.md: Update with new attributes.
2174 * config/arm/cortex-a5.md: Update with new attributes.
2175 * config/arm/cortex-a53.md: Update with new attributes.
2176 * config/arm/cortex-a7.md: Update with new attributes.
2177 * config/arm/cortex-a8-neon.md: Update with new attributes.
2178 * config/arm/cortex-a9.md: Update with new attributes.
2179 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2180 * config/arm/cortex-r4f.md: Update with new attributes.
2181 * config/arm/marvell-pj4.md: Update with new attributes.
2182 * config/arm/vfp11.md: Update with new attributes.
2184 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2186 * config/aarch64/aarch64.md
2187 (*madd<mode>): Fix type attribute.
2188 (*maddsi_uxtw): Likewise.
2189 (*msub<mode>): Likewise.
2190 (*msubsi_uxtw): Likewise.
2191 (<su_optab>maddsidi4): Likewise.
2192 (<su_optab>msubsidi4): Likewise.
2194 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2196 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2197 * config/arm/arm.md (core_cycles): Remove fdiv.
2198 * config/arm/vfp.md:
2199 (*sqrtsf2_vfp): Update for attribute changes.
2200 (*sqrtdf2_vfp): Likewise.
2201 * config/aarch64/aarch64.md:
2202 (sqrt<mode>2): Update for attribute changes.
2203 * config/arm/arm1020e.md: Update with new attributes.
2204 * config/arm/cortex-a15-neon.md: Update with new attributes.
2205 * config/arm/cortex-a5.md: Update with new attributes.
2206 * config/arm/cortex-a53.md: Update with new attributes.
2207 * config/arm/cortex-a7.md: Update with new attributes.
2208 * config/arm/cortex-a8-neon.md: Update with new attributes.
2209 * config/arm/cortex-a9.md: Update with new attributes.
2210 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2211 * config/arm/cortex-r4f.md: Update with new attributes.
2212 * config/arm/marvell-pj4.md: Update with new attributes.
2213 * config/arm/vfp11.md: Update with new attributes.
2215 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2217 * config/arm/types.md
2218 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2219 * config/aarch64/aarch64.md
2220 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2222 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2223 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2224 (float<GPI:mode><GPF:mode>2): Likewise.
2226 (*truncsisf2_vfp): Update with new attributes.
2227 (*truncsidf2_vfp): Likewise.
2228 (fixuns_truncsfsi2): Likewise.
2229 (fixuns_truncdfsi2): Likewise.
2230 (*floatsisf2_vfp): Likewise.
2231 (*floatsidf2_vfp): Likewise.
2232 (floatunssisf2): Likewise.
2233 (floatunssidf2): Likewise.
2234 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2235 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2236 * config/arm/arm1020e.md: Update with new attributes.
2237 * config/arm/cortex-a15-neon.md: Update with new attributes.
2238 * config/arm/cortex-a5.md: Update with new attributes.
2239 * config/arm/cortex-a53.md: Update with new attributes.
2240 * config/arm/cortex-a7.md: Update with new attributes.
2241 * config/arm/cortex-a8-neon.md: Update with new attributes.
2242 * config/arm/cortex-a9.md: Update with new attributes.
2243 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2244 * config/arm/cortex-r4f.md: Update with new attributes.
2245 * config/arm/marvell-pj4.md: Update with new attributes.
2246 * config/arm/vfp11.md: Update with new attributes.
2248 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2250 * config/aarch64/arm_neon.h
2251 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2252 (vqtbx<1,2,3,4><q>_s8): Likewise.
2254 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2256 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2257 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2258 (add<mode>3): Add type attribute.
2259 (add<mode>3): Likewise.
2260 (usadd<mode>3): Likewise.
2261 (ssadd<mode>3): Likewise.
2262 (sub<mode>3): Likewise.
2263 (sub<mode>3): Likewise.
2264 (ussub<mode>3): Likewise.
2265 (sssub<mode>3): Likewise.
2266 (ssmulsa3): Likewise.
2267 (usmulusa3): Likewise.
2268 (arm_usatsihi): Likewise.
2270 (*movdi_vfp): Add types for all instructions.
2271 (*movdi_vfp_cortexa8): Likewise.
2272 (*movhf_vfp_neon): Likewise.
2273 (*movhf_vfp): Likewise.
2274 (*movdf_vfp): Likewise.
2275 (*thumb2_movdf_vfp): Likewise.
2276 (*thumb2_movdfcc_vfp): Likewise.
2277 * config/arm/arm.md: Add type attribute to all insn patterns.
2278 (*thumb1_adddi3): Add type attribute.
2279 (*arm_adddi3): Likewise.
2280 (*adddi_sesidi_di): Likewise.
2281 (*adddi_zesidi_di): Likewise.
2282 (*thumb1_addsi3): Likewise.
2283 (addsi3_compare0): Likewise.
2284 (*addsi3_compare0_scratch): Likewise.
2285 (*compare_negsi_si): Likewise.
2286 (cmpsi2_addneg): Likewise.
2287 (*addsi3_carryin_<optab>): Likewise.
2288 (*addsi3_carryin_alt2_<optab>): Likewise.
2289 (*addsi3_carryin_clobercc_<optab>): Likewise.
2290 (*subsi3_carryin): Likewise.
2291 (*subsi3_carryin_const): Likewise.
2292 (*subsi3_carryin_compare): Likewise.
2293 (*subsi3_carryin_compare_const): Likewise.
2294 (*arm_subdi3): Likewise.
2295 (*thumb_subdi3): Likewise.
2296 (*subdi_di_zesidi): Likewise.
2297 (*subdi_di_sesidi): Likewise.
2298 (*subdi_zesidi_di): Likewise.
2299 (*subdi_sesidi_di): Likewise.
2300 (*subdi_zesidi_ze): Likewise.
2301 (thumb1_subsi3_insn): Likewise.
2302 (*arm_subsi3_insn): Likewise.
2303 (*anddi3_insn): Likewise.
2304 (*anddi_zesidi_di): Likewise.
2305 (*anddi_sesdi_di): Likewise.
2306 (*ne_zeroextracts): Likewise.
2307 (*ne_zeroextracts): Likewise.
2308 (*ite_ne_zeroextr): Likewise.
2309 (*ite_ne_zeroextr): Likewise.
2310 (*anddi_notdi_di): Likewise.
2311 (*anddi_notzesidi): Likewise.
2312 (*anddi_notsesidi): Likewise.
2313 (andsi_notsi_si): Likewise.
2314 (thumb1_bicsi3): Likewise.
2315 (*iordi3_insn): Likewise.
2316 (*iordi_zesidi_di): Likewise.
2317 (*iordi_sesidi_di): Likewise.
2318 (*thumb1_iorsi3_insn): Likewise.
2319 (*xordi3_insn): Likewise.
2320 (*xordi_zesidi_di): Likewise.
2321 (*xordi_sesidi_di): Likewise.
2322 (*arm_xorsi3): Likewise.
2323 (*andsi_iorsi3_no): Likewise.
2324 (*smax_0): Likewise.
2325 (*smax_m1): Likewise.
2326 (*arm_smax_insn): Likewise.
2327 (*smin_0): Likewise.
2328 (*arm_smin_insn): Likewise.
2329 (*arm_umaxsi3): Likewise.
2330 (*arm_uminsi3): Likewise.
2331 (*minmax_arithsi): Likewise.
2332 (*minmax_arithsi_): Likewise.
2333 (*satsi_<SAT:code>): Likewise.
2334 (arm_ashldi3_1bit): Likewise.
2335 (arm_ashrdi3_1bit): Likewise.
2336 (arm_lshrdi3_1bit): Likewise.
2337 (*arm_negdi2): Likewise.
2338 (*thumb1_negdi2): Likewise.
2339 (*arm_negsi2): Likewise.
2340 (*thumb1_negsi2): Likewise.
2341 (*negdi_extendsid): Likewise.
2342 (*negdi_zero_extend): Likewise.
2343 (*arm_abssi2): Likewise.
2344 (*thumb1_abssi2): Likewise.
2345 (*arm_neg_abssi2): Likewise.
2346 (*thumb1_neg_abss): Likewise.
2347 (one_cmpldi2): Likewise.
2348 (extend<mode>di2): Likewise.
2349 (*compareqi_eq0): Likewise.
2350 (*arm_extendhisi2addsi): Likewise.
2351 (*arm_movdi): Likewise.
2352 (*thumb1_movdi_insn): Likewise.
2353 (*arm_movt): Likewise.
2354 (*thumb1_movsi_insn): Likewise.
2355 (pic_add_dot_plus_four): Likewise.
2356 (pic_add_dot_plus_eight): Likewise.
2357 (tls_load_dot_plus_eight): Likewise.
2358 (*thumb1_movhi_insn): Likewise.
2359 (*thumb1_movsf_insn): Likewise.
2360 (*movdf_soft_insn): Likewise.
2361 (*thumb_movdf_insn): Likewise.
2362 (cbranchsi4_insn): Likewise.
2363 (cbranchsi4_scratch): Likewise.
2364 (*negated_cbranchsi4): Likewise.
2365 (*tbit_cbranch): Likewise.
2366 (*tlobits_cbranch): Likewise.
2367 (*tstsi3_cbranch): Likewise.
2368 (*cbranchne_decr1): Likewise.
2369 (*addsi3_cbranch): Likewise.
2370 (*addsi3_cbranch_scratch): Likewise.
2371 (*arm_cmpdi_insn): Likewise.
2372 (*arm_cmpdi_unsig): Likewise.
2373 (*arm_cmpdi_zero): Likewise.
2374 (*thumb_cmpdi_zero): Likewise.
2375 (*deleted_compare): Likewise.
2376 (*mov_scc): Likewise.
2377 (*mov_negscc): Likewise.
2378 (*mov_notscc): Likewise.
2379 (*cstoresi_eq0_thumb1_insn): Likewise.
2380 (cstoresi_nltu_thumb1): Likewise.
2381 (cstoresi_ltu_thu): Likewise.
2382 (thumb1_addsi3_addgeu): Likewise.
2383 (*arm_jump): Likewise.
2384 (*thumb_jump): Likewise.
2385 (*check_arch2): Likewise.
2386 (arm_casesi_internal): Likewise.
2387 (thumb1_casesi_dispatch): Likewise.
2388 (*arm_indirect_jump): Likewise.
2389 (*thumb1_indirect_jump): Likewise.
2391 (*and_scc): Likewise.
2392 (*ior_scc): Likewise.
2393 (*compare_scc): Likewise.
2394 (*cond_move): Likewise.
2395 (*cond_arith): Likewise.
2396 (*cond_sub): Likewise.
2397 (*cmp_ite0): Likewise.
2398 (*cmp_ite1): Likewise.
2399 (*cmp_and): Likewise.
2400 (*cmp_ior): Likewise.
2401 (*ior_scc_scc): Likewise.
2402 (*ior_scc_scc_cmp): Likewise.
2403 (*and_scc_scc): Likewise.
2404 (*and_scc_scc_cmp): Likewise.
2405 (*and_scc_scc_nod): Likewise.
2406 (*negscc): Likewise.
2407 (movcond_addsi): Likewise.
2408 (movcond): Likewise.
2409 (*ifcompare_plus_move): Likewise.
2410 (*if_plus_move): Likewise.
2411 (*ifcompare_move_plus): Likewise.
2412 (*if_move_plus): Likewise.
2413 (*ifcompare_arith_arith): Likewise.
2414 (*if_arith_arith): Likewise.
2415 (*ifcompare_arith_move): Likewise.
2416 (*if_arith_move): Likewise.
2417 (*ifcompare_move_arith): Likewise.
2418 (*if_move_arith): Likewise.
2419 (*ifcompare_move_not): Likewise.
2420 (*if_move_not): Likewise.
2421 (*ifcompare_not_move): Likewise.
2422 (*if_not_move): Likewise.
2423 (*ifcompare_shift_move): Likewise.
2424 (*if_shift_move): Likewise.
2425 (*ifcompare_move_shift): Likewise.
2426 (*if_move_shift): Likewise.
2427 (*ifcompare_shift_shift): Likewise.
2428 (*ifcompare_not_arith): Likewise.
2429 (*ifcompare_arith_not): Likewise.
2430 (*if_arith_not): Likewise.
2431 (*ifcompare_neg_move): Likewise.
2432 (*if_neg_move): Likewise.
2433 (*ifcompare_move_neg): Likewise.
2434 (*if_move_neg): Likewise.
2435 (prologue_thumb1_interwork): Likewise.
2436 (*cond_move_not): Likewise.
2437 (*sign_extract_onebit): Likewise.
2438 (*not_signextract_onebit): Likewise.
2439 (stack_tie): Likewise.
2440 (align_4): Likewise.
2441 (align_8): Likewise.
2442 (consttable_end): Likewise.
2443 (consttable_1): Likewise.
2444 (consttable_2): Likewise.
2445 (consttable_4): Likewise.
2446 (consttable_8): Likewise.
2447 (consttable_16): Likewise.
2448 (*thumb1_tablejump): Likewise.
2449 (prefetch): Likewise.
2450 (force_register_use): Likewise.
2451 (thumb_eh_return): Likewise.
2452 (load_tp_hard): Likewise.
2453 (load_tp_soft): Likewise.
2454 (tlscall): Likewise.
2455 (*arm_movtas_ze): Likewise.
2456 (*arm_rev): Likewise.
2457 (*arm_revsh): Likewise.
2458 (*arm_rev16): Likewise.
2459 * config/arm/thumb2.md
2460 (*thumb2_smaxsi3): Likewise.
2461 (*thumb2_sminsi3): Likewise.
2462 (*thumb32_umaxsi3): Likewise.
2463 (*thumb2_uminsi3): Likewise.
2464 (*thumb2_negdi2): Likewise.
2465 (*thumb2_abssi2): Likewise.
2466 (*thumb2_neg_abss): Likewise.
2467 (*thumb2_movsi_insn): Likewise.
2468 (tls_load_dot_plus_four): Likewise.
2469 (*thumb2_movhi_insn): Likewise.
2470 (*thumb2_mov_scc): Likewise.
2471 (*thumb2_mov_negs): Likewise.
2472 (*thumb2_mov_negs): Likewise.
2473 (*thumb2_mov_nots): Likewise.
2474 (*thumb2_mov_nots): Likewise.
2475 (*thumb2_movsicc_): Likewise.
2476 (*thumb2_movsfcc_soft_insn): Likewise.
2477 (*thumb2_indirect_jump): Likewise.
2478 (*thumb2_and_scc): Likewise.
2479 (*thumb2_ior_scc): Likewise.
2480 (*thumb2_ior_scc_strict_it): Likewise.
2481 (*thumb2_cond_move): Likewise.
2482 (*thumb2_cond_arith): Likewise.
2483 (*thumb2_cond_ari): Likewise.
2484 (*thumb2_cond_sub): Likewise.
2485 (*thumb2_negscc): Likewise.
2486 (*thumb2_movcond): Likewise.
2487 (thumb2_casesi_internal): Likewise.
2488 (thumb2_casesi_internal_pic): Likewise.
2489 (*thumb2_alusi3_short): Likewise.
2490 (*thumb2_mov<mode>_shortim): Likewise.
2491 (*thumb2_addsi_short): Likewise.
2492 (*thumb2_subsi_short): Likewise.
2493 (thumb2_addsi3_compare0): Likewise.
2494 (*thumb2_cbz): Likewise.
2495 (*thumb2_cbnz): Likewise.
2496 (*thumb2_one_cmplsi2_short): Likewise.
2497 (*thumb2_negsi2_short): Likewise.
2498 (*orsi_notsi_si): Likewise.
2499 * config/arm/arm1020e.md: Update with new attributes.
2500 * config/arm/arm1026ejs.md: Update with new attributes.
2501 * config/arm/arm1136jfs.md: Update with new attributes.
2502 * config/arm/arm926ejs.md: Update with new attributes.
2503 * config/arm/cortex-a15.md: Update with new attributes.
2504 * config/arm/cortex-a5.md: Update with new attributes.
2505 * config/arm/cortex-a53.md: Update with new attributes.
2506 * config/arm/cortex-a7.md: Update with new attributes.
2507 * config/arm/cortex-a8.md: Update with new attributes.
2508 * config/arm/cortex-a9.md: Update with new attributes.
2509 * config/arm/cortex-m4.md: Update with new attributes.
2510 * config/arm/cortex-r4.md: Update with new attributes.
2511 * config/arm/fa526.md: Update with new attributes.
2512 * config/arm/fa606te.md: Update with new attributes.
2513 * config/arm/fa626te.md: Update with new attributes.
2514 * config/arm/fa726te.md: Update with new attributes.
2516 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2518 * config/aarch64/aarch64-simd.md
2519 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2520 <vwx> iterator to ensure correct register choice.
2521 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2522 (aarch64_sqdmull_n<mode>): Likewise.
2523 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2524 * config/aarch64/arm_neon.h
2525 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2526 (vml<as><q>_n_<su>16): Likewise.
2527 (vml<as>l_high_lane<q>_<su>16): Likewise.
2528 (vml<as>l_high_n_<su>16): Likewise.
2529 (vml<as>l_lane<q>_<su>16): Likewise.
2530 (vml<as>l_n_<su>16): Likewise.
2531 (vmul<q>_lane<q>_<su>16): Likewise.
2532 (vmul<q>_n_<su>16): Likewise.
2533 (vmull_lane<q>_<su>16): Likewise.
2534 (vmull_n_<su>16): Likewise.
2535 (vmull_high_lane<q>_<su>16): Likewise.
2536 (vmull_high_n_<su>16): Likewise.
2537 (vqrdmulh<q>_n_s16): Likewise.
2539 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2541 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2542 have the correct lane parameter.
2544 2013-09-06 Richard Biener <rguenther@suse.de>
2546 * cfganal.c (control_dependences::~control_dependences):
2547 Properly free all of the vector.
2549 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2552 * config/i386/i386.c (ix86_conditional_register_usage):
2553 Proper initialize extended SSE registers.
2555 2013-09-06 Jan Hubicka <jh@suse.cz>
2557 PR tree-optimization/58311
2558 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2560 2013-09-06 Jan Hubicka <jh@suse.cz>
2562 * Makefile.in (tree-sra.o): Update dependencies.
2563 * tree-sra.c: Include ipa-utils.h
2564 (scan_function): Use recursive_call_p.
2565 (has_caller_p): New function.
2566 (cgraph_for_node_and_aliases): Count also callers of aliases.
2568 2013-09-06 Jan Hubicka <jh@suse.cz>
2571 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2572 * tree-tailcall.c: Include ipa-utils.h.
2573 (find_tail_calls): Use it.
2574 * ipa-pure-const.c (check_call): Likewise.
2575 * ipa-utils.c (recursive_call_p): New function.
2576 * ipa-utils.h (recursive_call_p): Dclare.
2577 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2578 (symtab_semantically_equivalent_p): New function.
2579 * Makefile.in (tree-tailcall.o): Update dependencies.
2581 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2583 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2586 2013-09-06 Richard Biener <rguenther@suse.de>
2588 * lto-streamer.h (lto_global_var_decls): Remove.
2589 * Makefile.in (OBJS): Remove lto-symtab.o.
2590 (lto-symtab.o): Remove.
2591 (GTFILES): Remove lto-symtab.c
2592 * lto-symtab.c: Move to lto/
2594 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2596 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2597 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2598 (UNSPEC_FPINT_RINT): New constant definitions.
2599 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2600 definition with 2 attributes.
2601 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2602 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2605 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2607 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2608 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2609 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2610 attribute to "z196_alone".
2611 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2614 2013-09-06 Richard Biener <rguenther@suse.de>
2616 * basic-block.h (class control_dependences): New.
2617 * tree-ssa-dce.c (control_dependence_map): Remove.
2619 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2620 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2621 find_pdom, find_control_dependence, find_all_control_dependences):
2623 (mark_control_dependent_edges_necessary,
2624 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2625 tree_dce_done, perform_tree_ssa_dce): Adjust.
2626 * cfganal.c (set_control_dependence_map_bit,
2627 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2628 find_all_control_dependences): Move from tree-ssa-dce.c and
2629 implement as methods of control_dependences class.
2630 (control_dependences::control_dependences): New.
2631 (control_dependences::~control_dependences): Likewise.
2632 (control_dependences::get_edges_dependent_on): Likewise.
2633 (control_dependences::get_edge): Likewise.
2635 2013-09-04 Jan Hubicka <jh@suse.cz>
2637 * tree.c (types_same_for_odr): Drop overactive check.
2638 * ipa-devirt.c (hash_type_name): Likewise.
2640 2013-09-04 Jan Hubicka <jh@suse.cz>
2642 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2643 (analyze_functions): ... here.
2645 2013-09-04 Jan Hubicka <jh@suse.cz>
2648 * cgraphunit.c (analyze_functions): Clear AUX fields
2649 after processing; initialize assembler name has.
2651 2013-09-05 Jeff Law <law@redhat.com>
2653 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2654 from thread_around_empty_block. Record threading path into PATH.
2655 Recurse if threading through the initial block is successful.
2656 (thread_across_edge): Corresponding changes to slightly simplify.
2658 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2660 * config/aarch64/aarch64.md
2661 (type): Remove frecpe, frecps, frecpx.
2662 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2663 fix to be a TARGET_SIMD instruction.
2664 (aarch64_frecps): Remove.
2665 * config/aarch64/aarch64-simd.md
2666 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2667 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2669 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2670 Sofiane Naci <sofiane.naci@arm.com>
2672 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2673 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2674 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2675 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2676 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2677 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2678 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2679 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2680 into "clz, "rbit". Rename "shift" to "shift_imm".
2681 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2682 changes. Update for attribute changes all occurrences of arlo_* and
2684 * config/arm/arm-fixed.md: Update for attribute changes
2685 all occurrences of arlo_* types.
2686 * config/arm/thumb2.md: Update for attribute changes all occurrences
2688 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2689 (cortexa7_older_only): Likewise.
2690 (cortexa7_younger): Likewise.
2691 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2692 (1020alu_shift_op): Likewise.
2693 (1020alu_shift_reg_op): Likewise.
2694 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2695 (alu_shift_op): Likewise.
2696 (alu_shift_reg_op): Likewise.
2697 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2698 (11_alu_shift_op): Likewise.
2699 (11_alu_shift_reg_op): Likewise.
2700 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2701 (9_alu_shift_reg_op): Likewise.
2702 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2704 (cortex_a15_alu_shift): Likewise.
2705 (cortex_a15_alu_shift_reg): Likewise.
2706 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2708 (cortex_a5_alu_shift): Likewise.
2709 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2711 (cortex_a53_alu_shift): Likewise.
2712 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2714 (cortex_a7_alu_reg): Likewise.
2715 (cortex_a7_alu_shift): Likewise.
2716 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2718 (cortex_a8_alu_shift): Likewise.
2719 (cortex_a8_alu_shift_reg): Likewise.
2720 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2721 (cortex_a9_dp_shift): Likewise.
2722 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2724 * config/arm/cortex-r4.md
2725 (cortex_r4_alu): Update for attribute changes.
2726 (cortex_r4_mov): Likewise.
2727 (cortex_r4_alu_shift_reg): Likewise.
2728 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2729 (526_alu_shift_op): Likewise.
2730 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2731 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2732 (626te_alu_shift_op): Likewise.
2733 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2734 (726te_alu_shift_op): Likewise.
2735 (726te_alu_shift_reg_op): Likewise.
2736 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2737 (mp626_alu_shift_op): Likewise.
2738 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2739 (pj4_alu_conds): Likewise.
2740 (pj4_shift): Likewise.
2741 (pj4_shift_conds): Likewise.
2742 (pj4_alu_shift): Likewise.
2743 (pj4_alu_shift_conds): Likewise.
2744 * config/aarch64/aarch64.md: Update for attribute change
2745 all occurrences of arlo_* and shift* types.
2747 2013-09-05 Mike Stump <mikestump@comcast.net>
2749 * tree.h: Move documentation for tree_function_decl to tree-core.h
2750 with the declaration.
2752 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2755 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2756 looking for widest mode.
2758 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2760 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2762 2013-09-05 Richard Biener <rguenther@suse.de>
2764 PR tree-optimization/58137
2765 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2766 Do not create vectors of pointers.
2767 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2768 types for the components of the vector initializer.
2769 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2770 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2772 2013-09-05 Martin Jambor <mjambor@suse.cz>
2774 * ipa-prop.c (remove_described_reference): Accept missing references,
2775 return false if that hppens, otherwise return true.
2776 (cgraph_node_for_jfunc): New function.
2777 (try_decrement_rdesc_refcount): Likewise.
2778 (try_make_edge_direct_simple_call): Use them.
2779 (ipa_edge_removal_hook): Remove references from rdescs.
2780 (ipa_edge_duplication_hook): Clone rdescs and their references
2781 when the new edge has the same caller as the old one.
2782 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2783 reference before removing any edges.
2785 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2787 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2789 * thumb2.md (thumb2_storewb_parisi): New pattern.
2791 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2793 * config/aarch64/aarch64-option-extensions.def: Add
2794 AARCH64_OPT_EXTENSION of 'crc'.
2795 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2796 (AARCH64_ISA_CRC): Ditto.
2797 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2798 description of the CRC extension.
2800 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2802 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2803 * config/rs6000/linux.h: Ditto.
2804 * alpha/linux.h: Ditto.
2805 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2806 no_c99_libc_has_function.
2807 * config/c6x/uclinux-elf.h: Ditto.
2808 * config/lm32/uclinux-elf.h: Ditto.
2809 * config/m68k/uclinux.h: Ditto.
2810 * config/moxie/uclinux.h: Ditto.
2811 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2812 (crisv32-*-linux*, cris-*-linux*): Ditto.
2813 * config/bfin/bfin.c: Include "tm_p.h".
2815 2013-09-05 Richard Biener <rguenther@suse.de>
2817 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2818 check for a definition without a basic-block.
2820 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2821 Sofiane Naci <sofiane.naci@arm.com>
2823 * config/aarch64/aarch64.md
2824 (*movti_aarch64): Rename r_2_f and f_2_r.
2825 (*movsf_aarch64): Likewise.
2826 (*movdf_aarch64): Likewise.
2827 (*movtf_aarch64): Likewise.
2828 (aarch64_movdi_<mode>low): Likewise.
2829 (aarch64_movdi_<mode>high): Likewise.
2830 (aarch64_mov<mode>high_di): Likewise.
2831 (aarch64_mov<mode>low_di): Likewise.
2832 (aarch64_movtilow_tilow): Likewise.
2833 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2834 values to config/arm/types.md
2835 (attribute "conds"): Update for attribute change.
2836 (anddi3_insn): Likewise.
2837 (iordi3_insn): Likewise.
2838 (xordi3_insn): Likewise.
2839 (one_cmpldi2): Likewise.
2840 * config/arm/types.md (type): Add Neon types.
2841 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2842 use "type" attribute.
2843 (movmisalign<mode>_neon_store): Likewise.
2844 (movmisalign<mode>_neon_load): Likewise.
2845 (vec_set<mode>_internal): Likewise.
2846 (vec_setv2di_internal): Likewise.
2847 (vec_extract<mode>): Likewise.
2848 (vec_extractv2di): Likewise.
2849 (add<mode>3_neon): Likewise.
2850 (adddi3_neon): Likewise.
2851 (sub<mode>3_neon): Likewise.
2852 (subdi3_neon): Likewise.
2853 (mul<mode>3_neon): Likewise.
2854 (mul<mode>3add<mode>_neon): Likewise.
2855 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
2856 (fma<VCVTF:mode>4)): Likewise.
2857 (fma<VCVTF:mode>4_intrinsic): Likewise.
2858 (fmsub<VCVTF:mode>4)): Likewise.
2859 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2860 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2861 (ior<mode>3): Likewise.
2862 (and<mode>3): Likewise.
2863 (anddi3_neon): Likewise.
2864 (orn<mode>3_neon): Likewise.
2865 (orndi3_neon): Likewise.
2866 (bic<mode>3_neon): Likewise.
2867 (bicdi3_neon): Likewise.
2868 (xor<mode>3): Likewise.
2869 (one_cmpl<mode>2): Likewise.
2870 (abs<mode>2): Likewise.
2871 (neg<mode>2): Likewise.
2872 (umin<mode>3_neon): Likewise.
2873 (umax<mode>3_neon): Likewise.
2874 (smin<mode>3_neon): Likewise.
2875 (smax<mode>3_neon): Likewise.
2876 (vashl<mode>3): Likewise.
2877 (vashr<mode>3_imm): Likewise.
2878 (vlshr<mode>3_imm): Likewise.
2879 (ashl<mode>3_signed): Likewise.
2880 (ashl<mode>3_unsigned): Likewise.
2881 (neon_load_count): Likewise.
2882 (ashldi3_neon_noclobber): Likewise.
2883 (signed_shift_di3_neon): Likewise.
2884 (unsigned_shift_di3_neon): Likewise.
2885 (ashrdi3_neon_imm_noclobber): Likewise.
2886 (lshrdi3_neon_imm_noclobber): Likewise.
2887 (widen_ssum<mode>3): Likewise.
2888 (widen_usum<mode>3): Likewise.
2889 (quad_halves_<code>v4si): Likewise.
2890 (quad_halves_<code>v4sf): Likewise.
2891 (quad_halves_<code>v8hi): Likewise.
2892 (quad_halves_<code>v16qi): Likewise.
2893 (reduc_splus_v2di): Likewise.
2894 (neon_vpadd_internal<mode>): Likewise.
2895 (neon_vpsmin<mode>): Likewise.
2896 (neon_vpsmax<mode>): Likewise.
2897 (neon_vpumin<mode>): Likewise.
2898 (neon_vpumax<mode>): Likewise.
2899 (ss_add<mode>_neon): Likewise.
2900 (us_add<mode>_neon): Likewise.
2901 (ss_sub<mode>_neon): Likewise.
2902 (us_sub<mode>_neon): Likewise.
2903 (neon_vadd<mode>_unspec): Likewise.
2904 (neon_vaddl<mode>): Likewise.
2905 (neon_vaddw<mode>): Likewise.
2906 (neon_vhadd<mode>): Likewise.
2907 (neon_vqadd<mode>): Likewise.
2908 (neon_vaddhn<mode>): Likewise.
2909 (neon_vmul<mode>): Likewise.
2910 (neon_vmla<mode>): Likewise.
2911 (neon_vmlal<mode>): Likewise.
2912 (neon_vmls<mode>): Likewise.
2913 (neon_vmlsl<mode>): Likewise.
2914 (neon_vqdmulh<mode>): Likewise.
2915 (neon_vqdmlal<mode>): Likewise.
2916 (neon_vqdmlsl<mode>): Likewise.
2917 (neon_vmull<mode>): Likewise.
2918 (neon_vqdmull<mode>): Likewise.
2919 (neon_vsub<mode>_unspec): Likewise.
2920 (neon_vsubl<mode>): Likewise.
2921 (neon_vsubw<mode>): Likewise.
2922 (neon_vqsub<mode>): Likewise.
2923 (neon_vhsub<mode>): Likewise.
2924 (neon_vsubhn<mode>): Likewise.
2925 (neon_vceq<mode>): Likewise.
2926 (neon_vcge<mode>): Likewise.
2927 (neon_vcgeu<mode>): Likewise.
2928 (neon_vcgt<mode>): Likewise.
2929 (neon_vcgtu<mode>): Likewise.
2930 (neon_vcle<mode>): Likewise.
2931 (neon_vclt<mode>): Likewise.
2932 (neon_vcage<mode>): Likewise.
2933 (neon_vcagt<mode>): Likewise.
2934 (neon_vtst<mode>): Likewise.
2935 (neon_vabd<mode>): Likewise.
2936 (neon_vabdl<mode>): Likewise.
2937 (neon_vaba<mode>): Likewise.
2938 (neon_vabal<mode>): Likewise.
2939 (neon_vmax<mode>): Likewise.
2940 (neon_vmin<mode>): Likewise.
2941 (neon_vpaddl<mode>): Likewise.
2942 (neon_vpadal<mode>): Likewise.
2943 (neon_vpmax<mode>): Likewise.
2944 (neon_vpmin<mode>): Likewise.
2945 (neon_vrecps<mode>): Likewise.
2946 (neon_vrsqrts<mode>): Likewise.
2947 (neon_vqabs<mode>): Likewise.
2948 (neon_vqneg<mode>): Likewise.
2949 (neon_vcls<mode>): Likewise.
2950 (clz<mode>2): Likewise.
2951 (popcount<mode>2): Likewise.
2952 (neon_vrecpe): Likewise.
2953 (neon_vrsqrte): Likewise.
2954 (neon_vget_lane<mode>_sext_internal): Likewise.
2955 (neon_vget_lane<mode>_zext_internal): Likewise.
2956 (neon_vdup_n<mode>): Likewise.
2957 (neon_vdup_nv2di): Likewise.
2958 (neon_vdpu_lane<mode>_internal): Likewise.
2959 (neon_vswp<mode>): Likewise.
2960 (float<mode><V_cvtto>2): Likewise.
2961 (floatuns<mode><V_cvtto>2): Likewise.
2962 (fix_trunc<mode><V_cvtto>)2): Likewise
2963 (fixuns_trunc<mode><V_cvtto)2): Likewise.
2964 (neon_vcvt<mode>): Likewise.
2965 (neon_vcvtv4sfv4hf): Likewise.
2966 (neon_vcvtv4hfv4sf): Likewise.
2967 (neon_vcvt_n<mode>): Likewise.
2968 (neon_vmovn<mode>): Likewise.
2969 (neon_vqmovn<mode>): Likewise.
2970 (neon_vqmovun<mode>): Likewise.
2971 (neon_vmovl<mode>): Likewise.
2972 (neon_vmul_lane<mode>): Likewise.
2973 (neon_vmull_lane<mode>): Likewise.
2974 (neon_vqdmull_lane<mode>): Likewise.
2975 (neon_vqdmulh_lane<mode>): Likewise.
2976 (neon_vmla_lane<mode>): Likewise.
2977 (neon_vmlal_lane<mode>): Likewise.
2978 (neon_vqdmlal_lane<mode>): Likewise.
2979 (neon_vmls_lane<mode>): Likewise.
2980 (neon_vmlsl_lane<mode>): Likewise.
2981 (neon_vqdmlsl_lane<mode>): Likewise.
2982 (neon_vext<mode>): Likewise.
2983 (neon_vrev64<mode>): Likewise.
2984 (neon_vrev32<mode>): Likewise.
2985 (neon_vrev16<mode>): Likewise.
2986 (neon_vbsl<mode>_internal): Likewise.
2987 (neon_vshl<mode>): Likewise.
2988 (neon_vqshl<mode>): Likewise.
2989 (neon_vshr_n<mode>): Likewise.
2990 (neon_vshrn_n<mode>): Likewise.
2991 (neon_vqshrn_n<mode>): Likewise.
2992 (neon_vqshrun_n<mode>): Likewise.
2993 (neon_vshl_n<mode>): Likewise.
2994 (neon_vqshl_n<mode>): Likewise.
2995 (neon_vqshlu_n<mode>): Likewise.
2996 (neon_vshll_n<mode>): Likewise.
2997 (neon_vsra_n<mode>): Likewise.
2998 (neon_vsri_n<mode>): Likewise.
2999 (neon_vsli_n<mode>): Likewise.
3000 (neon_vtbl1v8qi): Likewise.
3001 (neon_vtbl2v8qi): Likewise.
3002 (neon_vtbl3v8qi): Likewise.
3003 (neon_vtbl4v8qi): Likewise.
3004 (neon_vtbx1v8qi): Likewise.
3005 (neon_vtbx2v8qi): Likewise.
3006 (neon_vtbx3v8qi): Likewise.
3007 (neon_vtbx4v8qi): Likewise.
3008 (neon_vtrn<mode>_internal): Likewise.
3009 (neon_vzip<mode>_internal): Likewise.
3010 (neon_vuzp<mode>_internal): Likewise.
3011 (neon_vld1<mode>): Likewise.
3012 (neon_vld1_lane<mode>): Likewise.
3013 (neon_vld1_dup<mode>): Likewise.
3014 (neon_vld1_dupv2di): Likewise.
3015 (neon_vst1<mode>): Likewise.
3016 (neon_vst1_lane<mode>): Likewise.
3017 (neon_vld2<mode>): Likewise.
3018 (neon_vld2_lane<mode>): Likewise.
3019 (neon_vld2_dup<mode>): Likewise.
3020 (neon_vst2<mode>): Likewise.
3021 (neon_vst2_lane<mode>): Likewise.
3022 (neon_vld3<mode>): Likewise.
3023 (neon_vld3qa<mode>): Likewise.
3024 (neon_vld3qb<mode>): Likewise.
3025 (neon_vld3_lane<mode>): Likewise.
3026 (neon_vld3_dup<mode>): Likewise.
3027 (neon_vst3<mode>): Likewise.
3028 (neon_vst3qa<mode>): Likewise.
3029 (neon_vst3qb<mode>): Likewise.
3030 (neon_vst3_lane<mode>): Likewise.
3031 (neon_vld4<mode>): Likewise.
3032 (neon_vld4qa<mode>): Likewise.
3033 (neon_vld4qb<mode>): Likewise.
3034 (neon_vld4_lane<mode>): Likewise.
3035 (neon_vld4_dup<mode>): Likewise.
3036 (neon_vst4<mode>): Likewise.
3037 (neon_vst4qa<mode>): Likewise.
3038 (neon_vst4qb<mode>): Likewise.
3039 (neon_vst4_lane<mode>): Likewise.
3040 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3041 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3042 (neon_vec_<US>mult_lo_<mode>): Likewise.
3043 (neon_vec_<US>mult_hi_<mode>): Likewise.
3044 (neon_vec_<US>shiftl_<mode>): Likewise.
3045 (neon_unpack<US>_<mode>): Likewise.
3046 (neon_vec_<US>mult_<mode>): Likewise.
3047 (vec_pack_trunc_<mode>): Likewise.
3048 (neon_vec_pack_trunk_<mode>): Likewise.
3049 (neon_vabd<mode>_2): Likewise.
3050 (neon_vabd<mode>_3): Likewise.
3051 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3052 (thumb2_movsi_vfp): Likewise.
3053 (movdi_vfp): Likewise.
3054 (movdi_vfp_cortexa8): Likewise.
3055 (movhf_vfp_neon): Likewise.
3056 (movhf_vfp): Likewiwse.
3057 (movsf_vfp): Likewiwse.
3058 (thumb2_movsf_vfp): Likewiwse.
3059 (movdf_vfp): Likewise.
3060 (thumb2_movdf_vfp): Likewise.
3061 (movsfcc_vfp): Likewise.
3062 (thumb2_movsfcc_vfp): Likewise.
3063 (movdfcc_vfp): Likewise.
3064 (thumb2_movdfcc_vfp): Likewise.
3065 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3066 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3067 (v10_v2c): Likewise.
3068 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3070 (cortex_a15_neon_int_2): Likewise.
3071 (cortex_a15_neon_int_3): Likewise.
3072 (cortex_a15_neon_int_4): Likewise.
3073 (cortex_a15_neon_int_5): Likewise.
3074 (cortex_a15_neon_vqneg_vqabs): Likewise.
3075 (cortex_a15_neon_vmov): Likewise.
3076 (cortex_a15_neon_vaba): Likewise.
3077 (cortex_a15_neon_vaba_qqq): Likewise.
3078 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3079 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3080 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3081 scalar_64_32_long_scalar): Likewise.
3082 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3083 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3084 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3085 lotype_qdd_64_32_long): Likewise.
3086 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3087 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3088 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3089 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3090 (cortex_a15_neon_shift_1): Likewise.
3091 (cortex_a15_neon_shift_2): Likewise.
3092 (cortex_a15_neon_shift_3): Likewise.
3093 (cortex_a15_neon_vshl_ddd): Likewise.
3094 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3095 (cortex_a15_neon_vsra_vrsra): Likewise.
3096 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3097 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3098 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3099 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3100 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3101 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3102 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3103 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3104 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3105 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3106 (cortex_a15_neon_bp_simple): Likewise.
3107 (cortex_a15_neon_bp_2cycle): Likewise.
3108 (cortex_a15_neon_bp_3cycle): Likewise.
3109 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3110 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3111 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3112 (cortex_a15_neon_vld2_4_regs): Likewise.
3113 (cortex_a15_neon_vld3_vld4): Likewise.
3114 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3115 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3116 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3117 (cortex_a15_neon_vst3_vst4): Likewise.
3118 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3119 (cortex_a15_neon_vld3_vld4_lane" 10
3120 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3121 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3122 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3123 (cortex_a15_neon_ldm_2): Likewise.0
3124 (cortex_a15_neon_stm_2): Likewise.
3125 (cortex_a15_neon_mcr): Likewise.
3126 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3127 (cortex_a15_neon_mrc): Likewise.
3128 (cortex_a15_neon_mrrc): Likewise.
3129 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3131 (cortex_a15_alu_shift): Likewise.
3132 (cortex_a15_alu_shift_reg): Likewise.
3133 (cortex_a15_mult32): Likewise.
3134 (cortex_a15_mult64): Likewise.
3135 (cortex_a15_block): Likewise.
3136 (cortex_a15_branch): Likewise.
3137 (cortex_a15_load1): Likewise.
3138 (cortex_a15_load3): Likewise.
3139 (cortex_a15_store1): Likewise.
3140 (cortex_a15_store3): Likewise.
3141 (cortex_a15_call): Likewise.
3142 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3143 (cortex_a5_f2r): Likewise.
3144 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3146 (cortex_a53_f2r): Likewise.
3147 * config/arm/cortex-a7.md
3148 (cortex_a7_branch): Update for attribute change.
3149 (cortex_a7_call): Likewise.
3150 (cortex_a7_alu_imm): Likewise.
3151 (cortex_a7_alu_reg): Likewise.
3152 (cortex_a7_alu_shift): Likewise.
3153 (cortex_a7_mul): Likewise.
3154 (cortex_a7_load1): Likewise.
3155 (cortex_a7_store1): Likewise.
3156 (cortex_a7_load2): Likewise.
3157 (cortex_a7_store2): Likewise.
3158 (cortex_a7_load3): Likewise.
3159 (cortex_a7_store3): Likewise.
3160 (cortex_a7_load4): Likewise.
3161 (cortex_a7_store4): Likewise.
3162 (cortex_a7_fpalu): Likewise.
3163 (cortex_a7_fconst): Likewise.
3164 (cortex_a7_fpmuls): Likewise.
3165 (cortex_a7_neon_mul): Likewise.
3166 (cortex_a7_fpmacs): Likewise.
3167 (cortex_a7_neon_mla: Likewise.
3168 (cortex_a7_fpmuld: Likewise.
3169 (cortex_a7_fpmacd: Likewise.
3170 (cortex_a7_fpfmad: Likewise.
3171 (cortex_a7_fdivs: Likewise.
3172 (cortex_a7_fdivd: Likewise.
3173 (cortex_a7_r2f: Likewise.
3174 (cortex_a7_f2r: Likewise.
3175 (cortex_a7_f_flags: Likewise.
3176 (cortex_a7_f_loads: Likewise.
3177 (cortex_a7_f_loadd: Likewise.
3178 (cortex_a7_f_stores: Likewise.
3179 (cortex_a7_f_stored: Likewise.
3180 (cortex_a7_neon): Likewise.
3181 * config/arm/cortex-a8-neon.md
3182 (cortex_a8_neon_mrc): Update for attribute change.
3183 (cortex_a8_neon_mrrc): Likewise.
3184 (cortex_a8_neon_int_1): Likewise.
3185 (cortex_a8_neon_int_2): Likewise.
3186 (cortex_a8_neon_int_3): Likewise.
3187 (cortex_a8_neon_int_4): Likewise.
3188 (cortex_a8_neon_int_5): Likewise.
3189 (cortex_a8_neon_vqneg_vqabs): Likewise.
3190 (cortex_a8_neon_vmov): Likewise.
3191 (cortex_a8_neon_vaba): Likewise.
3192 (cortex_a8_neon_vaba_qqq): Likewise.
3193 (cortex_a8_neon_vsma): Likewise.
3194 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3195 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3196 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3197 long_scalar): Likewise.
3198 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3199 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3200 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3201 long_scalar_qdd_64_32_long): Likewise.
3202 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3203 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3204 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3205 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3206 (cortex_a8_neon_shift_1): Likewise.
3207 (cortex_a8_neon_shift_2): Likewise.
3208 (cortex_a8_neon_shift_3): Likewise.
3209 (cortex_a8_neon_vshl_ddd): Likewise.
3210 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3211 (cortex_a8_neon_vsra_vrsra): Likewise.
3212 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3213 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3214 (cortex_a8_neon_fp_vsum): Likewise.
3215 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3216 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3217 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3218 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3219 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3220 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3221 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3222 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3223 (cortex_a8_neon_bp_simple): Likewise.
3224 (cortex_a8_neon_bp_2cycle): Likewise.
3225 (cortex_a8_neon_bp_3cycle): Likewise.
3226 (cortex_a8_neon_ldr): Likewise.
3227 (cortex_a8_neon_str): Likewise.
3228 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3229 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3230 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3231 (cortex_a8_neon_vld2_4_regs): Likewise.
3232 (cortex_a8_neon_vld3_vld4): Likewise.
3233 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3234 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3235 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3236 (cortex_a8_neon_vst3_vst4): Likewise.
3237 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3238 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3239 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3240 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3241 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3242 (cortex_a8_neon_mcr): Likewise.
3243 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3244 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3245 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3247 (ca9_neon_mrrc): Likewise.
3248 (cortex_a9_neon_int_1): Likewise.
3249 (cortex_a9_neon_int_2): Likewise.
3250 (cortex_a9_neon_int_3): Likewise.
3251 (cortex_a9_neon_int_4): Likewise.
3252 (cortex_a9_neon_int_5): Likewise.
3253 (cortex_a9_neon_vqneg_vqabs): Likewise.
3254 (cortex_a9_neon_vmov): Likewise.
3255 (cortex_a9_neon_vaba): Likewise.
3256 (cortex_a9_neon_vaba_qqq): Likewise.
3257 (cortex_a9_neon_vsma): Likewise.
3258 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3259 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3260 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3261 long_scalar): Likewise.
3262 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3263 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3264 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3265 long_scalar_qdd_64_32_long): Likewise.
3266 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3267 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3268 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3269 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3270 (cortex_a9_neon_shift_1): Likewise.
3271 (cortex_a9_neon_shift_2): Likewise.
3272 (cortex_a9_neon_shift_3): Likewise.
3273 (cortex_a9_neon_vshl_ddd): Likewise.
3274 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3275 (cortex_a9_neon_vsra_vrsra): Likewise.
3276 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3277 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3278 (cortex_a9_neon_fp_vsum): Likewise.
3279 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3280 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3281 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3282 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3283 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3284 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3285 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3286 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3287 (cortex_a9_neon_bp_simple): Likewise.
3288 (cortex_a9_neon_bp_2cycle): Likewise.
3289 (cortex_a9_neon_bp_3cycle): Likewise.
3290 (cortex_a9_neon_ldr): Likewise.
3291 (cortex_a9_neon_str): Likewise.
3292 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3293 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3294 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3295 (cortex_a9_neon_vld2_4_regs): Likewise.
3296 (cortex_a9_neon_vld3_vld4): Likewise.
3297 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3298 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3299 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3300 (cortex_a9_neon_vst3_vst4): Likewise.
3301 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3302 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3303 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3304 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3305 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3306 (cortex_a9_neon_mcr): Likewise.
3307 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3308 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3309 (cortex_a9_fps): Likewise.
3310 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3312 (cortex_m4_fmuls): Likewise.
3313 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3315 (cortex_r4_mrc): Likewise.
3316 * config/arm/iterators.md: Update comment referring to neon_type.
3317 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3318 (iwmmxt_movsi_insn): Likewise.
3319 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3321 (pj4_core_to_vfp): Likewise.
3322 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3324 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3325 (vfp_fstore): Likewise.
3326 * doc/md.texi: Change references to neon_type to refer to type.
3328 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3330 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3332 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3334 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3337 2013-09-04 Easwaran Raman <eraman@google.com>
3340 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3341 (build_and_add_sum): Use it.
3342 (appears_later_in_bb): Simplify code.
3344 2013-09-04 Teresa Johnson <tejohnson@google.com>
3346 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3348 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3350 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3352 2013-09-04 Jan Hubicka <jh@suse.cz>
3354 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3355 * ipa-devirt.c: Include diganostic.h
3356 (odr_type_d): Add types and types_set.
3357 (hash_type_name): Work for types with vtables during LTO.
3358 (odr_hasher::remove): Fix comment; destroy types_set.
3359 (add_type_duplicate): New function,
3360 (get_odr_type): Use it.
3361 (dump_type_inheritance_graph): Dump type duplicates.
3362 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3364 * tree.c (types_same_for_odr): Give exact answers on types with
3367 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3369 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3370 explaining their differences.
3372 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3374 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3376 2013-09-03 Jeff Law <law@redhat.com>
3378 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3379 when not threading through a joiner block. Pass joiner/no joiner
3380 state to register_jump_thread.
3381 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3382 state from argument rather than implying on path length.
3383 Dump the entire jump thread path into debugging dump.
3384 * tree-flow.h (register_jump_thread): Update prototype.
3386 2013-08-29 Xinliang David Li <davidxl@google.com>
3388 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3389 Remove a trivial gcc_assert.
3391 2013-08-29 Xinliang David Li <davidxl@google.com>
3393 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3394 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3395 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3396 Delay base decl alignment adjustment.
3397 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3398 * tree-vectorizer.h: New data structure.
3399 (set_dr_misalignment): New function.
3400 (dr_misalignment): Ditto.
3401 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3402 (vectorizable_load): Ditto.
3403 (ensure_base_align): New function.
3404 (vectorize_loops): Add dbg_cnt support.
3405 (execute_vect_slp): Ditto.
3406 * dbgcnt.def: New debug counter.
3407 * Makefile: New dependency.
3409 2013-09-03 Meador Inge <meadori@codesourcery.com>
3413 2013-08-30 Meador Inge <meadori@codesourcery.com>
3415 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3417 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3419 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3420 function descriptor.
3422 2013-09-03 Richard Biener <rguenther@suse.de>
3424 * tree-affine.c (add_elt_to_tree): Fix association issue,
3425 avoid useless converts and make sure to always return a
3426 properly typed result.
3428 2013-09-03 Richard Biener <rguenther@suse.de>
3431 * fold-const.c (negate_expr_p): Fix division case.
3432 (negate_expr): Likewise.
3434 2013-09-03 Richard Biener <rguenther@suse.de>
3437 * tree-streamer-out.c: Include tm.h.
3438 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3440 2013-09-03 Jan Hubicka <jh@suse.cz>
3442 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3444 2013-09-03 Alan Modra <amodra@gmail.com>
3446 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3447 multilibs to big-endian target and vice versa.
3448 * config/rs6000/t-linux64: Use := assignment on all vars.
3449 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3450 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3451 * config/rs6000/t-linux64le: New file.
3452 * config/rs6000/t-linux64bele: New file.
3453 * config/rs6000/t-linux64lebe: New file.
3455 2013-09-02 Jan Hubicka <jh@suse.cz>
3457 * ipa-inline-transform.c (inline_transform): Do not
3458 optimize_inline_calls when not optimizing.
3460 2013-09-02 Jan Hubicka <jh@suse.cz>
3462 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3463 duplicated nodes for assembler names.
3464 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3465 hard registers from assembler name hash.
3467 2013-09-02 Jan Hubicka <jh@suse.cz>
3469 * ipa-split.c (execute_split_functions): Split externally visible
3470 functions called once.
3472 2013-09-02 Martin Jambor <mjambor@suse.cz>
3475 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3476 linked list. When finding the correct duplicate, also consider also
3477 the caller in additon to its inlined_to node.
3479 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3481 * config/aarch64/aarch64-simd-builtins.def
3482 (dup_lane_scalar): Remove.
3483 * config/aarch64/aarch64-simd.md
3484 (aarch64_simd_dup): Add 'w->w' alternative.
3485 (aarch64_dup_lane<mode>): Allow for VALL.
3486 (aarch64_dup_lane_scalar<mode>): Remove.
3487 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3488 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3489 (aarch64_get_lane_unsigned<mode>): Likewise.
3490 (aarch64_get_lane<mode>): Likewise.
3491 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3492 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3493 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3494 (VCON): Change container of V2SF.
3495 (vswap_width_name): Likewise.
3496 * config/aarch64/arm_neon.h
3497 (__aarch64_vdup_lane_any): New.
3498 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3499 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3500 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3502 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3505 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3506 the source or the destination is a single hard register.
3508 2013-09-02 Richard Biener <rguenther@suse.de>
3511 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3514 2013-09-02 Richard Biener <rguenther@suse.de>
3516 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3517 arithmetic to sizetype.
3519 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3521 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3522 Find auto-increment use both before and after candidate.
3524 2013-09-02 Marek Polacek <polacek@redhat.com>
3526 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3528 2013-09-01 Jan Hubicka <jh@suse.cz>
3530 * Makefile.in: Add ipa-profile.o
3531 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3532 * cgraph.c (struct cgraph_propagate_frequency_data,
3533 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3534 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3535 * cgraph.h (cgraph_propagate_frequency): Remove.
3536 * ipa-inline-analysis.c: Include ipa-utils.h;
3537 drop duplicated cfgloop.h.
3538 (inline_update_callee_summaries): Update.
3539 * ipa-profile.c: New file.
3540 * ipa-utils.h (ipa_propagate_frequency): Declare.
3541 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3542 data-streamer.h, value-prof.h.
3543 (symtab_remove_unreachable_nodes): Update profile.
3544 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3545 account_time_size, cmp_counts, dump_histogram,
3546 ipa_profile_generate_summary, ipa_profile_write_summary,
3547 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3548 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3549 Move to ipa-profile.c.
3551 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3553 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3555 2013-09-01 Jan Hubicka <jh@suse.cz>
3557 * common.opt (fdevirtualize-speculatively): New function.
3558 * invoke.texi (fdevirtualize-speculatively): Document.
3559 * ipa-devirt.c: Include ipa-inline.h
3560 (likely_target_p): New function.
3561 (ipa_devirt): New function.
3562 (gate_ipa_devirt): New function.
3563 (pass_data_ipa_devirt): New static var.
3564 (pass_ipa_devirt): Likewise.
3565 (make_pass_ipa_devirt): New function.
3566 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3567 (common_handle_option): Disable devirtualization when
3568 value range profiling is available.
3569 * passes.def (pass_ipa_devirt): Add.
3570 * timever.def (TV_IPA_DEVIRT): New timevar.
3571 * tree-pass.h (make_pass_ipa_devirt):
3573 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3575 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3576 include sanitize(undefined).
3578 2013-08-31 Diego Novillo <dnovillo@google.com>
3580 * Makefile.in (TREE_CORE_H): Define.
3582 (GTFILES): Add tree-core.h.
3583 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3585 * tree-core.h: New file.
3586 Move all data structures, enum, typedefs, global
3587 declarations and constants from ...
3590 2013-08-31 Jan Hubicka <jh@suse.cz>
3592 * bulitins.c (expand_builtin): Do not early exit for gcov
3593 instrumented functions.
3595 2013-08-31 Marek Polacek <polacek@redhat.com>
3597 * ubsan.c: Include tm_p.h.
3599 2013-08-31 Jan Hubicka <jh@suse.cz>
3601 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3604 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3605 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3606 * tree-cfg.c (verify_gimple_label): ... here.
3607 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3608 (ipa_merge_profiles): New function.
3609 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3610 (lto_input_function_body): Likewise.
3611 * ipa-utils.h (ipa_merge_profiles): Declare.
3612 * lto-streamer.h (lto_input_function_body): Update prototype.
3613 (emit_label_in_global_context_p): Remove.
3614 * lto-symtab.c: Include ipa-utils.h
3615 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3617 2013-08-31 Jan Hubicka <jh@suse.cz>
3619 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3621 2013-08-31 Jan Hubicka <jh@suse.cz>
3623 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3625 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3627 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3628 "cmp" RTX before signed_comparison_operator check to account
3631 2013-08-30 Jan Hubicka <jh@suse.cz>
3633 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3634 (detect_type_change_1): Rename to ...
3635 (detect_type_change): ... this one; early return on non-polymorphic
3637 (detect_type_change_ssa): Add comp_type parameter; update
3638 use of detect_type_change.
3639 (compute_complex_assign_jump_func): Add param_type parameter;
3640 update use of detect_type_change_ssa.
3641 (compute_complex_ancestor_jump_func): Likewise.
3642 (ipa_get_callee_param_type): New function.
3643 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3644 update calls to the jump function computation functions.
3646 2013-08-30 Teresa Johnson <tejohnson@google.com>
3647 Steven Bosscher <steven@gcc.gnu.org>
3649 * cfgrtl.c (fixup_new_cold_bb): New routine.
3650 (commit_edge_insertions): Invoke fixup_partitions.
3651 (find_partition_fixes): New routine.
3652 (fixup_partitions): Ditto.
3653 (verify_hot_cold_block_grouping): Update comments.
3654 (rtl_verify_edges): Invoke find_partition_fixes.
3655 (rtl_verify_bb_pointers): Update comments.
3656 (rtl_verify_bb_layout): Ditto.
3657 * basic-block.h (probably_never_executed_edge_p): Declare.
3658 (fixup_partitions): Ditto.
3659 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3660 * bb-reorder.c (sanitize_hot_paths): New function.
3661 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3663 * predict.c (probably_never_executed_edge_p): New routine.
3664 * cfg.c (check_bb_profile): Add partition insanity warnings.
3666 2013-08-30 Meador Inge <meadori@codesourcery.com>
3668 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3670 2013-08-30 Marek Polacek <polacek@redhat.com>
3672 * Makefile.in (ubsan.o): Add.
3673 (c-family/c-ubsan.o): Add.
3674 (builtins.o): Add ubsan.h dependency.
3675 * ubsan.h: New file.
3676 * ubsan.c: New file.
3677 * common.opt: Add -fsanitize=undefined option.
3678 (flag_sanitize): Add variable.
3679 (fsanitize=): Add option. Add Driver.
3680 (fsanitize=thread): Remove option.
3681 (fsanitize=address): Likewise.
3682 (static-libubsan): New option.
3683 * doc/invoke.texi: Document the new flag and -static-libubsan.
3684 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3685 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3686 * builtin-attrs.def (ATTR_COLD): Define.
3687 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3688 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3689 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3690 * flag-types.h (sanitize_code): New enum.
3691 * opts.c (common_handle_option): Parse command line arguments
3692 of -fsanitize=. Add -fsanitize=unreachable option.
3693 * varasm.c (get_variable_section): Adjust.
3694 (assemble_noswitch_variable): Likewise.
3695 (assemble_variable): Likewise.
3696 (output_constant_def_contents): Likewise.
3697 (categorize_decl_for_section): Likewise.
3698 (place_block_symbol): Likewise.
3699 (output_object_block): Likewise.
3700 * builtins.def: Likewise.
3701 * toplev.c (compile_file): Likewise.
3702 (process_options): Likewise.
3703 * cppbuiltin.c: Likewise.
3704 * tsan.c (tsan_pass): Likewise.
3705 (tsan_gate): Likewise.
3706 (tsan_gate_O0): Likewise.
3707 * cfgexpand.c (partition_stack_vars): Likewise.
3708 (expand_stack_vars): Likewise.
3709 (defer_stack_allocation): Likewise.
3710 (expand_used_vars): Likewise.
3711 * cfgcleanup.c (old_insns_match_p): Likewise.
3712 * asan.c (asan_finish_file): Likewise.
3713 (asan_instrument): Likewise.
3714 (gate_asan): Likewise.
3715 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3716 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3717 (asan_global_struct): Use pointer_sized_int_node instead
3718 calling build_nonstandard_integer_type.
3719 (initialize_sanitizer_builtins): Likewise.
3720 (asan_finish_file): Likewise.
3721 * gcc.c: Document %{%:function(args):X}.
3722 (static_spec_functions): Add sanitize.
3723 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3724 store funcval != NULL there.
3725 (do_spec_1): Adjust handle_spec_function caller.
3726 (handle_braces): Allow %:function(args) as condition.
3727 (sanitize_spec_function): New function.
3728 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3729 (LIBUBSAN_SPEC): Likewise.
3730 (LIBUBSAN_EARLY_SPEC): Likewise.
3731 (SANITIZER_SPEC): Handle libubsan.
3732 (SANITIZER_EARLY_SPEC): Likewise.
3733 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3734 instead of fsanitize=address.
3735 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3736 instead of fsanitize=address*.
3737 * builtins.c: Include ubsan.h.
3738 (fold_builtin_0): Instrument __builtin_unreachable.
3739 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3740 instead of flag_asan.
3741 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3742 (pointer_sized_int_node): Define.
3743 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3745 2013-08-30 Mike Stump <mikestump@comcast.net>
3747 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3750 2013-08-29 Jan Hubicka <jh@suse.cz>
3752 * cgraph.c (cgraph_function_body_availability): Handle weakref
3754 * passes.def: Remove pass_fixup_cfg.
3755 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3756 track when we need to remove functions.
3757 (gate_ipa_inline): Execute inlining always; add comment why.
3758 (pass_data_ipa_inline): Remove TODO_remove_functions.
3759 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3760 do not produce summaries.
3761 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3762 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3763 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3764 constant pool and vtable.
3766 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3768 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3769 New arm_neon.h's internal macros to specify 64-bit constants.
3770 Avoid using stdint.h's macros.
3772 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3774 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3775 before applying REG_P.
3777 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3779 PR tree-optimization/58277
3780 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3781 after seeing too many stmts with vdef in between dombb and current
3782 bb, invalidate everything.
3784 2013-08-30 Richard Biener <rguenther@suse.de>
3786 * fold-const.c (fold_single_bit_test): Fix overflow test.
3788 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3790 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3791 and which can live in a register, always retrieve the value on entry.
3792 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3793 passed by invisible reference specially.
3794 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3795 (vt_add_function_parameter): Correctly deal with a parameter passed by
3796 invisible reference.
3798 2013-08-30 Jan Hubicka <jh@suse.cz>
3800 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3802 2013-08-30 Richard Biener <rguenther@suse.de>
3804 PR tree-optimization/58228
3805 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3806 allow invariant loads in nested loop vectorization.
3808 2013-08-30 Richard Biener <rguenther@suse.de>
3810 PR tree-optimization/58223
3811 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3812 (has_anti_or_output_dependence): ... this and adjust to also
3813 look for output dependences.
3814 (mark_nodes_having_upstream_mem_writes): Adjust.
3815 (rdg_flag_uses): Likewise.
3817 2013-08-30 Richard Biener <rguenther@suse.de>
3819 PR tree-optimization/58010
3820 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3821 assert that we have a loop-closed PHI.
3823 2013-08-29 Jan Hubicka <jh@suse.cz>
3825 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3826 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3827 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3828 (lto_free_function_in_decl_state_for_node): New function.
3830 2013-08-29 Xinliang David Li <davidxl@google.com>
3832 * loop-unroll.c (report_unroll_peel): Minor message change.
3833 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3834 Emit alignment peeling message with default -fopt-info.
3835 (vect_loop_versioning): Emit loop version info message.
3836 * tree-vectorizer.c (vectorize_loops): Minor message change.
3837 (execute_vect_slp): Ditto.
3839 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3841 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3842 of the clone from the DECL_NAME of the original function.
3844 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3846 * passes.c (register_pass): Add overload.
3847 * tree-pass.h (register_pass): Forward declare it. Add comment.
3849 2013-08-29 Jan Hubicka <jh@suse.cz>
3851 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
3852 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
3853 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
3854 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3855 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
3856 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
3857 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3858 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
3860 2013-08-29 Teresa Johnson <tejohnson@google.com>
3862 * dumpfile.c (dump_loc): Output column number.
3863 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
3864 * doc/invoke.texi: Document optall -fopt-info flag.
3865 * profile.c (read_profile_edge_counts): Use new dump framework.
3866 (compute_branch_probabilities): Ditto.
3867 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
3868 when pass not in any opt group.
3869 * pass_manager.h (pass_manager::get_pass_profile): New method.
3870 * value-prof.c (check_counter): Use new dump framework.
3871 (check_ic_target): Ditto.
3872 * coverage.c (get_coverage_counts): Ditto.
3873 (coverage_init): Setup new dump framework.
3875 2013-08-29 Richard Biener <rguenther@suse.de>
3877 PR tree-optimization/58246
3878 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
3879 handle the dominance check inside a basic-block.
3881 2013-08-29 Richard Biener <rguenther@suse.de>
3884 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
3885 of default defs that appear in abnormal PHI nodes.
3887 2013-08-29 Richard Biener <rguenther@suse.de>
3889 PR tree-optimization/57685
3890 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
3891 single-use operands to avoid exponential complexity.
3893 2013-08-28 Dehao Chen <dehao@google.com>
3895 * ipa-inline.c (edge_badness): Fix integer underflow.
3897 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
3899 * gtm-builtins.def (_ITM_free): Declare leaf.
3901 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3904 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
3905 (*tls_local_dynamic_base_64_largepic): Likewise.
3906 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
3907 Remove predicate from call operand.
3908 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
3909 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
3911 2013-08-28 Jeff Law <law@redhat.com>
3913 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
3914 checks for the number of predecessors and successors allowed.
3915 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
3916 which require copying a joiner block if there is a request which
3917 is a subpath that requires no joiner block copying.
3919 2013-08-28 Jan Hubicka <jh@suse.cz>
3921 * lto-streamer-out.c (DFS_write_tree_body): Drop
3922 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
3923 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
3924 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
3925 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
3926 Do not read DECL_ERROR_ISSUED.
3927 (unpack_ts_decl_with_vis_value_fields): Do not read
3929 (lto_input_ts_binfo_tree_pointers): Do not read
3930 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
3931 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
3932 write DECL_ERROR_ISSUED..
3933 (pack_ts_decl_with_vis_value_fields): Do not write
3935 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
3936 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
3937 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
3938 * tree.h (tree_decl_common): Update comment.
3939 (DECL_ERROR_ISSUED): Remove.
3941 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3944 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
3946 2013-08-28 Jan Hubicka <jh@suse.cz>
3948 * builtins.def (free): Declare leaf.
3950 2013-08-27 David Malcolm <dmalcolm@redhat.com>
3953 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
3954 * configure: Regenerate.
3956 2013-08-27 Martin Jambor <mjambor@suse.cz>
3958 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
3959 (ipa_ancestor_jf_data): Likewise.
3960 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
3961 (ipa_get_jf_pass_through_type_preserved): New function.
3962 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
3963 (ipa_get_jf_ancestor_type_preserved): New function.
3964 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
3965 (ipa_get_jf_ancestor_result): Likewise.
3966 (propagate_vals_accross_pass_through): Use
3967 ipa_get_jf_pass_through_result to do all the value mappings.
3968 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
3969 type_preserved flag.
3970 (ipa_set_jf_cst_copy): New function.
3971 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
3972 (ipa_set_jf_arith_pass_through): Likewise.
3973 (ipa_set_ancestor_jf): Likewise.
3974 (compute_complex_assign_jump_func): Set type_preserved instead of
3976 (ipa_compute_jump_functions_for_edge): Likewise.
3977 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
3978 (update_jump_functions_after_inlining): Update type_preserved.
3979 Explicitely create jump functions when combining one with pass_through.
3980 (ipa_write_jump_function): Stream the type_preserved flags.
3981 (ipa_read_jump_function): Likewise.
3983 2013-08-27 Jakub Jelinek <jakub@redhat.com>
3984 Aldy Hernandez <aldyh@redhat.com>
3986 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
3987 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
3988 * function.h (struct function): Add has_force_vect_loops and
3990 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
3991 * gimple.c (gimple_build_omp_critical): Add KIND argument and
3993 * gimple.def: Update CLAUSES comments.
3994 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
3995 (gimple_build_omp_for): Add argument to prototype.
3996 (gimple_omp_for_kind): New.
3997 (gimple_omp_for_set_kind): New.
3998 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
3999 GOVD_DATA_SHARE_CLASS.
4000 (enum omp_region_type): Add ORT_SIMD.
4001 (gimple_add_tmp_var): Handle ORT_SIMD.
4002 (gimplify_var_or_parm_decl): Same.
4003 (is_gimple_stmt): Same.
4004 (omp_firstprivatize_variable): Same.
4005 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4006 (omp_notice_variable): Handle ORT_SIMD.
4007 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4008 (omp_check_private): Handle ORT_SIMD.
4009 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4011 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4012 Handle OMP_CLAUSE_LASTPRIVATE.
4013 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4015 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4016 (gimplify_expr): Handle OMP_SIMD.
4017 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4018 (expand_GOMP_SIMD_VF): New.
4019 (expand_GOMP_SIMD_LAST_LANE): New.
4020 * internal-fn.def (GOMP_SIMD_LANE): New.
4021 (GOMP_SIMD_VF): New.
4022 (GOMP_SIMD_LAST_LANE): New.
4023 * omp-low.c: Include target.h.
4024 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4026 (check_omp_nesting_restrictions): Same.
4028 (lower_rec_simd_input_clauses): New.
4029 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4031 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4032 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4033 (expand_omp_build_assign): New.
4034 (expand_omp_for_init_counts): New.
4035 (expand_omp_for_init_vars): New.
4036 (extract_omp_for_update_vars): New.
4037 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4038 and rewrite accordingly.
4039 (expand_omp_simd): New.
4040 (expand_omp_for): Use expand_omp_simd.
4041 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4042 (lower_omp_for): Do not lower the body.
4043 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4045 * tree-flow.h (find_omp_clause): Remove prototype.
4046 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4047 forcing vectorization of the loop, or if flag_tree_vectorize.
4048 (gate_tree_if_conversion): Similarly.
4049 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4050 gimple_build_omp_for.
4051 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4052 * tree-parloops (create_parallel_loop): Pass kind argument to
4053 gimple_build_omp_for.
4054 * tree-pretty-print.c (dump_omp_clause): Add cases for
4055 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4056 OMP_CLAUSE__SIMDUID_.
4057 (dump_generic_node): Handle OMP_SIMD.
4058 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4059 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4060 unroll OMP_SIMD loops here.
4061 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4062 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4064 (vect_analyze_data_refs): Handle simd loops.
4065 * tree-vect-loop.c (vectorizable_live_operation): Handle
4067 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4068 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4069 (vectorizable_load): Same.
4070 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4071 (struct simduid_to_vf): New.
4072 (simduid_to_vf::hash): New.
4073 (simduid_to-vf::equal): New.
4074 (struct simd_array_to_simduid): New.
4075 (simd_array_to_simduid::hash): New.
4076 (simd_array_to_simduid::equal): New.
4077 (adjust_simduid_builtins): New.
4078 (struct note_simd_array_uses_struct): New.
4079 (note_simd_array_uses_cb): New.
4080 (note_simd_array_uses): New.
4081 (vectorize_loops): Handle simd hints and adjust simd builtins
4083 * tree-vectorizer.h (struct _stmt_vec_info): Add
4084 simd_lane_access_p field.
4085 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4086 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4087 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4088 (omp_clause_code_name): Same.
4089 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4090 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4091 * tree.def (OMP_SIMD): New entry.
4092 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4093 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4094 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4095 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4096 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4097 (OMP_CLAUSE_LINEAR_STEP): New.
4098 (OMP_CLAUSE_SAFELEN_EXPR): New.
4099 (OMP_CLAUSE__SIMDUID__DECL): New.
4100 (find_omp_clause): New prototype.
4102 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4104 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4105 Haswell processor detection.
4107 2013-08-27 Christian Widmer <shadow@umbrox.de>
4110 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4111 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4112 AVX2 capable processors.
4114 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4116 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4117 of vget_low_* with implementations in terms of other intrinsics.
4119 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4122 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4123 values to -1, 0 and 1.
4125 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4127 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4128 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4129 (<optab><mode>3_insn): Remove.
4130 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4131 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4132 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4133 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4134 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4135 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4136 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4137 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4138 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4139 (ror<mode>3_insn): Likewise.
4140 * config/aarch64/predicates.md (aarch64_simd_register): New.
4142 2013-08-27 Richard Biener <rguenther@suse.de>
4144 PR tree-optimization/57521
4145 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4146 one edge is non-critical.
4147 (find_phi_replacement_condition): Make sure to use a non-critical
4148 edge. Cleanup and remove old bug workarounds.
4149 (bb_postdominates_preds): Remove.
4150 (if_convertible_loop_p_1): Do not compute post-dominators.
4151 (combine_blocks): Do not free post-dominators.
4152 (main_tree_if_conversion): Likewise.
4153 (pass_data_if_conversion): Add TODO_verify_ssa.
4155 2013-08-27 DJ Delorie <dj@redhat.com>
4157 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4159 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4161 * function.c (assign_parm_find_data_types): Set passed_mode and
4162 nominal_mode to the TYPE_MODE of nominal_type for the built
4163 pointer type in case of the struct-pass-by-reference.
4165 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4167 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4168 (UINT16_TYPE): Change default to "unsigned int".
4170 * config/avr/avr.opt (mfract-convert-truncate): New option.
4171 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4172 is set, round negative fractional integers according to n1169
4173 when converting to integer types.
4175 2013-08-26 Jan Hubicka <jh@suse.cz>
4177 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4178 methods can not be called indirectly when their address is not taken.
4180 2013-08-26 Jan Hubicka <jh@suse.cz>
4182 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4185 2013-08-26 Jan Hubicka <jh@suse.cz>
4187 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4190 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4192 * reload.c (find_valid_class): Allow classes that do not include
4193 FIRST_PSEUDO_REGISTER - 1.
4195 2013-08-26 Jan Hubicka <jh@suse.cz>
4197 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4198 fix edge count/frequency when speculation failed; fix type check
4199 for the direct call.
4201 2013-08-26 Jan Hubicka <jh@suse.cz>
4203 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4205 2013-08-26 Jan Hubicka <jh@suse.cz>
4207 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4208 to be changed by edge redirection.
4210 2013-08-26 Jan Hubicka <jh@suse.cz>
4212 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4213 formating; add sanity check.
4214 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4215 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4216 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4217 (resolve_noninline_speculation): Update callee keys, too.
4219 2013-08-26 Jan Hubicka <jh@suse.cz>
4221 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4222 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4224 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4226 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4229 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4231 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4232 (x86_64_elf_select_section): Ditto.
4233 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4235 (ix86_pass_by_reference): Ditto.
4236 (output_set_got): Ditto.
4237 (ix86_unary_operator_ok): Ditto.
4238 (ix86_expand_builtin): Ditto.
4240 2013-08-23 Jan Hubicka <jh@suse.cz>
4242 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4244 2013-08-23 Jan Hubicka <jh@suse.cz>
4246 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4247 (tree_decl_with_vis): Add FINAL field.
4249 2013-08-23 Jeff Law <law@redhat.com>
4251 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4252 the debugging dump when the expression is fully redundant.
4254 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4256 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4257 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4258 * pretty-print.c (pp_formatted_text_data): Likewise.
4259 (pp_write_text_to_stream): Likewise.
4260 (pp_write_text_as_dot_label_to_stream): Likewise.
4261 (pp_append_r): Likewise.
4262 (pp_format): Likewise.
4263 (pp_flush): Likewise.
4264 (pp_clear_output_area): Likewise.
4265 (pp_append_text): Likewise.
4266 (pp_formatted_text): Likewise.
4267 (pp_remaining_character_count_for_line): Likewise.
4268 (pp_newline): Likewise.
4269 (pp_character): Likewise.
4270 (output_buffer::~output_buffer): Define.
4271 (pretty_printer::~pretty_printer): Destruct output buffer.
4272 * pretty-print.h (output_buffer::~output_buffer): Declare.
4273 (pretty_printer::~pretty_printer): Declare virtual.
4275 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4278 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4279 HOST_WIDE_INT_M1U): New macros.
4280 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4281 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4282 unsigned -1 for lshift.
4283 * cse.c (cse_insn): Likewise.
4284 * double-int.c (rshift_double, lshift_double): Likewise.
4285 * builtins.c (fold_builtin_bitop): Likewise.
4286 * combine.c (force_to_mode): Likewise.
4287 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4288 * simplify-rtx.c (simplify_const_unary_operation,
4289 simplify_const_binary_operation): Likewise.
4290 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4291 check_va_list_escapes): Likewise.
4292 * rtlanal.c (nonzero_bits1): Likewise.
4293 * expmed.c (expand_smod_pow2): Likewise.
4294 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4296 2013-08-23 Jan Hubicka <jh@suse.cz>
4298 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4299 as having address taken.
4301 2013-08-23 Jan Hubicka <jh@suse.cz>
4303 * ipa-utils.h (method_class_type): Declare.
4304 * ipa-devirt.c (method_class_type): Export.
4306 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4307 do not walk base classes of anonymous types.
4309 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4311 PR rtl-optimization/58220
4313 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4314 handle SEQUENCE insns properly.
4316 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4318 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4320 (pp_newline_and_indent): Likewise.
4321 (pp_separate_with): Likewise.
4322 * pretty-print.c (pp_newline_and_flush): Define.
4323 (pp_newline_and_indent): Likewise.
4324 (pp_separate_with): Likewise.
4326 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4329 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4330 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4332 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4334 * config/i386/predicates.md (ext_sse_reg_operand): New.
4335 * config/i386/i386.md (*movti_internal): Use
4336 predicate to determine if EVEX is needed.
4337 (*movsi_internal): Ditto.
4338 (*movdf_internal): Ditto.
4339 (*movsf_internal): Ditto.
4340 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4342 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4344 PR tree-optimization/58209
4345 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4346 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4347 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4348 emit POINTER_PLUS_EXPR.
4349 (create_tailcall_accumulator): For pointer result type accumulate in
4352 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4354 * configure.ac: Add backslashes missing from the last change.
4355 * configure: Regenerate.
4357 2013-08-22 Jan Hubicka <jh@susue.cz>
4359 * ipa.c (function_and_variable_visibility): First remember function
4360 was global and then make it local.
4362 2013-08-22 Julian Brown <julian@codesourcery.com>
4364 * configure.ac: Add aarch64 to list of arches which use "nop" in
4366 * configure: Regenerate.
4368 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4370 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4371 gnu_libc_has_function.
4372 * config/s390/tpf.h: Likewise.
4374 2013-08-22 Jan Hubicka <jh@susue.cz>
4376 * timevar.c (validate_phases): Add cast.
4378 2013-08-22 Jan Hubicka <jh@susue.cz>
4380 * timevar.c (validate_phases): Use size_t for memory.
4381 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4383 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4385 * pretty-print.h (output_buffer::output_buffer): Declare.
4386 (pretty_printer::pretty_printer): Likewise.
4387 (pp_construct): Remove.
4388 * pretty-print.c (output_buffer::output_buffer): Define.
4389 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4390 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4391 (print_gimple_expr): Likewise.
4392 (print_gimple_seq): Likewise.
4393 (gimple_dump_bb): Likewise.
4394 * sched-vis.c (dump_value_slim): Likewise.
4395 (dump_insn_slim): Likewise.
4396 (dump_rtl_slim): Likewise.
4397 (str_pattern_slim): Likewise.
4398 * tree-mudflap.c (mf_varname_tree): Likewise.
4399 * graph.c (print_graph_cfg): Likewise.
4400 (start_graph_dump): Likewise.
4401 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4403 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4404 pretty printer initialization.
4405 * coretypes.h (diagnostic_context): Remove superflous type alias
4407 (pretty_printer): Likewise. Declare directly as a class.
4408 (pretty_print_info): Remove declaration as class.
4409 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4410 and pp_clear_output_area.
4411 (asan_add_global): Likewise.
4413 2013-08-22 Jan Hubicka <jh@suse.cz>
4415 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4416 * ipa-utils.h (update_type_inheritance_graph): Declare.
4417 (possible_polymorphic_call_target_p): Declare.
4418 (possible_polymorphic_call_target_p): New.
4419 * ipa-devirt.c: Update toplevel comments.
4420 (cached_polymorphic_call_targets): Move up.
4421 (odr_type_d): Move ID down.
4422 (polymorphic_type_binfo_p): Update comment.
4423 (odr_hasher::remove): Likewise;
4424 (get_odr_type): Set anonymous_namespace.
4425 (dump_odr_type): Dump it.
4426 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4427 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4428 (record_binfo): Add comment.
4429 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4431 (devirt_node_removal_hook): Do not iCE when cache is freed.
4432 (possible_polymorphic_call_target_p): New predicate.
4433 (update_type_inheritance_graph): New function.
4435 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4436 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4437 Sergey Lega <sergey.s.lega@intel.com>
4438 Anna Tikhonova <anna.tikhonova@intel.com>
4439 Ilya Tocar <ilya.tocar@intel.com>
4440 Andrey Turetskiy <andrey.turetskiy@intel.com>
4441 Ilya Verbin <ilya.verbin@intel.com>
4442 Kirill Yukhin <kirill.yukhin@intel.com>
4443 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4445 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4446 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4447 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4448 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4449 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4450 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4451 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4452 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4453 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4454 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4455 OPT_mavx512pf, OPT_mavx512er cases.
4456 * config/i386/constraints.md (v): New constraint.
4457 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4458 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4459 (bit_AVX512CD): New.
4460 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4461 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4462 * config/i386/i386-c.c (ix86_target_macros_internal):
4463 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4465 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4466 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4467 * config/i386/i386.c (regclass_map, dbx_register_map)
4468 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4469 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4470 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4472 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4473 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4474 -mavx512pf options. Fix formatting.
4475 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4476 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4477 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4478 -mavx512cd, -mavx512pf options.
4479 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4480 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4481 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4483 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4484 (ix86_set_reg_reg_cost): Ditto.
4485 (x86_order_regs_for_local_alloc): Ditto.
4486 (MAX_VECT_LEN): Extend to 64-byte.
4487 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4488 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4489 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4490 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4491 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4492 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4493 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4494 (SSE_REG_MODE_P): Support new modes.
4495 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4496 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4497 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4498 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4499 (REG_CLASS_CONTENTS): Add new registers.
4500 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4501 (EXT_REX_SSE_REGNO_P): New.
4502 (HI_REGISTER_NAMES): Add new registers.
4503 * config/i386/i386.md: Define constants for new registers.
4504 (mode): Add new 512-bit modes.
4505 (prefix): Support evex prefix.
4506 (isa): Support avx512f, noavx512f, fma_avx512f.
4507 (ssemodesuffix): Add new 512-bit modes.
4509 (*movxi_internal_avx512f): Ditto.
4510 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4512 (*movsi_internal): Likewise.
4513 (*movdf_internal): Likewise.
4514 (*movsf_internal): Likewise.
4515 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4516 (<code><mode>3): Likewise.
4517 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4519 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4520 with the new constraint "v".
4521 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4523 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4524 with the new constraint "v".
4525 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4526 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4527 (*<plusminus_insn><mode>3): Likewise.
4528 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4529 (*mul<mode>3): Likewise.
4530 (<sse>_vmmul<mode>3): Likewise.
4531 (<sse>_div<mode>3): Likewise.
4532 (<sse>_vmdiv<mode>3): Likewise.
4533 (<sse>_sqrt<mode>2): Likewise.
4534 (<sse>_vmsqrt<mode>2): Likewise.
4535 (*<code><mode>3_finite): Likewise.
4536 (*<code><mode>3) <smaxmin>: Likewise.
4537 (<sse>_vm<code><mode>3): Likewise.
4538 (*<code><mode>3) <any_logic>: Likewise.
4539 (*fma_fmadd_<mode>): Likewise.
4540 (*fma_fmsub_<mode>): Likewise.
4541 (*fma_fnmadd_<mode>): Likewise.
4542 (*fma_fnmsub_<mode>): Likewise.
4543 (*fma_fmaddsub_<mode>): Likewise.
4544 (*fma_fmsubadd_<mode>): Likewise.
4545 (*fmai_fmadd_<mode>): Likewise.
4546 (*fmai_fmsub_<mode>): Likewise.
4547 (*fmai_fnmadd_<mode>): Likewise.
4548 (*fmai_fnmsub_<mode>): Likewise.
4549 (sse_cvtsi2ss): Likewise.
4550 (sse_cvtsi2ssq): Likewise.
4551 (sse_cvtss2si): Likewise.
4552 (sse_cvtss2si_2): Likewise.
4553 (sse_cvtss2siq): Likewise.
4554 (sse_cvtss2siq_2): Likewise.
4555 (sse_cvttss2si): Likewise.
4556 (sse_cvtss2siq_2): Likewise.
4557 (float<sseintvecmodelower><mode>2): Likewise.
4558 (sse2_cvtsd2si_2): Likewise.
4559 (sse2_cvtsd2siq_2): Likewise.
4560 (*<plusminus_insn><mode>3): Likewise.
4561 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4562 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4563 (ashr<mode>3): Likewise.
4564 (<shift_insn><mode>3): Likewise.
4565 (avx2_<code><mode>3): Likewise.
4566 (*avx2_<code><mode>3): Likewise.
4567 (*andnot<mode>3): Likewise.
4568 (*<code><mode>3) <any_logic>: Likewise.
4569 (abs<mode>2): Likewise.
4570 (avx2_permvar<mode>): Likewise.
4571 (avx2_perm<mode>_1): Likewise.
4572 (*avx_vpermilp<mode>): Likewise.
4573 (avx_vpermilvar<mode>3): Likewise.
4574 (avx2_ashrv<mode>): Likewise.
4575 (avx2_<shift_insn>v<mode>): Likewise.
4576 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4578 * doc/rtl.texi: Document XImode.
4580 2013-08-21 Jeff Law <law@redhat.com>
4582 * tree-flow.h (register_jump_thread): Pass vector of edges
4583 instead of each important edge.
4584 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4585 thread path into a vector and pass that to register_jump_thread.
4586 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4587 passed in edge vector to the current 3-edge form.
4590 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4592 * dce.c (fini_dce): Call df_analyze again just in case
4593 delete_unmarked_insns removed anything.
4595 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4597 * reload.h (struct reg_equivs): Rename to ..
4598 (struct reg_equivs_s): .. this.
4600 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4602 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4604 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4606 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4608 2013-08-21 Jeff Law <law@redhat.com>
4610 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4611 simplify assignments too. If the RHS collapses to a singleton
4612 range, then return the value for the range.
4614 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4616 * config/i386/sse.md (V16): Rename to...
4618 (mov<mode>): Update iterator name.
4619 (*mov<mode>_internal): Ditto.
4620 (push<mode>1): Ditto.
4621 (movmisalign<mode>): Ditto.
4623 2013-08-20 Jan Hubicka <jh@suse.cz>
4626 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4627 entry for direct edges.
4628 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4630 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4632 Revert my last two changes, r201865 and r201864:
4635 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4637 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4638 instances can own GC refs.
4640 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4641 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4642 (gcc::context::gt_pch_nx): Likewise.
4643 (gcc::context::gt_pch_nx): Likewise.
4644 * ggc.h (gt_ggc_mx <T>): New.
4645 (gt_pch_nx_with_op <T>): New.
4646 (gt_pch_nx <T>): New.
4647 * passes.c (opt_pass::gt_ggc_mx): New.
4648 (opt_pass::gt_pch_nx): New.
4649 (opt_pass::gt_pch_nx_with_op): New.
4650 (pass_manager::gt_ggc_mx): New.
4651 (pass_manager::gt_pch_nx): New.
4652 (pass_manager::gt_pch_nx_with_op): New.
4653 (pass_manager::operator new): Use
4654 ggc_internal_cleared_alloc_stat rather than xcalloc.
4655 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4656 (pass_manager::gt_ggc_mx): New.
4657 (pass_manager::gt_pch_nx): New.
4658 (pass_manager::gt_pch_nx_with_op): New.
4659 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4660 (opt_pass::operator new): New.
4661 (opt_pass::gt_ggc_mx): New.
4662 (opt_pass::gt_pch_nx): New.
4663 (opt_pass::gt_pch_nx_with_op): New.
4666 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4668 * Makefile.in (GTFILES): Add context.h.
4669 * context.c (gcc::context::operator new): New.
4670 (gcc::context::gt_ggc_mx): New.
4671 (gcc::context::gt_pch_nx): New.
4672 (gcc::context::gt_pch_nx): New.
4673 * context.h (gcc::context): Add GTY((user)) marking.
4674 (gcc::context::operator new): New.
4675 (gcc::context::gt_ggc_mx): New.
4676 (gcc::context::gt_pch_nx): New.
4677 (gcc::context::gt_pch_nx): New.
4678 (g): Add GTY marking.
4679 (gt_ggc_mx (gcc::context *)): New.
4680 (gt_pch_nx (gcc::context *)): New.
4681 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4682 void *cookie)): New.
4683 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4685 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4687 * dce.c (fini_dce): Call df_analyze again just in case
4688 delete_unmarked_insns removed anything.
4690 2013-08-20 Teresa Johnson <tejohnson@google.com>
4692 PR rtl-optimizations/57451
4693 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4694 from crossing split section boundaries.
4696 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4698 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4699 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4700 with MULTLIB_DEFAULTS.
4702 2013-08-20 Nick Clifton <nickc@redhat.com>
4704 * target.def (narrow_volatile_bitfield): Note that the default
4705 value is false, not !TARGET_STRICT_ALIGN.
4706 * doc/tm.texi: Regenerate.
4708 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4710 Fix LIB_SPEC for systems without libpthread.
4712 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4713 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4715 * config/i386/linux-common.h: Likewise.
4716 * config/mips/linux-common.h: Likewise.
4718 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4720 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4723 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4725 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4726 instances can own GC refs.
4728 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4729 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4730 (gcc::context::gt_pch_nx): Likewise.
4731 (gcc::context::gt_pch_nx): Likewise.
4732 * ggc.h (gt_ggc_mx <T>): New.
4733 (gt_pch_nx_with_op <T>): New.
4734 (gt_pch_nx <T>): New.
4735 * passes.c (opt_pass::gt_ggc_mx): New.
4736 (opt_pass::gt_pch_nx): New.
4737 (opt_pass::gt_pch_nx_with_op): New.
4738 (pass_manager::gt_ggc_mx): New.
4739 (pass_manager::gt_pch_nx): New.
4740 (pass_manager::gt_pch_nx_with_op): New.
4741 (pass_manager::operator new): Use
4742 ggc_internal_cleared_alloc_stat rather than xcalloc.
4743 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4744 (pass_manager::gt_ggc_mx): New.
4745 (pass_manager::gt_pch_nx): New.
4746 (pass_manager::gt_pch_nx_with_op): New.
4747 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4748 (opt_pass::operator new): New.
4749 (opt_pass::gt_ggc_mx): New.
4750 (opt_pass::gt_pch_nx): New.
4751 (opt_pass::gt_pch_nx_with_op): New.
4753 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4755 * Makefile.in (GTFILES): Add context.h.
4756 * context.c (gcc::context::operator new): New.
4757 (gcc::context::gt_ggc_mx): New.
4758 (gcc::context::gt_pch_nx): New.
4759 (gcc::context::gt_pch_nx): New.
4760 * context.h (gcc::context): Add GTY((user)) marking.
4761 (gcc::context::operator new): New.
4762 (gcc::context::gt_ggc_mx): New.
4763 (gcc::context::gt_pch_nx): New.
4764 (gcc::context::gt_pch_nx): New.
4765 (g): Add GTY marking.
4766 (gt_ggc_mx (gcc::context *)): New.
4767 (gt_pch_nx (gcc::context *)): New.
4768 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4769 void *cookie)): New.
4770 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4772 2013-08-20 Alan Modra <amodra@gmail.com>
4775 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4776 (rs6000_emit_epilogue): Likewise.
4778 2013-08-19 Dehao Chen <dehao@google.com>
4780 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4782 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4783 Jakub Jelinek <jakub@redhat.com>
4785 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4786 (BUILT_IN_FABSD64): Likewise.
4787 (BUILT_IN_FABSD128): Likewise.
4788 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4789 (fold_builtin_1): Likewise.
4790 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4791 destination and source operands.
4792 (*abstd2_fpr): Likewise.
4793 (*nabstd2_fpr): Likewise.
4795 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4797 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4800 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4802 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4805 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4807 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4808 * builtins.c (default_libc_has_function): New.
4809 (gnu_libc_has_function): Ditto.
4810 (no_c99_libc_has_function): Ditto.
4811 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4812 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4813 (fold_builtin_sincos): Likewise.
4814 (fold_builtin_cexp): Likewise.
4815 * builtins.def (DEF_C94_BUILTIN): Likewise.
4816 (DEF_C99_BUILTIN): Likewise.
4817 (DEF_C99_C90RES_BUILTIN): Likewise.
4818 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4819 definitions to using this define.
4820 * config/darwin-protos.h (darwin_libc_has_function): New.
4821 * config/darwin.c (darwin_libc_has_function): Ditto.
4822 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4823 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4824 * config/darwin.h: Ditto.
4825 * config/elfos.h: Ditto.
4826 * config/freebsd.h: Ditto.
4827 * config/i386/cygming.h: Ditto.
4828 * config/i386/djgpp.h: Ditto.
4829 * config/i386/i386-interix.h: Ditto.
4830 * config/microblaze/microblaze.h: Ditto.
4831 * config/mmix/mmix.h: Ditto.
4832 * config/gnu-user.h: Ditto.
4833 * config/ia64/hpux.h: Ditto.
4834 * config/pa/pa-hpux.h: Ditto.
4835 * config/pdp11/pdp11.h: Ditto.
4836 * config/picochip/picochip.h: Ditto.
4837 * config/linux.h: Ditto.
4838 * config/netbsd.h: Ditto.
4839 * config/openbsd.h: Ditto.
4840 * config/rs6000/aix43.h: Ditto.
4841 * config/rs6000/aix51.h: Ditto.
4842 * config/rs6000/aix52.h: Ditto.
4843 * config/rs6000/aix53.h: Ditto.
4844 * config/rs6000/aix61.h: Ditto.
4845 * config/rs6000/darwin.h: Ditto.
4846 * config/rs6000/linux.h: Ditto.
4847 * config/rs6000/linux64.h: Ditto.
4848 * config/s390/tpf.h: Ditto.
4849 * config/sol2-10.h: Ditto.
4850 * config/sol2.h: Ditto.
4851 * config/vms/vms.h: Ditto.
4852 * config/vxworks.h: Ditto.
4853 * config/linux-android.c (linux_android_libc_has_function):
4854 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
4855 * config/linux-protos.h (linux_android_libc_has_function):
4857 * config/i386/i386.c (ix86_libc_has_function): New.
4858 * config/i386/i386-protos.h
4859 (ix86_libc_has_function): New declaration.
4860 * config/i386/i386.md
4861 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
4862 ("isinf<mode>2): Likewise.
4863 * convert.c (convert_to_integer): Using new target hook
4864 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
4865 TARGET_C99_FUNCTIONS.
4866 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
4867 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
4868 * coretypes.h (function_class): New enum for different
4869 classes of functions.
4870 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
4871 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
4872 (TARGET_HAS_SINCOS): Likewise.
4873 (TARGET_LIBC_HAS_FUNCTION): New.
4874 * doc/tm.texi: Regenerated.
4875 * targhooks.h (default_libc_has_function): New declaration.
4876 (no_c99_libc_has_function): Ditto.
4877 (gnu_libc_has_function): Ditto.
4878 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
4879 and TARGET_HAS_SINCOS.
4881 2013-08-18 Jan Hubicka <jh@suse.cz>
4883 * Makeifle-in (ipa-devirt.o): New.
4884 (GTFILES): Add ipa-utils.h and ipa-devirt.c
4885 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
4886 (analyze_functions): Look into possible targets of polymorphic call.
4887 * dumpfile.c (dump_files): Add type-inheritance dump.
4888 * dumpfile.h (TDI_inheritance): New.
4889 * ipa-devirt.c: New file.
4890 * ipa-utils.h (odr_type_d): Forward declare.
4891 (odr_type): New type.
4892 (build_type_inheritance_graph): Declare.
4893 (possible_polymorphic_call_targets): Declare and introduce inline
4894 variant when only edge is pased.
4895 (dump_possible_polymorphic_call_targets): Likewise.
4896 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
4897 * tree.c (type_in_anonymous_namespace_p): Break out from ...
4898 (types_same_for_odr): ... here.
4899 * tree.h (type_in_anonymous_namespace_p): Declare.
4901 2013-08-18 Jakub Jelinek <jakub@redhat.com>
4903 PR tree-optimization/58006
4904 * tree-parloops.c (take_address_of): Don't ICE if get_name
4906 (eliminate_local_variables_stmt): Remove clobber stmts.
4908 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
4910 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
4911 error message is issued for an alias to undefined symbol.
4913 2013-08-18 Jan Hubicka <jh@suse.cz>
4915 * cgraph.c (cgraph_create_indirect_edge): Discover
4916 polymorphic calls and record basic info into indirect_info.
4917 * gimple-fold.c (gimple_fold_call): When doing BINFO based
4918 devirtualization, ignore objc function calls.
4919 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
4920 call with no parm index info.
4921 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
4922 * tree.c (virtual_method_call_p): New function.
4923 * tree.h (virtual_method_call_p): Declare.
4925 2013-08-16 Jan Hubicka <jh@suse.cz>
4928 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
4930 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
4932 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
4935 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4937 * gengtype.c (type_for_name): Add special-case support for
4938 locating types within the "gcc::" namespace.
4939 (open_base_files): Emit a "using namespace gcc" directive.
4941 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4944 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
4945 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
4947 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
4948 array instead of each individual operand as a separate argument.
4949 (emit_fusion_gpr_load): Likewise.
4950 (expand_fusion_gpr_load): Add new function declaration.
4952 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
4953 signature to have the operands passed as an array, instead of as
4954 separate arguments. Allow ZERO_EXTEND to be in the memory
4955 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
4956 depend on the register live/dead flags when peepholes are run.
4957 (expand_fusion_gpr_load): New function to be called from the
4958 peephole2 pass, to change the register that addis sets to be the
4960 (emit_fusion_gpr_load): Change the calling signature to have the
4961 operands passed as an array, instead of as separate arguments.
4962 Allow ZERO_EXTEND to be in the memory address, and also
4963 SIGN_EXTEND if -mpower8-fusion-sign.
4965 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
4967 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
4968 adjust the register addis loads up in the peephole2 pass. Do not
4969 depend on the register live/dead state when the peephole pass is done.
4971 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4973 * gengtype.c (create_user_defined_type): Ensure that the kind
4974 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
4975 declaration is seen before the GTY((user)) marking.
4977 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
4980 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
4982 2013-08-16 Jan Hubicka <jh@suse.cz>
4984 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
4985 arugment expected_type.
4986 (gimple_fold_call): Use it.
4987 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
4988 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
4989 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
4990 (try_make_edge_direct_virtual_call): Likewise.
4991 * tree.c (obj_type_ref_class): New.
4992 * tree.h (obj_type_ref_class): Use it.
4994 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
4996 * sched-vis.c (rtl_slim_pp_initialized): Remove.
4997 (rtl_slim_pp): Likewise.
4998 (init_rtl_slim_pretty_print): Likewise.
4999 (dump_value_slim): Don't call it. Use local pretty printer.
5000 (dump_insn_slim): Likewise.
5001 (dump_rtl_slim): Likewise.
5002 (str_pattern_slim): Likewise.
5003 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5006 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5008 PR tree-optimization/58164
5009 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5010 walk gimple_goto_dest of GIMPLE_GOTO.
5012 PR tree-optimization/58165
5013 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5014 bi_call must be the last stmt in a bb, don't split_block, instead
5015 use fallthru edge from it and give up if there is none.
5016 Release conds vector when returning early.
5018 2013-08-14 Xinliang David Li <davidxl@google.com>
5020 * config/i386/i386.c (ix86_option_override_internal):
5021 Remove unused variable and field.
5023 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5026 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5027 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5028 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5029 and Linux, correct BLKmode alignment when 128-bit alignment is
5030 required and compatibility flag is not set.
5031 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5032 for zero-size arguments when compatibility flag is not set.
5034 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5036 PR tree-optimization/58145
5037 * tree-sra.c (build_ref_for_offset): If prev_base has
5038 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5040 2013-08-14 Xinliang David Li <davidxl@google.com>
5042 * config/i386/i386.c (ix86_option_override_internal):
5043 Fix uninitialized variable error.
5045 2013-08-14 Xinliang David Li <davidxl@google.com>
5047 * config/i386/i386.opt: Define two new options.
5048 * config/i386/x86-tune.def: Add arch selector field in macros.
5049 * config/i386/i386.h: Adjust macro definition.
5050 * config/i386/i386.c (ix86_option_override_internal):
5052 (parse_mtune_ctrl_str): New function.
5053 (set_ix86_tune_features): New function.
5054 (ix86_function_specific_restore): Call the new helper function.
5056 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5058 PR rtl-optimization/57662
5059 * sel-sched.c (code_motion_process_successors): When the current insn
5060 is removed after the recursive traversal, break from the loop.
5061 Add comments and debug printouts.
5063 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5064 Alexandre Oliva <aoliva@redhat.com>
5067 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5068 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5069 there also UNSPEC_PLTOFF.
5071 2013-08-14 Marek Polacek <polacek@redhat.com>
5073 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5074 NUM_CONDITIONS bit positions.
5076 2013-08-13 Cary Coutant <ccoutant@google.com>
5078 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5079 (attr_checksum): Hash vector contents instead of pointer.
5080 (attr_checksum_ordered): Likewise.
5082 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5084 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5085 when Pmode != word_mode. Add length_address attribute.
5086 (sse3_monitor_<mode>): Merge from sse3_monitor and
5087 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5088 Pmode != word_mode. Update insn length attribute.
5089 * config/i386/i386.c (ix86_option_override_internal): Update
5090 ix86_gen_monitor selection for merged sse3_monitor insn.
5092 2013-08-13 Julian Brown <julian@codesourcery.com>
5094 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5095 perform invalid legitimization on greater-than-word-size modes for
5098 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5100 * ira.c (setup_class_translate_array): Use aclass instead of cl
5101 for classes not fully covered by allocno classes.
5103 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5105 PR tree-optimization/57661
5106 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5107 * tree-inline.c (tree_function_versioning): Initialize it.
5108 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5109 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5110 that is not being copied.
5113 * asan.c (instrument_strlen_call): Fix typo in comment.
5114 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5116 2013-08-13 Steve Ellcey <sellcey@mips.com>
5118 * config/mips/mips.md (prefetch): Use lw instead of ld on
5119 loongson in 32bit mode.
5121 2013-08-13 Nick Clifton <nickc@redhat.com>
5123 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5125 2013-08-13 Jan Hubicka <jh@suse.cz>
5127 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5128 introduced edge; fix typo in sanity check.
5129 (cgraph_resolve_speculation): Export; improve diagnostic.
5130 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5131 speculation at type mismatch.
5132 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5133 (cgraph_resolve_speculation): Declare.
5134 (symtab_can_be_discarded): New function.
5135 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5136 * ipa-inline-transform.c (speculation_removed): New global var.
5137 (clone_inlined_nodes): See if speculation can be removed.
5138 (inline_call): If speculations was removed, we growths may not match.
5139 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5140 (speculation_useful_p): New function.
5141 (resolve_noninline_speculation): New function.
5142 (inline_small_functions): Resolve useless speculations.
5143 * ipa-inline.h (speculation_useful_p): Declare
5144 * ipa.c (can_replace_by_local_alias): Simplify.
5145 (ipa_profile): Produce speculative calls in non-lto, too;
5146 add simple cost model; produce local aliases.
5148 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5150 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5153 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5155 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5157 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5159 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5160 iterator instead of X87MODEF.
5162 2013-08-12 Perez Read <netfirewall@gmail.com>
5165 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5166 operand 0 for intel asm alternative.
5167 (*movabs<mode>_2): Ditto for operand 1.
5169 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5171 * config/aarch64/arm_none.h
5172 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5174 2013-08-12 Nick Clifton <nickc@redhat.com>
5176 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5178 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5180 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5181 expand for QI/HImode operand to produce more effictive code for
5182 unsigned char(short) --> float(double) conversion.
5184 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5186 * doc/invoke.texi: Mention that -ftls-model does not force the final
5189 2013-08-12 Marek Polacek <polacek@redhat.com>
5190 Marc Glisse <marc.glisse@inria.fr>
5192 PR tree-optimization/57980
5193 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5194 when creating -1 constant.
5196 2013-08-10 Jan Hubicka <jh@suse.cz>
5198 Workaround binutils PR14342.
5199 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5200 (gimple_init_edge_profiler): Likewise.
5201 (gimple_gen_ic_func_profiler): Likewise.
5203 2013-08-09 Jan Hubicka <jh@suse.cz>
5205 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5207 2013-08-09 Xinliang David Li <davidxl@google.com>
5209 * config/i386/stringop.def: New file.
5210 * config/i386/stringop.opt: New file.
5211 * config/i386/i386-opts.h: Include stringopt.def.
5212 * config/i386/i386.opt: Include stringopt.opt.
5213 * config/i386/i386.c (ix86_option_override_internal):
5214 Override default size based stringop inline strategies with options.
5215 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5218 2013-08-09 Jan Hubicka <jh@suse.cz>
5220 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5222 2013-08-09 Jan Hubicka <jh@suse.cz>
5224 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5226 (dump_cgraph_node): Dump profile-id.
5227 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5228 and common_target_probability.
5229 * lto-cgraph.c (lto_output_edge): Stream common targets.
5230 (lto_output_node): Stream profile ids.
5231 (input_node): Stream profile ids.
5232 (input_edge): Stream common targets.
5233 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5234 * ipa.c: Include value-prof.h
5235 (ipa_profile_generate_summary): Turn indirect call statement histograms
5236 into common targets.
5237 (ipa_profile): Turn common targets into speculative edges.
5239 2013-08-09 Jan Hubicka <jh@suse.cz>
5241 * cgraph.h (cgraph_node): Add profile_id.
5242 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5243 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5244 (del_node_map): Update.
5245 (find_func_by_funcdef_no): Replace by ...
5246 (find_func_by_profile_id): ... this one.
5247 (gimple_ic_transform): Do not remove useful histograms when
5248 speculation is not done; dump info when indirect call removal
5250 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5251 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5252 (__gcov_indirect_call_profiler_v2): .. this one.
5253 * profile.h (init_node_map): Update.
5254 * coverage.c (coverage_compute_profile_id): New function.
5255 * coverage.h (coverage_compute_profile_id): Declare.
5256 * tree-profile.c (init_ic_make_global_vars): Make
5257 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5258 (gimple_init_edge_profiler): Update prototype of
5259 __gcov_indirect_call_profiler.
5260 (gimple_gen_ic_func_profiler): Simplify.
5261 (tree_profiling): Use init_node_map
5263 2013-08-09 Jan Hubicka <jh@suse.cz>
5265 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5266 non-speculative refs.
5267 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5268 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5269 (cgraph_set_call_stmt): Likewise.
5270 (cgraph_create_edge_1): Fix release checking compilatoin;
5272 (cgraph_free_edge): Free indirect info.
5273 (cgraph_turn_edge_to_speculative): New function.
5274 (cgraph_speculative_call_info): New function.
5275 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5276 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5277 (dump_cgraph_node): Dump speculation.
5278 (verify_edge_count_and_frequency): Accept speculative edges.
5279 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5280 (verify_cgraph_node): Handle speculation.
5281 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5282 (cgraph_set_call_stmt): Update prototype.
5283 (cgraph_make_edge_direct): Update prototype.
5284 (cgraph_speculative_call_info): Declare.
5285 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5286 to change; update call of ipa_find_references.
5287 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5288 lto_stmt_uid and speculative flags.
5289 (ipa_dump_references): Dump speculation.
5290 (ipa_clone_references): Clone speculative flag.
5291 (ipa_clone_referring): Likewise.
5292 (ipa_clone_ref): New function.
5293 (ipa_find_reference): Look into lto_stmt_uids
5294 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5295 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5296 (ipa_find_reference): Update declaration.
5297 (ipa_clone_ref): Declare.
5298 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5299 stream speculative flag.
5300 (lto_output_ref): Stream statements uids and speculation.
5301 (input_ref): Likewise.
5302 (input_edge): Stream speuclation.
5303 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5304 (cgraph_set_call_stmt_including_clones): Handle speculation.
5305 * ipa-inline.c (heap_edge_removal_hook): New function.
5306 (inline_small_functions): Register it.
5307 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5308 also initialize refs.
5309 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5311 (try_make_edge_direct_simple_call): Likewise.
5312 (try_make_edge_direct_simple_call): Likewise.
5313 (update_indirect_edges_after_inlining): Likewise.
5314 (remove_described_reference): Look proper lto_stmt_uid.
5315 (propagate_controlled_uses): Likewise.
5316 (propagate_controlled_uses): Liekwise.
5317 * tree-inline.c (copy_bb): Copy speculative edges.
5318 (redirect_all_calls): New function.
5319 (copy_cfg_body): Do redirection after loop info is updated.
5320 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5322 2013-08-09 Jan Hubicka <jh@suse.cz>
5324 * lto-streamer-out.c (output_function): Renumber PHIs.
5325 * lto-streamer-in.c (input_function): Likewise.
5327 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5329 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5330 (get_lane_unsigned): Likewise.
5331 (dup_lane_scalar): Likewise.
5332 (get_lane): enable for VALL.
5333 * config/aarch64/aarch64-simd.md
5334 (aarch64_dup_lane_scalar<mode>): Remove.
5335 (aarch64_get_lane_signed<mode>): Likewise.
5336 (aarch64_get_lane_unsigned<mode>): Likewise.
5337 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5338 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5339 (aarch64_get_lane<mode>): Enable for all vector modes.
5340 (aarch64_get_lanedi): Remove misleading constraints.
5341 * config/aarch64/arm_neon.h
5342 (__aarch64_vget_lane_any): Define.
5343 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5344 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5345 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5346 * config/aarch64/iterators.md (VDQQH): New.
5350 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5352 * configure.ac: Add GAS check for LEON instructions on SPARC.
5353 * configure: Regenerate.
5354 * config.in: Likewise.
5355 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5357 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5358 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5360 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5361 (AS_LEON_FLAG): New macro.
5362 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5363 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5364 Deal with LEON and LEON3 for the memory model.
5365 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5366 (atomic_compare_and_swap<mode>_1): Likewise.
5367 (*atomic_compare_and_swap<mode>_1): Likewise.
5369 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5371 * config/arm/neon.md (vcond): Fix floating-point vector
5372 comparisons against 0.
5374 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5376 * lra-constraints.c (emit_spill_move): Remove assert.
5377 (process_alt_operands): Add more debugging
5378 output. Increase reject for spilling into memory. Decrease
5379 reject for reloading scratch.
5380 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5382 2013-08-08 Steve Ellcey <sellcey@mips.com>
5384 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5385 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5386 micromips incompatible. Add nan2008.
5387 (MULTILIB_DIRNAMES): Add nan2008.
5388 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5389 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5390 and micromips incompatible. Add nan2008.
5391 (MULTILIB_DIRNAMES): Add nan2008.
5392 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5394 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5396 PR rtl-optimization/58079
5397 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5398 simplify_comparison has widened a comparison with an integer.
5400 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5402 * config/arm/neon.md (movmisalign<mode>): Disable when we
5403 don't allow unaligned accesses.
5404 (*movmisalign<mode>_neon_store): Likewise.
5405 (*movmisalign<mode>_neon_load): Likewise.
5406 (*movmisalign<mode>_neon_store): Likewise.
5407 (*movmisalign<mode>_neon_load): Likewise.
5409 2013-08-08 Jan Hubicka <jh@suse.cz>
5411 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5412 (make_pass_rebuild_cgraph_edges): Also clear references.
5413 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5414 * ipa-inline-transform.c (inline_transform): Remove all references
5416 * cgraphunit.c (expand_function): Remove all references after
5418 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5419 (ipa_find_reference): Rewrite to iterator.
5420 (remove_stmt_references): Likewise.
5421 (ipa_clear_stmts_in_references): New function.
5422 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5423 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5425 * ipa-split.c (split_function): Remove references in split function.
5427 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5430 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5431 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5433 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5436 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5437 suggested mode for the assignment isn't compatible with the
5440 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5443 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5445 2013-08-07 Xinliang David Li <davidxl@google.com>
5447 * config/i386/i386.opt: New option -mtune-ctrl=.
5448 * config/i386/x86-tune.def: New file.
5449 * config/i386/i386.h: include x86-tune.def.
5450 * config/i386/i386.c (ix86_option_override_internal):
5451 Parsing -mtune-ctrl= option and set tune features.
5453 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5456 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5457 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5460 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5462 * rtl.h (update_alignments): Declare.
5463 * final.c (grow_label_align): New function extracted from...
5464 (shorten_branches): ...here. Call it.
5465 (update_alignments): New function.
5466 * reorg.c (sibling_labels): New variable.
5467 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5468 the new label along with it onto the sibling_labels vector.
5469 (fill_simple_delay_slots): Adjust call to get_label_before.
5470 (fill_slots_from_thread): Likewise.
5471 (relax_delay_slots): Likewise.
5472 (make_return_insns): Likewise.
5473 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5475 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5477 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5478 document its semantics.
5479 (diagnostic_report_diagnostic): Adjust accordingly.
5481 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5483 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5484 (sparc_option_override): ...and port to new C++ pass API.
5485 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5487 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5489 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5491 2013-08-06 Caroline Tice <cmtice@google.com>
5493 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5494 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5495 * tree-pass.h: Add pass_vtable_verify.
5496 * varasm.c (assemble_variable): Add code to properly set the comdat
5497 section and name for the .vtable_map_vars section.
5498 (assemble_vtyv_preinit_initializer): New function.
5499 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5501 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5502 * vtable-verify.c: New file.
5503 * vtable-verify.h: New file.
5504 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5505 initialiation levels.
5506 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5507 * passes.def: Insert pass_vtable_verify.
5508 * aclocal.m4: Reorder includes.
5509 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5510 -fvtv-counts options.
5511 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5512 as appropriate, if -fvtable-verify=... is used.
5513 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5514 -fvtable-verify=... is used.
5515 * Makefile.in (OBJS): Add vtable-verify.o to list.
5516 (vtable-verify.o): Add new build rule.
5517 (GTFILES): Add vtable-verify.c to list.
5518 * common.opt (fvtable-verify=): New flag.
5519 (vtv_priority): Values for fvtable-verify= flag.
5520 (fvtv-counts): New flag.
5521 (fvtv-debug): New flag.
5522 * tree.h (save_vtable_map_decl): New extern function decl.
5524 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5526 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5527 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5528 (pass_data_rl78_devirt): ...new pass_data instance and...
5529 (make_pass_rl78_devirt): ...new function.
5530 (rl78_asm_file_start): Port pass registration to new C++ API.
5532 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5534 * coretypes.h (rtl_opt_pass): Add.
5535 (gcc::context): Add.
5536 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5537 (epiphany_init): Port to new C++ pass API.
5538 (epiphany_optimize_mode_switching): Likewise.
5539 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5540 (pass_manager::get_pass_mode_switching): New.
5541 (pass_manager::get_pass_peephole2): New.
5542 * mode-switching.c (pass_mode_switching): Add clone method.
5543 * recog.c (pass_peephole2): Add clone method.
5544 (pass_split_all_insns): Add clone method.
5546 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5548 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5549 (mips_option_override): ...here, porting to new C++ API for passes.
5551 2013-08-06 Jan Hubicka <jh@suse.cz>
5553 * cgraph.c (cgraph_get_body): New function based on lto.c
5555 * cgraph.h (cgraph_get_body): Declare.
5556 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5558 * cgraphunit.c (expand_function): Get body prior expanding.
5559 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5560 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5562 * passes.c (do_per_function_toporder): Get body.
5563 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5564 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5566 2013-08-06 Martin Jambor <mjambor@suse.cz>
5569 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5570 is not re-finalized. Rename second parameter to no_collect.
5572 2013-08-06 Martin Jambor <mjambor@suse.cz>
5575 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5576 MEM_REF has proper alignment information.
5578 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5581 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5583 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5584 rtx (*) (rtx, ...) with insn_gen_fn.
5585 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5586 insn_gen_fn::stored_funcptr. Add initializer braces.
5588 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5590 Rewrite how instances of passes are cloned to remove assumptions
5591 about their sizes (thus allowing pass subclasses to have
5592 additional data fields, albeit non-GC-managed ones at this point).
5594 * passes.c (make_pass_instance): Now that passes have clone
5595 methods, rewrite this function to eliminate XNEW and memcpy
5596 calls that used hardcoded sizes. Since this function no longer
5597 creates pass instances, rename it to...
5598 (add_pass_instance): ...this. Document the old way that passes were
5599 numbered and flagged, and rework this function to continue using it.
5600 (next_pass_1): Add an initial_pass argument for use by
5602 (position_pass): When adding multiple instances of a pass, use
5603 the pass's clone method, rather than relying on the XNEW/memcpy
5604 within the former make_pass_instance (now add_pass_instance).
5605 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5606 the initial instance of the current pass within the pass manager.
5608 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5610 This is the automated part of the conversion of passes from C
5611 structs to C++ classes.
5613 Patch autogenerated by refactor_passes.py from
5614 https://github.com/davidmalcolm/gcc-refactoring-scripts
5615 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5617 * asan.c (pass_asan): Convert from a global struct to a subclass of
5618 gimple_opt_pass along with...
5619 (pass_data_asan): ...new pass_data instance and...
5620 (make_pass_asan): ...new function.
5621 (pass_asan_O0): Convert from a global struct to a subclass of
5622 gimple_opt_pass along with...
5623 (pass_data_asan_O0): ...new pass_data instance and...
5624 (make_pass_asan_O0): ...new function.
5625 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5626 subclass of rtl_opt_pass along with...
5627 (pass_data_inc_dec): ...new pass_data instance and...
5628 (make_pass_inc_dec): ...new function.
5629 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5630 a subclass of rtl_opt_pass along with...
5631 (pass_data_reorder_blocks): ...new pass_data instance and...
5632 (make_pass_reorder_blocks): ...new function.
5633 (pass_duplicate_computed_gotos): Convert from a global struct to a
5634 subclass of rtl_opt_pass along with...
5635 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5636 (make_pass_duplicate_computed_gotos): ...new function.
5637 (pass_partition_blocks): Convert from a global struct to a subclass of
5638 rtl_opt_pass along with...
5639 (pass_data_partition_blocks): ...new pass_data instance and...
5640 (make_pass_partition_blocks): ...new function.
5641 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5642 struct to a subclass of rtl_opt_pass along with...
5643 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5645 (make_pass_branch_target_load_optimize1): ...new function.
5646 (pass_branch_target_load_optimize2): Convert from a global struct to a
5647 subclass of rtl_opt_pass along with...
5648 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5650 (make_pass_branch_target_load_optimize2): ...new function.
5651 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5652 of rtl_opt_pass along with...
5653 (pass_data_jump): ...new pass_data instance and...
5654 (make_pass_jump): ...new function.
5655 (pass_jump2): Convert from a global struct to a subclass of
5656 rtl_opt_pass along with...
5657 (pass_data_jump2): ...new pass_data instance and...
5658 (make_pass_jump2): ...new function.
5659 * cfgexpand.c (pass_expand): Convert from a global struct to a
5660 subclass of rtl_opt_pass along with...
5661 (pass_data_expand): ...new pass_data instance and...
5662 (make_pass_expand): ...new function.
5663 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5664 of rtl_opt_pass along with...
5665 (pass_data_free_cfg): ...new pass_data instance and...
5666 (make_pass_free_cfg): ...new function.
5667 (pass_into_cfg_layout_mode): Convert from a global struct to a
5668 subclass of rtl_opt_pass along with...
5669 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5670 (make_pass_into_cfg_layout_mode): ...new function.
5671 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5672 subclass of rtl_opt_pass along with...
5673 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5674 (make_pass_outof_cfg_layout_mode): ...new function.
5675 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5676 struct to a subclass of gimple_opt_pass along with...
5677 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5678 (make_pass_build_cgraph_edges): ...new function.
5679 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5680 subclass of gimple_opt_pass along with...
5681 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5682 (make_pass_rebuild_cgraph_edges): ...new function.
5683 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5684 subclass of gimple_opt_pass along with...
5685 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5687 (make_pass_remove_cgraph_callee_edges): ...new function.
5688 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5689 struct to a subclass of rtl_opt_pass along with...
5690 (pass_data_stack_adjustments): ...new pass_data instance and...
5691 (make_pass_stack_adjustments): ...new function.
5692 * combine.c (pass_combine): Convert from a global struct to a subclass
5693 of rtl_opt_pass along with...
5694 (pass_data_combine): ...new pass_data instance and...
5695 (make_pass_combine): ...new function.
5696 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5697 global struct to a subclass of rtl_opt_pass along with...
5698 (pass_data_compare_elim_after_reload): ...new pass_data instance
5700 (make_pass_compare_elim_after_reload): ...new function.
5701 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5702 of rtl_opt_pass along with...
5703 (pass_data_rtl_cprop): ...new pass_data instance and...
5704 (make_pass_rtl_cprop): ...new function.
5705 * cse.c (pass_cse): Convert from a global struct to a subclass of
5706 rtl_opt_pass along with...
5707 (pass_data_cse): ...new pass_data instance and...
5708 (make_pass_cse): ...new function.
5709 (pass_cse2): Convert from a global struct to a subclass of
5710 rtl_opt_pass along with...
5711 (pass_data_cse2): ...new pass_data instance and...
5712 (make_pass_cse2): ...new function.
5713 (pass_cse_after_global_opts): Convert from a global struct to a
5714 subclass of rtl_opt_pass along with...
5715 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5716 (make_pass_cse_after_global_opts): ...new function.
5717 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5718 of rtl_opt_pass along with...
5719 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5720 (make_pass_ud_rtl_dce): ...new function.
5721 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5722 rtl_opt_pass along with...
5723 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5724 (make_pass_fast_rtl_dce): ...new function.
5725 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5726 a subclass of rtl_opt_pass along with...
5727 (pass_data_df_initialize_opt): ...new pass_data instance and...
5728 (make_pass_df_initialize_opt): ...new function.
5729 (pass_df_initialize_no_opt): Convert from a global struct to a
5730 subclass of rtl_opt_pass along with...
5731 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5732 (make_pass_df_initialize_no_opt): ...new function.
5733 (pass_df_finish): Convert from a global struct to a subclass of
5734 rtl_opt_pass along with...
5735 (pass_data_df_finish): ...new pass_data instance and...
5736 (make_pass_df_finish): ...new function.
5737 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5738 rtl_opt_pass along with...
5739 (pass_data_rtl_dse1): ...new pass_data instance and...
5740 (make_pass_rtl_dse1): ...new function.
5741 (pass_rtl_dse2): Convert from a global struct to a subclass of
5742 rtl_opt_pass along with...
5743 (pass_data_rtl_dse2): ...new pass_data instance and...
5744 (make_pass_rtl_dse2): ...new function.
5745 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5746 subclass of rtl_opt_pass along with...
5747 (pass_data_dwarf2_frame): ...new pass_data instance and...
5748 (make_pass_dwarf2_frame): ...new function.
5749 * except.c (pass_set_nothrow_function_flags): Convert from a global
5750 struct to a subclass of rtl_opt_pass along with...
5751 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5753 (make_pass_set_nothrow_function_flags): ...new function.
5754 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5755 subclass of rtl_opt_pass along with...
5756 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5758 (make_pass_convert_to_eh_region_ranges): ...new function.
5759 * final.c (pass_compute_alignments): Convert from a global struct to a
5760 subclass of rtl_opt_pass along with...
5761 (pass_data_compute_alignments): ...new pass_data instance and...
5762 (make_pass_compute_alignments): ...new function.
5763 (pass_final): Convert from a global struct to a subclass of
5764 rtl_opt_pass along with...
5765 (pass_data_final): ...new pass_data instance and...
5766 (make_pass_final): ...new function.
5767 (pass_shorten_branches): Convert from a global struct to a subclass of
5768 rtl_opt_pass along with...
5769 (pass_data_shorten_branches): ...new pass_data instance and...
5770 (make_pass_shorten_branches): ...new function.
5771 (pass_clean_state): Convert from a global struct to a subclass of
5772 rtl_opt_pass along with...
5773 (pass_data_clean_state): ...new pass_data instance and...
5774 (make_pass_clean_state): ...new function.
5775 * function.c (pass_instantiate_virtual_regs): Convert from a global
5776 struct to a subclass of rtl_opt_pass along with...
5777 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5778 (make_pass_instantiate_virtual_regs): ...new function.
5779 (pass_leaf_regs): Convert from a global struct to a subclass of
5780 rtl_opt_pass along with...
5781 (pass_data_leaf_regs): ...new pass_data instance and...
5782 (make_pass_leaf_regs): ...new function.
5783 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5784 subclass of rtl_opt_pass along with...
5785 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5787 (make_pass_thread_prologue_and_epilogue): ...new function.
5788 (pass_match_asm_constraints): Convert from a global struct to a
5789 subclass of rtl_opt_pass along with...
5790 (pass_data_match_asm_constraints): ...new pass_data instance and...
5791 (make_pass_match_asm_constraints): ...new function.
5792 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5793 subclass of rtl_opt_pass along with...
5794 (pass_data_rtl_fwprop): ...new pass_data instance and...
5795 (make_pass_rtl_fwprop): ...new function.
5796 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5797 rtl_opt_pass along with...
5798 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5799 (make_pass_rtl_fwprop_addr): ...new function.
5800 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5801 rtl_opt_pass along with...
5802 (pass_data_rtl_pre): ...new pass_data instance and...
5803 (make_pass_rtl_pre): ...new function.
5804 (pass_rtl_hoist): Convert from a global struct to a subclass of
5805 rtl_opt_pass along with...
5806 (pass_data_rtl_hoist): ...new pass_data instance and...
5807 (make_pass_rtl_hoist): ...new function.
5808 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5809 subclass of gimple_opt_pass along with...
5810 (pass_data_lower_cf): ...new pass_data instance and...
5811 (make_pass_lower_cf): ...new function.
5812 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5813 from a global struct to a subclass of gimple_opt_pass along with...
5814 (pass_data_strength_reduction): ...new pass_data instance and...
5815 (make_pass_strength_reduction): ...new function.
5816 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5817 of rtl_opt_pass along with...
5818 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5819 (make_pass_rtl_ifcvt): ...new function.
5820 (pass_if_after_combine): Convert from a global struct to a subclass of
5821 rtl_opt_pass along with...
5822 (pass_data_if_after_combine): ...new pass_data instance and...
5823 (make_pass_if_after_combine): ...new function.
5824 (pass_if_after_reload): Convert from a global struct to a subclass of
5825 rtl_opt_pass along with...
5826 (pass_data_if_after_reload): ...new pass_data instance and...
5827 (make_pass_if_after_reload): ...new function.
5828 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5829 a subclass of rtl_opt_pass along with...
5830 (pass_data_initialize_regs): ...new pass_data instance and...
5831 (make_pass_initialize_regs): ...new function.
5832 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5833 of ipa_opt_pass_d along with...
5834 (pass_data_ipa_cp): ...new pass_data instance and...
5835 (make_pass_ipa_cp): ...new function.
5836 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5837 global struct to a subclass of gimple_opt_pass along with...
5838 (pass_data_inline_parameters): ...new pass_data instance and...
5839 (make_pass_inline_parameters): ...new function.
5840 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5841 subclass of gimple_opt_pass along with...
5842 (pass_data_early_inline): ...new pass_data instance and...
5843 (make_pass_early_inline): ...new function.
5844 (pass_ipa_inline): Convert from a global struct to a subclass of
5845 ipa_opt_pass_d along with...
5846 (pass_data_ipa_inline): ...new pass_data instance and...
5847 (make_pass_ipa_inline): ...new function.
5848 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5849 struct to a subclass of gimple_opt_pass along with...
5850 (pass_data_local_pure_const): ...new pass_data instance and...
5851 (make_pass_local_pure_const): ...new function.
5852 (pass_ipa_pure_const): Convert from a global struct to a subclass of
5853 ipa_opt_pass_d along with...
5854 (pass_data_ipa_pure_const): ...new pass_data instance and...
5855 (make_pass_ipa_pure_const): ...new function.
5856 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
5857 to a subclass of ipa_opt_pass_d along with...
5858 (pass_data_ipa_reference): ...new pass_data instance and...
5859 (make_pass_ipa_reference): ...new function.
5860 * ipa-split.c (pass_split_functions): Convert from a global struct to
5861 a subclass of gimple_opt_pass along with...
5862 (pass_data_split_functions): ...new pass_data instance and...
5863 (make_pass_split_functions): ...new function.
5864 (pass_feedback_split_functions): Convert from a global struct to a
5865 subclass of gimple_opt_pass along with...
5866 (pass_data_feedback_split_functions): ...new pass_data instance and...
5867 (make_pass_feedback_split_functions): ...new function.
5868 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
5869 global struct to a subclass of simple_ipa_opt_pass along with...
5870 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
5872 (make_pass_ipa_function_and_variable_visibility): ...new function.
5873 (pass_ipa_free_inline_summary): Convert from a global struct to a
5874 subclass of simple_ipa_opt_pass along with...
5875 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
5876 (make_pass_ipa_free_inline_summary): ...new function.
5877 (pass_ipa_whole_program_visibility): Convert from a global struct to a
5878 subclass of ipa_opt_pass_d along with...
5879 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
5881 (make_pass_ipa_whole_program_visibility): ...new function.
5882 (pass_ipa_profile): Convert from a global struct to a subclass of
5883 ipa_opt_pass_d along with...
5884 (pass_data_ipa_profile): ...new pass_data instance and...
5885 (make_pass_ipa_profile): ...new function.
5886 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
5887 ipa_opt_pass_d along with...
5888 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
5889 (make_pass_ipa_cdtor_merge): ...new function.
5890 * ira.c (pass_ira): Convert from a global struct to a subclass of
5891 rtl_opt_pass along with...
5892 (pass_data_ira): ...new pass_data instance and...
5893 (make_pass_ira): ...new function.
5894 (pass_reload): Convert from a global struct to a subclass of
5895 rtl_opt_pass along with...
5896 (pass_data_reload): ...new pass_data instance and...
5897 (make_pass_reload): ...new function.
5898 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
5899 subclass of rtl_opt_pass along with...
5900 (pass_data_cleanup_barriers): ...new pass_data instance and...
5901 (make_pass_cleanup_barriers): ...new function.
5902 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
5903 of rtl_opt_pass along with...
5904 (pass_data_loop2): ...new pass_data instance and...
5905 (make_pass_loop2): ...new function.
5906 (pass_rtl_loop_init): Convert from a global struct to a subclass of
5907 rtl_opt_pass along with...
5908 (pass_data_rtl_loop_init): ...new pass_data instance and...
5909 (make_pass_rtl_loop_init): ...new function.
5910 (pass_rtl_loop_done): Convert from a global struct to a subclass of
5911 rtl_opt_pass along with...
5912 (pass_data_rtl_loop_done): ...new pass_data instance and...
5913 (make_pass_rtl_loop_done): ...new function.
5914 (pass_rtl_move_loop_invariants): Convert from a global struct to a
5915 subclass of rtl_opt_pass along with...
5916 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
5917 (make_pass_rtl_move_loop_invariants): ...new function.
5918 (pass_rtl_unswitch): Convert from a global struct to a subclass of
5919 rtl_opt_pass along with...
5920 (pass_data_rtl_unswitch): ...new pass_data instance and...
5921 (make_pass_rtl_unswitch): ...new function.
5922 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
5923 subclass of rtl_opt_pass along with...
5924 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
5926 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5927 (pass_rtl_doloop): Convert from a global struct to a subclass of
5928 rtl_opt_pass along with...
5929 (pass_data_rtl_doloop): ...new pass_data instance and...
5930 (make_pass_rtl_doloop): ...new function.
5931 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
5932 a subclass of rtl_opt_pass along with...
5933 (pass_data_lower_subreg): ...new pass_data instance and...
5934 (make_pass_lower_subreg): ...new function.
5935 (pass_lower_subreg2): Convert from a global struct to a subclass of
5936 rtl_opt_pass along with...
5937 (pass_data_lower_subreg2): ...new pass_data instance and...
5938 (make_pass_lower_subreg2): ...new function.
5939 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
5940 struct to a subclass of ipa_opt_pass_d along with...
5941 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
5942 (make_pass_ipa_lto_gimple_out): ...new function.
5943 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
5944 of ipa_opt_pass_d along with...
5945 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
5946 (make_pass_ipa_lto_finish_out): ...new function.
5947 * mode-switching.c (pass_mode_switching): Convert from a global struct
5948 to a subclass of rtl_opt_pass along with...
5949 (pass_data_mode_switching): ...new pass_data instance and...
5950 (make_pass_mode_switching): ...new function.
5951 * modulo-sched.c (pass_sms): Convert from a global struct to a
5952 subclass of rtl_opt_pass along with...
5953 (pass_data_sms): ...new pass_data instance and...
5954 (make_pass_sms): ...new function.
5955 * omp-low.c (pass_expand_omp): Convert from a global struct to a
5956 subclass of gimple_opt_pass along with...
5957 (pass_data_expand_omp): ...new pass_data instance and...
5958 (make_pass_expand_omp): ...new function.
5959 (pass_lower_omp): Convert from a global struct to a subclass of
5960 gimple_opt_pass along with...
5961 (pass_data_lower_omp): ...new pass_data instance and...
5962 (make_pass_lower_omp): ...new function.
5963 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
5964 of gimple_opt_pass along with...
5965 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
5966 (make_pass_diagnose_omp_blocks): ...new function.
5967 * passes.c (pass_early_local_passes): Convert from a global struct to
5968 a subclass of simple_ipa_opt_pass along with...
5969 (pass_data_early_local_passes): ...new pass_data instance and...
5970 (make_pass_early_local_passes): ...new function.
5971 (pass_all_early_optimizations): Convert from a global struct to a
5972 subclass of gimple_opt_pass along with...
5973 (pass_data_all_early_optimizations): ...new pass_data instance and...
5974 (make_pass_all_early_optimizations): ...new function.
5975 (pass_all_optimizations): Convert from a global struct to a subclass
5976 of gimple_opt_pass along with...
5977 (pass_data_all_optimizations): ...new pass_data instance and...
5978 (make_pass_all_optimizations): ...new function.
5979 (pass_all_optimizations_g): Convert from a global struct to a subclass
5980 of gimple_opt_pass along with...
5981 (pass_data_all_optimizations_g): ...new pass_data instance and...
5982 (make_pass_all_optimizations_g): ...new function.
5983 (pass_rest_of_compilation): Convert from a global struct to a subclass
5984 of rtl_opt_pass along with...
5985 (pass_data_rest_of_compilation): ...new pass_data instance and...
5986 (make_pass_rest_of_compilation): ...new function.
5987 (pass_postreload): Convert from a global struct to a subclass of
5988 rtl_opt_pass along with...
5989 (pass_data_postreload): ...new pass_data instance and...
5990 (make_pass_postreload): ...new function.
5991 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
5992 subclass of rtl_opt_pass along with...
5993 (pass_data_gcse2): ...new pass_data instance and...
5994 (make_pass_gcse2): ...new function.
5995 * postreload.c (pass_postreload_cse): Convert from a global struct to
5996 a subclass of rtl_opt_pass along with...
5997 (pass_data_postreload_cse): ...new pass_data instance and...
5998 (make_pass_postreload_cse): ...new function.
5999 * predict.c (pass_profile): Convert from a global struct to a subclass
6000 of gimple_opt_pass along with...
6001 (pass_data_profile): ...new pass_data instance and...
6002 (make_pass_profile): ...new function.
6003 (pass_strip_predict_hints): Convert from a global struct to a subclass
6004 of gimple_opt_pass along with...
6005 (pass_data_strip_predict_hints): ...new pass_data instance and...
6006 (make_pass_strip_predict_hints): ...new function.
6007 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6008 of rtl_opt_pass along with...
6009 (pass_data_peephole2): ...new pass_data instance and...
6010 (make_pass_peephole2): ...new function.
6011 (pass_split_all_insns): Convert from a global struct to a subclass of
6012 rtl_opt_pass along with...
6013 (pass_data_split_all_insns): ...new pass_data instance and...
6014 (make_pass_split_all_insns): ...new function.
6015 (pass_split_after_reload): Convert from a global struct to a subclass
6016 of rtl_opt_pass along with...
6017 (pass_data_split_after_reload): ...new pass_data instance and...
6018 (make_pass_split_after_reload): ...new function.
6019 (pass_split_before_regstack): Convert from a global struct to a
6020 subclass of rtl_opt_pass along with...
6021 (pass_data_split_before_regstack): ...new pass_data instance and...
6022 (make_pass_split_before_regstack): ...new function.
6023 (pass_split_before_sched2): Convert from a global struct to a subclass
6024 of rtl_opt_pass along with...
6025 (pass_data_split_before_sched2): ...new pass_data instance and...
6026 (make_pass_split_before_sched2): ...new function.
6027 (pass_split_for_shorten_branches): Convert from a global struct to a
6028 subclass of rtl_opt_pass along with...
6029 (pass_data_split_for_shorten_branches): ...new pass_data instance
6031 (make_pass_split_for_shorten_branches): ...new function.
6032 * ree.c (pass_ree): Convert from a global struct to a subclass of
6033 rtl_opt_pass along with...
6034 (pass_data_ree): ...new pass_data instance and...
6035 (make_pass_ree): ...new function.
6036 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6037 subclass of rtl_opt_pass along with...
6038 (pass_data_stack_regs): ...new pass_data instance and...
6039 (make_pass_stack_regs): ...new function.
6040 (pass_stack_regs_run): Convert from a global struct to a subclass of
6041 rtl_opt_pass along with...
6042 (pass_data_stack_regs_run): ...new pass_data instance and...
6043 (make_pass_stack_regs_run): ...new function.
6044 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6045 subclass of rtl_opt_pass along with...
6046 (pass_data_cprop_hardreg): ...new pass_data instance and...
6047 (make_pass_cprop_hardreg): ...new function.
6048 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6049 subclass of rtl_opt_pass along with...
6050 (pass_data_reginfo_init): ...new pass_data instance and...
6051 (make_pass_reginfo_init): ...new function.
6052 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6053 of rtl_opt_pass along with...
6054 (pass_data_regmove): ...new pass_data instance and...
6055 (make_pass_regmove): ...new function.
6056 * regrename.c (pass_regrename): Convert from a global struct to a
6057 subclass of rtl_opt_pass along with...
6058 (pass_data_regrename): ...new pass_data instance and...
6059 (make_pass_regrename): ...new function.
6060 * reorg.c (pass_delay_slots): Convert from a global struct to a
6061 subclass of rtl_opt_pass along with...
6062 (pass_data_delay_slots): ...new pass_data instance and...
6063 (make_pass_delay_slots): ...new function.
6064 (pass_machine_reorg): Convert from a global struct to a subclass of
6065 rtl_opt_pass along with...
6066 (pass_data_machine_reorg): ...new pass_data instance and...
6067 (make_pass_machine_reorg): ...new function.
6068 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6069 of rtl_opt_pass along with...
6070 (pass_data_sched): ...new pass_data instance and...
6071 (make_pass_sched): ...new function.
6072 (pass_sched2): Convert from a global struct to a subclass of
6073 rtl_opt_pass along with...
6074 (pass_data_sched2): ...new pass_data instance and...
6075 (make_pass_sched2): ...new function.
6076 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6077 to a subclass of rtl_opt_pass along with...
6078 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6079 (make_pass_stack_ptr_mod): ...new function.
6080 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6081 to a subclass of rtl_opt_pass along with...
6082 (pass_data_rtl_store_motion): ...new pass_data instance and...
6083 (make_pass_rtl_store_motion): ...new function.
6084 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6085 of gimple_opt_pass along with...
6086 (pass_data_tracer): ...new pass_data instance and...
6087 (make_pass_tracer): ...new function.
6088 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6089 to a subclass of gimple_opt_pass along with...
6090 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6091 (make_pass_diagnose_tm_blocks): ...new function.
6092 (pass_lower_tm): Convert from a global struct to a subclass of
6093 gimple_opt_pass along with...
6094 (pass_data_lower_tm): ...new pass_data instance and...
6095 (make_pass_lower_tm): ...new function.
6096 (pass_tm_init): Convert from a global struct to a subclass of
6097 gimple_opt_pass along with...
6098 (pass_data_tm_init): ...new pass_data instance and...
6099 (make_pass_tm_init): ...new function.
6100 (pass_tm_mark): Convert from a global struct to a subclass of
6101 gimple_opt_pass along with...
6102 (pass_data_tm_mark): ...new pass_data instance and...
6103 (make_pass_tm_mark): ...new function.
6104 (pass_tm_edges): Convert from a global struct to a subclass of
6105 gimple_opt_pass along with...
6106 (pass_data_tm_edges): ...new pass_data instance and...
6107 (make_pass_tm_edges): ...new function.
6108 (pass_tm_memopt): Convert from a global struct to a subclass of
6109 gimple_opt_pass along with...
6110 (pass_data_tm_memopt): ...new pass_data instance and...
6111 (make_pass_tm_memopt): ...new function.
6112 (pass_ipa_tm): Convert from a global struct to a subclass of
6113 simple_ipa_opt_pass along with...
6114 (pass_data_ipa_tm): ...new pass_data instance and...
6115 (make_pass_ipa_tm): ...new function.
6116 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6117 subclass of gimple_opt_pass along with...
6118 (pass_data_call_cdce): ...new pass_data instance and...
6119 (make_pass_call_cdce): ...new function.
6120 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6121 subclass of gimple_opt_pass along with...
6122 (pass_data_build_cfg): ...new pass_data instance and...
6123 (make_pass_build_cfg): ...new function.
6124 (pass_split_crit_edges): Convert from a global struct to a subclass of
6125 gimple_opt_pass along with...
6126 (pass_data_split_crit_edges): ...new pass_data instance and...
6127 (make_pass_split_crit_edges): ...new function.
6128 (pass_warn_function_return): Convert from a global struct to a
6129 subclass of gimple_opt_pass along with...
6130 (pass_data_warn_function_return): ...new pass_data instance and...
6131 (make_pass_warn_function_return): ...new function.
6132 (pass_warn_function_noreturn): Convert from a global struct to a
6133 subclass of gimple_opt_pass along with...
6134 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6135 (make_pass_warn_function_noreturn): ...new function.
6136 (pass_warn_unused_result): Convert from a global struct to a subclass
6137 of gimple_opt_pass along with...
6138 (pass_data_warn_unused_result): ...new pass_data instance and...
6139 (make_pass_warn_unused_result): ...new function.
6140 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6141 a subclass of gimple_opt_pass along with...
6142 (pass_data_merge_phi): ...new pass_data instance and...
6143 (make_pass_merge_phi): ...new function.
6144 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6145 a subclass of gimple_opt_pass along with...
6146 (pass_data_lower_complex): ...new pass_data instance and...
6147 (make_pass_lower_complex): ...new function.
6148 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6149 gimple_opt_pass along with...
6150 (pass_data_lower_complex_O0): ...new pass_data instance and...
6151 (make_pass_lower_complex_O0): ...new function.
6152 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6153 subclass of gimple_opt_pass along with...
6154 (pass_data_lower_eh): ...new pass_data instance and...
6155 (make_pass_lower_eh): ...new function.
6156 (pass_refactor_eh): Convert from a global struct to a subclass of
6157 gimple_opt_pass along with...
6158 (pass_data_refactor_eh): ...new pass_data instance and...
6159 (make_pass_refactor_eh): ...new function.
6160 (pass_lower_resx): Convert from a global struct to a subclass of
6161 gimple_opt_pass along with...
6162 (pass_data_lower_resx): ...new pass_data instance and...
6163 (make_pass_lower_resx): ...new function.
6164 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6165 of gimple_opt_pass along with...
6166 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6167 (make_pass_lower_eh_dispatch): ...new function.
6168 (pass_cleanup_eh): Convert from a global struct to a subclass of
6169 gimple_opt_pass along with...
6170 (pass_data_cleanup_eh): ...new pass_data instance and...
6171 (make_pass_cleanup_eh): ...new function.
6172 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6173 to a subclass of simple_ipa_opt_pass along with...
6174 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6175 (make_pass_ipa_lower_emutls): ...new function.
6176 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6177 a subclass of gimple_opt_pass along with...
6178 (pass_data_if_conversion): ...new pass_data instance and...
6179 (make_pass_if_conversion): ...new function.
6180 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6181 subclass of gimple_opt_pass along with...
6182 (pass_data_build_ssa): ...new pass_data instance and...
6183 (make_pass_build_ssa): ...new function.
6184 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6185 global struct to a subclass of gimple_opt_pass along with...
6186 (pass_data_loop_distribution): ...new pass_data instance and...
6187 (make_pass_loop_distribution): ...new function.
6188 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6189 subclass of gimple_opt_pass along with...
6190 (pass_data_mudflap_1): ...new pass_data instance and...
6191 (make_pass_mudflap_1): ...new function.
6192 (pass_mudflap_2): Convert from a global struct to a subclass of
6193 gimple_opt_pass along with...
6194 (pass_data_mudflap_2): ...new pass_data instance and...
6195 (make_pass_mudflap_2): ...new function.
6196 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6197 subclass of gimple_opt_pass along with...
6198 (pass_data_mudflap_1): ...new pass_data instance and...
6199 (make_pass_mudflap_1): ...new function.
6200 (pass_mudflap_2): Convert from a global struct to a subclass of
6201 gimple_opt_pass along with...
6202 (pass_data_mudflap_2): ...new pass_data instance and...
6203 (make_pass_mudflap_2): ...new function.
6204 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6205 gimple_opt_pass along with...
6206 (pass_data_nrv): ...new pass_data instance and...
6207 (make_pass_nrv): ...new function.
6208 (pass_return_slot): Convert from a global struct to a subclass of
6209 gimple_opt_pass along with...
6210 (pass_data_return_slot): ...new pass_data instance and...
6211 (make_pass_return_slot): ...new function.
6212 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6213 to a subclass of gimple_opt_pass along with...
6214 (pass_data_object_sizes): ...new pass_data instance and...
6215 (make_pass_object_sizes): ...new function.
6216 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6217 global struct to a subclass of gimple_opt_pass along with...
6218 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6220 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6221 (pass_fixup_cfg): Convert from a global struct to a subclass of
6222 gimple_opt_pass along with...
6223 (pass_data_fixup_cfg): ...new pass_data instance and...
6224 (make_pass_fixup_cfg): ...new function.
6225 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6226 (make_pass_mudflap_1): ...new function.
6227 (pass_mudflap_2): Replace declaration with that of...
6228 (make_pass_mudflap_2): ...new function.
6229 (pass_asan): Replace declaration with that of...
6230 (make_pass_asan): ...new function.
6231 (pass_asan_O0): Replace declaration with that of...
6232 (make_pass_asan_O0): ...new function.
6233 (pass_tsan): Replace declaration with that of...
6234 (make_pass_tsan): ...new function.
6235 (pass_tsan_O0): Replace declaration with that of...
6236 (make_pass_tsan_O0): ...new function.
6237 (pass_lower_cf): Replace declaration with that of...
6238 (make_pass_lower_cf): ...new function.
6239 (pass_refactor_eh): Replace declaration with that of...
6240 (make_pass_refactor_eh): ...new function.
6241 (pass_lower_eh): Replace declaration with that of...
6242 (make_pass_lower_eh): ...new function.
6243 (pass_lower_eh_dispatch): Replace declaration with that of...
6244 (make_pass_lower_eh_dispatch): ...new function.
6245 (pass_lower_resx): Replace declaration with that of...
6246 (make_pass_lower_resx): ...new function.
6247 (pass_build_cfg): Replace declaration with that of...
6248 (make_pass_build_cfg): ...new function.
6249 (pass_early_tree_profile): Replace declaration with that of...
6250 (make_pass_early_tree_profile): ...new function.
6251 (pass_cleanup_eh): Replace declaration with that of...
6252 (make_pass_cleanup_eh): ...new function.
6253 (pass_sra): Replace declaration with that of...
6254 (make_pass_sra): ...new function.
6255 (pass_sra_early): Replace declaration with that of...
6256 (make_pass_sra_early): ...new function.
6257 (pass_early_ipa_sra): Replace declaration with that of...
6258 (make_pass_early_ipa_sra): ...new function.
6259 (pass_tail_recursion): Replace declaration with that of...
6260 (make_pass_tail_recursion): ...new function.
6261 (pass_tail_calls): Replace declaration with that of...
6262 (make_pass_tail_calls): ...new function.
6263 (pass_tree_loop): Replace declaration with that of...
6264 (make_pass_tree_loop): ...new function.
6265 (pass_tree_loop_init): Replace declaration with that of...
6266 (make_pass_tree_loop_init): ...new function.
6267 (pass_lim): Replace declaration with that of...
6268 (make_pass_lim): ...new function.
6269 (pass_tree_unswitch): Replace declaration with that of...
6270 (make_pass_tree_unswitch): ...new function.
6271 (pass_predcom): Replace declaration with that of...
6272 (make_pass_predcom): ...new function.
6273 (pass_iv_canon): Replace declaration with that of...
6274 (make_pass_iv_canon): ...new function.
6275 (pass_scev_cprop): Replace declaration with that of...
6276 (make_pass_scev_cprop): ...new function.
6277 (pass_empty_loop): Replace declaration with that of...
6278 (make_pass_empty_loop): ...new function.
6279 (pass_record_bounds): Replace declaration with that of...
6280 (make_pass_record_bounds): ...new function.
6281 (pass_graphite): Replace declaration with that of...
6282 (make_pass_graphite): ...new function.
6283 (pass_graphite_transforms): Replace declaration with that of...
6284 (make_pass_graphite_transforms): ...new function.
6285 (pass_if_conversion): Replace declaration with that of...
6286 (make_pass_if_conversion): ...new function.
6287 (pass_loop_distribution): Replace declaration with that of...
6288 (make_pass_loop_distribution): ...new function.
6289 (pass_vectorize): Replace declaration with that of...
6290 (make_pass_vectorize): ...new function.
6291 (pass_slp_vectorize): Replace declaration with that of...
6292 (make_pass_slp_vectorize): ...new function.
6293 (pass_complete_unroll): Replace declaration with that of...
6294 (make_pass_complete_unroll): ...new function.
6295 (pass_complete_unrolli): Replace declaration with that of...
6296 (make_pass_complete_unrolli): ...new function.
6297 (pass_parallelize_loops): Replace declaration with that of...
6298 (make_pass_parallelize_loops): ...new function.
6299 (pass_loop_prefetch): Replace declaration with that of...
6300 (make_pass_loop_prefetch): ...new function.
6301 (pass_iv_optimize): Replace declaration with that of...
6302 (make_pass_iv_optimize): ...new function.
6303 (pass_tree_loop_done): Replace declaration with that of...
6304 (make_pass_tree_loop_done): ...new function.
6305 (pass_ch): Replace declaration with that of...
6306 (make_pass_ch): ...new function.
6307 (pass_ccp): Replace declaration with that of...
6308 (make_pass_ccp): ...new function.
6309 (pass_phi_only_cprop): Replace declaration with that of...
6310 (make_pass_phi_only_cprop): ...new function.
6311 (pass_build_ssa): Replace declaration with that of...
6312 (make_pass_build_ssa): ...new function.
6313 (pass_build_alias): Replace declaration with that of...
6314 (make_pass_build_alias): ...new function.
6315 (pass_build_ealias): Replace declaration with that of...
6316 (make_pass_build_ealias): ...new function.
6317 (pass_dominator): Replace declaration with that of...
6318 (make_pass_dominator): ...new function.
6319 (pass_dce): Replace declaration with that of...
6320 (make_pass_dce): ...new function.
6321 (pass_dce_loop): Replace declaration with that of...
6322 (make_pass_dce_loop): ...new function.
6323 (pass_cd_dce): Replace declaration with that of...
6324 (make_pass_cd_dce): ...new function.
6325 (pass_call_cdce): Replace declaration with that of...
6326 (make_pass_call_cdce): ...new function.
6327 (pass_merge_phi): Replace declaration with that of...
6328 (make_pass_merge_phi): ...new function.
6329 (pass_split_crit_edges): Replace declaration with that of...
6330 (make_pass_split_crit_edges): ...new function.
6331 (pass_pre): Replace declaration with that of...
6332 (make_pass_pre): ...new function.
6333 (pass_profile): Replace declaration with that of...
6334 (make_pass_profile): ...new function.
6335 (pass_strip_predict_hints): Replace declaration with that of...
6336 (make_pass_strip_predict_hints): ...new function.
6337 (pass_lower_complex_O0): Replace declaration with that of...
6338 (make_pass_lower_complex_O0): ...new function.
6339 (pass_lower_complex): Replace declaration with that of...
6340 (make_pass_lower_complex): ...new function.
6341 (pass_lower_vector): Replace declaration with that of...
6342 (make_pass_lower_vector): ...new function.
6343 (pass_lower_vector_ssa): Replace declaration with that of...
6344 (make_pass_lower_vector_ssa): ...new function.
6345 (pass_lower_omp): Replace declaration with that of...
6346 (make_pass_lower_omp): ...new function.
6347 (pass_diagnose_omp_blocks): Replace declaration with that of...
6348 (make_pass_diagnose_omp_blocks): ...new function.
6349 (pass_expand_omp): Replace declaration with that of...
6350 (make_pass_expand_omp): ...new function.
6351 (pass_expand_omp_ssa): Replace declaration with that of...
6352 (make_pass_expand_omp_ssa): ...new function.
6353 (pass_object_sizes): Replace declaration with that of...
6354 (make_pass_object_sizes): ...new function.
6355 (pass_strlen): Replace declaration with that of...
6356 (make_pass_strlen): ...new function.
6357 (pass_fold_builtins): Replace declaration with that of...
6358 (make_pass_fold_builtins): ...new function.
6359 (pass_stdarg): Replace declaration with that of...
6360 (make_pass_stdarg): ...new function.
6361 (pass_early_warn_uninitialized): Replace declaration with that of...
6362 (make_pass_early_warn_uninitialized): ...new function.
6363 (pass_late_warn_uninitialized): Replace declaration with that of...
6364 (make_pass_late_warn_uninitialized): ...new function.
6365 (pass_cse_reciprocals): Replace declaration with that of...
6366 (make_pass_cse_reciprocals): ...new function.
6367 (pass_cse_sincos): Replace declaration with that of...
6368 (make_pass_cse_sincos): ...new function.
6369 (pass_optimize_bswap): Replace declaration with that of...
6370 (make_pass_optimize_bswap): ...new function.
6371 (pass_optimize_widening_mul): Replace declaration with that of...
6372 (make_pass_optimize_widening_mul): ...new function.
6373 (pass_warn_function_return): Replace declaration with that of...
6374 (make_pass_warn_function_return): ...new function.
6375 (pass_warn_function_noreturn): Replace declaration with that of...
6376 (make_pass_warn_function_noreturn): ...new function.
6377 (pass_cselim): Replace declaration with that of...
6378 (make_pass_cselim): ...new function.
6379 (pass_phiopt): Replace declaration with that of...
6380 (make_pass_phiopt): ...new function.
6381 (pass_forwprop): Replace declaration with that of...
6382 (make_pass_forwprop): ...new function.
6383 (pass_phiprop): Replace declaration with that of...
6384 (make_pass_phiprop): ...new function.
6385 (pass_tree_ifcombine): Replace declaration with that of...
6386 (make_pass_tree_ifcombine): ...new function.
6387 (pass_dse): Replace declaration with that of...
6388 (make_pass_dse): ...new function.
6389 (pass_nrv): Replace declaration with that of...
6390 (make_pass_nrv): ...new function.
6391 (pass_rename_ssa_copies): Replace declaration with that of...
6392 (make_pass_rename_ssa_copies): ...new function.
6393 (pass_sink_code): Replace declaration with that of...
6394 (make_pass_sink_code): ...new function.
6395 (pass_fre): Replace declaration with that of...
6396 (make_pass_fre): ...new function.
6397 (pass_check_data_deps): Replace declaration with that of...
6398 (make_pass_check_data_deps): ...new function.
6399 (pass_copy_prop): Replace declaration with that of...
6400 (make_pass_copy_prop): ...new function.
6401 (pass_vrp): Replace declaration with that of...
6402 (make_pass_vrp): ...new function.
6403 (pass_uncprop): Replace declaration with that of...
6404 (make_pass_uncprop): ...new function.
6405 (pass_return_slot): Replace declaration with that of...
6406 (make_pass_return_slot): ...new function.
6407 (pass_reassoc): Replace declaration with that of...
6408 (make_pass_reassoc): ...new function.
6409 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6410 (make_pass_rebuild_cgraph_edges): ...new function.
6411 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6412 (make_pass_remove_cgraph_callee_edges): ...new function.
6413 (pass_build_cgraph_edges): Replace declaration with that of...
6414 (make_pass_build_cgraph_edges): ...new function.
6415 (pass_local_pure_const): Replace declaration with that of...
6416 (make_pass_local_pure_const): ...new function.
6417 (pass_tracer): Replace declaration with that of...
6418 (make_pass_tracer): ...new function.
6419 (pass_warn_unused_result): Replace declaration with that of...
6420 (make_pass_warn_unused_result): ...new function.
6421 (pass_diagnose_tm_blocks): Replace declaration with that of...
6422 (make_pass_diagnose_tm_blocks): ...new function.
6423 (pass_lower_tm): Replace declaration with that of...
6424 (make_pass_lower_tm): ...new function.
6425 (pass_tm_init): Replace declaration with that of...
6426 (make_pass_tm_init): ...new function.
6427 (pass_tm_mark): Replace declaration with that of...
6428 (make_pass_tm_mark): ...new function.
6429 (pass_tm_memopt): Replace declaration with that of...
6430 (make_pass_tm_memopt): ...new function.
6431 (pass_tm_edges): Replace declaration with that of...
6432 (make_pass_tm_edges): ...new function.
6433 (pass_split_functions): Replace declaration with that of...
6434 (make_pass_split_functions): ...new function.
6435 (pass_feedback_split_functions): Replace declaration with that of...
6436 (make_pass_feedback_split_functions): ...new function.
6437 (pass_strength_reduction): Replace declaration with that of...
6438 (make_pass_strength_reduction): ...new function.
6439 (pass_ipa_lower_emutls): Replace declaration with that of...
6440 (make_pass_ipa_lower_emutls): ...new function.
6441 (pass_ipa_function_and_variable_visibility): Replace declaration with
6443 (make_pass_ipa_function_and_variable_visibility): ...new function.
6444 (pass_ipa_tree_profile): Replace declaration with that of...
6445 (make_pass_ipa_tree_profile): ...new function.
6446 (pass_early_local_passes): Replace declaration with that of...
6447 (make_pass_early_local_passes): ...new function.
6448 (pass_ipa_whole_program_visibility): Replace declaration with that
6450 (make_pass_ipa_whole_program_visibility): ...new function.
6451 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6452 (make_pass_ipa_lto_gimple_out): ...new function.
6453 (pass_ipa_increase_alignment): Replace declaration with that of...
6454 (make_pass_ipa_increase_alignment): ...new function.
6455 (pass_ipa_inline): Replace declaration with that of...
6456 (make_pass_ipa_inline): ...new function.
6457 (pass_ipa_free_lang_data): Replace declaration with that of...
6458 (make_pass_ipa_free_lang_data): ...new function.
6459 (pass_ipa_free_inline_summary): Replace declaration with that of...
6460 (make_pass_ipa_free_inline_summary): ...new function.
6461 (pass_ipa_cp): Replace declaration with that of...
6462 (make_pass_ipa_cp): ...new function.
6463 (pass_ipa_reference): Replace declaration with that of...
6464 (make_pass_ipa_reference): ...new function.
6465 (pass_ipa_pure_const): Replace declaration with that of...
6466 (make_pass_ipa_pure_const): ...new function.
6467 (pass_ipa_pta): Replace declaration with that of...
6468 (make_pass_ipa_pta): ...new function.
6469 (pass_ipa_lto_finish_out): Replace declaration with that of...
6470 (make_pass_ipa_lto_finish_out): ...new function.
6471 (pass_ipa_tm): Replace declaration with that of...
6472 (make_pass_ipa_tm): ...new function.
6473 (pass_ipa_profile): Replace declaration with that of...
6474 (make_pass_ipa_profile): ...new function.
6475 (pass_ipa_cdtor_merge): Replace declaration with that of...
6476 (make_pass_ipa_cdtor_merge): ...new function.
6477 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6479 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6480 (pass_init_datastructures): Replace declaration with that of...
6481 (make_pass_init_datastructures): ...new function.
6482 (pass_fixup_cfg): Replace declaration with that of...
6483 (make_pass_fixup_cfg): ...new function.
6484 (pass_expand): Replace declaration with that of...
6485 (make_pass_expand): ...new function.
6486 (pass_instantiate_virtual_regs): Replace declaration with that of...
6487 (make_pass_instantiate_virtual_regs): ...new function.
6488 (pass_rtl_fwprop): Replace declaration with that of...
6489 (make_pass_rtl_fwprop): ...new function.
6490 (pass_rtl_fwprop_addr): Replace declaration with that of...
6491 (make_pass_rtl_fwprop_addr): ...new function.
6492 (pass_jump): Replace declaration with that of...
6493 (make_pass_jump): ...new function.
6494 (pass_jump2): Replace declaration with that of...
6495 (make_pass_jump2): ...new function.
6496 (pass_lower_subreg): Replace declaration with that of...
6497 (make_pass_lower_subreg): ...new function.
6498 (pass_cse): Replace declaration with that of...
6499 (make_pass_cse): ...new function.
6500 (pass_fast_rtl_dce): Replace declaration with that of...
6501 (make_pass_fast_rtl_dce): ...new function.
6502 (pass_ud_rtl_dce): Replace declaration with that of...
6503 (make_pass_ud_rtl_dce): ...new function.
6504 (pass_rtl_dce): Replace declaration with that of...
6505 (make_pass_rtl_dce): ...new function.
6506 (pass_rtl_dse1): Replace declaration with that of...
6507 (make_pass_rtl_dse1): ...new function.
6508 (pass_rtl_dse2): Replace declaration with that of...
6509 (make_pass_rtl_dse2): ...new function.
6510 (pass_rtl_dse3): Replace declaration with that of...
6511 (make_pass_rtl_dse3): ...new function.
6512 (pass_rtl_cprop): Replace declaration with that of...
6513 (make_pass_rtl_cprop): ...new function.
6514 (pass_rtl_pre): Replace declaration with that of...
6515 (make_pass_rtl_pre): ...new function.
6516 (pass_rtl_hoist): Replace declaration with that of...
6517 (make_pass_rtl_hoist): ...new function.
6518 (pass_rtl_store_motion): Replace declaration with that of...
6519 (make_pass_rtl_store_motion): ...new function.
6520 (pass_cse_after_global_opts): Replace declaration with that of...
6521 (make_pass_cse_after_global_opts): ...new function.
6522 (pass_rtl_ifcvt): Replace declaration with that of...
6523 (make_pass_rtl_ifcvt): ...new function.
6524 (pass_into_cfg_layout_mode): Replace declaration with that of...
6525 (make_pass_into_cfg_layout_mode): ...new function.
6526 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6527 (make_pass_outof_cfg_layout_mode): ...new function.
6528 (pass_loop2): Replace declaration with that of...
6529 (make_pass_loop2): ...new function.
6530 (pass_rtl_loop_init): Replace declaration with that of...
6531 (make_pass_rtl_loop_init): ...new function.
6532 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6533 (make_pass_rtl_move_loop_invariants): ...new function.
6534 (pass_rtl_unswitch): Replace declaration with that of...
6535 (make_pass_rtl_unswitch): ...new function.
6536 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6537 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6538 (pass_rtl_doloop): Replace declaration with that of...
6539 (make_pass_rtl_doloop): ...new function.
6540 (pass_rtl_loop_done): Replace declaration with that of...
6541 (make_pass_rtl_loop_done): ...new function.
6542 (pass_web): Replace declaration with that of...
6543 (make_pass_web): ...new function.
6544 (pass_cse2): Replace declaration with that of...
6545 (make_pass_cse2): ...new function.
6546 (pass_df_initialize_opt): Replace declaration with that of...
6547 (make_pass_df_initialize_opt): ...new function.
6548 (pass_df_initialize_no_opt): Replace declaration with that of...
6549 (make_pass_df_initialize_no_opt): ...new function.
6550 (pass_reginfo_init): Replace declaration with that of...
6551 (make_pass_reginfo_init): ...new function.
6552 (pass_inc_dec): Replace declaration with that of...
6553 (make_pass_inc_dec): ...new function.
6554 (pass_stack_ptr_mod): Replace declaration with that of...
6555 (make_pass_stack_ptr_mod): ...new function.
6556 (pass_initialize_regs): Replace declaration with that of...
6557 (make_pass_initialize_regs): ...new function.
6558 (pass_combine): Replace declaration with that of...
6559 (make_pass_combine): ...new function.
6560 (pass_if_after_combine): Replace declaration with that of...
6561 (make_pass_if_after_combine): ...new function.
6562 (pass_ree): Replace declaration with that of...
6563 (make_pass_ree): ...new function.
6564 (pass_partition_blocks): Replace declaration with that of...
6565 (make_pass_partition_blocks): ...new function.
6566 (pass_match_asm_constraints): Replace declaration with that of...
6567 (make_pass_match_asm_constraints): ...new function.
6568 (pass_regmove): Replace declaration with that of...
6569 (make_pass_regmove): ...new function.
6570 (pass_split_all_insns): Replace declaration with that of...
6571 (make_pass_split_all_insns): ...new function.
6572 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6573 (make_pass_fast_rtl_byte_dce): ...new function.
6574 (pass_lower_subreg2): Replace declaration with that of...
6575 (make_pass_lower_subreg2): ...new function.
6576 (pass_mode_switching): Replace declaration with that of...
6577 (make_pass_mode_switching): ...new function.
6578 (pass_sms): Replace declaration with that of...
6579 (make_pass_sms): ...new function.
6580 (pass_sched): Replace declaration with that of...
6581 (make_pass_sched): ...new function.
6582 (pass_ira): Replace declaration with that of...
6583 (make_pass_ira): ...new function.
6584 (pass_reload): Replace declaration with that of...
6585 (make_pass_reload): ...new function.
6586 (pass_clean_state): Replace declaration with that of...
6587 (make_pass_clean_state): ...new function.
6588 (pass_branch_prob): Replace declaration with that of...
6589 (make_pass_branch_prob): ...new function.
6590 (pass_value_profile_transformations): Replace declaration with that
6592 (make_pass_value_profile_transformations): ...new function.
6593 (pass_postreload_cse): Replace declaration with that of...
6594 (make_pass_postreload_cse): ...new function.
6595 (pass_gcse2): Replace declaration with that of...
6596 (make_pass_gcse2): ...new function.
6597 (pass_split_after_reload): Replace declaration with that of...
6598 (make_pass_split_after_reload): ...new function.
6599 (pass_branch_target_load_optimize1): Replace declaration with that
6601 (make_pass_branch_target_load_optimize1): ...new function.
6602 (pass_thread_prologue_and_epilogue): Replace declaration with that
6604 (make_pass_thread_prologue_and_epilogue): ...new function.
6605 (pass_stack_adjustments): Replace declaration with that of...
6606 (make_pass_stack_adjustments): ...new function.
6607 (pass_peephole2): Replace declaration with that of...
6608 (make_pass_peephole2): ...new function.
6609 (pass_if_after_reload): Replace declaration with that of...
6610 (make_pass_if_after_reload): ...new function.
6611 (pass_regrename): Replace declaration with that of...
6612 (make_pass_regrename): ...new function.
6613 (pass_cprop_hardreg): Replace declaration with that of...
6614 (make_pass_cprop_hardreg): ...new function.
6615 (pass_reorder_blocks): Replace declaration with that of...
6616 (make_pass_reorder_blocks): ...new function.
6617 (pass_branch_target_load_optimize2): Replace declaration with that
6619 (make_pass_branch_target_load_optimize2): ...new function.
6620 (pass_leaf_regs): Replace declaration with that of...
6621 (make_pass_leaf_regs): ...new function.
6622 (pass_split_before_sched2): Replace declaration with that of...
6623 (make_pass_split_before_sched2): ...new function.
6624 (pass_compare_elim_after_reload): Replace declaration with that of...
6625 (make_pass_compare_elim_after_reload): ...new function.
6626 (pass_sched2): Replace declaration with that of...
6627 (make_pass_sched2): ...new function.
6628 (pass_stack_regs): Replace declaration with that of...
6629 (make_pass_stack_regs): ...new function.
6630 (pass_stack_regs_run): Replace declaration with that of...
6631 (make_pass_stack_regs_run): ...new function.
6632 (pass_df_finish): Replace declaration with that of...
6633 (make_pass_df_finish): ...new function.
6634 (pass_compute_alignments): Replace declaration with that of...
6635 (make_pass_compute_alignments): ...new function.
6636 (pass_duplicate_computed_gotos): Replace declaration with that of...
6637 (make_pass_duplicate_computed_gotos): ...new function.
6638 (pass_variable_tracking): Replace declaration with that of...
6639 (make_pass_variable_tracking): ...new function.
6640 (pass_free_cfg): Replace declaration with that of...
6641 (make_pass_free_cfg): ...new function.
6642 (pass_machine_reorg): Replace declaration with that of...
6643 (make_pass_machine_reorg): ...new function.
6644 (pass_cleanup_barriers): Replace declaration with that of...
6645 (make_pass_cleanup_barriers): ...new function.
6646 (pass_delay_slots): Replace declaration with that of...
6647 (make_pass_delay_slots): ...new function.
6648 (pass_split_for_shorten_branches): Replace declaration with that of...
6649 (make_pass_split_for_shorten_branches): ...new function.
6650 (pass_split_before_regstack): Replace declaration with that of...
6651 (make_pass_split_before_regstack): ...new function.
6652 (pass_convert_to_eh_region_ranges): Replace declaration with that
6654 (make_pass_convert_to_eh_region_ranges): ...new function.
6655 (pass_shorten_branches): Replace declaration with that of...
6656 (make_pass_shorten_branches): ...new function.
6657 (pass_set_nothrow_function_flags): Replace declaration with that of...
6658 (make_pass_set_nothrow_function_flags): ...new function.
6659 (pass_dwarf2_frame): Replace declaration with that of...
6660 (make_pass_dwarf2_frame): ...new function.
6661 (pass_final): Replace declaration with that of...
6662 (make_pass_final): ...new function.
6663 (pass_rtl_seqabstr): Replace declaration with that of...
6664 (make_pass_rtl_seqabstr): ...new function.
6665 (pass_release_ssa_names): Replace declaration with that of...
6666 (make_pass_release_ssa_names): ...new function.
6667 (pass_early_inline): Replace declaration with that of...
6668 (make_pass_early_inline): ...new function.
6669 (pass_inline_parameters): Replace declaration with that of...
6670 (make_pass_inline_parameters): ...new function.
6671 (pass_update_address_taken): Replace declaration with that of...
6672 (make_pass_update_address_taken): ...new function.
6673 (pass_convert_switch): Replace declaration with that of...
6674 (make_pass_convert_switch): ...new function.
6675 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6676 to a subclass of simple_ipa_opt_pass along with...
6677 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6678 (make_pass_ipa_tree_profile): ...new function.
6679 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6680 subclass of gimple_opt_pass along with...
6681 (pass_data_sra_early): ...new pass_data instance and...
6682 (make_pass_sra_early): ...new function.
6683 (pass_sra): Convert from a global struct to a subclass of
6684 gimple_opt_pass along with...
6685 (pass_data_sra): ...new pass_data instance and...
6686 (make_pass_sra): ...new function.
6687 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6688 gimple_opt_pass along with...
6689 (pass_data_early_ipa_sra): ...new pass_data instance and...
6690 (make_pass_early_ipa_sra): ...new function.
6691 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6692 subclass of gimple_opt_pass along with...
6693 (pass_data_ccp): ...new pass_data instance and...
6694 (make_pass_ccp): ...new function.
6695 (pass_fold_builtins): Convert from a global struct to a subclass of
6696 gimple_opt_pass along with...
6697 (pass_data_fold_builtins): ...new pass_data instance and...
6698 (make_pass_fold_builtins): ...new function.
6699 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6700 subclass of gimple_opt_pass along with...
6701 (pass_data_copy_prop): ...new pass_data instance and...
6702 (make_pass_copy_prop): ...new function.
6703 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6704 global struct to a subclass of gimple_opt_pass along with...
6705 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6706 (make_pass_rename_ssa_copies): ...new function.
6707 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6708 subclass of gimple_opt_pass along with...
6709 (pass_data_dce): ...new pass_data instance and...
6710 (make_pass_dce): ...new function.
6711 (pass_dce_loop): Convert from a global struct to a subclass of
6712 gimple_opt_pass along with...
6713 (pass_data_dce_loop): ...new pass_data instance and...
6714 (make_pass_dce_loop): ...new function.
6715 (pass_cd_dce): Convert from a global struct to a subclass of
6716 gimple_opt_pass along with...
6717 (pass_data_cd_dce): ...new pass_data instance and...
6718 (make_pass_cd_dce): ...new function.
6719 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6720 subclass of gimple_opt_pass along with...
6721 (pass_data_dominator): ...new pass_data instance and...
6722 (make_pass_dominator): ...new function.
6723 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6724 gimple_opt_pass along with...
6725 (pass_data_phi_only_cprop): ...new pass_data instance and...
6726 (make_pass_phi_only_cprop): ...new function.
6727 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6728 subclass of gimple_opt_pass along with...
6729 (pass_data_dse): ...new pass_data instance and...
6730 (make_pass_dse): ...new function.
6731 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6732 a subclass of gimple_opt_pass along with...
6733 (pass_data_forwprop): ...new pass_data instance and...
6734 (make_pass_forwprop): ...new function.
6735 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6736 struct to a subclass of gimple_opt_pass along with...
6737 (pass_data_tree_ifcombine): ...new pass_data instance and...
6738 (make_pass_tree_ifcombine): ...new function.
6739 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6740 subclass of gimple_opt_pass along with...
6741 (pass_data_ch): ...new pass_data instance and...
6742 (make_pass_ch): ...new function.
6743 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6744 subclass of gimple_opt_pass along with...
6745 (pass_data_tree_loop): ...new pass_data instance and...
6746 (make_pass_tree_loop): ...new function.
6747 (pass_tree_loop_init): Convert from a global struct to a subclass of
6748 gimple_opt_pass along with...
6749 (pass_data_tree_loop_init): ...new pass_data instance and...
6750 (make_pass_tree_loop_init): ...new function.
6751 (pass_lim): Convert from a global struct to a subclass of
6752 gimple_opt_pass along with...
6753 (pass_data_lim): ...new pass_data instance and...
6754 (make_pass_lim): ...new function.
6755 (pass_tree_unswitch): Convert from a global struct to a subclass of
6756 gimple_opt_pass along with...
6757 (pass_data_tree_unswitch): ...new pass_data instance and...
6758 (make_pass_tree_unswitch): ...new function.
6759 (pass_predcom): Convert from a global struct to a subclass of
6760 gimple_opt_pass along with...
6761 (pass_data_predcom): ...new pass_data instance and...
6762 (make_pass_predcom): ...new function.
6763 (pass_vectorize): Convert from a global struct to a subclass of
6764 gimple_opt_pass along with...
6765 (pass_data_vectorize): ...new pass_data instance and...
6766 (make_pass_vectorize): ...new function.
6767 (pass_graphite): Convert from a global struct to a subclass of
6768 gimple_opt_pass along with...
6769 (pass_data_graphite): ...new pass_data instance and...
6770 (make_pass_graphite): ...new function.
6771 (pass_graphite_transforms): Convert from a global struct to a subclass
6772 of gimple_opt_pass along with...
6773 (pass_data_graphite_transforms): ...new pass_data instance and...
6774 (make_pass_graphite_transforms): ...new function.
6775 (pass_check_data_deps): Convert from a global struct to a subclass of
6776 gimple_opt_pass along with...
6777 (pass_data_check_data_deps): ...new pass_data instance and...
6778 (make_pass_check_data_deps): ...new function.
6779 (pass_iv_canon): Convert from a global struct to a subclass of
6780 gimple_opt_pass along with...
6781 (pass_data_iv_canon): ...new pass_data instance and...
6782 (make_pass_iv_canon): ...new function.
6783 (pass_scev_cprop): Convert from a global struct to a subclass of
6784 gimple_opt_pass along with...
6785 (pass_data_scev_cprop): ...new pass_data instance and...
6786 (make_pass_scev_cprop): ...new function.
6787 (pass_record_bounds): Convert from a global struct to a subclass of
6788 gimple_opt_pass along with...
6789 (pass_data_record_bounds): ...new pass_data instance and...
6790 (make_pass_record_bounds): ...new function.
6791 (pass_complete_unroll): Convert from a global struct to a subclass of
6792 gimple_opt_pass along with...
6793 (pass_data_complete_unroll): ...new pass_data instance and...
6794 (make_pass_complete_unroll): ...new function.
6795 (pass_complete_unrolli): Convert from a global struct to a subclass of
6796 gimple_opt_pass along with...
6797 (pass_data_complete_unrolli): ...new pass_data instance and...
6798 (make_pass_complete_unrolli): ...new function.
6799 (pass_parallelize_loops): Convert from a global struct to a subclass
6800 of gimple_opt_pass along with...
6801 (pass_data_parallelize_loops): ...new pass_data instance and...
6802 (make_pass_parallelize_loops): ...new function.
6803 (pass_loop_prefetch): Convert from a global struct to a subclass of
6804 gimple_opt_pass along with...
6805 (pass_data_loop_prefetch): ...new pass_data instance and...
6806 (make_pass_loop_prefetch): ...new function.
6807 (pass_iv_optimize): Convert from a global struct to a subclass of
6808 gimple_opt_pass along with...
6809 (pass_data_iv_optimize): ...new pass_data instance and...
6810 (make_pass_iv_optimize): ...new function.
6811 (pass_tree_loop_done): Convert from a global struct to a subclass of
6812 gimple_opt_pass along with...
6813 (pass_data_tree_loop_done): ...new pass_data instance and...
6814 (make_pass_tree_loop_done): ...new function.
6815 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6816 struct to a subclass of gimple_opt_pass along with...
6817 (pass_data_cse_reciprocals): ...new pass_data instance and...
6818 (make_pass_cse_reciprocals): ...new function.
6819 (pass_cse_sincos): Convert from a global struct to a subclass of
6820 gimple_opt_pass along with...
6821 (pass_data_cse_sincos): ...new pass_data instance and...
6822 (make_pass_cse_sincos): ...new function.
6823 (pass_optimize_bswap): Convert from a global struct to a subclass of
6824 gimple_opt_pass along with...
6825 (pass_data_optimize_bswap): ...new pass_data instance and...
6826 (make_pass_optimize_bswap): ...new function.
6827 (pass_optimize_widening_mul): Convert from a global struct to a
6828 subclass of gimple_opt_pass along with...
6829 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6830 (make_pass_optimize_widening_mul): ...new function.
6831 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6832 subclass of gimple_opt_pass along with...
6833 (pass_data_phiopt): ...new pass_data instance and...
6834 (make_pass_phiopt): ...new function.
6835 (pass_cselim): Convert from a global struct to a subclass of
6836 gimple_opt_pass along with...
6837 (pass_data_cselim): ...new pass_data instance and...
6838 (make_pass_cselim): ...new function.
6839 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6840 subclass of gimple_opt_pass along with...
6841 (pass_data_phiprop): ...new pass_data instance and...
6842 (make_pass_phiprop): ...new function.
6843 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6844 subclass of gimple_opt_pass along with...
6845 (pass_data_pre): ...new pass_data instance and...
6846 (make_pass_pre): ...new function.
6847 (pass_fre): Convert from a global struct to a subclass of
6848 gimple_opt_pass along with...
6849 (pass_data_fre): ...new pass_data instance and...
6850 (make_pass_fre): ...new function.
6851 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
6852 subclass of gimple_opt_pass along with...
6853 (pass_data_reassoc): ...new pass_data instance and...
6854 (make_pass_reassoc): ...new function.
6855 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
6856 subclass of gimple_opt_pass along with...
6857 (pass_data_sink_code): ...new pass_data instance and...
6858 (make_pass_sink_code): ...new function.
6859 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
6860 subclass of gimple_opt_pass along with...
6861 (pass_data_strlen): ...new pass_data instance and...
6862 (make_pass_strlen): ...new function.
6863 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
6864 struct to a subclass of gimple_opt_pass along with...
6865 (pass_data_build_alias): ...new pass_data instance and...
6866 (make_pass_build_alias): ...new function.
6867 (pass_build_ealias): Convert from a global struct to a subclass of
6868 gimple_opt_pass along with...
6869 (pass_data_build_ealias): ...new pass_data instance and...
6870 (make_pass_build_ealias): ...new function.
6871 (pass_ipa_pta): Convert from a global struct to a subclass of
6872 simple_ipa_opt_pass along with...
6873 (pass_data_ipa_pta): ...new pass_data instance and...
6874 (make_pass_ipa_pta): ...new function.
6875 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
6876 subclass of gimple_opt_pass along with...
6877 (pass_data_uncprop): ...new pass_data instance and...
6878 (make_pass_uncprop): ...new function.
6879 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
6880 global struct to a subclass of gimple_opt_pass along with...
6881 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
6882 (make_pass_late_warn_uninitialized): ...new function.
6883 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
6884 to a subclass of gimple_opt_pass along with...
6885 (pass_data_init_datastructures): ...new pass_data instance and...
6886 (make_pass_init_datastructures): ...new function.
6887 (pass_early_warn_uninitialized): Convert from a global struct to a
6888 subclass of gimple_opt_pass along with...
6889 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
6890 (make_pass_early_warn_uninitialized): ...new function.
6891 (pass_update_address_taken): Convert from a global struct to a
6892 subclass of gimple_opt_pass along with...
6893 (pass_data_update_address_taken): ...new pass_data instance and...
6894 (make_pass_update_address_taken): ...new function.
6895 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
6896 struct to a subclass of gimple_opt_pass along with...
6897 (pass_data_release_ssa_names): ...new pass_data instance and...
6898 (make_pass_release_ssa_names): ...new function.
6899 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
6900 subclass of gimple_opt_pass along with...
6901 (pass_data_stdarg): ...new pass_data instance and...
6902 (make_pass_stdarg): ...new function.
6903 * tree-switch-conversion.c (pass_convert_switch): Convert from a
6904 global struct to a subclass of gimple_opt_pass along with...
6905 (pass_data_convert_switch): ...new pass_data instance and...
6906 (make_pass_convert_switch): ...new function.
6907 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
6908 to a subclass of gimple_opt_pass along with...
6909 (pass_data_tail_recursion): ...new pass_data instance and...
6910 (make_pass_tail_recursion): ...new function.
6911 (pass_tail_calls): Convert from a global struct to a subclass of
6912 gimple_opt_pass along with...
6913 (pass_data_tail_calls): ...new pass_data instance and...
6914 (make_pass_tail_calls): ...new function.
6915 * tree-vect-generic.c (pass_lower_vector): Convert from a global
6916 struct to a subclass of gimple_opt_pass along with...
6917 (pass_data_lower_vector): ...new pass_data instance and...
6918 (make_pass_lower_vector): ...new function.
6919 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
6920 gimple_opt_pass along with...
6921 (pass_data_lower_vector_ssa): ...new pass_data instance and...
6922 (make_pass_lower_vector_ssa): ...new function.
6923 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
6924 to a subclass of gimple_opt_pass along with...
6925 (pass_data_slp_vectorize): ...new pass_data instance and...
6926 (make_pass_slp_vectorize): ...new function.
6927 (pass_ipa_increase_alignment): Convert from a global struct to a
6928 subclass of simple_ipa_opt_pass along with...
6929 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
6930 (make_pass_ipa_increase_alignment): ...new function.
6931 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
6932 gimple_opt_pass along with...
6933 (pass_data_vrp): ...new pass_data instance and...
6934 (make_pass_vrp): ...new function.
6935 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
6936 subclass of simple_ipa_opt_pass along with...
6937 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
6938 (make_pass_ipa_free_lang_data): ...new function.
6939 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
6940 gimple_opt_pass along with...
6941 (pass_data_tsan): ...new pass_data instance and...
6942 (make_pass_tsan): ...new function.
6943 (pass_tsan_O0): Convert from a global struct to a subclass of
6944 gimple_opt_pass along with...
6945 (pass_data_tsan_O0): ...new pass_data instance and...
6946 (make_pass_tsan_O0): ...new function.
6947 * var-tracking.c (pass_variable_tracking): Convert from a global
6948 struct to a subclass of rtl_opt_pass along with...
6949 (pass_data_variable_tracking): ...new pass_data instance and...
6950 (make_pass_variable_tracking): ...new function.
6951 * web.c (pass_web): Convert from a global struct to a subclass of
6952 rtl_opt_pass along with...
6953 (pass_data_web): ...new pass_data instance and...
6954 (make_pass_web): ...new function.
6955 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
6956 declaration with that of...
6957 (make_pass_mode_switch_use): ...new function.
6958 (pass_resolve_sw_modes): Replace declaration with that of...
6959 (make_pass_resolve_sw_modes): ...new function.
6960 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
6961 from a global struct to a subclass of rtl_opt_pass along with...
6962 (pass_data_mode_switch_use): ...new pass_data instance and...
6963 (make_pass_mode_switch_use): ...new function.
6964 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
6965 from a global struct to a subclass of rtl_opt_pass along with...
6966 (pass_data_resolve_sw_modes): ...new pass_data instance and...
6967 (make_pass_resolve_sw_modes): ...new function.
6968 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
6969 struct to a subclass of rtl_opt_pass along with...
6970 (pass_data_insert_vzeroupper): ...new pass_data instance and...
6971 (make_pass_insert_vzeroupper): ...new function.
6972 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
6973 global struct to a subclass of rtl_opt_pass along with...
6974 (pass_data_work_around_errata): ...new pass_data instance and...
6975 (make_pass_work_around_errata): ...new function.
6976 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
6977 struct to a subclass of rtl_opt_pass along with...
6978 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
6979 (make_pass_mips_machine_reorg2): ...new function.
6981 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6983 * passes.c (pass_manager::operator new): New.
6985 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6987 Handwritten part of conversion of passes to C++ classes.
6989 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
6990 (toplev.o): Add dep on PASS_MANAGER_H.
6991 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
6992 of early local pases to reflect this moving from a global to a
6993 member of gcc::pass_manager.
6994 (cgraph_add_new_function): Likewise.
6995 * lto-cgraph.c (lto_output_node): Update for conversion of
6996 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
6997 * passes.c (opt_pass::clone): New.
6998 (opt_pass::gate): New.
6999 (opt_pass::execute): New.
7000 (opt_pass::opt_pass): New.
7001 (pass_manager::execute_early_local_passes): New.
7002 (pass_manager::execute_pass_mode_switching): new.
7003 (finish_optimization_passes): Convert to...
7004 (pass_manager::finish_optimization_passes): ...this.
7005 (finish_optimization_passes): Update for conversion of passes to
7007 (register_dump_files_1): Use has_gate since we cannot portably
7008 check a vtable entry against NULL.
7009 (dump_one_pass): Likewise.
7010 (ipa_write_summaries_2): Likewise.
7011 (ipa_write_optimization_summaries_1): Likewise.
7012 (ipa_read_summaries_1): Likewise.
7013 (ipa_read_optimization_summaries_1): Likewise.
7014 (execute_ipa_stmt_fixups): Likewise.
7015 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7016 pass-creation functions rather than wiring up globals, and
7017 storing the results in fields of pass_manager generated using
7019 (pass_manager::dump_profile_report): Update for conversion of
7020 passes to C++ classes.
7021 (pass_manager::execute_ipa_summary_passes): Likewise.
7022 (execute_one_ipa_transform_pass): Likewise.
7023 (execute_one_pass): Use has_gate and has_execute since we cannot
7024 portably check a vtable entry against NULL.
7025 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7026 (pass_manager): Use pass-instances.def to add fields for the
7027 various pass instances.
7028 * toplev.c (finalize): Update for move of
7029 finish_optimization_passes to a method of gcc::pass_manager.
7030 * toplev.h (finish_optimization_passes): Move to method of class
7032 * tree-pass.h (struct pass_data): New.
7033 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7034 (opt_pass::gate): Convert to virtual function.
7035 (opt_pass::~opt_pass): New.
7036 (opt_pass::clone): New.
7037 (opt_pass::execute): Convert to virtual function.
7038 (opt_pass::opt_pass): New.
7039 (opt_pass::ctxt_): new.
7040 (gimple_opt_pass): Convert to subclass of opt_pass.
7041 (gimple_opt_pass::gimple_opt_pass): New.
7042 (rtl_opt_pass): Convert to subclass of opt_pass.
7043 (rtl_opt_pass::rtl_opt_pass): New.
7044 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7045 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7046 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7047 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7048 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7049 invocation of pass_mode_switching to reflect this moving from a
7050 global to a member of gcc::pass_manager.
7051 (ix86_option_override): Rework how pass_insert_vzeroupper is
7052 added to the pass_manager to reflect autogenerated changes.
7053 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7055 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7057 PR rtl-optimization/57708
7058 * recog.c (peep2_find_free_register): Validate all regs in a
7061 2013-08-05 Jan Hubicka <jh@suse.cz>
7064 * cgraph.c (verify_cgraph_node): Accept local flags from other
7066 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7067 (function_and_variable_visibility): Likewise.
7068 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7070 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7072 * graph.c (init_graph_slim_pretty_print): Remove.
7073 (print_graph_cfg): Do not call it. Use local pretty printer.
7074 (start_graph_dump): Likewise.
7076 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7078 * gimple-pretty-print.c (buffer): Remove.
7079 (initialized): Likewise.
7080 (maybe_init_pretty_print): Likewise.
7081 (print_gimple_stmt): Do not call it. Use non-static local
7082 pretty_printer variable.
7083 (print_gimple_expr): Likewise.
7084 (print_gimple_seq): Likewise.
7085 (gimple_dump_bb): Likewise.
7087 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7089 * asan.c (asan_pp): Remove.
7090 (asan_pp_initialized): Likewise.
7091 (asan_pp_initialize): Likewise.
7092 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7093 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7094 (asan_add_global): Likewise.
7096 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7098 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7099 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7100 * pretty-print.c (pp_formatted_text_data): Likewise.
7101 (pp_indent): Rename from pp_base_indent.
7102 (pp_format): Rename from pp_base_format.
7103 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7104 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7105 (pp_flush): Rename from pp_base_flush.
7106 (pp_set_line_maximum_length): Rename from
7107 pp_base_set_line_maximum_length.
7108 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7109 (pp_set_prefix): Rename from pp_base_set_prefix.
7110 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7111 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7112 (pp_append_text): Rename from pp_base_append_text.
7113 (pp_formatted_text): Rename from pp_base_formatted_text.
7114 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7115 (pp_remaining_character_count_for_line): Rename from
7116 pp_base_remaining_character_count_for_line.
7117 (pp_newline): Rename from pp_base_newline.
7118 (pp_character): Rename from pp_base_character.
7119 (pp_string): Rename from pp_base_string.
7120 (pp_maybe_space): Rename from pp_base_maybe_space.
7121 * asan.c (asan_pp_string): Adjust.
7122 (asan_emit_stack_protection): Likewise.
7123 (asan_add_global): Likewise.
7124 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7125 * tree-mudflap.c (mf_varname_tree): Likewise.
7126 * tree-pretty-print.c (pp_tree_identifier): Rename from
7127 pp_base_tree_identifier.
7128 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7129 Declare as function.
7131 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7133 * pretty-print.h (pp_bar_bar): New.
7134 (pp_ampersand_ampersand): Likewise.
7135 (pp_less_equal): Likewise.
7136 (pp_greater_equal): Likewise.
7137 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7138 printer functions instead of pp_string or operators and punctuators.
7139 (dump_gimple_call): Likewise.
7140 (dump_gimple_omp_for): Likewise.
7141 (dump_gimple_transaction): Likewise.
7142 (dump_gimple_phi): Likewise.
7143 (pp_gimple_stmt_1): Likewise.
7144 * sched-vis.c (print_insn): Likewise.
7145 * tree-mudflap.c (mf_varname_tree): Likewise.
7146 * tree-pretty-print.c (dump_block_node): Likewise.
7147 (dump_generic_node): Likewise.
7149 2013-08-02 Jan Hubicka <jh@suse.cz>
7151 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7153 * lto-streamer-out.c (tree_is_indexable): Results decls and
7154 parm decls are not indexable.
7155 (DFS_write_tree_body): Do not follow args and results.
7156 (hash_tree): Likewise.
7157 (output_functions): Rearrange so struct function is needed
7158 only when real body is output; be able to also ouptut abstract
7159 functions; output DECL_ARGUMENTS and DECL_RESULT.
7160 (lto_output): When not in WPA, ale store abstract functions.
7161 (write_symbol): Do not care about RESULT_DECL.
7162 (output_symbol_p): Handle correctly sbtract decls.
7163 * lto-streamer-in.c (input_function): Rearrange so struct
7164 function can be NULL at entry; allow streaming of
7165 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7166 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7167 sanity check during LTO.
7168 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7169 RESULT_DECl and DECL_ARGUMENTS.
7170 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7173 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7175 * pretty-print.h (pp_underscore): New.
7177 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7178 printer functions instead of pp_character.
7179 (dump_binary_rhs): Likewise.
7180 (dump_ternary_rhs): Likewise.
7181 (dump_gimple_call_args): Likewise.
7182 (pp_points_to_solution): Likewise.
7183 (dump_gimple_call): Likewise.
7184 (dump_gimple_switch): Likewise.
7185 (dump_gimple_cond): Likewise.
7186 (dump_gimple_bind): Likewise.
7187 (dump_gimple_try): Likewise.
7188 (dump_gimple_omp_for): Likewise.
7189 (dump_gimple_omp_continue): Likewise.
7190 (dump_gimple_omp_single): Likewise.
7191 (dump_gimple_omp_sections): Likewise.
7192 (dump_gimple_omp_block): Likewise.
7193 (dump_gimple_omp_critical): Likewise.
7194 (dump_gimple_transaction): Likewise.
7195 (dump_gimple_asm): Likewise.
7196 (dump_gimple_phi): Likewise.
7197 (dump_gimple_omp_parallel): Likewise.
7198 (dump_gimple_omp_task): Likewise.
7199 (dump_gimple_omp_atomic_load): Likewise.
7200 (dump_gimple_omp_atomic_store): Likewise.
7201 (dump_gimple_mem_ops): Likewise.
7202 (pp_gimple_stmt_1): Likewise.
7203 (pp_cfg_jump): Likewise.
7204 (dump_implicit_edges): Likewise.
7205 (gimple_dump_bb_for_graph): Likewise.
7206 * graph.c (draw_cfg_node): Likewise.
7207 * langhooks.c (lhd_print_error_function): Likewise.
7208 * sched-vis.c (print_exp): Likewise.
7209 (print_value): Likewise.
7210 (print_pattern): Likewise.
7211 (print_insn): Likewise.
7212 (rtl_dump_bb_for_graph): Likewise.
7213 * tree-pretty-print.c (dump_function_declaration): Likewise.
7214 (dump_array_domain): Likewise.
7215 (dump_omp_clause): Likewise.
7216 (dump_location): Likewise.
7217 (dump_generic_node): Likewise.
7218 (print_struct_decl): Likewise.
7219 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7221 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7223 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7224 candidate table when replacing a candidate statement.
7225 (replace_rhs_if_not_dup): Likewise.
7226 (replace_one_candidate): Likewise.
7228 2013-08-02 Jan Hubicka <jh@suse.cz>
7229 Martin Liska <marxin.liska@gmail.com>
7231 * cgraphunit.c (add_new_function): Fix logic when adding from
7233 (assemble_thunk): Rename to ...
7234 (expand_thunk); .. this one; export; get it working with
7235 general functions; make produced gimple valid.
7236 * cgraph.h (expand_thunk): Declare.
7238 2013-08-02 Jan Hubicka <jh@suse.cz>
7240 * ipa-cp.c (gather_context_independent_values): Use
7241 ipa_get_param_move_cost.
7242 (get_replacement_map): Remove PARAM; move parameter folding
7244 (create_specialized_node): Update.
7245 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7246 assert that we have gimple body; update move_cost.
7247 (count_formal_params): Assert that we have gimple body.
7248 (ipa_dump_param): New function.
7249 (ipa_alloc_node_params): Break out from ...
7250 (ipa_initialize_node_params): ... here.
7251 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7252 (ipa_write_node_info): Stream move costs.
7253 (ipa_read_node_info): Read move costs.
7254 (ipa_update_after_lto_read): Do not recompute node params.
7255 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7256 (ipa_get_param): Check we are not in WPA.
7257 (ipa_get_param_move_cost): New.
7258 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7259 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7260 parm numbers to be present.
7262 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7264 PR rtl-optimization/58048
7265 * lra-constraints.c (process_alt_operands): Don't check asm
7266 operand on register.
7268 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7270 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7271 the implied StoreLoad barrier for atomic operations if before.
7273 2013-08-02 Jan Hubicka <jh@suse.cz>
7274 Martin Liska <marxin.liska@gmail.com>
7276 * cgraph.c (cgraph_function_body_availability): Do not check
7278 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7279 symtab_node_availability): Declare.
7280 * ipa.c (can_replace_by_local_alias): New.
7281 (function_and_variable_visibility): Use it.
7282 * symtab.c (symtab_for_node_and_aliases,
7283 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7285 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7287 PR rtl-optimization/57963
7288 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7289 (lra_constraints): Use them.
7291 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7293 * config/arm/types.md (define_attr "type"): Add "load_acq"
7295 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7297 (cortex_a53_store1): Likewise.
7299 2013-08-01 Jan Hubicka <jh@suse.cz>
7301 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7302 partitions are not needed.
7304 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7306 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7307 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7308 MAYBE_NON_Q_CLASS_P where appropriate.
7310 2013-08-01 Jan Hubicka <jh@suse.cz>
7312 * cgraph.h (release_function_body): Declare.
7313 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7314 of unused delcarations.
7316 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7318 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7319 RTL form when subtracting a constant.
7321 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7323 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7324 Generate canonical plus rtx with negated immediate instead of minus
7326 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7328 2013-08-01 Jan Hubicka <jh@suse.cz>
7330 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7331 (cgraph_release_function_body): Likewise.
7332 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7333 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7334 to used_as_abstract_origin.
7335 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7336 symbols used as abstract origins.
7337 * cgraphunit.c (analyze_functions): Update.
7338 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7339 used_as_abstract_origin.
7340 * tree-inline.c (tree_function_versioning): Update
7341 used_as_abstract_origin; be ready for DECL_RESULT and
7342 DECL_ARGUMENTS to be NULL.
7344 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7345 for abstract functions.
7346 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7349 2013-08-01 Jan Hubicka <jh@suse.cz>
7351 * profile.c (compute_value_histograms): Fix thinko.
7353 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7355 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7356 aarch-common-protos.h to extra_headers.
7357 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7358 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7359 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7361 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7363 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7364 Include "../arm/types.md". Define "type" attribute for all patterns.
7365 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7368 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7370 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7371 to support power8 load fusion.
7372 (fusion_gpr_mem_load): Likewise.
7374 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7376 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7377 declarations for power8 load fusion.
7378 (emit_fusion_gpr_load): Likewise.
7380 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7381 tuning for power8, turn on fusion mode by default. Turn on sign
7382 extending fusion mode if normal fusion mode is on, and we are at
7384 (fusion_gpr_load_p): New function, return true if we can fuse an
7385 addis instruction with a dependent load to a GPR.
7386 (emit_fusion_gpr_load): Emit the instructions for power8 load
7389 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7390 (VSX load fusion peepholes): New peepholes to fuse together an
7391 addi instruction with a VSX load instruction.
7393 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7394 peepholes to fuse an addis instruction with a load to a GPR base
7395 register. If we are supporting sign extending fusions, convert
7396 sign extending loads to zero extending loads and add an explicit
7399 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7401 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7402 aarch-common-protos.h to extra_headers.
7403 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7404 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7405 (arm_early_store_addr_dep): Likewise.
7406 (arm_no_early_alu_shift_dep): Likewise.
7407 (arm_no_early_alu_shift_value_dep): Likewise.
7408 (arm_no_early_mul_dep): Likewise.
7409 (arm_no_early_store_addr_dep): Likewise.
7410 (arm_mac_accumulator_is_mul_result): Likewise.
7411 (arm_mac_accumulator_is_result): Likewise.
7412 * config/arm/aarch-common.c: ... here. New file.
7413 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7415 (arm_early_store_addr_dep): Likewise.
7416 (arm_no_early_alu_shift_dep): Likewise.
7417 (arm_no_early_alu_shift_value_dep): Likewise.
7418 (arm_no_early_mul_dep): Likewise.
7419 (arm_no_early_store_addr_dep): Likewise.
7420 (arm_mac_accumulator_is_mul_result): Likewise.
7421 (arm_mac_accumulator_is_result): Likewise.
7422 * config/arm/aarch-common-protos.h: ... here. New file.
7423 * config/arm/t-arm (aarch-common.o): Define.
7425 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7427 * config/arm/arm.md: Include new file "types.md".
7428 (define_attr "type"): Move from here to ...
7429 (define_attr "mul32"): Likewise.
7430 (define_attr "mul64"): Likewise.
7431 * config/arm/types.md: ... here. New file.
7433 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7435 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7436 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7438 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7440 * gen-pass-instances.awk: Fix offset of substr().
7442 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7444 * Makefile.in (pass-instances.def): New.
7445 (passes.o): Replace dependency on passes.def with one on
7448 * gen-pass-instances.awk: New.
7450 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7451 rather than passes.def, updating local definition of NEXT_PASS
7452 macro to add an extra NUM parameter (currently unused).
7454 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7456 * Makefile.in (PASS_MANAGER_H): New.
7457 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7458 (passes.o): Likewise.
7459 (statistics.o): Likewise.
7460 (cgraphunit.o): Likewise.
7461 (context.o): Depend on PASS_MANAGER_H.
7463 * pass_manager.h: New.
7465 * cgraphunit.c (cgraph_add_new_function): Update for moves
7466 of globals to fields of pass_manager.
7467 (analyze_function): Likewise.
7468 (expand_function): Likewise.
7469 (ipa_passes): Likewise.
7470 (compile): Likewise.
7472 * context.c (context::context): New.
7473 * context.h (context::context): New.
7474 (context::get_passes): New.
7475 (context::passes_): New.
7477 * lto-cgraph.c (input_node): Update for moves of globals to
7478 fields of pass_manager.
7480 * passes.c (all_passes): Remove, in favor of a field of the
7481 same name within the new class pass_manager.
7482 (all_small_ipa_passes): Likewise.
7483 (all_lowering_passes): Likewise.
7484 (all_regular_ipa_passes): Likewise.
7485 (all_late_ipa_passes): Likewise.
7486 (all_lto_gen_passes): Likewise.
7487 (passes_by_id): Likewise.
7488 (passes_by_id_size): Likewise.
7489 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7490 the new class pass_manager.
7491 (set_pass_for_id): Convert to...
7492 (pass_manager::set_pass_for_id): ...method.
7493 (get_pass_for_id): Convert to...
7494 (pass_manager::get_pass_for_id): ...method.
7495 (register_one_dump_file): Move body of implementation into...
7496 (pass_manager::register_one_dump_file): ...here.
7497 (register_dump_files_1): Convert to...
7498 (pass_manager::register_dump_files_1): ...method.
7499 (register_dump_files): Convert to...
7500 (pass_manager::register_dump_files): ...method.
7501 (create_pass_tab): Update for moves of globals to fields of
7503 (dump_passes): Move body of implementation into...
7504 (pass_manager::dump_passes): ...here.
7505 (register_pass): Move body of implementation into...
7506 (pass_manager::register_pass): ...here.
7507 (init_optimization_passes): Convert into...
7508 (pass_manager::pass_manager): ...constructor for new
7509 pass_manager class, and initialize the pass_lists array.
7510 (check_profile_consistency): Update for moves of globals to
7511 fields of pass_manager.
7512 (dump_profile_report): Move body of implementation into...
7513 (pass_manager::dump_profile_report): ...here.
7514 (ipa_write_summaries_1): Update for moves of pass lists from
7515 being globals to fields of pass_manager.
7516 (ipa_write_optimization_summaries): Likewise.
7517 (ipa_read_summaries): Likewise.
7518 (ipa_read_optimization_summaries): Likewise.
7519 (execute_all_ipa_stmt_fixups): Likewise.
7521 * statistics.c (statistics_fini): Update for moves of globals to
7522 fields of pass_manager.
7524 * toplev.c (general_init): Replace call to
7525 init_optimization_passes with construction of the pass_manager
7528 * tree-pass.h (all_passes): Remove, in favor of a field of the
7529 same name within the new class pass_manager.
7530 (all_small_ipa_passes): Likewise.
7531 (all_lowering_passes): Likewise.
7532 (all_regular_ipa_passes): Likewise.
7533 (all_lto_gen_passes): Likewise.
7534 (all_late_ipa_passes): Likewise.
7535 (passes_by_id): Likewise.
7536 (passes_by_id_size): Likewise.
7537 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7538 the new class pass_manager.
7539 (get_pass_for_id): Remove.
7541 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7543 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7546 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7548 * arm.md (mulhi3): New expand pattern.
7550 2013-07-30 Jan Hubicka <jh@suse.cz>
7551 Martin Liska <marxin.liska@gmail.com>
7553 * profile.c (compute_value_histograms): Do not ICE when
7554 there is mismatch only on some counters.
7556 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7558 PR rtl-optimization/57637
7559 * function.c (move_insn_for_shrink_wrap): Also check the
7560 GEN set of the LIVE problem for the liveness analysis
7561 if it exists, otherwise give up.
7563 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7565 PR tree-optimization/57993
7566 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7567 replaced statement in the candidate table.
7568 (phi_add_costs): Return infinite cost when the hidden basis does
7569 not dominate all phis on which the candidate is dependent.
7570 (replace_one_candidate): Record replaced statement in the
7573 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7575 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7576 (ashlv2si3): New expander.
7577 (*ashlv2si3_i): New define_insn_and_split.
7578 * predicates.md (float_operation): Allow patterns with three
7581 PR rtl-optimization/58021
7582 * mode-switching.c (create_pre_exit): Always split off preceding
7583 insns if we are not at the basic block head.
7585 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7587 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7588 (UCLIBC_DYNAMIC_LINKER): New macro.
7589 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7591 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7592 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7594 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7595 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7596 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7597 for SF and DF modes. Use ieee_quad_format for TF mode.
7598 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7599 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7600 (mips_option_override): Handle `-mnan=legacy'.
7601 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7602 `-mabs=2008' and `-mnan=2008'.
7603 (OPTION_DEFAULT_SPECS): Add "nan" default.
7604 (ASM_SPEC): Handle `-mnan='.
7605 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7606 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7607 comment accordingly.
7608 (neg<mode>2): Likewise.
7609 * config/mips/mips.opt (mabs, mnan): New options.
7610 * doc/install.texi (Configuration): Document `--with-nan=' option.
7611 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7613 (MIPS Options): Document them.
7614 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7615 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7616 * configure: Regenerate.
7617 * config.in: Regenerate.
7619 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7621 * config/i386/i386.md (float post-reload splitters): Do not check
7622 for subregs of SSE registers.
7624 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7625 H.J. Lu <hongjiu.lu@intel.com>
7629 * config/i386/i386.md (post-reload splitter
7630 to avoid partial SSE reg dependency stalls): New pattern.
7632 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7634 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7636 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7638 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7639 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7641 2013-07-26 Cary Coutant <ccoutant@google.com>
7643 * dwarf2out.c (die_checksum_ordered): Don't include template
7644 instantiations in signature.
7645 (is_template_parameter): New function.
7646 (is_template_instantiation): New function.
7647 (generate_skeleton_bottom_up): Don't include template instantiations
7649 (generate_skeleton): Likewise.
7650 (break_out_comdat_types): Move recursive call to break out nested
7652 (prune_unused_types_mark_generic_parms_dies): Call
7653 is_template_parameter.
7655 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7657 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7658 uses vector registers.
7659 * config/aarch64/iterators.md: Add attributes rtn and vas.
7661 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7662 Richard Earnshaw <richard.earnshaw@arm.com>
7664 * combine.c (simplify_comparison): Re-canonicalize operands
7666 * config/arm/arm.md (movcond_addsi): New splitter.
7668 2013-07-25 Sterling Augustine <saugustine@google.com>
7670 * dwarf2out.c (size_of_pubnames): Move code to...
7671 (include_pubname_in_output): ...here. New.
7672 (want_pubnames): Rearrange.
7673 (output_pubnames): Call include_pubname_in_output. Move assertion.
7675 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7677 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7679 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7682 * doc/extend.texi: Remove obsolete builtins. Fix
7683 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7685 2013-07-25 Jan Hubicka <jh@suse.cz>
7687 * cgraph.c (release_function_body): Break out from ...
7688 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7690 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7691 old_tree in the map.
7692 (create_specialized_node): Update.
7693 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7695 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7696 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7697 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7698 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7701 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7703 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7704 addsi3_carryin_alt2_<optab>): Correct output template.
7706 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7708 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7709 Adjust for arm_restrict_it.
7710 Remove trailing whitespace.
7712 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7714 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7715 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7717 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7719 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7721 PR rtl-optimization/57960
7722 * lra-constraints.c (process_alt_operands): Use the right mode
7723 when checking strict_low.
7725 2013-07-25 Jan Hubicka <jh@suse.cz>
7727 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7728 * cgraph.c (cgraph_remove_node): Do not release function body
7729 when in cgraph streaming.
7730 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7731 in other partitions are not considered reachable; fix handling of
7734 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7736 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7738 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7743 * config/arm/arm.md ("*sibcall_insn): Replace use of
7744 Ss with US. Adjust output for v5 and v4t.
7745 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7747 * config/arm/constraints.md ("Ss"): Rename to US.
7749 2013-07-25 Terry Guo <terry.guo@arm.com>
7751 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7752 shift_add/shift_sub0/shift_sub1 RTXs.
7754 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7755 Anton Blanchard <anton@au1.ibm.com>
7757 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7758 (altivec_vpks<VI_char>ss): Likewise.
7759 (altivec_vpks<VI_char>us): Likewise.
7760 (altivec_vpku<VI_char>us): Likewise.
7761 (altivec_vpku<VI_char>um): Likewise.
7763 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7765 Introduce context class.
7767 * Makefile.in (CONTEXT_H): New.
7768 (OBJS): Add context.o.
7769 (toplev.o): Add CONTEXT_H to dependencies.
7772 * toplev.c (general_init): Create the singleton gcc::context instance.
7778 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7780 PR rtl-optimization/57968
7781 * mode-switching.c (create_pre_exit): Allow instructions that
7782 don't set a return register to need a non-exit mode.
7784 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7785 Anton Blanchard <anton@au1.ibm.com>
7787 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7788 operands to vperm for little endian.
7789 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7790 of lvsl to create the control mask for a vperm for little endian.
7792 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7793 Anton Blanchard <anton@au1.ibm.com>
7795 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7796 two operands for little-endian.
7798 2013-07-23 Steve Ellcey <sellcey@mips.com>
7800 * config/mips/mips.c (mips_case_values_threshold): New.
7801 (TARGET_CASE_VALUES_THRESHOLD): Define.
7803 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7804 Anton Blanchard <anton@au1.ibm.com>
7806 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7807 selection of field for vector splat in little endian mode.
7809 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7811 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7812 expanders to rs6000.md.
7813 (ior<mode>3): Likewise.
7814 (and<mode>3): Likewise.
7815 (one_cmpl<mode>2): Likewise.
7816 (nor<mode>3): Likewise.
7817 (andc<mode>3): Likewise.
7818 (eqv<mode>3): Likewise.
7819 (nand<mode>3): Likewise.
7820 (orc<mode>3): Likewise.
7822 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7825 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7826 to split multi-word logical operations.
7827 (rs6000_split_logical_di): Likewise.
7828 (rs6000_split_logical): Likewise.
7830 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7831 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7832 and allow TImode operations in 32-bit.
7833 (vsx_and<mode>3_64bit): Likewise.
7834 (vsx_ior<mode>3_32bit): Likewise.
7835 (vsx_ior<mode>3_64bit): Likewise.
7836 (vsx_xor<mode>3_32bit): Likewise.
7837 (vsx_xor<mode>3_64bit): Likewise.
7838 (vsx_one_cmpl<mode>2_32bit): Likewise.
7839 (vsx_one_cmpl<mode>2_64bit): Likewise.
7840 (vsx_nor<mode>3_32bit): Likewise.
7841 (vsx_nor<mode>3_64bit): Likewise.
7842 (vsx_andc<mode>3_32bit): Likewise.
7843 (vsx_andc<mode>3_64bit): Likewise.
7844 (vsx_eqv<mode>3_32bit): Likewise.
7845 (vsx_eqv<mode>3_64bit): Likewise.
7846 (vsx_nand<mode>3_32bit): Likewise.
7847 (vsx_nand<mode>3_64bit): Likewise.
7848 (vsx_orc<mode>3_32bit): Likewise.
7849 (vsx_orc<mode>3_64bit): Likewise.
7851 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
7852 logical types in GPRs.
7854 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
7855 logical insns to rs6000.md, and allow TImode operations in
7857 (altivec_ior<mode>3): Likewise.
7858 (altivec_xor<mode>3): Likewise.
7859 (altivec_one_cmpl<mode>2): Likewise.
7860 (altivec_nor<mode>3): Likewise.
7861 (altivec_andc<mode>3): Likewise.
7863 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
7864 attributes for moving the 128-bit logical operations into
7866 (BOOL_REGS_OUTPUT): Likewise.
7867 (BOOL_REGS_OP1): Likewise.
7868 (BOOL_REGS_OP2): Likewise.
7869 (BOOL_REGS_UNARY): Likewise.
7870 (BOOL_REGS_AND_CR0): Likewise.
7871 (one_cmpl<mode>2): Add support for DI logical operations on
7872 32-bit, splitting the operations to 32-bit.
7876 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
7877 changes to combine the 32/64-bit code, allow logical operations on
7878 TI mode in 32-bit, and to use similar match_operator patterns like
7879 scalar mode uses. Combine the Altivec and VSX code for logical
7880 operations, and move it here.
7881 (ior<mode>3, 128-bit types): Likewise.
7882 (xor<mode>3, 128-bit types): Likewise.
7883 (one_cmpl<mode>3, 128-bit types): Likewise.
7884 (nor<mode>3, 128-bit types): Likewise.
7885 (andc<mode>3, 128-bit types): Likewise.
7886 (eqv<mode>3, 128-bit types): Likewise.
7887 (nand<mode>3, 128-bit types): Likewise.
7888 (orc<mode>3, 128-bit types): Likewise.
7889 (and<mode>3_internal): Likewise.
7890 (bool<mode>3_internal): Likewise.
7891 (boolc<mode>3_internal1): Likewise.
7892 (boolc<mode>3_internal2): Likewise.
7893 (boolcc<mode>3_internal1): Likewise.
7894 (boolcc<mode>3_internal2): Likewise.
7895 (eqv<mode>3_internal1): Likewise.
7896 (eqv<mode>3_internal2): Likewise.
7897 (one_cmpl1<mode>3_internal): Likewise.
7899 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7901 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7902 Rename flag_stack_usage to flag_stack_usage_info.
7904 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7906 * config/microblaze/sync.md: New file.
7907 * config/microblaze/microblaze.md: Include sync.md
7908 * config/microblaze/microblaze.c: Add print_operand 'y'.
7909 * config/microblaze/constraints.md: Add memory_contraint
7910 'Q' which is a single register.
7912 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
7914 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
7916 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7918 * reload.c (find_reloads): Exit loop once we find this operand
7919 cannot be reloaded somehow for this alternative.
7921 * reload.c (find_reloads): Exit loop once we find a hard register.
7923 * rtlanal.c (computed_jump_p): Exit loop once we find label
7926 * i386.c (ix86_pad_returns): Exit loop after setting replace.
7928 * cfgloopmanip.c (remove_path): Exit loop after setting
7931 * gensupport.c (subst_dup): Avoid loop if code is not
7932 MATCH_DUP nor MATCH_OP_DUP.
7934 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
7936 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
7938 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7940 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
7941 true for SP_REGNUM if mode == ptr_mode.
7942 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
7943 with value R0_REGNUM + 31.
7945 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7947 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
7948 pad pointer-typed argument downward.
7950 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7952 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
7953 and __ILP32__ when the ILP32 model is in use.
7955 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7957 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
7958 (aarch64_load_symref_appropriately): In the case of
7959 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
7960 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
7961 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
7962 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
7963 if 'mode' doesn't equal to 'ptr_mode'.
7964 (aarch64_output_mi_thunk): Add an assertion on the alignment of
7965 'vcall_offset'; change to call aarch64_emit_move differently depending
7966 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
7967 to calculate the upper bound of 'vcall_offset'.
7968 (aarch64_cannot_force_const_mem): Change to also return true if
7970 (aarch64_legitimize_reload_address): In the case of large
7971 displacements, add new local variable 'xmode' and an assertion
7972 based on it; change to use 'xmode' to generate the new rtx and
7974 (aarch64_asm_trampoline_template): Change to generate the template
7975 differently depending on TARGET_ILP32 or not; change to use
7976 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
7977 (aarch64_trampoline_size): Removed.
7978 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
7979 and replace immediate literals with it. Change to use 'ptr_mode'
7980 instead of 'DImode' and call convert_memory_address if the mode
7981 of 'fnaddr' doesn't equal to 'ptr_mode'.
7982 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
7984 (aarch64_elf_asm_destructor): Likewise.
7985 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
7986 on TARGET_ILP32 instead of aarch64_trampoline_size.
7987 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
7988 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
7989 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
7990 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
7991 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
7992 (storewb_pair<GPI:mode>_<P:mode>): ... this.
7993 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
7994 depending on the value of 'mode'.
7995 (add_losym_<mode>): New.
7996 (ldr_got_small_<mode>): New, based on ldr_got_small.
7997 (ldr_got_small): Remove.
7998 (ldr_got_small_sidi): New.
7999 * config/aarch64/iterators.md (P): New.
8000 (PTR): Change to 'ptr_mode' in the condition.
8002 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8004 * config.gcc (aarch64*-*-*): Support --with-abi.
8005 (aarch64*-*-elf): Support --with-multilib-list.
8006 (aarch64*-*-linux*): Likewise.
8007 (supported_defaults): Add abi to aarch64*-*-*.
8008 * configure.ac: Mention AArch64 for --with-multilib-list.
8009 * configure: Re-generated.
8010 * config/aarch64/biarchilp32.h: New file.
8011 * config/aarch64/biarchlp64.h: New file.
8012 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8014 (MULTILIB_DEFAULTS): Ditto.
8015 (DRIVER_SELF_SPECS): Ditto.
8016 (ASM_SPEC): Update to also substitute -mabi.
8017 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8018 file whose name depends on -mabi= and -mbig-endian.
8019 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8021 (POINTER_SIZE): New define.
8022 (POINTERS_EXTEND_UNSIGNED): Ditto.
8023 (enum aarch64_abi_type): New enumeration tag.
8024 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8025 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8026 (TARGET_ILP32): New define.
8027 * config/aarch64/aarch64.opt (mabi): New.
8029 (ilp32, lp64): New values for -mabi.
8030 * config/aarch64/t-aarch64 (comma): New define.
8031 (MULTILIB_OPTIONS): Ditto.
8032 (MULTILIB_DIRNAMES): Ditto.
8033 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8034 * doc/invoke.texi: Document -mabi for AArch64.
8036 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8038 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8040 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8041 Anton Blanchard <anton@au1.ibm.com>
8043 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8044 endianness when selecting field to splat.
8046 2013-07-22 Eric Christopher <echristo@gmail.com>
8048 * dwarf2out.c (die_odr_checksum): New function to use
8049 CHECKSUM_ macros and ULEB128 for DIE tag.
8050 (generate_type_signature): Use.
8052 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8054 * config.gcc (sparc*-*-*): Accept leon3 processor.
8055 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8056 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8057 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8058 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8059 (mfix-ut699): Adjust comment.
8060 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8061 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8062 (CPP_CPU_SPEC): Likewise.
8063 (ASM_CPU_SPEC): Likewise.
8064 * config/sparc/sparc.c (leon3_cost): New constant.
8065 (sparc_option_override): Add leon3 support.
8066 (mem_ref): New function.
8067 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8068 (sparc_do_work_around_errata): Look into the instruction in the delay
8069 slot and adjust accordingly. Add fix for the data cache nullify issues
8070 of the UT699. Change insertion position for the NOP.
8071 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8072 (leon3_load): New reservation.
8073 (leon_store): Bump latency to 2.
8074 (grfpu): New automaton.
8075 (grfpu_alu): New unit.
8076 (grfpu_ds): Likewise.
8077 (leon_fp_alu): Adjust.
8078 (leon_fp_mult): Delete.
8079 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8080 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8081 * config/sparc/sparc.md (cpu): Add leon3.
8082 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8084 (atomic_test_and_set): Likewise.
8087 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8089 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8090 default for R5900 targets.
8091 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8092 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8093 * config/mips/mips.c (mips_option_override): Report an error for
8094 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8095 for -march=r5900 -mhard-float.
8097 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8099 * df-problems.c (can_move_insns_across): Exit loop once we
8100 find a non-fixed, non-global register.
8102 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8105 * omega.c (omega_eliminate_red): Break after setting red_found.
8106 (omega_problem_has_red_equations): Similarly after setting found.
8107 (omega_query_variable): Similarly after setting coupled.
8109 2013-07-22 Marek Polacek <polacek@redhat.com>
8111 * gimplify.c: Don't include gimple.h twice.
8113 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8115 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8116 instead of TARGET_THUMB1.
8117 (Pz): New constraint.
8118 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8120 (compare_negsi_si): Likewise.
8121 (compare_addsi2_op0): Likewise.
8122 (compare_addsi2_op1): Likewise.
8123 (addsi3_carryin_<optab>): Likewise.
8124 (addsi3_carryin_alt2_<optab>): Likewise.
8125 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8126 for arm_restrict_it.
8127 (subsi3_carryin): Likewise.
8128 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8129 (minmax_arithsi): Disable for arm_restrict_it.
8130 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8131 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8132 (satsi_<SAT:code>_shift): Likewise.
8133 (arm_shiftsi3): Add alternative for 16-bit encoding.
8134 (arm32_movhf): Disable for arm_restrict_it.
8135 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8136 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8138 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8140 * config/arm/arm.md (attribute "insn"): Delete.
8141 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8142 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8143 (not_shiftsi): Update for attribute change.
8144 (not_shiftsi_compare0): Likewise.
8145 (not_shiftsi_compare0_scratch): Likewise.
8146 (arm_one_cmplsi2): Likewise.
8147 (thumb1_one_cmplsi2): Likewise.
8148 (notsi_compare0): Likewise.
8149 (notsi_compare0_scratch): Likewise.
8150 (thumb1_movdi_insn): Likewise.
8151 (arm_movsi_insn): Likewise.
8152 (movhi_insn_arch4): Likewise.
8153 (movhi_bytes): Likewise.
8154 (arm_movqi_insn): Likewise.
8155 (thumb1_movqi_insn): Likewise.
8156 (arm32_movhf): Likewise.
8157 (thumb1_movhf): Likewise.
8158 (arm_movsf_soft_insn): Likewise.
8159 (thumb1_movsf_insn): Likewise.
8160 (thumb_movdf_insn): Likewise.
8161 (movsicc_insn): Likewise.
8162 (movsfcc_soft_insn): Likewise.
8163 (and_scc): Likewise.
8164 (cond_move): Likewise.
8165 (if_move_not): Likewise.
8166 (if_not_move): Likewise.
8167 (if_shift_move): Likewise.
8168 (if_move_shift): Likewise.
8169 (if_shift_shift): Likewise.
8170 (if_not_arith): Likewise.
8171 (if_arith_not): Likewise.
8172 (cond_move_not): Likewise.
8173 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8174 (neon_mov<mode>): Likewise.
8175 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8176 (thumb2_movsi_vfp): Likewise.
8177 (movsf_vfp): Likewise.
8178 (thumb2_movsf_vfp): Likewise.
8179 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8181 (cortexa7_older_only): Likewise.
8182 (cortexa7_younger): Likewise.
8183 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8184 (1020alu_shift_op): Likewise.
8185 (1020alu_shift_reg_op): Likewise.
8186 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8187 (alu_shift_op): Likewise.
8188 (alu_shift_reg_op): Likewise.
8189 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8190 (11_alu_shift_op): Likewise.
8191 (11_alu_shift_reg_op): Likewise.
8192 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8193 (9_alu_shift_reg_op): Likewise.
8194 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8196 (cortex_a15_alu_shift): Likewise.
8197 (cortex_a15_alu_shift_reg): Likewise.
8198 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8199 (cortex_a5_alu_shift): Likewise.
8200 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8202 (cortex_a53_alu_shift): Likewise.
8203 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8205 (cortex_a7_alu_reg): Likewise.
8206 (cortex_a7_alu_shift): Likewise.
8207 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8208 (cortex_a8_alu_shift): Likewise.
8209 (cortex_a8_alu_shift_reg): Likewise.
8210 (cortex_a8_mov): Likewise.
8211 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8212 (cortex_a9_dp_shift): Likewise.
8213 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8214 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8215 (cortex_r4_mov): Likewise.
8216 (cortex_r4_alu_shift): Likewise.
8217 (cortex_r4_alu_shift_reg): Likewise.
8218 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8219 (526_alu_shift_op): Likewise.
8220 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8221 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8222 (626te_alu_shift_op): Likewise.
8223 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8224 (726te_alu_op): Likewise.
8225 (726te_alu_shift_op): Likewise.
8226 (726te_alu_shift_reg_op): Likewise.
8227 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8228 (mp626_alu_shift_op): Likewise.
8229 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8230 (pj4_alu_e1_conds): Likewise.
8231 (pj4_alu): Likewise.
8232 (pj4_alu_conds): Likewise.
8233 (pj4_shift): Likewise.
8234 (pj4_shift_conds): Likewise.
8235 (pj4_alu_shift): Likewise.
8236 (pj4_alu_shift_conds): Likewise.
8238 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8240 * config/arm/predicates.md (shiftable_operator_strict_it):
8242 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8243 Disable cond_exec version for arm_restrict_it.
8244 (thumb2_smaxsi3): Convert to generate cond_exec.
8245 (thumb2_sminsi3): Likewise.
8246 (thumb32_umaxsi3): Likewise.
8247 (thumb2_uminsi3): Likewise.
8248 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8249 (thumb2_neg_abssi2): Likewise.
8250 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8251 (thumb2_movsicc_insn): Adjust alternatives.
8252 (thumb2_mov_negscc): Disable for arm_restrict_it.
8253 (thumb2_mov_negscc_strict_it): New pattern.
8254 (thumb2_mov_notscc_strict_it): New pattern.
8255 (thumb2_mov_notscc): Disable for arm_restrict_it.
8256 (thumb2_ior_scc): Likewise.
8257 (thumb2_ior_scc_strict_it): New pattern.
8258 (thumb2_cond_move): Adjust for arm_restrict_it.
8259 (thumb2_cond_arith): Disable for arm_restrict_it.
8260 (thumb2_cond_arith_strict_it): New pattern.
8261 (thumb2_cond_sub): Adjust for arm_restrict_it.
8262 (thumb2_movcond): Likewise.
8263 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8264 (thumb2_zero_extendhisi2_v6): Likewise.
8265 (thumb2_zero_extendqisi2_v6): Likewise.
8266 (orsi_notsi_si): Likewise.
8267 (orsi_not_shiftsi_si): Likewise.
8269 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8271 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8272 instruction sequence is 1 byte shorter.
8274 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8276 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8277 it is not needed after split.
8279 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8282 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8283 second label for nonlocal goto receivers. Don't output pic base labels
8284 unless we're producing PIC; mark that action unreachable().
8285 (ix86_save_reg): If the function contains a nonlocal label, save the
8287 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8288 * config/darwin.c (emitted_pic_label_num): New GTY.
8289 (update_pic_label_number_if_needed): New.
8290 (machopic_output_function_base_name): Adjust for nonlocal receiver
8292 (machopic_should_output_picbase_label): New.
8293 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8294 (nonlocal_goto_receiver): New insn and split.
8296 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8298 * config/aarch64/aarch64-builtins.c
8299 (aarch64_fold_builtin): Fold abs in all modes.
8300 * config/aarch64/aarch64-simd-builtins.def
8301 (abs): Enable for all modes.
8302 * config/aarch64/arm_neon.h
8303 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8304 (vabs_f64): Add missing intrinsic.
8306 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8308 * config/aarch64/arm_neon.h (vabs_s64): New function
8310 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8313 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8314 * config/avr/avr.md (adjust_len): Add `round'.
8315 * config/avr/avr-protos.h (avr_out_round): New prototype.
8316 (avr_out_plus): Add `out_label' argument.
8317 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8318 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8319 Handle the case where `insn' is just a pattern.
8320 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8321 (avr_out_round): New function.
8322 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8324 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8326 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8327 Add check for flag_stack_usage to handle -fstack-usage support
8329 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8331 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8332 interaction for new Power8 flags and VSX.
8334 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8337 * tree-inline.c (expand_call_inline): Emit errors during
8338 early_inlining only if optimization is not turned on.
8340 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8344 * passes.c (init_optimization_passes): Move the construction of
8345 the pass hierarchy into a new passes.def file.
8347 * Makefile.in (passes.o): Add dependency on passes.def.
8349 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8351 * passes.c (init_optimization_passes): Introduce macros for
8352 constructing the tree of passes (INSERT_PASSES_AFTER,
8353 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8354 TERMINATE_PASS_LIST).
8356 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8357 Wei Mi <wmi@google.com>
8359 PR rtl-optimization/57878
8360 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8362 (reload_pseudo_compare_func): Check nregs first for reload
8365 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8367 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8369 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8371 * read-rtl.c (validate_const_int): Once an invalid character is
8372 seen, quit the loop.
8374 * gengtype.c (write_roots): Similarly once we find the "deletable"
8375 or "if_marked" option.
8377 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8379 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8380 "xtab" and "sat". Move value "clz" from here to ...
8381 (attriubte "type"): ... here.
8382 (satsi_<SAT:code>): Delete "insn" attribute.
8383 (satsi_<SAT:code>_shift): Likewise.
8384 (arm_zero_extendqisi2addsi): Likewise.
8385 (arm_extendqisi2addsi): Likewise.
8386 (clzsi2): Update for attribute changes.
8387 (rbitsi2): Likewise.
8388 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8390 (arm_usatsihi): Likewise.
8391 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8393 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8395 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8396 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8397 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8398 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8399 in alphabetical order.
8400 (attribute "core_cycles"): Update for attribute changes.
8401 (arm_addsi3): Likewise.
8402 (addsi3_compare0): Likewise.
8403 (addsi3_compare0_scratch): Likewise.
8404 (addsi3_compare_op1): Likewise.
8405 (addsi3_compare_op2): Likewise.
8406 (compare_addsi2_op0): Likewise.
8407 (compare_addsi2_op1): Likewise.
8408 (addsi3_carryin_shift_<optab>): Likewise.
8409 (subsi3_carryin_shift): Likewise.
8410 (rsbsi3_carryin_shift): Likewise.
8411 (arm_subsi3_insn): Likewise.
8412 (subsi3_compare0): Likewise.
8413 (subsi3_compare): Likewise.
8414 (arm_andsi3_insn): Likewise.
8415 (thumb1_andsi3_insn): Likewise.
8416 (andsi3_compare0): Likewise.
8417 (andsi3_compare0_scratch): Likewise.
8418 (zeroextractsi_compare0_scratch
8419 (andsi_not_shiftsi_si): Likewise.
8420 (iorsi3_insn): Likewise.
8421 (iorsi3_compare0): Likewise.
8422 (iorsi3_compare0_scratch): Likewise.
8423 (arm_xorsi3): Likewise.
8424 (thumb1_xorsi3_insn): Likewise.
8425 (xorsi3_compare0): Likewise.
8426 (xorsi3_compare0_scratch): Likewise.
8427 (satsi_<SAT:code>_shift): Likewise.
8429 (arm_shiftsi3): Likewise.
8430 (shiftsi3_compare0): Likewise.
8431 (not_shiftsi): Likewise.
8432 (not_shiftsi_compare0): Likewise.
8433 (not_shiftsi_compare0_scratch): Likewise.
8434 (arm_one_cmplsi2): Likewise.
8435 (thumb_one_complsi2): Likewise.
8436 (notsi_compare0): Likewise.
8437 (notsi_compare0_scratch): Likewise.
8438 (thumb1_zero_extendhisi2): Likewise.
8439 (arm_zero_extendhisi2): Likewise.
8440 (arm_zero_extendhisi2_v6): Likewise.
8441 (arm_zero_extendhisi2addsi): Likewise.
8442 (thumb1_zero_extendqisi2): Likewise.
8443 (thumb1_zero_extendqisi2_v6): Likewise.
8444 (arm_zero_extendqisi2): Likewise.
8445 (arm_zero_extendqisi2_v6): Likewise.
8446 (arm_zero_extendqisi2addsi): Likewise.
8447 (thumb1_extendhisi2): Likewise.
8448 (arm_extendhisi2): Likewise.
8449 (arm_extendhisi2_v6): Likewise.
8450 (arm_extendqisi): Likewise.
8451 (arm_extendqisi_v6): Likewise.
8452 (arm_extendqisi2addsi): Likewise.
8453 (thumb1_extendqisi2): Likewise.
8454 (thumb1_movdi_insn): Likewise.
8455 (arm_movsi_insn): Likewise.
8456 (movsi_compare0): Likewise.
8457 (movhi_insn_arch4): Likewise.
8458 (movhi_bytes): Likewise.
8459 (arm_movqi_insn): Likewise.
8460 (thumb1_movqi_insn): Likewise.
8461 (arm32_movhf): Likewise.
8462 (thumb1_movhf): Likewise.
8463 (arm_movsf_soft_insn): Likewise.
8464 (thumb1_movsf_insn): Likewise.
8465 (movdf_soft_insn): Likewise.
8466 (thumb_movdf_insn): Likewise.
8467 (arm_cmpsi_insn): Likewise.
8468 (cmpsi_shiftsi): Likewise.
8469 (cmpsi_shiftsi_swp): Likewise.
8470 (arm_cmpsi_negshiftsi_si): Likewise.
8471 (movsicc_insn): Likewise.
8472 (movsfcc_soft_insn): Likewise.
8473 (arith_shiftsi): Likewise.
8474 (arith_shiftsi_compare0
8475 (arith_shiftsi_compare0_scratch
8476 (sub_shiftsi): Likewise.
8477 (sub_shiftsi_compare0
8478 (sub_shiftsi_compare0_scratch
8479 (and_scc): Likewise.
8480 (cond_move): Likewise.
8481 (if_plus_move): Likewise.
8482 (if_move_plus): Likewise.
8483 (if_move_not): Likewise.
8484 (if_not_move): Likewise.
8485 (if_shift_move): Likewise.
8486 (if_move_shift): Likewise.
8487 (if_shift_shift): Likewise.
8488 (if_not_arith): Likewise.
8489 (if_arith_not): Likewise.
8490 (cond_move_not): Likewise.
8491 (thumb1_ashlsi3): Set type attribute.
8492 (thumb1_ashrsi3): Likewise.
8493 (thumb1_lshrsi3): Likewise.
8494 (thumb1_rotrsi3): Likewise.
8495 (shiftsi3_compare0_scratch): Likewise.
8496 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8497 (neon_mov<mode>): Likewise.
8498 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8500 (thumb2_movsi_insn): Likewise.
8501 (thumb2_cmpsi_neg_shiftsi): Likewise.
8502 (thumb2_extendqisi_v6): Likewise.
8503 (thumb2_zero_extendhisi2_v6): Likewise.
8504 (thumb2_zero_extendqisi2_v6): Likewise.
8505 (thumb2_shiftsi3_short): Likewise.
8506 (thumb2_addsi3_compare0_scratch): Likewise.
8507 (orsi_not_shiftsi_si): Likewise.
8508 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8509 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8511 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8512 (1020alu_shift_op): Likewise.
8513 (1020alu_shift_reg_op): Likewise.
8514 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8515 (alu_shift_op): Likewise.
8516 (alu_shift_reg_op): Likewise.
8517 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8518 (11_alu_shift_op): Likewise.
8519 (11_alu_shift_reg_op): Likewise.
8520 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8521 (9_alu_shift_reg_op): Likewise.
8522 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8524 (cortex_a15_alu_shift): Likewise.
8525 (cortex_a15_alu_shift_reg): Likewise.
8526 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8528 (cortex_a5_alu_shift): Likewise.
8529 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8531 (cortex_a53_alu_shift): Likewise.
8532 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8534 (cortex_a7_alu_reg): Likewise.
8535 (cortex_a7_alu_shift): Likewise.
8536 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8538 (cortex_a8_alu_shift): Likewise.
8539 (cortex_a8_alu_shift_reg): Likewise.
8540 (cortex_a8_mov): Likewise.
8541 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8542 (cortex_a9_dp_shift): Likewise.
8543 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8545 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8547 (cortex_r4_mov): Likewise.
8548 (cortex_r4_alu_shift): Likewise.
8549 (cortex_r4_alu_shift_reg): Likewise.
8550 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8551 (526_alu_shift_op): Likewise.
8552 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8553 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8554 (626te_alu_shift_op): Likewise.
8555 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8556 (726te_alu_op): Likewise.
8557 (726te_alu_shift_op): Likewise.
8558 (726te_alu_shift_reg_op): Likewise.
8559 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8560 (mp626_alu_shift_op): Likewise.
8561 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8562 (pj4_alu_e1_conds): Likewise.
8563 (pj4_alu): Likewise.
8564 (pj4_alu_conds): Likewise.
8565 (pj4_shift): Likewise.
8566 (pj4_shift_conds): Likewise.
8567 (pj4_alu_shift): Likewise.
8568 (pj4_alu_shift_conds): Likewise.
8569 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8571 (cortexa7_older_only): Likewise.
8572 (cortexa7_younger): Likewise.
8574 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8576 * ipa-pure-const.c (generate_summary): Rename to...
8577 (pure_const_generate_summary): ... this.
8579 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8581 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8583 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8586 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8589 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8591 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8592 enabled without -march=zEC12.
8593 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8596 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8598 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8599 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8600 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8601 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8602 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8603 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8604 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8606 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8607 and ISA_HAS_NMADD3_NMSUB3.
8608 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8609 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8610 (nmsub4<mode>, nmsub3<mode>): Likewise.
8611 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8613 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8615 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8616 TARGET_MIPS5400 checking.
8618 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8619 Peter Bergner <bergner@vnet.ibm.com>
8621 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8622 registers in the comment.
8623 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8624 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8625 rather than FIRST_PSEUDO_REGISTERS.
8627 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8629 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8630 enable extra ISA flags with TARGET_HTM.
8632 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8634 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8637 2013-07-15 Cong Hou <congh@google.com>
8639 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8640 in compare function for sorting.
8642 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8644 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8645 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8646 * config/rs6000/rs6000.opt: Add -mhtm option.
8647 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8648 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8649 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8650 __HTM__ if the HTM instructions are available.
8651 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8652 htm_spr_reg_operand): New define_predicates.
8653 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8654 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8656 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8657 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8658 HTM builtin functions.
8659 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8660 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8661 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8662 (rs6000_builtin_mask_calculate): Likewise.
8663 (rs6000_option_override_internal): Likewise.
8664 (bdesc_htm): Add new HTM builtin support.
8665 (htm_spr_num): New function.
8666 (htm_spr_regno): Likewise.
8667 (rs6000_htm_spr_icode): Likewise.
8668 (htm_expand_builtin): Likewise.
8669 (htm_init_builtins): Likewise.
8670 (rs6000_expand_builtin): Add support for HTM builtin functions.
8671 (rs6000_init_builtins): Likewise.
8672 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8674 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8675 (TARGET_HTM, MASK_HTM): Define macros.
8676 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8677 (FIXED_REGISTERS): Likewise.
8678 (CALL_USED_REGISTERS): Likewise.
8679 (CALL_REALLY_USED_REGISTERS): Likewise.
8680 (REG_ALLOC_ORDER): Likewise.
8681 (enum reg_class): Likewise.
8682 (REG_CLASS_NAMES): Likewise.
8683 (REG_CLASS_CONTENTS): Likewise.
8684 (REGISTER_NAMES): Likewise.
8685 (ADDITIONAL_REGISTER_NAMES): Likewise.
8686 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8687 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8688 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8689 * config/rs6000/htm.md: New file.
8690 * config/rs6000/htmintrin.h: New file.
8691 * config/rs6000/htmxlintrin.h: New file.
8693 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8695 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8696 Define SYMBOL_TINY_GOT, update comment.
8697 * config/aarch64/aarch64.c
8698 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8699 (aarch64_expand_mov_immediate): Likewise.
8700 (aarch64_print_operand): Likewise.
8701 (aarch64_classify_symbol): Likewise.
8702 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8703 (ldr_got_tiny): Define.
8705 2013-07-13 Tobias Grosser <tobias@grosser.es>
8707 PR tree-optimization/54094
8708 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8709 scheduling dimension for the parallelism check from the polyhedral
8710 information in the AST.
8711 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8714 2013-07-13 Jason Merrill <jason@redhat.com>
8716 * print-tree.c (debug_vec_tree): Use debug_raw.
8717 (debug_raw (vec<tree, va_gc> &)): New.
8718 (debug_raw (vec<tree, va_gc> *)): New.
8719 * tree.h: Declare them.
8721 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8723 * ifcvt.c (ifcvt_after_combine): New static variable.
8724 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8726 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8727 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8728 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8730 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8732 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8734 2013-07-12 Michael Matz <matz@suse.de>
8737 * convert.c (convert_to_real): Reject non-float inner types.
8739 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8741 * config/aarch64/aarch64-protos.h
8742 (aarch64_simd_immediate_valid_for_move): Remove.
8743 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8744 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8746 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8748 2013-07-11 Steve Ellcey <sellcey@mips.com>
8750 * config/mips/mips.c (mips_conditional_register_usage): Do not
8751 use t[0-7] registers in MIPS16 mode when optimizing for size.
8753 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8755 * config/i386/i386.c (dispatch_function_versions): Fix array
8756 indexing of function_version_info to match actual_versions.
8758 2013-07-11 Teresa Johnson <tejohnson@google.com>
8760 * vec.h (struct va_gc): Move release out-of-line.
8761 (va_gc::release): Call ggc_free on released vec.
8763 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8765 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8766 Require GOT register as additional operand in UNSPEC.
8767 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8768 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8769 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8770 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8771 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8772 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8773 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8775 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8778 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8779 name seen by assembler/linker rather if available.
8781 2013-07-11 Andreas Schwab <schwab@suse.de>
8783 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8785 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8787 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8789 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8791 * doc/tm.texi.in: Move hook documentation to ....
8792 * target.def: ... here.
8794 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8796 * doc/tm.texi: Regenerate.
8798 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8801 * doc/invoke.texi: Document -Wconditionally-supported.
8803 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8806 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8809 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8812 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8813 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8814 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8815 Remove duplicate devices.
8816 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8817 * config/avr/t-multilib: Regenerate.
8818 * config/avr/avr-tables.opt: Regenerate.
8819 * doc/avr-mmcu.texi: Regenerate.
8821 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8824 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8826 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8828 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8829 the cost of MULT when optimizing for size.
8831 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8833 * config/cr16/cr16-protos.h: Don't include target.h.
8835 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8837 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8838 adjust register size for TDmode and TFmode for VSX registers.
8840 2013-07-08 Kai Tietz <ktietz@redhat.com>
8843 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8844 hook_bool_const_tree_true.
8846 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8848 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8849 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
8850 * config/s390/s390.md: Define FPR*_REGNUM constants.
8851 Fix FPR2_REGNUM constant (18 -> 17).
8852 ("*trunc<BFP:mode><DFP_ALL:mode>2")
8853 ("*trunc<DFP_ALL:mode><BFP:mode>2")
8854 ("trunc<BFP:mode><DFP_ALL:mode>2")
8855 ("trunc<DFP_ALL:mode><BFP:mode>2")
8856 ("*extend<BFP:mode><DFP_ALL:mode>2")
8857 ("*extend<DFP_ALL:mode><BFP:mode>2")
8858 ("extend<BFP:mode><DFP_ALL:mode>2")
8859 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
8862 2013-07-08 Graham Stott <graham.stott@btinternet.com>
8864 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
8866 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8868 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
8869 and cfun_fpr_bit_p to cfun_fpr_save_p.
8870 (s390_frame_area, s390_register_info, s390_frame_info)
8871 (s390_emit_prologue, s390_emit_epilogue)
8872 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
8874 * config/s390/s390.h: Define *_REGNUM macros for floating point
8877 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
8879 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
8881 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
8883 PR rtl-optimization/57786
8884 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
8885 and break out of the loop when it is set to false.
8887 2013-07-08 Jakub Jelinek <jakub@redhat.com>
8890 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
8891 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
8892 (const_int 63)) 0)).
8893 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
8894 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
8895 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
8897 PR rtl-optimization/57829
8898 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
8899 mask bits outside of mode are just sign-extension from mode to HWI.
8901 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8903 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
8904 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
8905 adjust_address instead of change_address to keep info about alignment.
8906 (emit_strmov): Remove.
8907 (emit_memmov): New function.
8908 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
8909 (expand_movmem_epilogue): Likewise and return updated rtx for
8911 (expand_constant_movmem_prologue): Likewise and return updated rtx for
8912 destination and source.
8913 (decide_alignment): Refactor, handle vector_loop.
8914 (ix86_expand_movmem): Likewise.
8915 (ix86_expand_setmem): Likewise.
8916 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
8918 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
8920 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
8921 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
8923 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8925 * config/i386/sse.md (sse_movlhps): Change alternative 3
8926 of operand 2 to "m".
8928 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8931 * config/i386/sse.md (iptr): New mode attribute.
8932 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
8933 (<sse>_vm<plusminus_insn><mode>3): Ditto.
8934 (<sse>_vmmul<mode>3): Ditto.
8935 (<sse>_vmdiv<mode>3): Ditto.
8936 (sse_vmrcpv4sf2): Ditto.
8937 (<sse>_vmsqrt<mode>2): Ditto.
8938 (sse_vmrsqrtv4sf2): Ditto.
8939 (<sse>_vm<code><mode>3): Ditto.
8940 (avx_vmcmp<mode>3): Ditto.
8941 (<sse>_vmmaskcmp<mode>3): Ditto.
8942 (<sse>_comi): Ditto.
8943 (<sse>_ucomi): Ditto.
8944 (*xop_vmfrcz_<mode>): Ditto.
8945 (*fmai_fmadd_<mode>): Ditto.
8946 (*fmai_fmsub_<mode>): Ditto.
8947 (*fmai_fnmadd_<mode>): Ditto.
8948 (*fmai_fnmsub_<mode>): Ditto.
8949 (*fma4i_vmfmadd_<mode>): Ditto.
8950 (*fma4i_vmfmsub_<mode>): Ditto.
8951 (*fma4i_vmfnmadd_<mode>): Ditto.
8952 (*fma4i_vmfnmsub_<mode>): Ditto.
8953 (*xop_vmfrcz_<mode>): Ditto.
8954 (sse_cvtps2pi): Ditto.
8955 (sse_cvttps2pi): Ditto.
8956 (sse_cvtss2si): Ditto.
8957 (sse_cvtss2si_2): Ditto.
8958 (sse_cvtss2siq_2): Ditto.
8959 (sse_cvttss2si): Ditto.
8960 (sse_cvttss2siq): Ditto.
8961 (sse_cvtsd2si): Ditto.
8962 (sse_cvtsd2si_2): Ditto.
8963 (sse_cvtsd2siq_2): Ditto.
8964 (sse_cvttsd2si): Ditto.
8965 (sse_cvttsd2siq): Ditto.
8966 (sse_cvtsd2ss): Ditto.
8967 (sse_cvtss2sd): Ditto.
8968 (avx2_pbroadcast<mode>): Ditto.
8969 (avx2_pbroadcast<mode>_1): Ditto.
8970 (*avx_vperm_broadcast_v4sf): Ditto.
8972 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
8973 (sse_movlhps): Ditto.
8974 (sse_storehps): Ditto.
8975 (sse_loadhps): Ditto.
8976 (sse_storelps): Ditto.
8977 (sse_loadlps): Ditto.
8978 (*vec_concatv4sf): Ditto.
8979 (*vec_interleave_highv2df): Ditto.
8980 (*vec_interleave_lowv2df): Ditto.
8981 (*vec_extractv2df_1_sse): Ditto.
8982 (*vec_extractv2df_0_sse): Ditto.
8983 (sse2_storelpd): Ditto.
8984 (sse2_loadlpd): Ditto.
8985 (sse2_movsd): Ditto.
8986 (*vec_concatv4si): Ditto.
8987 (vec_concatv2di): Ditto.
8989 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
8990 for Intel asm dialect.
8991 (mmx_punpcklwd): Ditto.
8992 (mmx_punpckldq): Ditto.
8994 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
8995 for intel assembler dialect.
8997 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9000 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9001 for BUILT_IN_C{LZ,LRSB}*.
9002 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9003 * tree-vrp.c (extract_range_basic): Handle
9004 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9005 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9006 fall thru to code calling set_value*.
9007 * builtins.c (expand_builtin): Remove *IMAX cases.
9008 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9009 if width is bigger than 2*HWI.
9011 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9013 PR rtl-optimization/55342
9014 * lra-int.h (lra_subreg_reload_pseudos): New.
9015 * lra.c: Add undoing optional reloads to the block diagram.
9016 (lra_subreg_reload_pseudos): New.
9017 (lra_optional_reload_pseudos): Change comments.
9018 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9019 lra_optional_reload_pseudos after undo transformations.
9020 * lra-assigns.c (pseudo_prefix_title): New.
9021 (lra_setup_reg_renumber): Use it.
9022 (spill_for): Ditto. Check subreg reload pseudos too.
9023 (assign_by_spills): Consider subreg reload pseudos too.
9024 * lra-constraints.c (simplify_operand_subreg): Use
9025 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9026 (curr_insn_transform): Recognize and do optional reloads.
9027 (undo_optional_reloads): New.
9028 (lra_undo_inheritance): Call undo_optional_reloads.
9030 2013-07-05 Thomas Quinot <quinot@adacore.com>
9032 * tree-complex.c (expand_complex_operations_1): Fix typo.
9034 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9036 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9037 (tune_params): New member 'const vec_costs'.
9038 * config/aarch64/aarch64.c (generic_vector_cost): New.
9039 (generic_tunings): New member 'generic_vector_cost'.
9040 (aarch64_builtin_vectorization_cost): New.
9041 (aarch64_add_stmt_cost): New.
9042 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9043 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9045 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9048 * config/i386/predicates.md (vsib_address_operand): Disallow
9049 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9051 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9054 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9055 expand_builtin_setjmp_receiver.
9056 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9057 with NULL for the label parameter.
9058 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9059 the frame-pointer. Adjust comments.
9060 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9061 only if LABEL is non-NULL.
9063 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9065 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9066 (ARM_ABI_AAPCS64): Ditto.
9068 (ARM_DEFAULT_ABI): Ditto.
9070 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9072 * config/aarch64/aarch64-builtins.c
9073 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9074 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9076 * config/aarch64/aarch64-simd.md
9077 (aarch64_ld1<VALL:mode>): New.
9078 (aarch64_st1<VALL:mode>): Likewise.
9079 * config/aarch64/arm_neon.h
9080 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9082 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9084 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9086 (ix86_option_override_internal):Turn on all -mavx target flags by
9087 default as they are dependent on AVX anyway.
9089 2013-07-02 Cary Coutant <ccoutant@google.com>
9091 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9093 (loc_checksum_ordered): Likewise.
9094 (hash_loc_operands): Remove inline keyword.
9096 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9098 PR tree-optimization/57741
9099 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9100 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9101 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9102 Allow REAL_CST step_exprs if flag_associative_math.
9103 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9105 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9107 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9109 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9111 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9113 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9115 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9117 (iorsi3_insn): Likewise.
9118 (arm_xorsi3): Likewise.
9120 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9122 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9124 (attribute "type"): ... here, and prefix with "wmmx_".
9125 (attribute "core_cycles"): Update for attribute changes.
9126 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9127 (tbcstv4hi): Likewise.
9128 (tbcstv2si): Likewise.
9129 (iwmmxt_iordi3): Likewise.
9130 (iwmmxt_xordi3): Likewise.
9131 (iwmmxt_anddi3): Likewise.
9132 (iwmmxt_nanddi3): Likewise.
9133 (iwmmxt_arm_movdi): Likewise.
9134 (iwmmxt_movsi_insn): Likewise.
9135 (mov<mode>_internal): Likewise.
9136 (and<mode>3_iwmmxt): Likewise.
9137 (ior<mode>3_iwmmxt): Likewise.
9138 (xor<mode>3_iwmmxt): Likewise.
9139 (add<mode>3_iwmmxt): Likewise.
9140 (ssaddv8qi3): Likewise.
9141 (ssaddv4hi3): Likewise.
9142 (ssaddv2si3): Likewise.
9143 (usaddv8qi3): Likewise.
9144 (usaddv4hi3): Likewise.
9145 (usaddv2si3): Likewise.
9146 (sub<mode>3_iwmmxt): Likewise.
9147 (sssubv8qi3): Likewise.
9148 (sssubv4hi3): Likewise.
9149 (sssubv2si3): Likewise.
9150 (ussubv8qi3): Likewise.
9151 (ussubv4hi3): Likewise.
9152 (ussubv2si3): Likewise.
9153 (mulv4hi3_iwmmxt): Likewise.
9154 (smulv4hi3_highpart): Likewise.
9155 (umulv4hi3_highpart): Likewise.
9156 (iwmmxt_wmacs): Likewise.
9157 (iwmmxt_wmacsz): Likewise.
9158 (iwmmxt_wmacu): Likewise.
9159 (iwmmxt_wmacuz): Likewise.
9160 (iwmmxt_clrdi): Likewise.
9161 (iwmmxt_clrv8qi): Likewise.
9162 (iwmmxt_clr4hi): Likewise.
9163 (iwmmxt_clr2si): Likewise.
9164 (iwmmxt_uavgrndv8qi3): Likewise.
9165 (iwmmxt_uavgrndv4hi3): Likewise.
9166 (iwmmxt_uavgv8qi3): Likewise.
9167 (iwmmxt_uavgv4hi3): Likewise.
9168 (iwmmxt_tinsrb): Likewise.
9169 (iwmmxt_tinsrh): Likewise.
9170 (iwmmxt_tinsrw): Likewise.
9171 (iwmmxt_textrmub): Likewise.
9172 (iwmmxt_textrmsb): Likewise.
9173 (iwmmxt_textrmuh): Likewise.
9174 (iwmmxt_textrmsh): Likewise.
9175 (iwmmxt_textrmw): Likewise.
9176 (iwmxxt_wshufh): Likewise.
9177 (eqv8qi3): Likewise.
9178 (eqv4hi3): Likewise.
9179 (eqv2si3): Likewise.
9180 (gtuv8qi3): Likewise.
9181 (gtuv4hi3): Likewise.
9182 (gtuv2si3): Likewise.
9183 (gtv8qi3): Likewise.
9184 (gtv4hi3): Likewise.
9185 (gtv2si3): Likewise.
9186 (smax<mode>3_iwmmxt): Likewise.
9187 (umax<mode>3_iwmmxt): Likewise.
9188 (smin<mode>3_iwmmxt): Likewise.
9189 (umin<mode>3_iwmmxt): Likewise.
9190 (iwmmxt_wpackhss): Likewise.
9191 (iwmmxt_wpackwss): Likewise.
9192 (iwmmxt_wpackdss): Likewise.
9193 (iwmmxt_wpackhus): Likewise.
9194 (iwmmxt_wpackwus): Likewise.
9195 (iwmmxt_wpackdus): Likewise.
9196 (iwmmxt_wunpckihb): Likewise.
9197 (iwmmxt_wunpckihh): Likewise.
9198 (iwmmxt_wunpckihw): Likewise.
9199 (iwmmxt_wunpckilb): Likewise.
9200 (iwmmxt_wunpckilh): Likewise.
9201 (iwmmxt_wunpckilw): Likewise.
9202 (iwmmxt_wunpckehub): Likewise.
9203 (iwmmxt_wunpckehuh): Likewise.
9204 (iwmmxt_wunpckehuw): Likewise.
9205 (iwmmxt_wunpckehsb): Likewise.
9206 (iwmmxt_wunpckehsh): Likewise.
9207 (iwmmxt_wunpckehsw): Likewise.
9208 (iwmmxt_wunpckelub): Likewise.
9209 (iwmmxt_wunpckeluh): Likewise.
9210 (iwmmxt_wunpckeluw): Likewise.
9211 (iwmmxt_wunpckelsb): Likewise.
9212 (iwmmxt_wunpckelsh): Likewise.
9213 (iwmmxt_wunpckelsw): Likewise.
9214 (ror<mode>3): Likewise.
9215 (ashr<mode>3_iwmmxt): Likewise.
9216 (lshr<mode>3_iwmmxt): Likewise.
9217 (ashl<mode>3_iwmmxt): Likewise.
9218 (ror<mode>3_di): Likewise.
9219 (ashr<mode>3_di): Likewise.
9220 (lshr<mode>3_di): Likewise.
9221 (ashl<mode>3_di): Likewise.
9222 (iwmmxt_wmadds): Likewise.
9223 (iwmmxt_wmaddu): Likewise.
9224 (iwmmxt_tmia): Likewise.
9225 (iwmmxt_tmiaph): Likewise.
9226 (iwmmxt_tmiabb): Likewise.
9227 (iwmmxt_tmiatb): Likewise.
9228 (iwmmxt_tmiabt): Likewise.
9229 (iwmmxt_tmiatt): Likewise.
9230 (iwmmxt_tmovmskb): Likewise.
9231 (iwmmxt_tmovmskh): Likewise.
9232 (iwmmxt_tmovmskw): Likewise.
9233 (iwmmxt_waccb): Likewise.
9234 (iwmmxt_wacch): Likewise.
9235 (iwmmxt_waccw): Likewise.
9236 (iwmmxt_waligni): Likewise.
9237 (iwmmxt_walignr): Likewise.
9238 (iwmmxt_walignr0): Likewise.
9239 (iwmmxt_walignr1): Likewise.
9240 (iwmmxt_walignr2): Likewise.
9241 (iwmmxt_walignr3): Likewise.
9242 (iwmmxt_wsadb): Likewise.
9243 (iwmmxt_wsadh): Likewise.
9244 (iwmmxt_wsadbz): Likewise.
9245 (iwmmxt_wsadhz): Likewise.
9246 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9247 (iwmmxt_wabsdiffb): Likewise.
9248 (iwmmxt_wabsdiffh): Likewise.
9249 (iwmmxt_wabsdiffw): Likewise.
9250 (iwmmxt_waddsubhx): Likewise
9251 (iwmmxt_wsubaddhx): Likewise.
9252 (addc<mode>3): Likewise.
9253 (iwmmxt_avg4): Likewise.
9254 (iwmmxt_avg4r): Likewise.
9255 (iwmmxt_wmaddsx): Likewise.
9256 (iwmmxt_wmaddux): Likewise.
9257 (iwmmxt_wmaddsn): Likewise.
9258 (iwmmxt_wmaddun): Likewise.
9259 (iwmmxt_wmulwsm): Likewise.
9260 (iwmmxt_wmulwum): Likewise.
9261 (iwmmxt_wmulsmr): Likewise.
9262 (iwmmxt_wmulumr): Likewise.
9263 (iwmmxt_wmulwsmr): Likewise.
9264 (iwmmxt_wmulwumr): Likewise.
9265 (iwmmxt_wmulwl): Likewise.
9266 (iwmmxt_wqmulm): Likewise.
9267 (iwmmxt_wqmulwm): Likewise.
9268 (iwmmxt_wqmulmr): Likewise.
9269 (iwmmxt_wqmulwmr): Likewise.
9270 (iwmmxt_waddbhusm): Likewise.
9271 (iwmmxt_waddbhusl): Likewise.
9272 (iwmmxt_wqmiabb): Likewise.
9273 (iwmmxt_wqmiabt): Likewise.
9274 (iwmmxt_wqmiatb): Likewise.
9275 (iwmmxt_wqmiatt): Likewise.
9276 (iwmmxt_wqmiabbn): Likewise.
9277 (iwmmxt_wqmiabtn): Likewise.
9278 (iwmmxt_wqmiatbn): Likewise.
9279 (iwmmxt_wqmiattn): Likewise.
9280 (iwmmxt_wmiabb): Likewise.
9281 (iwmmxt_wmiabt): Likewise.
9282 (iwmmxt_wmiatb): Likewise.
9283 (iwmmxt_wmiatt): Likewise.
9284 (iwmmxt_wmiabbn): Likewise.
9285 (iwmmxt_wmiabtn): Likewise.
9286 (iwmmxt_wmiatbn): Likewise.
9287 (iwmmxt_wmiattn): Likewise.
9288 (iwmmxt_wmiawbb): Likewise.
9289 (iwmmxt_wmiawbt): Likewise.
9290 (iwmmxt_wmiawtb): Likewise.
9291 (iwmmxt_wmiawtt): Likewise.
9292 (iwmmxt_wmiawbbn): Likewise.
9293 (iwmmxt_wmiawbtn): Likewise.
9294 (iwmmxt_wmiawtbn): Likewise.
9295 (iwmmxt_wmiawttn): Likewise.
9296 (iwmmxt_wmerge): Likewise.
9297 (iwmmxt_tandc<mode>3): Likewise.
9298 (iwmmxt_torc<mode>3): Likewise.
9299 (iwmmxt_torvsc<mode>3): Likewise.
9300 (iwmmxt_textrc<mode>3): Likewise.
9301 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9302 (wmmxt_pack): Likewise.
9303 (wmmxt_mult_c1): Likewise.
9304 (wmmxt_mult_c2): Likewise.
9305 (wmmxt_alu_c1): Likewise.
9306 (wmmxt_alu_c2): Likewise.
9307 (wmmxt_alu_c3): Likewise.
9308 (wmmxt_transfer_c1): Likewise.
9309 (wmmxt_transfer_c2): Likewise.
9310 (wmmxt_transfer_c3): Likewise.
9311 (marvell_f_iwmmxt_wstr): Likewise.
9312 (marvell_f_iwmmxt_wldr): Likewise.
9314 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9316 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9318 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9321 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9322 * lra-constraints.c (need_for_split_p): Check call used hard regs
9323 living through calls.
9325 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9326 call used regs for call insn.
9328 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9331 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9332 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9334 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9336 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9338 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9340 * lra-constraints.c (need_for_split_p): Check call used hard regs
9341 living through calls.
9343 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9346 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9347 to tie with any other modes. Eliminate Altivec vector mode tests,
9348 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9349 code, to return 0 if testing MODE2 for a condition, if we've
9350 already tested MODE1 for the same condition.
9352 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9354 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9357 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9359 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9360 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9362 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9364 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9366 (aarch64_symbolic_constant_p): Remove.
9367 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9368 static. Fix line length and white space.
9369 (aarch64_symbolic_constant_p): Remove.
9370 * config/aarch64/predicates.md (aarch64_valid_symref):
9371 Use aarch64_classify_symbol_expression.
9373 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9375 * config/arm/constraints.md (Ts): New constraint.
9376 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9378 (compare_scc): Use "Ts" constraint for operand 0.
9379 (ior_scc_scc): Likewise.
9380 (and_scc_scc): Likewise.
9381 (and_scc_scc_nodom): Likewise.
9382 (ior_scc_scc_cmp): Likewise for operand 7.
9383 (and_scc_scc_cmp): Likewise.
9384 * config/arm/thumb2.md (thumb2_movsi_insn):
9385 Add alternatives for 16-bit encodings.
9386 (thumb2_movhi_insn): Likewise.
9387 (thumb2_movsicc_insn): Likewise.
9388 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9389 (thumb2_negscc): Use "Ts" constraint.
9390 Move mvn instruction outside cond_exec block.
9391 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9392 for 16-bit encodings.
9394 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9396 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9398 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9399 (mulsi3subsi): Likewise.
9400 (mulsidi3adddi): Likewise.
9401 (mulsidi3_v6): Likewise.
9402 (umulsidi3_v6): Likewise.
9403 (umulsidi3adddi_v6): Likewise.
9404 (smulsi3_highpart_v6): Likewise.
9405 (umulsi3_highpart_v6): Likewise.
9406 (mulhisi3tb): Likewise.
9407 (mulhisi3bt): Likewise.
9408 (mulhisi3tt): Likewise.
9409 (maddhisi4): Likewise.
9410 (maddhisi4tb): Likewise.
9411 (maddhisi4tt): Likewise.
9412 (maddhidi4): Likewise.
9413 (maddhidi4tb): Likewise.
9414 (maddhidi4tt): Likewise.
9415 (zeroextractsi_compare0_scratch): Likewise.
9416 (insv_zero): Likewise.
9417 (insv_t2): Likewise.
9418 (anddi_notzesidi_di): Likewise.
9419 (anddi_notsesidi_di): Likewise.
9420 (andsi_notsi_si): Likewise.
9421 (iordi_zesidi_di): Likewise.
9422 (xordi_zesidi_di): Likewise.
9423 (andsi_iorsi3_notsi): Likewise.
9425 (smax_m1): Likewise.
9427 (not_shiftsi): Likewise.
9428 (unaligned_loadsi): Likewise.
9429 (unaligned_loadhis): Likewise.
9430 (unaligned_loadhiu): Likewise.
9431 (unaligned_storesi): Likewise.
9432 (unaligned_storehi): Likewise.
9433 (extv_reg): Likewise.
9434 (extzv_t2): Likewise.
9436 (udivsi3): Likewise.
9437 (arm_zero_extendhisi2addsi): Likewise.
9438 (arm_zero_extendqisi2addsi): Likewise.
9439 (compareqi_eq0): Likewise.
9440 (arm_extendhisi2_v6): Likewise.
9441 (arm_extendqisi2addsi): Likewise.
9442 (arm_movt): Likewise.
9443 (thumb2_ldrd): Likewise.
9444 (thumb2_ldrd_base): Likewise.
9445 (thumb2_ldrd_base_neg): Likewise.
9446 (thumb2_strd): Likewise.
9447 (thumb2_strd_base): Likewise.
9448 (thumb2_strd_base_neg): Likewise.
9449 (arm_negsi2): Add alternative for 16-bit encoding.
9450 (arm_one_cmplsi2): Likewise.
9452 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9454 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9455 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9456 (movdfcc): Likewise.
9457 * config/arm/vfp.md (*thumb2_movsf_vfp):
9458 Disable predication for arm_restrict_it.
9459 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9460 (*thumb2_movdfcc_vfp): Likewise.
9461 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9462 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9463 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9464 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9465 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9466 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9467 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9468 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9469 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9470 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9471 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9472 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9473 Disable predication for arm_restrict_it.
9475 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9477 * config/i386/bmiintrin.h (_bextr_u32): New.
9478 (_bextr_u64): Ditto.
9480 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9482 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9483 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9484 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9485 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9486 * config/mips/n32-elf.h: ...this new file.
9488 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9491 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9492 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9494 2013-06-27 Catherine Moore <clm@codesourcery.com>
9496 * config/mips/mips-tables.opt: Regenerate.
9497 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9498 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9499 * doc/invoke.texi: Add -m14kc.
9501 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9504 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9505 constraints of operand 1 and 2.
9508 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9509 to match RTL canonicalization. Swap predicates and
9510 constraints of operand 1 and 2.
9512 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9514 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9515 Process OP_INOUT regs for splitting too.
9517 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9519 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9520 decl before the loop, initialize to NULL.
9521 (vectorizable_load): Initialize ptr_incr to NULL.
9523 2013-06-27 Martin Jambor <mjambor@suse.cz>
9526 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9527 * ipa-ref.c (ipa_maybe_record_reference): New function.
9528 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9529 * ipa-cp.c (create_specialized_node): Record potential references from
9531 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9533 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9535 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9536 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9537 'mode' to force_reg.
9538 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9539 (aarch64_expand_mov_immediate): Likewise.
9541 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9543 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9544 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9546 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9548 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9549 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9550 (struct machine_function): Add tbegin_p.
9551 (s390_canonicalize_comparison): Fold CC mode compares to
9552 conditional jump if possible.
9553 (s390_emit_jump): Return the emitted jump.
9554 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9555 Handle CCRAWmode compares.
9556 (s390_option_override): Default to -mhtm if available.
9557 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9558 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9559 FPRs instead of df_regs_ever_live_p.
9560 (s390_optimize_nonescaping_tx): New function.
9561 (s390_init_frame_layout): Extend clobbered_regs array to cover
9563 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9564 (s390_expand_tbegin): New function.
9565 (enum s390_builtin): New enum definition.
9566 (code_for_builtin): New array definition.
9567 (s390_init_builtins): New function.
9568 (s390_expand_builtin): New function.
9569 (TARGET_INIT_BUILTINS): Define.
9570 (TARGET_EXPAND_BUILTIN): Define.
9571 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9572 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9573 (s390_alc_comparison): Likewise.
9574 * config/s390/s390-modes.def: Add CCRAWmode.
9575 * config/s390/s390.h (processor_flags): Add PF_TX.
9576 (TARGET_CPU_HTM): Define macro.
9577 (TARGET_HTM): Define macro.
9578 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9579 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9580 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9581 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9582 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9583 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9584 ("*cc_to_int"): Move up.
9585 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9586 constants other than 0.
9587 ("*ccraw_to_int"): New insn and splitter definition.
9588 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9589 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9590 ("tx_assist"): New expander.
9591 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9592 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9593 * config/s390/s390.opt: Add -mhtm option.
9594 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9595 * config/s390/htmxlintrin.h: New file.
9596 * config/s390/htmintrin.h: New file.
9597 * config/s390/s390intrin.h: New file.
9598 * doc/extend.texi: Document htm builtins.
9599 * config.gcc: Add the new header files to extra_headers.
9601 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9603 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9604 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9606 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9607 Pat Haugen <pthaugen@us.ibm.com>
9608 Peter Bergner <bergner@vnet.ibm.com>
9610 * config/rs6000/power8.md: New.
9611 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9612 setting for power8 entry.
9613 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9614 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9615 test for Power4/Power5 only.
9616 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9618 (force_new_group): Adjust comment.
9619 * config/rs6000/rs6000.md: Include power8.md.
9621 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9623 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9624 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9626 (tune_params): Update comment.
9627 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9628 (arm_max_conditional_execute): New function.
9629 (thumb2_final_prescan_insn): Use max_insn_skipped and
9630 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9632 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9634 PR tree-optimization/57705
9635 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9636 SSA_NAME step, provided that it is not defined inside the loop.
9637 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9638 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9640 2013-06-25 Martin Jambor <mjambor@suse.cz>
9643 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9644 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9646 (ipa_note_param_call): Initialize member_ptr flag.
9647 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9648 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9649 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9650 (ipa_read_indirect_edge_info): Likewise.
9652 2013-06-25 Richard Biener <rguenther@suse.de>
9655 * passes.c (init_optimization_passes): Move pass_fold_builtins
9656 and pass_dce earlier with -Og.
9658 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9660 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9661 <BIT_FIELD_REF>: Remove trailing TAB.
9662 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9665 2013-06-24 Martin Jambor <mjambor@suse.cz>
9667 PR tree-optimization/57358
9668 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9669 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9670 (ipa_analyze_params_uses): Generate pessimistic info when true.
9672 2013-06-24 Martin Jambor <mjambor@suse.cz>
9674 PR tree-optimization/57539
9675 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9676 global.inlined_to of the new node to it. All callers changed.
9677 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9678 inlining_into, pass it to cgraph_clone_node.
9679 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9680 ipa_free_edge_args_substructures.
9681 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9682 rdesc linked list. Do not assert rdesc edges have inlined caller.
9683 Assert we have found an rdesc in the rdesc list.
9685 2013-06-24 Richard Biener <rguenther@suse.de>
9687 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9688 (pointer_set_lookup): Declare.
9689 (class pointer_map): New template class implementing a
9690 generic pointer to T map.
9691 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9692 pointer_map<T>::contains, pointer_map<T>::insert,
9693 pointer_map<T>::traverse): New functions.
9694 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9695 (pointer_set_lookup): New function.
9696 (pointer_set_contains): Use pointer_set_lookup.
9697 (pointer_set_insert): Likewise.
9698 (insert_aux): Remove.
9699 (struct pointer_map_t): Embed a pointer_set_t.
9700 (pointer_map_create): Adjust.
9701 (pointer_map_destroy): Likewise.
9702 (pointer_map_contains): Likewise.
9703 (pointer_map_insert): Likewise.
9704 (pointer_map_traverse): Likewise.
9705 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9706 pointer_map<unsigned> instead of a pointer_map_t.
9707 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9708 (streamer_tree_cache_lookup): Likewise.
9709 (streamer_tree_cache_create): Likewise.
9710 (streamer_tree_cache_delete): Likewise.
9711 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9712 pointer_map<unsigned> instead of a pointer_map_t.
9713 (lto_init_tree_ref_encoder): Adjust.
9714 (lto_destroy_tree_ref_encoder): Likewise.
9715 * lto-section-out.c (lto_output_decl_index): Likewise.
9716 (lto_record_function_out_decl_state): Likewise.
9717 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9719 2013-06-24 Richard Biener <rguenther@suse.de>
9721 PR tree-optimization/57488
9722 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9724 2013-06-24 Alan Modra <amodra@gmail.com>
9726 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9727 (gen_easy_altivec_constant): Likewise.
9728 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9729 easy_vector_constant_msb): Likewise.
9731 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9734 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9735 add missing return true.
9737 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9740 * config/sh/predicates.md (general_extend_operand): Invoke
9741 general_movsrc_operand for memory operands.
9742 (general_movsrc_operand): Allow reg+reg addressing, do not use
9743 general_operand for memory operands.
9745 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9747 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9748 when current target options does not apply.
9749 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9750 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9751 * config/i386/bmiintrin.h: Pass appropriate target
9752 attributes to header.
9753 * config/i386/mmintrin.h: Ditto.
9754 * config/i386/nmmintrin.h: Ditto.
9755 * config/i386/avx2intrin.h: Ditto.
9756 * config/i386/fxsrintrin.h: Ditto.
9757 * config/i386/tbmintrin.h: Ditto.
9758 * config/i386/xsaveintrin.h: Ditto.
9759 * config/i386/f16cintrin.h: Ditto.
9760 * config/i386/xtestintrin.h: Ditto.
9761 * config/i386/xsaveoptintrin.h: Ditto.
9762 * config/i386/bmi2intrin.h: Ditto.
9763 * config/i386/lzcntintrin.h: Ditto.
9764 * config/i386/smmintrin.h: Ditto.
9765 * config/i386/wmmintrin.h: Ditto.
9766 * config/i386/x86intrin.h: Remove all header include guards.
9767 * config/i386/prfchwintrin.h: Ditto.
9768 * config/i386/pmmintrin.h: Ditto.
9769 * config/i386/tmmintrin.h: Ditto.
9770 * config/i386/xmmintrin.h: Ditto.
9771 * config/i386/popcntintrin.h: Ditto.
9772 * config/i386/rdseedintrin.h: Ditto.
9773 * config/i386/ammintrin.h: Ditto.
9774 * config/i386/emmintrin.h: Ditto.
9775 * config/i386/immintrin.h: Remove all header include guards.
9776 * config/i386/fma4intrin.h: Ditto.
9777 * config/i386/lwpintrin.h: Ditto.
9778 * config/i386/xopintrin.h: Ditto.
9779 * config/i386/ia32intrin.h: Ditto.
9780 * config/i386/avxintrin.h: Ditto.
9781 * config/i386/rtmintrin.h: Ditto.
9782 * config/i386/fmaintrin.h: Ditto.
9783 * config/i386/mm3dnow.h: Ditto.
9785 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9787 * common/config/i386/i386-common.c: Handle LZCNT.
9789 2013-06-22 Andi Kleen <ak@linux.intel.com>
9791 * doc/extend.texi: Use __atomic_store_n instead of
9792 __atomic_store in HLE example.
9794 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9796 * config/sh/sh.c: Remove <cstdlib> workaround.
9798 2013-06-21 Andi Kleen <ak@linux.intel.com>
9800 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9802 2013-06-21 Andi Kleen <ak@linux.intel.com>
9804 * doc/extend.texi: Document that __atomic_clear and
9805 __atomic_test_and_set should only be used with bool.
9807 2013-06-20 Jan Hubicka <jh@suse.cz>
9809 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9811 * tree.c (decls_same_for_odr): New function.
9812 (same_for_edr): New function.
9813 (types_same_for_odr): New function.
9814 (get_binfo_at_offset): Use it.
9815 * tree.h (types_same_for_odr): Declare.
9817 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9818 Jason Merrill <jason@redhat.com>
9820 * system.h: Include <cstdlib> as well as <stdlib.h>.
9822 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9825 * config/i386/i386.c (construct_container): Report error if
9826 long double is used with disabled x87 float returns.
9828 2013-06-20 Jan Hubicka <jh@suse.cz>
9830 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9832 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9834 PR rtl-optimization/57425
9835 PR rtl-optimization/57569
9836 * alias.c (write_dependence_p): Remove parameters mem_mode and
9837 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9838 Changed all callers.
9839 (canon_anti_dependence): Get comments and semantics in sync.
9840 Add parameter mem_canonicalized. Changed all callers.
9841 * rtl.h (canon_anti_dependence): Update prototype.
9843 2013-06-20 Richard Biener <rguenther@suse.de>
9845 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9846 byte case, inline streamer_read_uchar and defer section
9849 2013-06-20 Richard Biener <rguenther@suse.de>
9851 PR tree-optimization/57584
9852 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
9853 SSA names into the expanded expression that take part in
9854 abnormal coalescing.
9856 2013-06-19 Sharad Singhai <singhai@google.com>
9858 * gcov.c (print_usage): Handle new option.
9859 (process_args): Ditto.
9860 (get_gcov_intermediate_filename): New function.
9861 (output_intermediate_file): New function.
9862 (output_gcov_file): New function
9863 (generate_results): Handle new option.
9864 (release_function): Relase demangled name.
9865 (read_graph_file): Handle demangled name.
9866 (output_lines): Ditto.
9867 * doc/gcov.texi: Document gcov intermediate format.
9869 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
9872 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
9873 (lra_emit_add): Use the functions. Add comment about Y as an
9876 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
9879 * collect2.c (collect_atexit): New.
9880 (collect_exit): Delete.
9881 (main): Register collect_atexit with atexit.
9882 (collect_wait): Change collect_exit to exit.
9884 * collect2.h (collect_exit): Delete.
9885 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
9887 2013-06-19 Wei Mi <wmi@google.com>
9889 PR rtl-optimization/57518
9890 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
9891 if regno is used in paradoxical subreg.
9892 (update_equiv_regs): Check pdx_subregs[regno] before
9893 set a reg to be equivalent with a mem.
9895 2013-06-19 Matthias Klose <doko@ubuntu.com>
9898 * file-find.h (find_a_file): Add a mode parameter.
9899 * file-find.c (find_a_file): Likewise.
9900 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
9901 with X_OK for the executables.
9902 * collect2.c (main): Call find_a_file with X_OK.
9904 2013-06-19 Steve Ellcey <sellcey@mips.com>
9907 * config/mips/mips.md (casesi_internal_mips16_<mode>):
9908 Use NEXT_INSN instead of next_real_insn.
9910 2013-06-19 Jan Hubicka <jh@suse.cz>
9912 * cgraph.h (const_value_known_p): Replace by ...
9913 (ctor_for_folding): .. this one.
9914 * cgraphunit.c (process_function_and_variable_attributes): Use it.
9915 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
9916 * expr.c (expand_expr_real_1): Likewise.
9917 (string_constant): Likewise.
9918 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
9919 * ipa.c (process_references): Likewise.
9920 (symtab_remove_unreachable_nodes): Likewise.
9921 * ipa-inline-analysis.c (param_change_prob): Likewise.
9922 * gimple-fold.c (canonicalize_constructor_val): Likewise.
9923 (get_base_constructor): Likwise.
9924 * varpool.c (varpool_remove_node): Likewise.
9925 (varpool_remove_initializer): LIkewise.
9926 (dump_varpool_node): LIkwise.
9927 (const_value_known_p): Rewrite to ...
9928 (ctor_for_folding): ... this one.
9930 2013-06-19 Jakub Jelinek <jakub@redhat.com>
9933 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
9934 PERSONALITY in $PATH derived prefixes.
9936 2013-06-19 Jeff Law <law@redhat.com>
9938 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
9941 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
9942 (simplify_bitwise_binary): Use it to simpify certain binary ops on
9945 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
9947 * config/arm/vfp.md: Move VFP instruction classification documentation
9949 * config/arm/arm.md: ... here. Update instruction classification
9952 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
9954 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
9955 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
9956 pattern. Use more efficient sequences on ARMv5 and Thumb2.
9958 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
9961 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9962 with NEXT_INSN. Use tablejump_p to check for jump table data
9965 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
9968 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
9969 that now in C++ the value is correct per the C++ standards.
9971 2013-06-19 Richard Biener <rguenther@suse.de>
9973 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
9976 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9979 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9982 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9983 with next_active_insn.
9985 2013-06-18 Sriraman Tallam <tmsriram@google.com>
9987 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
9988 functions are inlined during failures to flag an error.
9989 * tree-inline.c (expand_call_inline): Allow the error to be flagged
9990 in early inline pass.
9992 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
9994 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
9997 2013-06-18 Julian Brown <julian@codesourcery.com>
9999 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10000 Permit virtual register pre-reload if !strict.
10001 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10003 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10005 * config/arm/neon.md (movmisalign<mode>): Use
10006 neon_perm_struct_or_reg_operand instead of
10007 neon_struct_or_register_operand.
10008 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10009 neon_permissive_struct_operand instead of neon_struct_operand.
10010 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10011 neon_vector_mem_operand.
10012 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10013 neon_vector_mem_operand.
10014 (neon_permissive_struct_operand): New.
10015 (neon_struct_or_register_operand): Rename to...
10016 (neon_perm_struct_or_reg_operand): This. Adjust call to
10017 neon_vector_mem_operand.
10019 2013-06-18 Richard Biener <rguenther@suse.de>
10021 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10022 * lto-streamer.h: Include pointer-set.h.
10023 (struct lto_decl_slot): Remove.
10024 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10025 Remove next_index entry.
10026 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10027 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10028 (lto_init_tree_ref_encoder): Adjust.
10029 (lto_destroy_tree_ref_encoder): Likewise.
10030 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10031 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10032 (lto_output_decl_index): Adjust.
10033 (lto_new_out_decl_state): Likewise.
10034 (lto_record_function_out_decl_state): Likewise.
10035 * lto-streamer-out.c (copy_function): Likewise.
10037 2013-06-18 Richard Biener <rguenther@suse.de>
10039 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10040 * cgraphunit.c: Include cfgloop.h.
10041 (init_lowered_empty_function): Initialize the loop tree.
10042 (assemble_thunk): Insert new BBs into loops.
10044 2013-06-18 Richard Biener <rguenther@suse.de>
10046 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10047 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10048 the map from cache entry to cache index optional.
10049 (streamer_tree_cache_replace_tree): Adjust accordingly.
10050 (streamer_tree_cache_append): Likewise.
10051 (streamer_tree_cache_delete): Likewise.
10052 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10053 streamer cache map from cache entry to cache index.
10054 * lto-streamer-out.c (create_output_block): Adjust.
10056 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10058 * config/arm/arm.md (attribute "insn"): Move multiplication and
10059 division attributes to...
10060 (attribute "type"): ... here. Remove mult.
10061 (attribute "mul32"): New attribute.
10062 (attribute "mul64"): Add umaal.
10063 (*arm_mulsi3): Update attributes.
10064 (*arm_mulsi3_v6): Likewise.
10065 (*thumb_mulsi3): Likewise.
10066 (*thumb_mulsi3_v6): Likewise.
10067 (*mulsi3_compare0): Likewise.
10068 (*mulsi3_compare0_v6): Likewise.
10069 (*mulsi_compare0_scratch): Likewise.
10070 (*mulsi_compare0_scratch_v6): Likewise.
10071 (*mulsi3addsi): Likewise.
10072 (*mulsi3addsi_v6): Likewise.
10073 (*mulsi3addsi_compare0): Likewise.
10074 (*mulsi3addsi_compare0_v6): Likewise.
10075 (*mulsi3addsi_compare0_scratch): Likewise.
10076 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10077 (*mulsi3subsi): Likewise.
10078 (*mulsidi3adddi): Likewise.
10079 (*mulsi3addsi_v6): Likewise.
10080 (*mulsidi3adddi_v6): Likewise.
10081 (*mulsidi3_nov6): Likewise.
10082 (*mulsidi3_v6): Likewise.
10083 (*umulsidi3_nov6): Likewise.
10084 (*umulsidi3_v6): Likewise.
10085 (*umulsidi3adddi): Likewise.
10086 (*umulsidi3adddi_v6): Likewise.
10087 (*smulsi3_highpart_nov6): Likewise.
10088 (*smulsi3_highpart_v6): Likewise.
10089 (*umulsi3_highpart_nov6): Likewise.
10090 (*umulsi3_highpart_v6): Likewise.
10091 (mulhisi3): Likewise.
10092 (*mulhisi3tb): Likewise.
10093 (*mulhisi3bt): Likewise.
10094 (*mulhisi3tt): Likewise.
10095 (maddhisi4): Likewise.
10096 (*maddhisi4tb): Likewise.
10097 (*maddhisi4tt): Likewise.
10098 (maddhidi4): Likewise.
10099 (*maddhidi4tb): Likewise.
10100 (*maddhidi4tt): Likewise.
10101 (divsi3): Likewise.
10102 (udivsi3): Likewise.
10103 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10104 (thumb2_mulsi_short_compare0): Likewise.
10105 (thumb2_mulsi_short_compare0_scratch): Likewise.
10106 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10107 (1020mult2): Likewise.
10108 (1020mult3): Likewise.
10109 (1020mult4): Likewise.
10110 (1020mult5): Likewise.
10111 (1020mult6): Likewise.
10112 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10114 (cortex_a15_mult64): Likewise.
10115 (cortex_a15_sdiv): Likewise.
10116 (cortex_a15_udiv): Likewise.
10117 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10123 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10124 (pj4_ir_div): Likewise.
10125 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10126 (11_mult2): Likewise.
10127 (11_mult3): Likewise.
10128 (11_mult4): Likewise.
10129 (11_mult5): Likewise.
10130 (11_mult6): Likewise.
10131 (11_mult7): Likewise.
10132 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10133 (cortex_a8_mla): Likewise.
10134 (cortex_a8_mull): Likewise.
10135 (cortex_a8_smulwy): Likewise.
10136 (cortex_a8_smlald): Likewise.
10137 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10138 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10139 (cortex_r4_mul_3): Likewise.
10140 (cortex_r4_mla_4): Likewise.
10141 (cortex_r4_mla_3): Likewise.
10142 (cortex_r4_smlald): Likewise.
10143 (cortex_r4_mull): Likewise.
10144 (cortex_r4_sdiv): Likewise.
10145 (cortex_r4_udiv): Likewise.
10146 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10147 (cortex_a7_idiv): Likewise.
10148 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10149 (9_mult2): Likewise.
10150 (9_mult3): Likewise.
10151 (9_mult4): Likewise.
10152 (9_mult5): Likewise.
10153 (9_mult6): Likewise.
10154 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10155 (cortex_a53_sdiv): Likewise.
10156 (cortex_a53_udiv): Likewise.
10157 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10158 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10159 (mp626_mult2): Likewise.
10160 (mp626_mult3): Likewise.
10161 (mp626_mult4): Likewise.
10162 * config/arm/fa526.md (526_mult1): Update attribute change.
10163 (526_mult2): Likewise.
10164 * config/arm/arm-generic.md (mult): Update attribute change.
10165 (mult_ldsched_strongarm): Likewise.
10166 (mult_ldsched): Likewise.
10167 (multi_cycle): Likewise.
10168 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10169 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10170 (606te_mult2): Likewise.
10171 (606te_mult3): Likewise.
10172 (606te_mult4): Likewise.
10173 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10174 (cortex_a9_mac16): Likewise.
10175 (cortex_a9_multiply): Likewise.
10176 (cortex_a9_mac): Likewise.
10177 (cortex_a9_multiply_long): Likewise.
10178 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10179 (626te_mult2): Likewise.
10180 (626te_mult3): Likewise.
10181 (626te_mult4): Likewise.
10183 2013-06-18 Richard Biener <rguenther@suse.de>
10186 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10188 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10191 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10192 with next_active_insn.
10194 2013-06-18 Alan Modra <amodra@gmail.com>
10196 * config/rs6000/rs6000.h (enum data_align): New.
10197 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10198 (DATA_ABI_ALIGNMENT): Define.
10199 (CONSTANT_ALIGNMENT): Correct comment.
10200 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10201 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10203 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10205 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10206 ATTRIBUTE_UNUSED marking.
10208 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10210 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10211 alternative and update.
10212 (aarch64_dup_lanedi): Delete.
10213 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10214 * config/aarch64/aarch64-simd-builtins.def: Update.
10216 2013-06-17 Richard Biener <rguenther@suse.de>
10218 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10219 (lto_input_scc): Declare.
10220 (lto_input_tree_1): Likewise.
10221 (struct lto_stats_d): Add num_tree_bodies_output and
10222 num_pickle_refs_output.
10223 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10224 (lto_read_tree_1): Split out from ...
10225 (lto_read_tree): ... this.
10226 (lto_input_scc): New function.
10227 (lto_input_tree_1): Split out from ...
10228 (lto_input_tree): ... this. Handle LTO_tree_scc.
10229 (lto_data_in_create): Create the streamer cache without hashes.
10230 * lto-streamer-out.c (create_output_block): Create the streamer
10231 cache with hashes when not doing WPA.
10232 (lto_write_tree_1): Split out from ...
10233 (lto_write_tree): ... this.
10234 (get_symbol_initial_value): New function.
10235 (lto_output_tree_1): Split out from ...
10236 (lto_output_tree): ... this. Write trees as series of SCCs
10237 using a DFS walk via DFS_write_tree.
10238 (struct sccs, struct scc_entry): New types.
10239 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10240 (DFS_write_tree_body): New function.
10241 (DFS_write_tree): Likewise.
10242 (hash_tree): Likewise.
10243 (scc_entry_compare): Likewise.
10244 (hash_scc): Likewise.
10245 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10246 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10247 TREE_CHAIN as regular reference.
10248 (streamer_read_integer_cst): Remove.
10249 (streamer_get_pickled_tree): Adjust.
10250 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10251 of DECL_EXTERNALs in BLOCK_VARS for now.
10252 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10254 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10255 Add hash value argument and record that if hashes are recorded
10257 (streamer_tree_cache_insert_1): Adjust.
10258 (streamer_tree_cache_insert): Likewise.
10259 (streamer_tree_cache_insert_at): Rename to ...
10260 (streamer_tree_cache_replace_tree): ... this and adjust.
10261 (streamer_tree_cache_append): Adjust.
10262 (record_common_node): Likewise.
10263 (streamer_tree_cache_create): Add argument whether to
10264 record hash values together with trees.
10265 (streamer_tree_cache_delete): Adjust.
10266 * tree-streamer.h (struct streamer_tree_cache_d): Add
10268 (streamer_read_integer_cst): Remove.
10269 (streamer_tree_cache_insert): Adjust.
10270 (streamer_tree_cache_append): Likewise.
10271 (streamer_tree_cache_insert_at): Rename to ...
10272 (streamer_tree_cache_replace_tree): ... this and adjust.
10273 (streamer_tree_cache_create): Add argument whether to record hashes.
10274 (streamer_tree_cache_get): Rename to ...
10275 (streamer_tree_cache_get_tree): ... this.
10276 (streamer_tree_cache_get_hash): New function.
10277 * tree.c (cache_integer_cst): New function.
10278 * tree.h (cache_integer_cst): Declare.
10279 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10280 * lto-symtab.c (lto_varpool_replace_node): Only release
10281 DECL_INITIAL of non-prevailing decls.
10282 * varpool.c (varpool_remove_initializer): Do not release
10283 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10285 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10287 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10288 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10289 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10290 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10291 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10292 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10293 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10294 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10295 instead of TARGET_64BIT.
10296 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10297 Require ISA_HAS_<D>DIV.
10299 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10301 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10302 (mips*-*-linux*): Move default with_llsc setting to where other
10304 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10306 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10307 Likewise. Remove default with_tune setting. Move default float
10308 setting to its own block. Handle with_llsc in the same block as above.
10310 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10312 PR rtl-optimization/57425
10313 PR rtl-optimization/57569
10314 * alias.c (write_dependence_p): Add new parameters mem_mode,
10315 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10316 Changed all callers.
10317 (canon_anti_dependence): New function.
10318 * cse.c (check_dependence): Use canon_anti_dependence.
10319 * cselib.c (cselib_invalidate_mem): Likewise.
10320 * rtl.h (canon_anti_dependence): Declare.
10322 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10324 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10325 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10326 ".set mips3" for 64-bit targets.
10328 2013-06-15 Dehao Chen <dehao@google.com>
10330 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10331 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10332 (gimple_check_call_args): Likewise.
10333 * value-prof.c (check_ic_target): Likewise.
10334 * ipa-inline.c (early_inliner): Likewise.
10335 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10336 * cgraph.c (cgraph_create_edge_1): Likewise.
10337 (cgraph_make_edge_direct): Likewise.
10339 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10342 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10343 rs6000_output_move_128bit to handle emitting quad memory
10344 operations. Set attribute length to 8 bytes.
10346 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10348 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10350 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10351 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10352 (aarch64_<su>mlsl<mode>): Likewise.
10354 2013-06-14 Mike Stump <mikestump@comcast.net>
10356 * Makefile.in (TARGET_H): Add insn-codes.h.
10358 2013-06-14 Alan Modra <amodra@gmail.com>
10360 PR middle-end/57134
10361 PR middle-end/57586
10362 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10363 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10364 bitfield expansion when EXPAND_MEMORY.
10365 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10367 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10369 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10370 test for clearing quad memory on 32-bit later.
10372 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10374 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10375 (fold_negate_expr): Likewise.
10376 (fold_real_zero_addition_p): Handle vectors.
10377 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10379 2013-06-14 Alan Modra <amodra@gmail.com>
10381 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10383 2013-06-13 Jan Hubicka <jh@suse.cz>
10385 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10386 Local comdats are not externally visible.
10387 * symtab.c (dump_symtab_base): Dump externally visible.
10388 (verify_symtab_base): Verify back links in the symtab hash.
10390 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10392 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10393 CONVERT_EXPR as equal nodes.
10395 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10397 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10399 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10401 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10402 Generalize to complex and vector.
10403 * tree.c (build_all_ones_cst): New function.
10404 * tree.h (build_all_ones_cst): Declare it.
10406 2013-06-13 Alan Modra <amodra@gmail.com>
10408 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10409 * config/rs6000/rs6000.md (signbittf2): New insn.
10410 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10411 (abstf2_internal, cmptf_internal2): Likewise.
10412 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10414 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10415 Pat Haugen <pthaugen@us.ibm.com>
10416 Peter Bergner <bergner@vnet.ibm.com>
10418 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10419 power8 byte, half-word, and quad-word atomic instructions.
10420 (emit_store_conditional): Likewise.
10421 (rs6000_expand_atomic_compare_and_swap): Likewise.
10422 (rs6000_expand_atomic_op): Likewise.
10424 * config/rs6000/sync.md (larx): Add new modes for power8.
10426 (AINT): New mode iterator to include TImode as well as normal
10427 integer modes on power8.
10428 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10429 that VSX registers are not considered. Use AINT mode iterator
10430 instead of INT1 to allow inclusion of quad word atomic operations
10432 (load_locked<mode>): Likewise.
10433 (store_conditional<mode>): Likewise.
10434 (atomic_compare_and_swap<mode>): Likewise.
10435 (atomic_exchange<mode>): Likewise.
10436 (atomic_nand<mode>): Likewise.
10437 (atomic_fetch_<fetchop_name><mode>): Likewise.
10438 (atomic_nand_fetch<mode>): Likewise.
10439 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10441 (ATOMIC): On power8, add QImode, HImode modes.
10442 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10443 modes that promote to SImode.
10444 (load_lockedti): Convert TImode arguments to PTImode, so that we
10445 get a guaranteed even/odd register pair.
10446 (load_lockedpti): Likewise.
10447 (store_conditionalti): Likewise.
10448 (store_conditionalpti): Likewise.
10450 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10451 atomic load/store instructions.
10454 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10456 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10458 (insn_count): New attribute, with most cases extracted from...
10459 (length): ...here. Redefine most cases in terms of insn_count.
10460 (single_insn): Delete.
10461 (can_delay): Use insn_count to check for single instructions.
10462 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10463 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10464 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10465 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10466 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10467 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10468 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10469 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10470 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10471 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10472 rather than "length".
10473 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10474 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10475 Use "insn_count" rather than "length".
10476 * config/mips/mips-dsp.md
10477 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10478 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10481 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10483 PR tree-optimization/57361
10484 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10486 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10488 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10490 (aarch64_simd_combine<mode>): New instruction expansion.
10491 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10492 function prototype.
10493 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10494 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10496 2013-06-12 Jan Hubicka <jh@suse.cz>
10498 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10499 decl has when in streaming stage.
10500 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10501 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10503 2013-06-12 Roland Stigge <stigge@antcom.de>
10506 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10508 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10510 PR tree-optimization/57537
10511 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10512 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10514 2013-06-12 Richard Biener <rguenther@suse.de>
10516 * data-streamer.h (streamer_write_char_stream): CSE
10517 obs->current_pointer.
10518 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10519 streamer_write_char_stream manually and optimize the resulting loop.
10520 (streamer_write_hwi_stream): Likewise.
10522 2013-06-12 Jan Hubicka <jh@suse.cz>
10524 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10525 * cgraph.h (varpool_create_empty_node): Declare.
10526 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10528 * symtab.c (symtab_unregister_node): Be lax about missin entries
10530 (symtab_get_node): Update comment.
10531 * varpool.c (varpool_create_empty_node): Break out from ...
10532 (varpool_node_for_decl): ... here.
10533 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10535 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10537 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10538 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10539 part. Use straight-line flow at the end.
10540 <COMPONENT_REF>: Remove superfluous else.
10541 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10543 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10546 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10547 target hook even for !TREE_PUBLIC decls. If no resolution info
10548 is available, return false for common and external decls.
10550 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10552 * config/rl78/constraints.md (U): New constraint.
10553 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10556 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10559 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10560 to allow returning address to AT_PLATFORM name.
10562 2013-06-11 Jan Hubicka <jh@suse.cz>
10564 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10565 * cgraph.h (symtab_node_base): Add weakref flag.
10566 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10567 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10568 (output_weakrefs): Use weakref flag.
10569 * fold-const.c (simple_operand_p): Handle WEAK.
10570 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10571 * ipa.c (varpool_externally_visible_p): Drop weakref.
10572 (function_and_variable_visibility): Update comment; fix weakref
10573 sanity checks; do not clear DECL_WEAK on them.
10574 * lto-cgraph.c (lto_output_node): update.
10575 (lto_output_varpool_node): Update.
10576 (input_overwrite_node): Update.
10577 (input_node): Update.
10578 (input_varpool_node): Update.
10579 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10580 (lto_symtab_merge_symbols): Add sanity check.
10581 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10582 * passes.c (rest_of_decl_compilation): Set static flag, too.
10583 * symtab.c (dump_symtab_base): Dump weakref.
10584 (verify_symtab_base): Sanity check weakrefs.
10585 (symtab_make_decl_local): Remove duplicated code.
10586 (symtab_alias_ultimate_target): Simplify.
10587 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10589 2013-06-11 Tom de Vries <tom@codesourcery.com>
10591 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10592 sequence_vect == NULL.
10594 2013-06-11 DJ Delorie <dj@redhat.com>
10596 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10597 (rl78_unwind_word_mode): New.
10599 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10601 * final.c (debug_prefix_maps): Make static.
10603 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10605 * function.c (initial_trampoline): Remove stray copy.
10607 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10609 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10611 2013-06-11 Martin Jambor <mjambor@suse.cz>
10613 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10614 within bounds at the beginning of the function.
10616 2013-06-11 Alan Modra <amodra@gmail.com>
10618 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10620 (get_named_section): Don't NULL !DECL_P decl.
10622 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10624 * doc/invoke.texi (core-avx2): Document.
10626 (atom): Updated with MOVBE.
10628 2013-06-11 Richard Biener <rguenther@suse.de>
10630 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10632 2013-06-11 Anton Blanchard <anton@samba.org>
10634 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10635 correct shift value in little-endian mode.
10637 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10640 * varasm.c (get_variable_align): Move #endif to the right place.
10642 2013-06-10 Cary Coutant <ccoutant@google.com>
10644 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10645 for hash so that hash table traversal order is deterministic.
10647 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10648 Pat Haugen <pthaugen@us.ibm.com>
10649 Peter Bergner <bergner@vnet.ibm.com>
10651 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10652 of vectors in GPRS if they are direct moves or quad word load or
10655 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10657 (direct_move_p): Likewise.
10658 (quad_load_store_p): Likewise.
10660 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10661 classes into bins based on the physical register type.
10662 (reg_class_to_reg_type): Likewise.
10663 (IS_STD_REG_TYPE): Likewise.
10664 (IS_FP_VECT_REG_TYPE): Likewise.
10665 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10666 use direct move instructions.
10667 (reload_gpr_vsx): Likewise.
10668 (reload_vsx_gpr): Likewise.
10669 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10670 information that is a simplification of register classes. Also
10671 precalculate direct move reload helpers.
10672 (direct_move_p): New function to return true if the operation can
10673 be done as a direct move instruciton.
10674 (quad_load_store_p): New function to return true if the operation
10675 is a quad memory operation.
10676 (rs6000_legitimize_address): If quad memory, only allow register
10677 indirect for TImode addresses.
10678 (rs6000_legitimate_address_p): Likewise.
10679 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10680 (rs6000_reload_register_type): Likewise.
10681 (register_to_reg_type): Return register type.
10682 (rs6000_secondary_reload_simple_move): New helper function for
10683 secondary reload and secondary memory needed to identify anything
10684 that is a simple move, and does not need reloading.
10685 (rs6000_secondary_reload_direct_move): New helper function for
10686 secondary reload to identify cases that can be done with several
10687 instructions via the direct move instructions.
10688 (rs6000_secondary_reload_move): New helper function for secondary
10689 reload to identify moves between register types that can be done.
10690 (rs6000_secondary_reload): Add support for quad memory operations
10691 and for direct move.
10692 (rs6000_secondary_memory_needed): Likewise.
10693 (rs6000_debug_secondary_memory_needed): Change argument names.
10694 (rs6000_output_move_128bit): New function to return the move to
10695 use for 128-bit moves, including knowing about the various
10696 limitations of quad memory operations.
10698 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10699 memory operations. call rs6000_output_move_128bit for the actual
10700 instruciton(s) to generate.
10701 (vsx_movti_64bit): Likewise.
10703 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10704 (UNSPEC_P8V_MTVSRWZ): Likewise.
10705 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10706 (UNSPEC_P8V_MTVSRD): Likewise.
10707 (UNSPEC_P8V_XXPERMDI): Likewise.
10708 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10709 (UNSPEC_FUSION_GPR): Likewise.
10710 (FMOVE128_GPR): New iterator for direct move.
10711 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10712 (f32_sv): Likewise.
10713 (f32_dm): Likewise.
10714 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10715 loads and direct move instructions.
10716 (zero_extendsidi2_lfiwzx): Likewise.
10717 (extendsidi2_lfiwax): Likewise.
10718 (extendsidi2_nocell): Likewise.
10719 (floatsi<mode>2_lfiwax): Likewise.
10720 (lfiwax): Likewise.
10721 (floatunssi<mode>2_lfiwzx): Likewise.
10722 (lfiwzx): Likewise.
10723 (fix_trunc<mode>_stfiwx): Likewise.
10724 (fixuns_trunc<mode>_stfiwx): Likewise.
10725 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10726 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10727 (parity<mode>2_cmpb): Set length/type attr.
10728 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10729 for 'mr.' to fast_compare.
10730 (bpermd_<mode>): Change type attr to popcnt.
10731 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10732 (p8_mtvsrwz_1): Likewise.
10733 (p8_mtvsrwz_2): Likewise.
10734 (reload_fpr_from_gpr<mode>): Likewise.
10735 (p8_mtvsrd_1): Likewise.
10736 (p8_mtvsrd_2): Likewise.
10737 (p8_xxpermdi_<mode>): Likewise.
10738 (reload_vsx_from_gpr<mode>): Likewise.
10739 (reload_vsx_from_gprsf): Likewise.
10740 (p8_mfvsrd_3_<mode>): LIkewise.
10741 (reload_gpr_from_vsx<mode>): Likewise.
10742 (reload_gpr_from_vsxsf): Likewise.
10743 (p8_mfvsrd_4_disf): Likewise.
10744 (multi-word GPR splits): Do not split direct moves or quad memory
10747 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10749 * tree-into-ssa.c (interesting_blocks): Make static.
10751 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10754 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10755 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10756 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10757 (get_variable_align): New function.
10758 (get_variable_section, emit_bss, emit_common,
10759 assemble_variable_contents, place_block_symbol): Use
10760 get_variable_align instead of DECL_ALIGN.
10761 (assemble_noswitch_variable): Add align argument, use it
10762 instead of DECL_ALIGN.
10763 (assemble_variable): Adjust caller. Use get_variable_align
10764 instead of DECL_ALIGN.
10765 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10767 (DATA_ABI_ALIGNMENT): Define.
10768 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10769 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10770 opt is false, only return the psABI mandated alignment increase.
10771 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10772 (DATA_ABI_ALIGNMENT): ... this.
10773 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10774 (DATA_ABI_ALIGNMENT): ... this.
10775 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10776 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10777 (DATA_ABI_ALIGNMENT): ... this.
10778 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10779 * doc/tm.texi: Regenerated.
10781 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10783 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10784 cmp_code to construct REG_EQUAL note.
10786 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10789 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10790 that operands[2] doesn't overlap with operands[0].
10792 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10793 Jan Hubicka <jh@suse.cz>
10795 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10796 hack to mark symbols as used.
10798 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10800 PR rtl-optimization/57559
10801 * lra-constraints.c (process_alt_operands): Don't discourage
10802 memory with known offset for offsetable memory constraint.
10803 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10805 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10807 * varasm.c (struct oc_local_state): Reorder fields.
10808 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10809 and adjust accordingly.
10810 (output_constructor): Reorder initialization code and adjust call to
10811 output_constructor_bitfield.
10813 2013-06-07 Jan Hubicka <jh@suse.cz>
10815 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10817 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10819 * tree-object-size.c (unknown): Make const.
10821 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10823 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10824 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10825 for last alternative in the cpu_facility attribute.
10827 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10830 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10831 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10832 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10833 * config/arm/constraints.md (Dg): New constraint.
10834 * config/arm/neon.md (xordi3_neon): Remove.
10835 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10836 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10838 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10840 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10841 Clean up alternatives.
10843 2013-06-07 Alan Modra <amodra@gmail.com>
10845 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10848 2013-06-07 Alan Modra <amodra@gmail.com>
10850 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
10852 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10854 * config/arm/constraints.md (Df): New constraint.
10855 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
10856 Correct length attribute for last two alternatives.
10858 2013-06-07 Alan Modra <amodra@gmail.com>
10860 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10861 override user -mfp-in-toc.
10862 (offsettable_ok_by_alignment): Consider just the current access
10863 rather than the whole object, unless BLKmode. Handle
10864 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
10865 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
10866 for -mcmodel=medium.
10867 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
10868 override user -mfp-in-toc or -msum-in-toc. Default to
10869 -mno-fp-in-toc for -mcmodel=medium.
10871 2013-06-06 DJ Delorie <dj@redhat.com>
10873 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
10874 TARGET_VALID_POINTER_MODE.
10876 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
10877 Pat Haugen <pthaugen@us.ibm.com>
10878 Peter Bergner <bergner@vnet.ibm.com>
10880 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
10881 Document new power8 builtins.
10883 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
10884 condition code register, to allow 128-bit logical operations to be
10885 done in the VSX or GPR registers.
10886 (nor<mode>3): Use the canonical form for nor.
10887 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
10888 vclz*, and vpopcnt* vector instructions.
10889 (nand<mode>3): Likewise.
10890 (orc<mode>3): Likewise.
10891 (clz<mode>2): LIkewise.
10892 (popcount<mode>2): Likewise.
10894 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
10895 that only the GPRs are recognized.
10897 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10898 support for new power8 builtins.
10900 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
10902 (xscvdpspn): Likewise.
10908 (vpopcnt): Likewise.
10909 (vpopcntb): Likewise.
10910 (vpopcnth): Likewise.
10911 (vpopcntw): Likewise.
10912 (vpopcntd): Likewise.
10914 (vmrgew): Likewise.
10915 (vmrgow): Likewise.
10917 (eqv_v16qi3): Likewise.
10918 (eqv_v8hi3): Likewise.
10919 (eqv_v4si3): Likewise.
10920 (eqv_v2di3): Likewise.
10921 (eqv_v4sf3): Likewise.
10922 (eqv_v2df3): Likewise.
10924 (nand_v16qi3): Likewise.
10925 (nand_v8hi3): Likewise.
10926 (nand_v4si3): Likewise.
10927 (nand_v2di3): Likewise.
10928 (nand_v4sf3): Likewise.
10929 (nand_v2df3): Likewise.
10931 (orc_v16qi3): Likewise.
10932 (orc_v8hi3): Likewise.
10933 (orc_v4si3): Likewise.
10934 (orc_v2di3): Likewise.
10935 (orc_v4sf3): Likewise.
10936 (orc_v2df3): Likewise.
10938 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
10939 allow power8 quad mode in 64-bit.
10940 (rs6000_builtin_vectorized_function): Add support to vectorize
10941 ISA 2.07 count leading zeros, population count builtins.
10942 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
10943 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
10944 (builtin_function_type): Add vgbbd builtin function which takes an
10946 (altivec_expand_vec_perm_const): Add support for new power8 merge
10949 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
10950 that does not include TImdoe for use with 32-bit.
10951 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
10953 (UNSPEC_VSX_CVDPSPN): Likewise.
10954 (vsx_xscvdpspn): Likewise.
10955 (vsx_xscvspdpn): Likewise.
10956 (vsx_xscvdpspn_scalar): Likewise.
10957 (vsx_xscvspdpn_directmove): Likewise.
10958 (vsx_and<mode>3): Split logical operations into 32-bit and
10959 64-bit. Add support to do logical operations on TImode as well as
10960 VSX vector types. Allow logical operations to be done in either
10961 VSX registers or in general purpose registers in 64-bit mode. Add
10962 splitters if GPRs were used. For AND, add clobber of CCmode to
10963 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
10965 (vsx_and<mode>3_32bit): Likewise.
10966 (vsx_and<mode>3_64bit): Likewise.
10967 (vsx_ior<mode>3): Likewise.
10968 (vsx_ior<mode>3_32bit): Likewise.
10969 (vsx_ior<mode>3_64bit): Likewise.
10970 (vsx_xor<mode>3): Likewise.
10971 (vsx_xor<mode>3_32bit): Likewise.
10972 (vsx_xor<mode>3_64bit): Likewise.
10973 (vsx_one_cmpl<mode>2): Likewise.
10974 (vsx_one_cmpl<mode>2_32bit): Likewise.
10975 (vsx_one_cmpl<mode>2_64bit): Likewise.
10976 (vsx_nor<mode>3): Likewise.
10977 (vsx_nor<mode>3_32bit): Likewise.
10978 (vsx_nor<mode>3_64bit): Likewise.
10979 (vsx_andc<mode>3): Likewise.
10980 (vsx_andc<mode>3_32bit): Likewise.
10981 (vsx_andc<mode>3_64bit): Likewise.
10982 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
10983 and xxlorc instructions.
10984 (vsx_eqv<mode>3_64bit): Likewise.
10985 (vsx_nand<mode>3_32bit): Likewise.
10986 (vsx_nand<mode>3_64bit): Likewise.
10987 (vsx_orc<mode>3_32bit): Likewise.
10988 (vsx_orc<mode>3_64bit): Likewise.
10990 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
10992 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
10994 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
10995 (p8_vmrgow): Likewise.
10996 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
10997 GPRs to be split under VSX.
10998 (p8v_clz<mode>2): Add power8 count leading zero support.
10999 (p8v_popcount<mode>2): Add power8 population count support.
11000 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11003 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11006 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11008 (vec_nand): Likewise.
11009 (vec_vclz): Likewise.
11010 (vec_vclzb): Likewise.
11011 (vec_vclzd): Likewise.
11012 (vec_vclzh): Likewise.
11013 (vec_vclzw): Likewise.
11014 (vec_vgbbd): Likewise.
11015 (vec_vmrgew): Likewise.
11016 (vec_vmrgow): Likewise.
11017 (vec_vpopcnt): Likewise.
11018 (vec_vpopcntb): Likewise.
11019 (vec_vpopcntd): Likewise.
11020 (vec_vpopcnth): Likewise.
11021 (vec_vpopcntw): Likewise.
11023 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11025 PR rtl-optimization/57468
11026 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11029 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11031 PR rtl-optimization/57459
11032 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11033 type when setting live regs.
11035 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11037 * config/s390/s390.opt (mlra): New option.
11038 * config/s390/s390.c (s390_decompose_address): Check displacement
11039 for all registers for LRA.
11040 (s390_secondary_reload): Don't used secondary reloads for LRA.
11041 (s390_lra_p): New function.
11042 (TARGET_LRA_P): Define.
11043 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11044 of attribute cpu_facility to zarch for the last alternative.
11045 (*cmpmem_short): Ditto.
11047 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11049 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11050 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11051 (arm_expand_prologue): Likewise.
11053 2013-06-06 Teresa Johnson <tejohnson@google.com>
11056 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11057 as this is now done by redirect_edge_and_branch_force.
11058 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11059 barriers, and fix interaction with splitting.
11060 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11061 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11062 reflect changes made in the routine.
11063 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11064 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11065 since this is called in cfglayout mode, and replace partition fixup
11066 with assert as that is now done by force_nonfallthru_and_redirect.
11067 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11068 already be marked with region crossing note.
11069 (insert_section_boundary_note): Make non-static, gate on flag
11070 has_bb_partition, rewrite to also check for multiple partitions.
11071 (rest_of_handle_reorder_blocks): Remove call to
11072 insert_section_boundary_note, now done later during free_cfg.
11073 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11074 * bb-reorder.h (insert_section_boundary_note): Declare.
11075 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11076 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11077 invoke insert_section_boundary_note.
11078 (try_redirect_by_replacing_jump): Remove unnecessary
11079 check for region crossing note.
11080 (fixup_partition_crossing): New function.
11081 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11082 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11083 in non-cfglayout mode.
11084 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11085 remove old code that tried to do this. Emit barrier correctly
11086 when we are in cfglayout mode.
11087 (last_bb_in_partition): New function.
11088 (rtl_split_edge): Correctly fixup partition boundaries.
11089 (commit_one_edge_insertion): Remove old code that tried to
11090 fixup region crossing edge since this is now handled in
11091 split_block, and set up insertion point correctly since
11092 block may now end in a jump.
11093 (verify_hot_cold_block_grouping): Guard against checking when not in
11094 linearized RTL mode.
11095 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11097 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11098 rtl_verify_flow_info, so not called in cfglayout mode.
11099 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11100 (fixup_reorder_chain): Remove old code that attempted to fixup region
11101 crossing note as this is now handled in force_nonfallthru_and_redirect.
11102 (duplicate_insn_chain): Don't duplicate switch section notes.
11103 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11105 * basic-block.h (emit_barrier_after_bb): Declare.
11107 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11109 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11110 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11111 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11113 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11115 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11117 * config/arm/ldmstm.md: Regenerate.
11119 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11121 * config/arm/sync.md (atomic_loaddi_1):
11122 Disable predication for arm_restrict_it.
11123 (arm_load_exclusive<mode>): Likewise.
11124 (arm_load_exclusivesi): Likewise.
11125 (arm_load_exclusivedi): Likewise.
11126 (arm_load_acquire_exclusive<mode>): Likewise.
11127 (arm_load_acquire_exclusivesi): Likewise.
11128 (arm_load_acquire_exclusivedi): Likewise.
11129 (arm_store_exclusive<mode>): Likewise.
11130 (arm_store_exclusive<mode>): Likewise.
11131 (arm_store_release_exclusivedi): Likewise.
11132 (arm_store_release_exclusive<mode>): Likewise.
11134 2013-06-06 Richard Biener <rguenther@suse.de>
11136 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11138 (lto_tag_is_tree_code_p): Adjust.
11139 (lto_tag_is_gimple_code_p): Likewise.
11140 (lto_gimple_code_to_tag): Likewise.
11141 (lto_tag_to_gimple_code): Likewise.
11142 (lto_tree_code_to_tag): Likewise.
11143 (lto_tag_to_tree_code): Likewise.
11144 * data-streamer.h (streamer_write_hwi_in_range): Use
11145 uhwi streaming to stream the normalized range.
11146 (streamer_read_hwi_in_range): Likewise.
11148 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11150 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11151 (predicable_short_it): Likewise.
11152 (predicated): Likewise.
11153 (enabled): Handle above.
11154 (define_cond_exec): Set predicated attribute to yes.
11156 2013-06-05 Mike Stump <mikestump@comcast.net>
11158 * gdbinit.in (__FUNCTION__): Add.
11160 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11162 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11163 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11165 2013-06-05 Jan Hubicka <jh@suse.cz>
11167 * varasm.c (mark_decl_referenced): Revert the removal until targets
11170 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11172 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11173 instead of mark_decl_referenced.
11175 2013-06-05 Jan Hubicka <jh@suse.cz>
11177 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11178 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11179 and symtab_used_from_object_file_p.
11180 (cgraph_make_node_local_1): Clear forced_by_abi.
11181 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11182 * cgraph.h (symtab_node_base): Add forced_by_abi.
11183 (decide_is_variable_needed): Remove.
11184 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11185 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11186 (decide_is_symbol_needed): ... this one; handle symbols in general;
11187 always analyze virtuals; honnor forced_by_abi.
11188 (cgraph_finalize_function): Update.
11189 (varpool_finalize_decl): Update.
11190 (symbol_defined_and_needed): Remove.
11191 (analyze_functions): Update.
11192 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11193 output_refs, input_overwrite_node): Handle forced_by_abi.
11194 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11195 (address_taken_from_non_vtable_p): ... this one.
11196 (comdat_can_be_unshared_p_1): New function.
11197 (cgraph_comdat_can_be_unshared_p): Rename to ...
11198 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11199 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11200 (function_and_variable_visibility): Clear forced_by_abi as needed.
11201 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11202 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11203 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11204 * varpool.c (decide_is_variable_needed): Remove.
11206 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11208 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11209 (arm_option_override): Override arm_restrict_it where appropriate.
11210 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11211 * config/arm/arm.opt (mrestrict-it): New command-line option.
11212 * doc/invoke.texi: Document -mrestrict-it.
11214 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11216 * tsan.c (tsan_atomic_table): Make const.
11218 2013-06-05 Richard Biener <rguenther@suse.de>
11220 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11221 index associated with the tree we are supposed to replace.
11222 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11223 TREE_ASM_WRITTEN as zero for everything but SSA names.
11225 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11227 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11229 2013-06-04 Jan Hubicka <jh@suse.cz>
11231 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11232 (input_node, input_varpool_node): Handle correctly external same
11234 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11235 nodes at ltrans stage.
11237 2013-06-04 Jan Hubicka <jh@suse.cz>
11239 * ipa-inline.c (update_caller_keys): Fix availability test.
11240 (update_callee_keys): Likewise.
11241 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11242 to follow ELF standard.
11244 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11246 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11247 (mips64r5900el-*-elf*): New configurations.
11248 * config/mips/mips-cpus.def (r5900): New processor.
11249 * config/mips/mips-tables.opt: Regenerate.
11250 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11251 (mips_issue_rate): Handle PROCESSOR_R5900.
11252 (mips_reorg_process_insns): Force reorder mode for the R5900.
11253 * config/mips/mips.h (TARGET_MIPS5900): Define.
11254 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11256 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11258 * config/mips/mips.md (processor): Add r5900.
11259 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11261 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11263 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11264 into function to generate MOVI instruction.
11265 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11266 (aarch64_preferred_simd_mode): Turn into wrapper.
11267 (aarch64_output_scalar_simd_mov_immediate): New function.
11268 * config/aarch64/aarch64-protos.h: Add prototype for above.
11270 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11272 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11273 element_char member.
11274 (sizetochar): Return signed char.
11275 (aarch64_simd_valid_immediate): Remove elchar and other
11276 unnecessary variables.
11277 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11278 Calculate element_char as required.
11279 * config/aarch64/aarch64-protos.h: Update and move prototype
11280 for aarch64_output_simd_mov_immediate.
11281 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11284 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11286 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11287 information completed by aarch64_simd_valid_immediate.
11288 (aarch64_legitimate_constant_p): Update arguments.
11289 (aarch64_simd_valid_immediate): Work with struct rather than many
11291 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11292 (aarch64_simd_make_constant): Update arguments.
11293 (aarch64_output_simd_mov_immediate): Work with struct rather than
11294 many pointers. Output immediate directly rather than as operand.
11295 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11297 * config/aarch64/constraints.md (Dn): Update arguments.
11299 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11301 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11303 (aarch64_simd_immediate_valid_for_move): Remove.
11304 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11305 (aarch64_simd_make_constant): Update call.
11306 (aarch64_output_simd_mov_immediate): Update call.
11307 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11309 * config/aarch64/constraints.md (Dn): Update call.
11311 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11313 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11314 return type to bool for prototype.
11315 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11316 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11317 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11319 2013-06-04 Catherine Moore <clm@codesourcery.com>
11321 * config/mips/mips.opt (meva): New.
11322 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11323 (ASM_SPEC): Handle -meva.
11324 * doc/invoke.texi (meva): Document.
11326 2013-06-04 Alan Modra <amodra@gmail.com>
11328 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11331 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11333 * rtl.def: Add extra fourth optional field to define_cond_exec.
11334 * gensupport.c (process_one_cond_exec): Process attributes from
11336 * doc/md.texi: Document fourth field in define_cond_exec.
11338 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11340 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11341 out the processing order as in store_bit_field_1.
11343 2013-06-04 Jan Hubicka <jh@suse.cz>
11345 PR middle-end/57500
11346 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11347 non-VAR_DECL node if it does not exist yet.
11349 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11351 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11352 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11353 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11354 target_cpu_default setting.
11356 2013-06-03 Teresa Johnson <tejohnson@google.com>
11358 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11359 default to -fopt-info=optimized instead of all.
11360 * doc/invoke.texi: Ditto.
11361 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11362 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11363 (execute_vect_slp): Emit BB vectorization success under
11364 MSG_OPTIMIZED_LOCATIONS.
11365 * tree-vect-slp.c (vect_slp_transform_bb): Change
11366 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11367 * tree-vect-loop.c (vect_transform_loop): Ditto.
11369 2013-06-03 Jason Merrill <jason@redhat.com>
11372 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11373 Use TARGET_EXPR for C++.
11375 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11377 PR rtl-optimization/57268
11378 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11379 if DEBUG_INSN_P (insn).
11382 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11384 PR rtl-optimization/57268
11385 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11386 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11388 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11390 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11391 (ix86_avoid_lea_for_addr): Likewise.
11392 (exact_dependency_1): Likewise.
11393 (ix86_adjust_cost): Likewise.
11394 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11396 (do_reorder_for_imul): Fix typo, formatting and
11397 !reload_completed check removed.
11398 (ix86_sched_reorder): Fix typo and formatting.
11399 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11402 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11404 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11406 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11408 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11409 <STRING_CST>: Likewise.
11410 <VECTOR_CST>: Likewise.
11412 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11413 Mikael Morin <mikael@gcc.gnu.org>
11415 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11416 * config.in: Regenerated.
11417 * configure: Regenerated.
11419 2013-06-01 Jan Hubicka <jh@suse.cz>
11421 PR middle-end/57366
11422 * cgraphunit.c (compile): When weakref is not supported,
11423 set up transparent aliases before final output pass.
11424 * varasm.c (assemble_alias): Do not try to do it here.
11426 2013-06-01 Jan Hubicka <jh@suse.cz>
11428 PR middle-end/57467
11429 * passes.c (for_per_function): Skip unanalyzed functions.
11431 2013-06-01 Jan Hubicka <jh@suse.cz>
11433 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11434 (lto_symtab_merge_symbols_1): ... this one.
11435 (lto_symtab_merge_cgraph_nodes): Rename to ...
11436 (lto_symtab_merge_symbols): ... this one; simplify.
11437 * cgraph.c (same_body_aliases_done): Rename to ...
11438 (cpp_implicit_aliases_done): ... this one.
11439 (cgraph_create_function_alias): Update.
11440 (cgraph_same_body_alias): Update.
11441 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11442 (verify_edge_corresponds_to_fndecl): Simplify.
11443 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11444 (cgraph_node): Remove same_body_alias.
11445 (varpool_node): Remove alias_of and extra_name_alias.
11446 (same_body_aliases_done): Rename to ..
11447 (cpp_implicit_aliases_done): ... this one.
11448 (symtab_alias_ultimate_target): Add default parameter.
11449 (symtab_resolve_alias): New function.
11450 (fixup_same_cpp_alias_visibility): Declare.
11451 (cgraph_function_node): Add default parameter.
11452 (cgraph_node_asm_name): Likewise.
11453 (cgraph_function_or_thunk_node): Add default parameter; do
11454 not ICE when it is NULL.
11455 (varpool_variable_node): Likewise.
11456 * tree-emutls.c (create_emultls_var): Update.
11457 (ipa_lower_emutls): Update.
11458 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11459 (cgraph_reset_node): Reset alias info.
11460 (cgraph_finalize_function): Update.
11461 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11462 (analyze_function): Simplify.
11463 (cgraph_process_same_body_aliases): Simplify.
11464 (analyze_functions): Fixup same body aliases.
11465 (handle_alias_pairs): Simplify.
11466 (assemble_thunk): Update.
11467 (assemble_thunks_and_aliases): Update.
11468 (output_weakrefs): Rewrite.
11469 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11470 (lto_output_varpool_node): Likewise.
11471 (compute_ltrans_boundary): Remve assert.
11472 (get_alias_symbol): New functoin.
11473 (input_node): Rewrite alias handling.
11474 (input_varpool_node): Likewise.
11475 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11476 * ipa.c (process_references): Handle weakrefs correctly.
11477 (symtab_remove_unreachable_nodes): Likewise.
11478 * trans-mem.c (get_cg_data): Update.
11479 (ipa_tm_create_version_alias): Update.
11480 (ipa_tm_execute): Update.
11481 * symtab.c (dump_symtab_base): Dump aliases.
11482 (verify_symtab_base): Verify aliases.
11483 (symtab_node_availability): New function.
11484 (symtab_alias_ultimate_target): Simplify.
11485 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11486 handle all the fixup cases.
11487 (symtab_resolve_alias): New function.
11488 * passes.c (ipa_write_summaries): Handle weakrefs.
11489 * varpool.c (varpool_analyze_node): Simplify.
11490 (assemble_aliases): Update.
11491 (varpool_create_variable_alias): Simplify.
11492 (varpool_extra_name_alias): Simplify.
11493 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11494 (lto_symtab_merge_symbols): ... this one.
11496 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11499 PR rtl-optimization/57268
11500 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11501 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11503 2013-06-01 Tobias Burnus <burnus@net-b.de>
11505 Partially reverted:
11506 2013-05-31 Tobias Burnus <burnus@net-b.de>
11508 PR middle-end/57073
11509 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11512 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11514 PR rtl-optimization/57268
11515 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11516 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11518 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11520 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11521 unordered comparison operators when -fno-trapping-math is in effect
11523 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11524 and implement unordered comparison operators properly on the e500.
11526 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11528 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11529 for constant scalar integers.
11530 (simplify_relational_operation_1): Likewise.
11532 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11534 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11535 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11538 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11539 Igor Zamyatin <igor.zamyatin@intel.com>
11541 Silvermont (SLM) architecture performance tuning.
11542 * config/i386/i386.h (enum ix86_tune_indices): Add
11543 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11544 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11546 * config/i386/i386.c (initial_ix86_tune_features)
11547 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11548 (ix86_lea_outperforms): Handle Silvermont tuning.
11549 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11551 (ix86_use_lea_for_mov): Likewise.
11552 (ix86_avoid_lea_for_addr): Likewise.
11553 (ix86_lea_for_add_ok): Likewise.
11554 (exact_dependency_1): New function.
11555 (exact_store_load_dependency): Likewise.
11556 (ix86_adjust_cost): Handle Silvermont tuning.
11557 (do_reoder_for_imul): Likewise.
11558 (swap_top_of_ready_list): New function.
11559 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11561 * config/i386/i386.md (peepholes that split memory operand in fp
11564 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11566 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11567 Remove un-necessary braces.
11569 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11571 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11572 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11574 2013-05-31 Tobias Burnus <burnus@net-b.de>
11576 PR middle-end/57073
11577 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11579 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11582 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11583 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11584 * config/arm/neon.md (iordi3_neon): Remove.
11585 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11586 * config/arm/predicates.md (imm_for_neon_logic_operand):
11587 Move to earlier in the file.
11588 (neon_logic_op2): Likewise.
11589 (arm_iordi_operand_neon): New predicate.
11591 2013-05-31 Richard Biener <rguenther@suse.de>
11593 PR tree-optimization/57478
11594 PR tree-optimization/57453
11595 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11598 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11600 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11601 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11603 2013-05-30 Tobias Burnus <burnus@net-b.de>
11604 Thomas Koenig <tkoenig@gcc.gnu.org>
11606 PR middle-end/57073
11607 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11608 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11610 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11612 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11614 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11616 * target.def (register_usage_leveling_p): New hook.
11617 * targhooks.c (default_register_usage_leveling_p): New.
11618 * targhooks.h (default_register_usage_leveling_p): New prototype.
11619 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11620 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11621 * doc/tm.texi: Update.
11622 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11624 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11626 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11627 (*insv_reg<mode>): New define_insn.
11629 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11631 PR rtl-optimization/57439
11632 * postreload.c (move2add_valid_value_p): Check that we have
11633 a zero subreg_regno_offset when accessing the register in
11634 the requested mode.
11636 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11637 Igor Zamyatin <igor.zamyatin@intel.com>
11639 Silvermont (SLM) architecture pipeline model, tuning and
11641 * config.gcc: Add slm config options and target.
11643 * config/i386/slm.md: New.
11645 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11647 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11649 (ix86_target_macros_internal): Likewise.
11651 * config/i386/i386.c (slm_cost): New cost.
11652 (m_SLM): New macro flag.
11653 (initial_ix86_tune_features): Set m_SLM.
11654 (x86_accumulate_outgoing_args): Likewise.
11655 (x86_arch_always_fancy_math_387): Likewise.
11656 (processor_target_table): Add slm cost.
11657 (cpu_names): Add slm cpu name.
11658 (x86_option_override_internal): Set SLM ISA.
11659 (ix86_issue_rate): New case PROCESSOR_SLM.
11660 (ia32_multipass_dfa_lookahead): Likewise.
11661 (fold_builtin_cpu): Add slm.
11663 * config/i386/i386.h (TARGET_SLM): New target macro.
11664 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11665 (processor_type): Add PROCESSOR_SLM.
11667 * config/i386/i386.md (cpu): Add new value "slm".
11668 (slm.md): Include slm.md.
11670 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11671 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11673 * config/arm/arm-protos.h: Add and update function protos.
11674 * config/arm/arm.c (use_simple_return_p): New added.
11675 (thumb2_expand_return): Check simple_return flag.
11676 * config/arm/arm.md: Add simple_return and conditional simple_return.
11677 * config/arm/iterators.md: Add iterator for return and simple_return.
11679 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11681 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11682 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11683 (arm_emit_vfp_multi_reg_pop): Likewise.
11684 (thumb2_emit_ldrd_pop): Likewise.
11685 (arm_expand_epilogue): Add misc REG_CFA notes.
11686 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11688 2013-05-29 Lawrence Crowl <crowl@google.com>
11690 * config/arm/t-arm: Update for below.
11692 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11693 Change type to hash_table. Update dependent calls and types.
11695 * config/i386/t-cygming: Update for below.
11697 * config/i386/t-interix: Update for below.
11699 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11700 Change type to hash_table. Update dependent calls and types.
11701 (i386_find_on_wrapper_list::wrappers): Likewise.
11703 * config/ia64/t-ia64: Update for below.
11705 * config/ia64/ia64.c (bundle_state_table):
11706 Change type to hash_table. Update dependent calls and types.
11708 * config/mips/mips.c (mips_reorg_process_insns::htab):
11709 Change type to hash_table. Update dependent calls and types.
11711 * config/sol2.c (solaris_comdat_htab):
11712 Change type to hash_table. Update dependent calls and types.
11714 * config/t-sol2: Update for above.
11716 2013-05-29 Teresa Johnson <tejohnson@google.com>
11718 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11719 functions are not yet marked as defined.
11721 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11722 Pat Haugen <pthaugen@us.ibm.com>
11723 Peter Bergner <bergner@vnet.ibm.com>
11725 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11729 (vrotl<mode>3): Likewise.
11730 (vashl<mode>3): Likewise.
11731 (vlshr<mode>3): Likewise.
11732 (vashr<mode>3): Likewise.
11734 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11735 support for power8 V2DI builtins.
11737 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11738 power8 V2DI builtins.
11739 (vupkhsw): Likewise.
11740 (vupklsw): Likewise.
11741 (vaddudm): Likewise.
11742 (vminsd): Likewise.
11743 (vmaxsd): Likewise.
11744 (vminud): Likewise.
11745 (vmaxud): Likewise.
11746 (vpkudum): Likewise.
11747 (vpksdss): Likewise.
11748 (vpkudus): Likewise.
11749 (vpksdus): Likewise.
11754 (vsubudm): Likewise.
11755 (vcmpequd): Likewise.
11756 (vcmpgtsd): Likewise.
11757 (vcmpgtud): Likewise.
11758 (vcmpequd_p): Likewise.
11759 (vcmpgtsd_p): Likewise.
11760 (vcmpgtud_p): Likewise.
11761 (vupkhsw): Likewise.
11762 (vupklsw): Likewise.
11763 (vaddudm): Likewise.
11764 (vmaxsd): Likewise.
11765 (vmaxud): Likewise.
11766 (vminsd): Likewise.
11767 (vminud): Likewise.
11768 (vpksdss): Likewise.
11769 (vpksdus): Likewise.
11770 (vpkudum): Likewise.
11771 (vpkudus): Likewise.
11776 (vsubudm): Likewise.
11778 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11779 support for power8 V2DI instructions.
11781 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11782 power8 V2DI instructions. Combine pack and unpack insns to use an
11783 iterator for each mode. Check whether a particular mode supports
11784 Altivec instructions instead of just checking TARGET_ALTIVEC.
11785 (UNSPEC_VPKUWUM): Likewise.
11786 (UNSPEC_VPKSHSS): Likewise.
11787 (UNSPEC_VPKSWSS): Likewise.
11788 (UNSPEC_VPKUHUS): Likewise.
11789 (UNSPEC_VPKSHUS): Likewise.
11790 (UNSPEC_VPKUWUS): Likewise.
11791 (UNSPEC_VPKSWUS): Likewise.
11792 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11793 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11794 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11795 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11796 (UNSPEC_VUPKHSB): Likewise.
11797 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11798 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11799 (UNSPEC_VUPKHSH): Likewise.
11800 (UNSPEC_VUPKLSB): Likewise.
11801 (UNSPEC_VUPKLSH): Likewise.
11803 (VI_char): Likewise.
11804 (VI_scalar): Likewise.
11805 (VI_unit): Likewise.
11807 (VP_small): Likewise.
11808 (VP_small_lc): Likewise.
11809 (VU_char): Likewise.
11810 (add<mode>3): Likewise.
11811 (altivec_vaddcuw): Likewise.
11812 (altivec_vaddu<VI_char>s): Likewise.
11813 (altivec_vadds<VI_char>s): Likewise.
11814 (sub<mode>3): Likewise.
11815 (altivec_vsubcuw): Likewise.
11816 (altivec_vsubu<VI_char>s): Likewise.
11817 (altivec_vsubs<VI_char>s): Likewise.
11818 (altivec_vavgs<VI_char>): Likewise.
11819 (altivec_vcmpbfp): Likewise.
11820 (altivec_eq<mode>): Likewise.
11821 (altivec_gt<mode>): Likewise.
11822 (altivec_gtu<mode>): Likewise.
11823 (umax<mode>3): Likewise.
11824 (smax<mode>3): Likewise.
11825 (umin<mode>3): Likewise.
11826 (smin<mode>3): Likewise.
11827 (altivec_vpkuhum): Likewise.
11828 (altivec_vpkuwum): Likewise.
11829 (altivec_vpkshss): Likewise.
11830 (altivec_vpkswss): Likewise.
11831 (altivec_vpkuhus): Likewise.
11832 (altivec_vpkshus): Likewise.
11833 (altivec_vpkuwus): Likewise.
11834 (altivec_vpkswus): Likewise.
11835 (altivec_vpks<VI_char>ss): Likewise.
11836 (altivec_vpks<VI_char>us): Likewise.
11837 (altivec_vpku<VI_char>us): Likewise.
11838 (altivec_vpku<VI_char>um): Likewise.
11839 (altivec_vrl<VI_char>): Likewise.
11840 (altivec_vsl<VI_char>): Likewise.
11841 (altivec_vsr<VI_char>): Likewise.
11842 (altivec_vsra<VI_char>): Likewise.
11843 (altivec_vsldoi_<mode>): Likewise.
11844 (altivec_vupkhsb): Likewise.
11845 (altivec_vupkhs<VU_char>): Likewise.
11846 (altivec_vupkls<VU_char>): Likewise.
11847 (altivec_vupkhsh): Likewise.
11848 (altivec_vupklsb): Likewise.
11849 (altivec_vupklsh): Likewise.
11850 (altivec_vcmpequ<VI_char>_p): Likewise.
11851 (altivec_vcmpgts<VI_char>_p): Likewise.
11852 (altivec_vcmpgtu<VI_char>_p): Likewise.
11853 (abs<mode>2): Likewise.
11854 (vec_unpacks_hi_v16qi): Likewise.
11855 (vec_unpacks_hi_v8hi): Likewise.
11856 (vec_unpacks_lo_v16qi): Likewise.
11857 (vec_unpacks_hi_<VP_small_lc>): Likewise.
11858 (vec_unpacks_lo_v8hi): Likewise.
11859 (vec_unpacks_lo_<VP_small_lc>): Likewise.
11860 (vec_pack_trunc_v8h): Likewise.
11861 (vec_pack_trunc_v4si): Likewise.
11862 (vec_pack_trunc_<mode>): Likewise.
11864 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
11866 (vec_vmaxsd): Likewise.
11867 (vec_vmaxud): Likewise.
11868 (vec_vminsd): Likewise.
11869 (vec_vminud): Likewise.
11870 (vec_vpksdss): Likewise.
11871 (vec_vpksdus): Likewise.
11872 (vec_vpkudum): Likewise.
11873 (vec_vpkudus): Likewise.
11874 (vec_vrld): Likewise.
11875 (vec_vsld): Likewise.
11876 (vec_vsrad): Likewise.
11877 (vec_vsrd): Likewise.
11878 (vec_vsubudm): Likewise.
11879 (vec_vupkhsw): Likewise.
11880 (vec_vupklsw): Likewise.
11882 2013-05-29 Jan Hubicka <jh@suse.cz>
11884 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
11885 flags; reorder rest of fields in more consistent way.
11886 (varpool_node): Remove analyzed, finalized and alias.
11887 (cgraph_ndoe): Likewise.
11888 (symtab_alias_ultimate_target): New function.
11889 (cgraph_function_node): Move offline.
11890 (cgraph_reset_node): Declare.
11891 (cgraph_comdat_can_be_unshared_p): Remove.
11892 (varpool_remove_initializer): Declare.
11893 (varpool_first_defined_variable, varpool_next_defined_variable
11894 cgraph_first_defined_function, cgraph_next_defined_function): Update.
11895 (cgraph_function_with_gimple_body_p): Update.
11896 (varpool_all_refs_explicit_p): Update.
11897 (symtab_alias_target): New function.
11898 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
11899 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
11900 (cgraph_function_or_thunk_node): Simplify using
11901 symtab_alias_ultimate_target.
11902 (varpool_variable_node): Likewise.
11903 * cgraph.c (cgraph_create_function_alias): Update.
11904 (cgraph_add_thunk): Update.
11905 (cgraph_remove_node): Update.
11906 (dump_cgraph_node): Do not dump removed flags.
11907 (cgraph_function_body_availability): Update.
11908 (cgraph_propagate_frequency): Update.
11909 (verify_cgraph_node): Check sanity of local flag.
11910 (cgraph_function_node): Move here from cgraph.h; revamp for
11911 cgraph_function_or_thunk_node.
11912 * lto-symtab.c (lto_varpool_replace_node): Update.
11913 (lto_symtab_resolve_can_prevail_p): Update.
11914 (lto_symtab_merge_cgraph_nodes): Update.
11915 * ipa-cp.c (determine_versionability, initialize_node_lattices,
11916 propagate_constants_accross_call, devirtualization_time_bonus,
11917 ipcp_propagate_stage): Update.
11918 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
11919 * ipa-inline-transform.c (clone_inlined_nodes,
11920 preserve_function_body_p): Update.
11921 * ipa-reference.c (propagate): Update.
11922 (write_node_summary_p): Update.
11923 * toplev.c (wrapup_global_declaration_2): Update.
11924 * cgraphunit.c (cgraph_analyze_function): Rename to ...
11925 (analyze_function) ... this one.
11926 (cgraph_process_new_functions): Update.
11927 (cgraph_reset_node): Export.
11928 (cgraph_finalize_function): Update.
11929 (cgraph_add_new_function): Update.
11930 (process_function_and_variable_attributes): Update.
11931 (varpool_finalize_decl): Update.
11932 (symbol_finalized): Remove.
11933 (symbol_finalized_and_needed): Rename to ...
11934 (symbol_defined_and_needed): ... update.
11935 (cgraph_analyze_functions): Update.
11936 (handle_alias_pairs): Update.
11937 (mark_functions_to_output): Update.
11938 (assemble_thunk): Update.
11939 (output_in_order): Update.
11940 (output_weakrefs): Update.
11941 (finalize_compilation_unit): Update.
11942 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
11943 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
11944 input_node, input_varpool_node): Update.
11945 * dbxout.c (dbxout_expand_expr): Update.
11946 * cgraphclones.c (cgraph_clone_node): Update.
11947 (cgraph_copy_node_for_versioning): Update.
11948 (cgraph_materialize_clone): Update.
11949 (cgraph_materialize_all_clones): Update.
11950 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
11951 propagate_pure_const, propagate_nothrow): Update.
11952 * lto-streamer-out.c (lto_output, write_symbol): Update.
11953 * ipa-utils.c (ipa_reverse_postorder): Update.
11954 * ipa-inline.c (can_inline_edge_p): Update.
11955 (update_caller_keys, ipa_inline): Update.
11956 * dwarf2out.c (reference_to_unused,
11957 premark_types_used_by_global_vars_helper): Update.
11958 * tree-eh.c (tree_could_trap_p): Update.
11959 * ipa-split.c (consider_split, execute_split_functions): Update.
11960 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
11961 has_addr_references_p): Update; move ahead in file for better
11963 (process_references): Simplify.
11964 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
11965 bodies are removed.
11966 (cgraph_comdat_can_be_unshared_p): Make static.
11967 (cgraph_externally_visible_p): Update.
11968 (varpool_externally_visible_p): Update.
11969 (function_and_variable_visibility): Update.
11970 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
11971 ipa_tm_mark_force_output_node): Update.
11972 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
11973 estimate_edge_devirt_benefit, inline_generate_summary,
11974 inline_write_summary): Update.
11975 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
11976 * ipa-prop.c (ipa_compute_jump_functions): Update.
11977 (ipa_print_node_params, ipa_prop_read_section,
11978 ipa_update_after_lto_read, read_replacements_section): Update.
11979 * varasm.c (mark_decl_referenced): Update.
11980 (assemble_alias, dump_tm_clone_pairs): Update.
11981 * tree-inline.c (copy_bb): Update.
11982 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
11984 * symtab.c (dump_symtab_base): Print new flags.
11985 (verify_symtab_base): Verify new flags.
11986 (symtab_alias_ultimate_target): New function.
11987 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
11988 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
11990 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
11992 * i386.c (ix86_get_function_versions_dispatcher,
11993 ix86_generate_version_dispatcher_body): Update.
11994 (fold_builtin_cpu): Use varpool_add_new_variable.
11995 * varpool.c (varpool_remove_initializer): Break out from ...
11996 (varpool_remove_node): ... this one.
11997 (dump_varpool_node, varpool_node_for_asm,
11998 cgraph_variable_initializer_availability, varpool_analyze_node,
11999 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12000 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12003 2013-05-29 Jan Hubicka <jh@suse.cz>
12005 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12007 2013-05-29 Easwaran Raman <eraman@google.com>
12009 PR tree-optimization/57442
12010 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12011 when control exits the main loop.
12013 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12015 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12017 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12018 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12019 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12022 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12024 PR tree-optimization/57441
12025 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12026 Don't limit size of incr_vec to number of candidates.
12028 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12030 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12031 and mips16 directories.
12032 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12033 (MULTILIB_DIRNAMES): Ditto.
12034 (MULTILIB_EXCEPTIONS): Add new exceptions.
12035 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12036 (MULTILIB_DIRNAMES): Ditto.
12037 (MULTILIB_EXCEPTIONS): Add new exceptions.
12039 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12040 Marcus Shawcroft <marcus.shawcroft@arm.com>
12042 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12043 SYMBOL_TINY_ABSOLUTE.
12044 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12045 SYMBOL_TINY_ABSOLUTE.
12046 (aarch64_expand_mov_immediate): Likewise.
12047 (aarch64_classify_symbol): Likewise.
12048 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12049 Permit SYMBOL_TINY_ABSOLUTE.
12050 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12052 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12053 Marcus Shawcroft <marcus.shawcroft@arm.com>
12055 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12056 Refactor if/switch. Replace gcc_assert with if.
12058 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12060 * config/i386/i386.c (initial_ix86_tune_features): Enable
12061 FP Reassociation for AMD bdver1 and bdver2.
12063 2013-05-29 Martin Jambor <mjambor@suse.cz>
12065 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12066 and IMAGPART_EXPR do not occur within other handled_components.
12068 2013-05-29 Richard Biener <rguenther@suse.de>
12070 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12071 access on whether the use is in the BB we currently try to
12073 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12074 vectorize to vect_bb_slp_scalar_cost.
12076 2013-05-29 Richard Biener <rguenther@suse.de>
12078 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12079 computing scalar cost offsetted by stmts that are kept live
12081 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12082 for computation of scalar cost.
12084 2013-05-28 Steve Ellcey <sellcey@mips.com>
12086 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12088 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12090 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12091 array notation built-in reduction functions.
12092 * doc/passes.texi (Passes): Added documentation about changes done
12094 * doc/invoke.texi (C Dialect Options): Added documentation about
12095 the -fcilkplus flag.
12096 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12097 (BUILTINS_DEF): Depend on cilkplus.def.
12098 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12099 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12100 * cilkplus.def: New file.
12102 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12104 PR rtl-optimization/57439
12105 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12107 2013-05-28 Easwaran Raman <eraman@google.com>
12109 PR tree-optimization/57337
12110 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12111 (find_insert_point): Correctly identify the insertion point
12112 when two statements with the same UID is compared.
12114 2013-05-28 Richard Biener <rguenther@suse.de>
12116 PR tree-optimization/56787
12117 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12118 from the list of data references.
12119 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12121 (vect_analyze_loop_operations): Likewise.
12122 (vect_transform_loop): Remove clobbers.
12124 2013-05-28 Martin Jambor <mjambor@suse.cz>
12126 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12127 and REALPART_EXPRs have scalar type.
12129 2013-05-28 Richard Biener <rguenther@suse.de>
12131 PR tree-optimization/57411
12132 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12134 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12135 virtual operand propagation.
12137 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12139 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12140 destination register for bmasksi_vis.
12141 (vector_init_bshuffle): Likewise.
12142 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12144 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12146 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12147 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12148 mode if the instruction isn't available in the original mode.
12149 * config/sparc/sparc.opt (mfix-ut699): New option.
12150 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12151 (divdf3): Turn into expander.
12152 (divdf3_nofix): New insn.
12153 (divdf3_fix): Likewise.
12154 (divsf3): Disable if -mfix-ut699.
12155 (sqrtdf2): Turn into expander.
12156 (sqrtdf2_nofix): New insn.
12157 (sqrtdf2_fix): Likewise.
12158 (sqrtsf2): Disable if -mfix-ut699.
12160 2013-05-27 Richard Biener <rguenther@suse.de>
12162 PR middle-end/57412
12163 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12164 block for the new loop.
12166 2013-05-27 Richard Biener <rguenther@suse.de>
12168 PR tree-optimization/57343
12169 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12170 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12171 (number_of_iterations_cond): Do not build the folded tree.
12173 2013-05-27 Richard Biener <rguenther@suse.de>
12176 PR middle-end/57381
12177 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12178 OEP_CONSTANT_ADDRESS_OF retained.
12180 PR tree-optimization/57417
12181 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12182 for unchanged base.
12183 (set_ssa_val_to): Compare addresses using
12184 get_addr_base_and_unit_offset.
12186 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12188 PR rtl-optimization/56833
12189 * postreload.c (move2add_record_mode): New function.
12190 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12191 (move2add_use_add2_insn): Use move2add_record_sym_value.
12192 (move2add_use_add3_insn): Likewise.
12193 (reload_cse_move2add): Use move2add_valid_value_p and
12194 move2add_record_mode. Invalidate call-clobbered and REG_INC
12195 affected regs by setting reg_mode to VOIDmode.
12196 (move2add_note_store): Don't pretend the inside of a SUBREG is
12197 the actual destination. Invalidate single/leading registers by
12198 setting reg_mode to VOIDmode.
12199 Use move2add_record_sym_value, move2add_valid_value_p and
12200 move2add_record_mode.
12202 2013-05-27 Richard Biener <rguenther@suse.de>
12204 PR tree-optimization/57396
12205 * tree-affine.c (double_int_constant_multiple_p): Properly
12206 return false for val == 0 and div != 0.
12208 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12210 * config/mips/mips.h: Use #elif in preprocessor conditions.
12212 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12215 * config/mips/constraints.md (kl): New constraint.
12216 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12217 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12218 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12219 from LO for MIPS16.
12220 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12222 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12225 * config/mips/mips.c (mips_can_inline_p): New function.
12226 (TARGET_CAN_INLINE_P): Define.
12228 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12230 * sched-int.h (ds_t, dw_t): Make unsigned int.
12231 Fix documentation that describes how all the ds_t bits are used.
12232 Reserve the last bit for delayed-branch scheduling.
12233 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12234 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12235 (gen_dep_weak_1): Remove prototype.
12236 * sched-deps.c (get_dep_weak_1): Make static.
12237 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12238 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12239 * doc/tm.texi: Regenerate.
12240 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12242 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12245 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12247 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12248 Sandra Loosemore <sandra@codesourcery.com>
12250 * config.gcc (powerpc-*): Allow native for with-cpu.
12252 2013-05-24 Jeff Law <law@redhat.com>
12254 PR tree-optimization/57124
12255 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12256 conversion feeding a condition if the range has an overflow
12257 if -fstrict-overflow. Add warnings for when we do make the
12260 2013-05-24 Dehao Chen <dehao@google.com>
12262 * tree-cfg.c (locus_discrim_map): Fix the typo.
12263 (locus_discrim_hasher): Likewise.
12264 (locus_discrim_hasher::hash): Likewise.
12265 (locus_discrim_hasher::equal): Likewise.
12267 2013-05-24 Martin Jambor <mjambor@suse.cz>
12269 PR tree-optimization/57294
12270 * cgraph.h (ipa_record_stmt_references): Declare.
12271 * cgraphbuild.c (ipa_record_stmt_references): New function.
12272 (build_cgraph_edges): Use ipa_record_stmt_references.
12273 (rebuild_cgraph_edges): Likewise.
12274 (cgraph_rebuild_references): Likewise.
12275 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12276 associated with the old statement and build references from the
12277 newly built statements.
12278 * ipa-ref.c (ipa_remove_stmt_references): New function.
12279 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12281 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12283 * lra-constraints.c (emit_spill_move): Use smaller mode for
12285 (check_and_process_move): Consider mem-reg moves for secondary
12287 (curr_insn_transform): Don't lose insns emitted before for
12288 secondary memory moves.
12289 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12290 reg set up in the current insn.
12292 2013-05-24 Dehao Chen <dehao@google.com>
12294 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12296 (locus_descrim_hasher::equal): Likewise.
12297 (build_gimple_cfg): New discriminator assignment algorithm.
12298 (make_edges): Likewise.
12299 (next_discriminator_for_locus): Likewise.
12300 (same_line_p): Likewise.
12301 (assign_discriminators): Likewise.
12302 (make_cond_expr_edges): Likewise.
12303 (make_gimple_switch_edges): Likewise.
12304 (make_goto_expr_edges): Likewise.
12305 (make_gimple_asm_edges): Likewise.
12307 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12309 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12310 X format specifier to only display bottom 16 bits.
12311 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12312 immediate to match for operand 2, since it will be masked.
12314 2013-05-24 Richard Biener <rguenther@suse.de>
12316 PR tree-optimization/57287
12317 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12318 all SSA names that occur in abnormal PHIs.
12320 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12323 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12324 that index is not negative.
12326 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12328 PR rtl-optimization/55177
12329 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12330 (simplify_byte_swapping_operation): New.
12331 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12332 (simplify_relational_operation_1): Deal with BSWAP.
12334 2013-05-23 Richard Henderson <rth@redhat.com>
12337 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12338 (ix86_reorg): Call it.
12340 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12343 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12344 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12345 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12347 2013-05-23 Christian Bruel <christian.bruel@st.com>
12350 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12352 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12353 Marcus Shawcroft <marcus.shawcroft@arm.com>
12355 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12356 * config/aarch64/constraints.md (Usa): Remove.
12357 * doc/md.texi (AArch64 Usa): Remove.
12359 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12360 Marcus Shawcroft <marcus.shawcroft@arm.com>
12362 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12363 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12364 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12365 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12367 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12369 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12370 instruction (AdvSIMD).
12371 * config/aarch64/aarch64-builtins.c
12372 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12373 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12375 2013-05-23 Martin Jambor <mjambor@suse.cz>
12377 PR middle-end/57347
12378 * tree.h (contains_bitfld_component_ref_p): Declare.
12379 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12380 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12382 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12383 not access a bit-field. Assert all final offsets are byte-aligned.
12385 2013-05-23 Richard Biener <rguenther@suse.de>
12387 PR tree-optimization/57380
12388 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12389 least one invariant or re-used load.
12390 * passes.c (init_optimization_passes): Move pass_phiprop before
12393 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12395 * config/aarch64/aarch64-simd.md
12396 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12398 2013-05-23 Richard Biener <rguenther@suse.de>
12400 PR middle-end/57381
12401 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12402 OEP_CONSTANT_ADDRESS_OF retained.
12404 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12406 PR middle-end/57344
12407 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12408 don't lower unit. Handle unit not being always BITS_PER_WORD.
12410 2013-05-23 Richard Biener <rguenther@suse.de>
12412 PR rtl-optimization/57341
12413 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12414 instead of true_dependence.
12416 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12418 * bb-reorder.c (branch_threshold): Make const.
12419 (exec_threshold): Ditto.
12421 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12422 Pat Haugen <pthaugen@us.ibm.com>
12423 Peter Bergner <bergner@vnet.ibm.com>
12425 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12426 documentation for the power8 crypto builtins.
12428 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12430 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12431 macros for defining power8 builtin functions.
12432 (BU_P8V_AV_2): Likewise.
12433 (BU_P8V_AV_P): Likewise.
12434 (BU_P8V_VSX_1): Likewise.
12435 (BU_P8V_OVERLOAD_1): Likewise.
12436 (BU_P8V_OVERLOAD_2): Likewise.
12437 (BU_CRYPTO_1): Likewise.
12438 (BU_CRYPTO_2): Likewise.
12439 (BU_CRYPTO_3): Likewise.
12440 (BU_CRYPTO_OVERLOAD_1): Likewise.
12441 (BU_CRYPTO_OVERLOAD_2): Likewise.
12442 (XSCVSPDP): Fix typo, point to the correct instruction.
12443 (VCIPHER): Add power8 crypto builtins.
12444 (VCIPHERLAST): Likewise.
12445 (VNCIPHER): Likewise.
12446 (VNCIPHERLAST): Likewise.
12447 (VPMSUMB): Likewise.
12448 (VPMSUMH): Likewise.
12449 (VPMSUMW): Likewise.
12450 (VPERMXOR_V2DI): Likewise.
12451 (VPERMXOR_V4SI: Likewise.
12452 (VPERMXOR_V8HI: Likewise.
12453 (VPERMXOR_V16QI: Likewise.
12454 (VSHASIGMAW): Likewise.
12455 (VSHASIGMAD): Likewise.
12456 (VPMSUM): Likewise.
12457 (VPERMXOR): Likewise.
12458 (VSHASIGMA): Likewise.
12460 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12461 __CRYPTO__ if the crypto instructions are available.
12462 (altivec_overloaded_builtins): Add support for overloaded power8
12465 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12466 support for power8 crypto builtins.
12467 (builtin_function_type): Likewise.
12468 (altivec_init_builtins): Add support for builtins that take vector
12469 long long (V2DI) arguments.
12471 * config/rs6000/crypto.md: New file, define power8 crypto
12474 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12475 Pat Haugen <pthaugen@us.ibm.com>
12476 Peter Bergner <bergner@vnet.ibm.com>
12478 * doc/invoke.texi (Option Summary): Add power8 options.
12479 (RS/6000 and PowerPC Options): Likewise.
12481 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12482 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12483 wm, wn, wr documentation.
12485 * config/rs6000/constraints.md (wm): New constraint for VSX
12486 registers if direct move instructions are enabled.
12487 (wn): New constraint for no registers.
12488 (wq): New constraint for quad word even GPR registers.
12489 (wr): New constraint if 64-bit instructions are enabled.
12490 (wv): New constraint if power8 vector instructions are enabled.
12491 (wQ): New constraint for quad word memory locations.
12493 * config/rs6000/predicates.md (const_0_to_15_operand): New
12494 constraint for 0..15 for crypto instructions.
12495 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12496 as GPR and floating point registers.
12497 (int_reg_operand): New predicate to match only GPR registers.
12498 (base_reg_operand): New predicate to match base registers.
12499 (quad_int_reg_operand): New predicate to match even GPR registers
12500 for quad memory operations.
12501 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12502 operations in both GPR and VSX registers.
12503 (quad_memory_operand): New predicate for quad memory operations.
12504 (reg_or_indexed_operand): New predicate for direct move support.
12506 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12507 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12508 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12509 (POWERPC_MASKS): Add power8 options.
12510 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12513 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12514 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12516 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12517 (-mpower8-fusion): New power8 options.
12518 (-mpower8-fusion-sign): Likewise.
12519 (-mpower8-vector): Likewise.
12520 (-mcrypto): Likewise.
12521 (-mdirect-move): Likewise.
12522 (-mquad-memory): Likewise.
12524 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12525 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12527 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12528 (rs6000_debug_vector_unit): Add p8_vector.
12529 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12530 definitions. Also print fusion state.
12531 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12532 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12533 (rs6000_option_override_internal): Add support for power8.
12534 (rs6000_common_init_builtins): Add debugging for skipped builtins
12535 if -mdebug=builtin.
12536 (rs6000_adjust_cost): Add power8 support.
12537 (rs6000_issue_rate): Likewise.
12538 (insn_must_be_first_in_group): Likewise.
12539 (insn_must_be_last_in_group): Likewise.
12540 (force_new_group): Likewise.
12541 (rs6000_register_move_cost): Likewise.
12542 (rs6000_opt_masks): Likewise.
12544 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12545 power8 capable assembler, default to power7 options.
12546 (TARGET_DIRECT_MOVE): Likewise.
12547 (TARGET_CRYPTO): Likewise.
12548 (TARGET_P8_VECTOR): Likewise.
12549 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12550 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12551 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12552 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12553 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12554 (TARGET_XSCVDPSPN): Likewise.
12555 (TARGET_XSCVSPDPN): Likewsie.
12556 (TARGET_SYNC_HI_QI): Likewise.
12557 (TARGET_SYNC_TI): Likewise.
12558 (MASK_CRYPTO): Likewise.
12559 (MASK_DIRECT_MOVE): Likewise.
12560 (MASK_P8_FUSION): Likewise.
12561 (MASK_P8_VECTOR): Likewise.
12562 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12563 temporary used by some of the direct move instructions to get two FP
12564 temporary registers does not force creation of a stack frame.
12565 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12566 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12567 that any VSX registers are tieable, even if they are also an
12568 Altivec vector mode.
12569 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12570 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12571 (RS6000_BTM_CRYPTO): Likewise.
12572 (RS6000_BTM_COMMON): Likewise.
12574 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12575 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12576 (enum rs6000_vector): Add power8 vector support.
12578 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12582 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12583 (any_sibcall_could_use_r3): this and handle indirect calls.
12584 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12586 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12588 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12590 2013-05-22 Richard Biener <rguenther@suse.de>
12592 PR middle-end/57349
12593 * profile.c (branch_prob): Do not split blocks that are
12594 abnormally receiving from ECF_RETURNS_TWICE functions.
12596 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12598 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12599 address mode. Move pointer mode initialization to the same place.
12601 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12603 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12604 while it has any effect.
12606 2013-05-21 Easwaran Raman <eraman@google.com>
12608 PR tree-optimization/57322
12609 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12610 UID of the statement added to the BB to be 1.
12612 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12614 PR tree-optimization/57331
12615 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12616 of conversion from pointer type to integral type with integer.
12618 2013-05-21 Martin Jambor <mjambor@suse.cz>
12621 * ipa-prop.c (ipa_read_node_info): Process param_used and
12622 controlled_uses in the same order as when writing.
12624 2013-05-21 Magnus Granberg <baldrick@free.fr>
12627 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12629 2013-05-21 Richard Biener <rguenther@suse.de>
12631 PR tree-optimization/57318
12632 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12633 estimate stmts with side-effects as likely eliminated.
12635 2013-05-21 Richard Biener <rguenther@suse.de>
12637 PR tree-optimization/57330
12638 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12639 preserve the call stmts fntype.
12641 2013-05-21 Richard Biener <rguenther@suse.de>
12643 PR tree-optimization/57303
12644 * tree-ssa-sink.c (statement_sink_location): Improve killing
12645 stmt detection and properly handle self-assignments.
12647 2013-05-21 Christian Bruel <christian.bruel@st.com>
12649 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12650 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12651 registers. Set register size out of the PARALLEL loop.
12653 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12656 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12657 (*fmasf4, *fmasf4_media): New insns.
12659 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12661 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12662 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12663 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12664 (mips_idiv_insns): Update the comments to say that the returned
12665 instruction counts are in units of BASE_INSN_LENGTH.
12666 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12667 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12668 using 2 rather than 4 as the length of indirect MIPS16 and
12669 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12670 length of a NOP. Don't divide MIPS16 lengths by 2.
12671 (mips16_split_long_branches): Assume a branch is long if the
12672 length is greater than 4 rather than 8.
12673 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12674 rather than multiplying them by 2. Multiply instruction counts
12675 by BASE_INSN_LENGTH rather than 4.
12676 (*jump_mips16, tls_get_tp_mips16_<mode>)
12677 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12679 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12681 * config/mips/mips.md (extended_mips16): Remove branch case.
12682 (length): Remove duplicated extended_mips16 test.
12684 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12686 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12688 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12690 * recog.h (Recog_data): Rename to...
12691 (recog_data_d): ...this.
12692 (recog_data): Update accordingly.
12693 * recog.c (recog_data): Likewise.
12694 * reload.c (save_recog_data): Likewise.
12695 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12696 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12698 2013-05-17 Julian Brown <julian@codesourcery.com>
12700 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12701 found in a REG_EQUAL note, invalidate it.
12703 2013-05-17 Easwaran Raman <eraman@google.com>
12705 * tree-ssa-reassoc.c (find_insert_point): New function.
12706 (insert_stmt_after): Likewise.
12707 (get_def_stmt): Likewise.
12708 (ensure_ops_are_available): Likewise.
12709 (not_dominated_by): Likewise.
12710 (rewrite_expr_tree): Do not move statements beyond what is
12711 necessary. Remove call to swap_ops_for_binary_stmt...
12712 (reassociate_bb): ... and move it here.
12713 (build_and_add_sum): Assign UIDs for new statements.
12714 (linearize_expr): Likewise.
12715 (do_reassoc): Renumber gimple statement UIDs.
12717 2013-05-17 Jan Hubicka <jh@suse.cz>
12719 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12721 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12722 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12724 (output_weakrefs): Update.
12726 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12727 Martin Jambor <mjambor@suse.cz>
12729 PR middle-end/57276
12730 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12731 value that corresponds to the given aggval is found in values vector.
12733 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12735 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12736 sse, sse2, sse3, ssse3 and sse4a flags to options.
12738 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12740 * gengtype-state.c: (s_expr_writer): New class, to handle
12741 prettifying of output layout of s-expressions.
12742 (state_writer): New class, to write out gtype.state.
12743 (state_written_type_count): Move this variable into member data of
12745 (s_expr_writer::s_expr_writer): New code: constructor for new class
12746 (state_writer::state_writer(): ditto
12747 (s_expr_writer::write_new_line): New function
12748 (s_expr_writer::write_any_indent): ditto
12749 (s_expr_writer::begin_s_expr): ditto
12750 (s_expr_writer::end_s_expr): ditto
12751 (write_state_fileloc): convert to method of state_writer...
12752 (state_writer:: write_state_fileloc): ...and use methods of
12753 s_expr_writer to write indentation into the gtype.state output file
12754 to visually represent the hierarchical structure of the list
12756 (write_state_fields): ditto, renaming to...
12757 (state_writer::write_state_fields)
12758 (write_state_a_string): ditto, renaming to...
12759 (state_writer::write_state_a_string)
12760 (write_state_string_option): ditto, renaming to...
12761 (state_writer::write_state_string_option)
12762 (write_state_type_option): ditto, renaming to...
12763 (state_writer::write_state_type_option)
12764 (write_state_nested_option): ditto, renaming to...
12765 (state_writer::write_state_nested_option)
12766 (write_state_option): ditto, renaming to...
12767 (state_writer::write_state_option)
12768 (write_state_options): ditto, renaming to...
12769 (state_writer::write_state_options)
12770 (write_state_lang_bitmap): ditto, renaming to...
12771 (state_writer::write_state_lang_bitmap)
12772 (write_state_version): ditto, renaming to...
12773 (state_writer::write_state_version)
12774 (write_state_scalar_type): ditto, renaming to...
12775 (state_writer::write_state_scalar_type)
12776 (write_state_string_type): ditto, renaming to...
12777 (state_writer::write_state_string_type)
12778 (write_state_undefined_type): ditto, renaming to...
12779 (state_writer::write_state_undefined_type)
12780 (write_state_struct_union_type): ditto, renaming to...
12781 (state_writer::write_state_struct_union_type)
12782 (write_state_struct_type): ditto, renaming to...
12783 (state_writer::write_state_struct_type)
12784 (write_state_user_struct_type): ditto, renaming to...
12785 (state_writer::write_state_user_struct_type)
12786 (write_state_lang_struct_type): ditto, renaming to...
12787 (state_writer::write_state_lang_struct_type)
12788 (write_state_param_struct_type): ditto, renaming to...
12789 (state_writer::write_state_param_struct_type)
12790 (write_state_pointer_type): ditto, renaming to...
12791 (state_writer::write_state_pointer_type)
12792 (write_state_array_type): ditto, renaming to...
12793 (state_writer::write_state_array_type)
12794 (write_state_gc_used): ditto, renaming to...
12795 (state_writer::write_state_gc_used)
12796 (write_state_common_type_content): ditto, renaming to...
12797 (state_writer::write_state_common_type_content)
12798 (write_state_type): ditto, renaming to...
12799 (state_writer::write_state_type)
12800 (write_state_pair_list): ditto, renaming to...
12801 (state_writer::write_state_pair_list)
12802 (write_state_pair): ditto, renaming to...
12803 (state_writer::write_state_pair)
12804 (write_state_typedefs): ditto, renaming to...
12805 (state_writer::write_state_typedefs)
12806 (write_state_structures): ditto, renaming to...
12807 (state_writer::write_state_structures)
12808 (write_state_param_structs): ditto, renaming to...
12809 (state_writer::write_state_param_structs)
12810 (write_state_variables): ditto, renaming to...
12811 (state_writer::write_state_variables)
12812 (write_state_srcdir): ditto, renaming to...
12813 (state_writer::write_state_srcdir)
12814 (write_state_files_list): ditto, renaming to...
12815 (state_writer::write_state_files_list)
12816 (write_state_languages): ditto, renaming to...
12817 (state_writer::write_state_languages)
12818 (write_state): create a state_writer instance and use it when
12819 writing out the state file
12821 2013-05-17 Mike Stump <mikestump@comcast.net>
12823 PR rtl-optimization/57304
12824 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12825 accessing DF_REF_REAL_LOC.
12827 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12829 PR rtl-optimization/57281
12830 PR rtl-optimization/57300
12831 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12832 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12833 what the other splitter did if the registers are dead.
12835 2013-05-17 Richard Biener <rguenther@suse.de>
12837 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12840 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12842 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12845 2013-05-17 Marek Polacek <polacek@redhat.com>
12847 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12848 length when doing non-zero store of storing '\0' to '\0'.
12850 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12852 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
12853 vect_external_def oprnd1 with loop_vinfo, try to emit
12854 optional cast, negation and and stmts on the loop preheader
12855 edge instead of into the pattern def seq.
12857 PR tree-optimization/57051
12858 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
12859 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
12861 2013-05-16 Nick Clifton <nickc@redhat.com>
12863 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
12864 (rl78_is_naked_func): New function.
12865 (rl78_expand_prologue): Skip prologue generation for naked functions.
12866 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
12867 * doc/extend.texi (naked): Add RL78 to the list of processors
12868 that supports this attribute.
12870 2013-05-16 Jeff Law <law@redhat.com>
12872 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
12874 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12876 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
12877 cache parameters using detect_caches_amd also for CYRIX,
12878 NSC and TM2 signatures.
12880 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12881 Dzianis Kahanovich <mahatma@eu.by>
12885 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12886 VIA/Centaur processors and determine their cache parameters
12887 using detect_caches_amd.
12889 2013-05-16 Teresa Johnson <tejohnson@google.com>
12891 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
12892 (rtl_verify_edges): New function.
12893 (rtl_verify_bb_insns): Ditto.
12894 (rtl_verify_bb_pointers): Ditto.
12895 (rtl_verify_bb_insn_chain): Ditto.
12896 (rtl_verify_fallthru): Ditto.
12897 (rtl_verify_bb_layout): Ditto.
12898 (rtl_verify_flow_info_1): Outline checks into new functions.
12899 (rtl_verify_flow_info): Ditto.
12901 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
12903 * cfghooks.c (copy_bbs): Add update_dominance argument.
12904 * cfghooks.h (copy_bbs): Update prototype.
12905 * tree-cfg.c (gimple_duplicate_sese_region):
12906 Add update_dominance argument.
12907 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
12908 * tree-ssa-loop-ch.c (copy_loop_headers): Update
12909 gimple_duplicate_sese_region call.
12910 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
12911 Update copy_bbs call.
12912 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
12913 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
12915 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12917 * tree-vectorizer.h (NUM_PATTERNS): Increment.
12918 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
12919 vect_recog_rotate_pattern.
12920 (vect_recog_rotate_pattern): New function.
12922 2013-05-16 Jason Merrill <jason@redhat.com>
12924 * Makefile.in (LLINKER): New variable.
12925 (mostlyclean): Remove link mutex.
12926 * configure.ac: Handle --enable-link-mutex.
12927 * lock-and-run.sh: New script.
12929 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12932 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
12935 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12937 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
12939 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
12941 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
12942 * config/arm/arm.c (next_consecutive_mem): New function.
12943 (gen_movmem_ldrd_strd): Likewise.
12944 * config/arm/arm.md (movmemqi): Update condition and code.
12945 (unaligned_loaddi, unaligned_storedi): New patterns.
12947 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12949 * config.gcc: Obsolete *-*-solaris2.9*.
12950 * doc/install.texi (Specific, *-*-solaris2*): Document it.
12952 2013-05-16 Richard Biener <rguenther@suse.de>
12954 * passes.c (init_optimization_passes): Move pass_parallelize_loops
12955 earlier, after GRAPHITE transforms and IV canonicalization.
12957 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12959 * omp-low.c (extract_omp_for_data): For collapsed loops,
12960 if at least one of the loops is known at compile time to
12961 iterate zero times, set count to 0.
12962 (expand_omp_regimplify_p): New function.
12963 (expand_omp_for_generic): For collapsed loops, if at least
12964 one of the loops isn't known to iterate at least once,
12965 add runtime check with setting count to 0.
12966 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
12967 For unsigned types if it isn't known at compile time that
12968 the loop will iterate at least once, add runtime check to bypass
12969 the whole loop if initial condition isn't true.
12971 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
12973 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
12975 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
12977 PR middle-end/57286
12978 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
12979 transformations to avoid an infinite loop.
12981 2013-05-16 Marek Polacek <polacek@redhat.com>
12983 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
12985 2013-05-15 Leif Ekblad <leif@rdos.net>
12987 * config/i386/i386.c (ix86_decompose_address): Use
12988 DEFAULT_TLS_SEG_REG to access TLS segment register.
12989 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
12990 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
12991 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
12993 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
12996 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
12997 sibling calls to functions that would normally be lazily bound,
12998 unless $gp is call-clobbered.
13000 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13002 * config/i386/i386.c (ix86_option_override_internal): Update
13003 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13004 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13005 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13006 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13008 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13010 2013-05-15 Andreas Schwab <schwab@suse.de>
13012 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13013 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13016 2013-05-15 Teresa Johnson <tejohnson@google.com>
13018 * loop-unroll.c (report_unroll_peel): Check decision before
13019 emitting unroll/peel message.
13021 2013-05-15 Teresa Johnson <tejohnson@google.com>
13023 * function.h (has_bb_partition): New rtl_data flag.
13024 (bb_reorder_complete): Ditto.
13025 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13026 instead of flag_reorder_blocks_and_partition.
13027 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13028 with some enhancements.
13029 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13030 * bb-reorder.c (connect_traces): Check for has_bb_partition
13031 instead of flag_reorder_blocks_and_partition.
13032 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13033 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13034 verify_hot_cold_block_grouping.
13035 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13037 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13040 * config/arm/predicates.md (call_insn_operand): New predicate.
13041 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13042 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13043 if insn is not a tail call.
13044 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13046 * config/arm/arm.h (enum reg_class): New caller save register class.
13047 (REG_CLASS_NAMES): Likewise.
13048 (REG_CLASS_CONTENTS): Likewise.
13049 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13052 2013-05-15 Richard Biener <rguenther@suse.de>
13054 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13055 of MSG_OPTIMIZED_LOCATIONS.
13056 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13057 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13059 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13060 of MSG_OPTIMIZED_LOCATIONS.
13061 (execute_vect_slp): Likewise.
13062 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13063 (vect_create_cond_for_alias_checks): Likewise.
13064 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13065 (vect_recog_widen_mult_pattern): Likewise.
13066 (vect_recog_widen_sum_pattern): Likewise.
13067 (vect_recog_over_widening_pattern): Likewise.
13068 (vect_recog_widen_shift_pattern): Likewise.
13069 (vect_recog_vector_vector_shift_pattern): Likewise.
13070 (vect_recog_divmod_pattern): Likewise.
13071 (vect_recog_mixed_size_cond_pattern): Likewise.
13072 (vect_recog_bool_pattern): Likewise.
13073 (vect_pattern_recog_1): Likewise.
13075 2013-05-15 Martin Jambor <mjambor@suse.cz>
13077 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13078 non-functions to builtin_unreachable.
13079 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13080 correct when new direct edges were discovered.
13082 2013-05-15 Martin Jambor <mjambor@suse.cz>
13084 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13085 header, print symbol order instead of node uid, print more information
13086 about indirect edge targets.
13087 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13089 (ipa_make_edge_direct_to_target): Likewise.
13090 (remove_described_reference): Likewise.
13091 (propagate_controlled_uses): Likewise.
13092 (ipa_print_node_params): Also print symbol order.
13093 (ipcp_transform_function): Print symbol order instead of node uids.
13094 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13095 (cgraph_get_create_real_symbol_node): Likewise.
13096 * ipa-cp.c (print_lattice): Likewise.
13097 (print_all_lattices): Likewise.
13098 (determine_versionability): Likewise.
13099 (initialize_node_lattices): Likewise.
13100 (estimate_local_effects): Likewise.
13101 (update_profiling_info): Likewise.
13102 (create_specialized_node): Likewise.
13103 (perhaps_add_new_callers): Likewise.
13104 (decide_about_value): Likewise.
13105 (decide_whether_version_node): Likewise.
13106 (identify_dead_nodes): Likewise.
13107 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13108 (dump_inline_summary): Likewise.
13109 (estimate_node_size_and_time): Likewise.
13110 (inline_analyze_function): Likewise.
13111 * ipa-inline.c (report_inline_failed_reason): Likewise.
13112 (want_early_inline_function_p): Likewise.
13113 (edge_badness): Likewise.
13114 (update_edge_key): Likewise.
13115 (inline_small_functions): Likewise. Add dumping of order to two other
13117 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13118 instead of node uids.
13119 (propagate_pure_const): Likewise.
13120 (propagate_pure_const): Likewise.
13121 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13122 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13123 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13125 * tree-pretty-print.c (dump_function_header): Likewise.
13126 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13127 Print symbol order instead of node uids.
13129 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13131 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13132 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13134 2013-05-15 Richard Biener <rguenther@suse.de>
13136 PR tree-optimization/57275
13137 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13138 return value for fail to do runtime alias checks for gather loads.
13140 2013-05-15 Jan Hubicka <jh@suse.cz>
13144 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13145 weakrefs are not external.
13146 (lto_symtab_merge_decls): Fix thinko when dealing with
13147 non-lto_symtab decls.
13148 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13149 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13150 * varpool.c (dump_varpool_node): Dump more flags.
13152 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13154 * config/i386/i386.c (processor_alias_table): Add instruction
13155 FSGSBASE for AMD bdver3 architecture.
13157 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13159 * tree.c (warn_deprecated_use): Print file:line using locus color.
13160 * diagnostic.c (diagnostic_report_current_module): Print file:line
13161 and file:line:column using locus color.
13163 2013-05-14 Mike Stump <mikestump@comcast.net>
13165 * gdbinit.in: Add __null.
13167 2013-05-14 Mike Stump <mikestump@comcast.net>
13169 * recog.h: Rename struct recog_data to Recog_data.
13170 * recog.c: Likewise.
13171 * reload.c (can_reload_into): Likewise.
13172 * config/picochip/picochip.c: Likewise.
13174 2013-05-14 Mike Stump <mikestump@comcast.net>
13176 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13178 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13180 * resource.h (struct resources): Remove unch_memory member.
13181 (CLEAR_RESOURCE): Don't clear unch_memory.
13182 * resource.c (mark_referenced_resources): Don't set it.
13183 (mark_set_resources): Likewise.
13184 (mark_target_live_regs): Don't clear it.
13185 (init_resource_info): Likewise.
13186 * reorg.c (resource_conflicts_p): Don't compare it.
13187 (redundant_insn): Don't set it.
13189 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13191 * emit-rtl.c (next_label): Remove unused function.
13192 (skip_consecutive_labels, link_cc0_insns): Move to ...
13193 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13194 only place where these functions are used, and make them static.
13196 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13198 * fold-const.c (fold_negate_expr): Handle vectors.
13199 (fold_truth_not_expr): Make it static.
13200 (fold_invert_truthvalue): New static function.
13201 (invert_truthvalue_loc): Handle vectors. Do not call
13202 fold_truth_not_expr directly.
13203 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13204 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13205 (fold_comparison): Handle vector constants.
13206 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13207 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13208 * tree.h (fold_truth_not_expr): Remove declaration.
13210 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13212 * config/aarch64/aarch64-simd.md
13213 (aarch64_vcond_internal<mode>): Rename to...
13214 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13215 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13216 float modes. Clarify all iterator modes.
13217 (vcond<mode><mode>): Use new name for vcond expanders.
13218 (vcond<v_cmp_result><mode>): Likewise.
13219 (vcondu<mode><mode>: Likewise.
13220 * config/aarch64/iterators.md (VDQF_COND): New.
13222 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13225 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13226 variable for the shift amount. Check that we shift by non-negative
13229 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13232 * config/arm/arm.h (EPILOGUE_USES): Only return true
13233 for LR_REGNUM after epilogue_completed.
13235 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13237 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13238 is error_mark_node.
13240 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13243 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13244 and Solaris 11+/x86 with gld.
13245 * configure: Regenerate.
13247 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13249 * expmed.c (expand_shift_1): Canonicalize rotates by
13250 constant bitsize / 2 to bitsize - 1.
13251 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13252 case ROTATERT>: Likewise.
13255 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13257 * config/i386/i386.md (rotateinv): New code attr.
13258 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13259 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13260 roll $31, %eax, etc.
13262 2013-05-14 Richard Biener <rguenther@suse.de>
13264 PR middle-end/57235
13265 * tree-eh.c (sink_clobbers): Give up for successors with
13266 multiple predecessors and no virtual uses.
13268 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13270 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13271 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13273 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13275 PR middle-end/57251
13276 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13277 the case when both op0 and op1 have VOIDmode.
13279 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13281 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13282 in multiply-accumulate mode.
13284 2013-05-13 Guozhi Wei <carrot@google.com>
13286 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13288 2013-05-13 Kai Tietz <ktietz@redhat.com>
13291 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13292 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13293 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13294 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13295 (ix86_expand_prologue): Likewise.
13296 (ix86_expand_split_stack_prologue): Likewise.
13297 (legitimate_pic_address_disp_p): Likewise.
13298 (legitimize_pic_address): Likewise.
13299 (legitimize_tls_address): Likewise.
13300 (legitimize_pe_coff_symbol): Likewise.
13301 (output_pic_addr_const): Likewise.
13302 (construct_plt_address): Likewise.
13303 (ix86_expand_call): Likewise.
13304 (x86_output_mi_thunk): Likewise.
13305 (x86_function_profiler): Likewise.
13307 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13309 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13310 similar switch cases.
13311 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13312 (aarch64_simd_mov_to_<mode>low): Delete.
13313 (aarch64_simd_mov_to_<mode>high): Delete.
13314 (move_lo_quad_<mode>): Add w<-r alternative.
13315 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13316 (aarch64_simd_mov_from_*): Update type attribute.
13317 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13320 2013-05-13 Jan Hubicka <jh@suse.cz>
13322 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13323 * config/i386/i386.c (ix86_compute_frame_layout,
13324 ix86_expand_epilogue, emit_i387_cw_initialization,
13325 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13326 ix86_local_alignment): Fix use of size/speed predicates.
13328 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13330 PR tree-optimization/45216
13331 PR tree-optimization/57157
13332 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13333 the (-Y) & (B - 1) variant if OP is |.
13334 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13335 return shifted. Use (-op1) & (prec - 1) as other_amount
13336 instead of prec - op1.
13338 2013-05-13 Martin Jambor <mjambor@suse.cz>
13340 PR middle-end/42371
13341 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13342 (ipa_constant_data): New type.
13343 (ipa_jump_func): Use ipa_constant_data to hold information about
13344 constant jump functions.
13345 (ipa_get_jf_constant): Adjust to jump function type changes.
13346 (ipa_get_jf_constant_rdesc): New function.
13347 (ipa_param_descriptor): New field controlled_uses.
13348 (ipa_get_controlled_uses): New function.
13349 (ipa_set_controlled_uses): Likewise.
13350 * ipa-ref.h (ipa_find_reference): Declare.
13351 * ipa-prop.c (ipa_cst_ref_desc): New type.
13352 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13354 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13355 New parameter cs. Adjust all callers.
13356 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13357 (remove_described_reference): New function.
13358 (jfunc_rdesc_usable): Likewise.
13359 (try_make_edge_direct_simple_call): Decrement controlled use count,
13360 attempt to remove reference if it hits zero.
13361 (combine_controlled_uses_counters): New function.
13362 (propagate_controlled_uses): Likewise.
13363 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13364 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13365 (ipa_print_node_params): Print described use counter.
13366 (ipa_write_jump_function): Adjust to jump function type changes.
13367 (ipa_read_jump_function): New parameter CS, pass it to
13368 ipa_set_jf_constant. Adjust caller.
13369 (ipa_write_node_info): Stream controlled use count
13370 (ipa_read_node_info): Likewise.
13371 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13373 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13374 count. Remove cloning-added reference if it reaches zero.
13375 * ipa-ref.c (ipa_find_reference): New function.
13377 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13379 * config/i386/i386.c (processor_target_table): Modified default
13380 alignment values for AMD BD and BT architectures.
13382 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13384 * tree-vect-generic.c (uniform_vector_p): Move ...
13385 * tree.c (uniform_vector_p): ... here.
13386 * tree.h (uniform_vector_p): Declare it.
13387 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13390 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13392 PR tree-optimization/57230
13393 * tree-ssa-strlen.c (handle_char_store): Record length for
13394 array store from STRING_CST.
13396 PR tree-optimization/57230
13397 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13400 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13402 * config/epiphany/epiphany.c (epiphany_init): Check size of
13403 NUM_MODES_FOR_MODE_SWITCHING.
13404 (epiphany_expand_prologue):
13405 Remove CONFIG_REGNUM initial value handling code.
13406 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13407 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13408 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13409 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13410 Don't return 1 for FP_MODE_NONE.
13411 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13412 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13413 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13414 * config/epiphany/epiphany.md (save_config): New pattern.
13416 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13418 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13420 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13422 * config/i386/i386.md (memory): Handle sseishft1.
13423 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13424 (*vec_extractv2di_1): Ditto.
13426 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13428 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13431 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13433 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13434 Add mthumb/march=armv7-a multilib.
13435 Add mthumb/march=armv7-r multilib.
13436 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13438 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13440 * config/v850/t-rtems: Add more multilibs.
13442 2013-05-10 Richard Biener <rguenther@suse.de>
13444 PR tree-optimization/57214
13445 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13446 not propagate from SSA names that occur in abnormal PHI nodes.
13448 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13450 * stor-layout.c (element_precision): New function.
13451 * machmode.h (element_precision): Declare it.
13452 * tree.c (build_minus_one_cst): New function.
13453 (element_precision): Likewise.
13454 * tree.h (build_minus_one_cst): Declare new function.
13455 (element_precision): Likewise.
13456 * fold-const.c (operand_equal_p): Use element_precision.
13457 (fold_binary_loc): Handle vector types.
13458 * convert.c (convert_to_integer): Use element_precision.
13459 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13462 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13464 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13465 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13466 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13467 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13468 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13469 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13470 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13471 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13472 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13473 (Uuw8): New constraints.
13474 (Usb4): Move into alphabetical order.
13475 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13476 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13477 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13478 previously unnamed patterns.
13479 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13480 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13481 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13482 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13483 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13484 of specific lengths.
13486 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13488 * config/i386/i386.md (rotateinv): New code attr.
13489 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13490 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13491 roll $31, %eax, etc.
13493 PR tree-optimization/45216
13494 PR tree-optimization/57157
13495 * tree-ssa-forwprop.c (simplify_rotate): New function.
13496 (ssa_forward_propagate_and_combine): Call it.
13498 2013-05-10 Richard Biener <rguenther@suse.de>
13500 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13501 disable peeling when we version for aliasing.
13502 (vector_alignment_reachable_p): Honor explicit user alignment.
13503 (vect_supportable_dr_alignment): Likewise.
13504 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13505 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13506 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13507 then peeling to arrange for the cost-model check to come first.
13509 2013-05-10 Alan Modra <amodra@gmail.com>
13511 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13512 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13513 * configure: Regenerate.
13515 2013-05-10 Alan Modra <amodra@gmail.com>
13518 * varasm.c (default_elf_select_section): Move !DECL_P check..
13519 (get_named_section): ..to here before calling get_section_name.
13521 (default_section_type_flags): Add DECL_P check.
13522 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13523 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13525 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13527 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13528 When using gen_stack_adjust_str with a register offset, add a
13529 REG_FRAME_RELATED_EXPR note.
13531 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13533 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13534 (*vec_extractv4si_zext_mem): Ditto.
13535 (*vec_extractv2di): Add 0->x and x->x alternatives.
13536 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13537 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13539 2013-05-09 Jason Merrill <jason@redhat.com>
13541 N3639 C++1y VLA support
13542 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13545 * tree.c (build_constructor_va): New.
13546 * tree.h: Declare it.
13548 2013-05-09 Martin Jambor <mjambor@suse.cz>
13551 * gimple-fold.c (canonicalize_constructor_val): Call
13552 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13554 2013-05-09 Jan Hubicka <jh@suse.cz>
13555 Richard Biener <rguenther@suse.de>
13558 * symtab.c (symtab_make_decl_local): Do not add private names.
13560 2013-05-09 Jan Hubicka <jh@suse.cz>
13563 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13564 (unlink_from_assembler_name_hash): Likewise.
13565 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13566 symtab_unregister_node, symtab_initialize_asm_name_hash,
13567 change_decl_assembler_name): Update.
13569 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13571 * config/aarch64/aarch64.md: New movtf split.
13572 (*movtf_aarch64): Update.
13573 (aarch64_movdi_tilow): Handle TF modes and rename to
13574 aarch64_movdi_<mode>low.
13575 (aarch64_movdi_tihigh): Handle TF modes and rename to
13576 aarch64_movdi_<mode>high
13577 (aarch64_movtihigh_di): Handle TF modes and rename to
13578 aarch64_mov<mode>high_di
13579 (aarch64_movtilow_di): Handle TF modes and rename to
13580 aarch64_mov<mode>low_di
13581 (aarch64_movtilow_tilow): Remove spurious whitespace.
13582 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13584 (aarch64_print_operand): Update.
13586 2013-05-09 Alan Modra <amodra@gmail.com>
13588 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13590 * configure: Regenerate.
13592 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13594 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13595 splitter preparation statements.
13596 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13597 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13600 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13602 * gimple-ssa-strength-reduction.c (count_candidates): Change
13603 return value to int.
13604 (analyze_candidates_and_replace): Change type of length to int.
13606 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13608 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13609 (*vec_extract<mode>): Use VI12_128 mode iterator.
13610 (*vec_extract<mode>_mem): Ditto.
13611 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13614 2013-05-08 Diego Novillo <dnovillo@google.com>
13619 2012-08-17 Diego Novillo <dnovillo@google.com>
13622 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13623 * config.in: Regenerate.
13624 * configure: Regenerate.
13625 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13627 2013-05-08 Jan Hubicka <jh@suse.cz>
13630 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13631 * cgraph.h (symtab_node_base): Add unique_name.
13632 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13633 input_overwrite_node, input_varpool_node): Stream unique_name.
13634 * cgraphclones.c (cgraph_create_virtual_clone,
13635 cgraph_function_versioning): Set unique_name.
13636 * ipa.c (function_and_variable_visibility): Set unique_name.
13638 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13640 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13641 (alloc_cand_and_find_basis): Restrict conditional candidate
13642 processing to CAND_MULTs.
13644 2013-05-08 Jan Hubicka <jh@suse.cz>
13647 lto-symtab.c (lto_symtab_symbol_p): New function.
13648 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13649 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13650 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13651 Skip static symbols.
13653 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13655 PR tree-optimization/57200
13656 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13657 Only call inform if the preceding warning_at returns true.
13659 2013-05-07 Han Shen <shenhan@google.com>
13661 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13662 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13663 * common.opt (fstack-protector-strong): New option.
13664 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13665 * doc/invoke.texi (Optimization Options): Document
13666 "-fstack-protector-strong".
13667 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13669 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13671 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13673 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13675 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13676 it is not lhs of the insn.
13677 (setup_reg_equiv): Remove insn having equiv memory which it is not
13679 * lra-constraints.c (process_address): Try to improve generation
13680 code for address base + disp.
13681 (lra_constraints): Make correct the code for checking insn setting
13682 up backward equivalence. Remove insn only if it is in the init
13684 * lra-eliminations.c (update_reg_eliminate): Change return value.
13685 (lra_eliminate): Use the result.
13687 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13689 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13690 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13691 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13692 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13693 register target operands.
13694 (*vec_extractv8hi_sse2): New pattern.
13695 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13696 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13697 (*vec_extract<mode>_mem): New insn and split pattern.
13699 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13701 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13702 (TARGET_ASAN_SHADOW_OFFSET): Define.
13703 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13704 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13706 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13708 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13709 (incr_vec_index): Return -1 if increment not found.
13710 (create_add_on_incoming_edge): Assert if increment not found.
13711 (record_increment): Limit number of increments recorded.
13712 (all_phi_incrs_profitable): Return false if an increment not found.
13713 (replace_profitable_candidates): Don't process increments that were
13715 (analyze_candidates_and_replace): Limit size of incr_vec.
13717 2013-05-07 Richard Biener <rguenther@suse.de>
13719 * calls.c (special_function_p): setjmp-like functions are leaf.
13720 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13721 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13723 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13725 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13726 (aarch64_simd_mov<mode>): New expander.
13727 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13728 (aarch64_simd_mov_to_<mode>high): Likewise.
13729 (aarch64_simd_mov_from_<mode>low): Likewise.
13730 (aarch64_simd_mov_from_<mode>high): Likewise.
13731 (aarch64_dup_lane<mode>): Update.
13732 (aarch64_dup_lanedi): New instruction pattern.
13733 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13734 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13736 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13738 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13739 (replace_mult_candidate): Remove unnecessary argument; remove
13740 unnecessary parameter from call to introduce_cast_before_cand.
13741 (replace_unconditional_candidate): Remove unnecessary parameter
13742 from call to replace_mult_candidate.
13743 (replace_conditional_candidate): Likewise.
13744 (insert_initializers): Use make_temp_ssa_name.
13745 (introduce_cast_before_cand): Remove unnecessary argument; use
13746 make_temp_ssa_name.
13747 (replace_one_candidate): Remove unnecessary argument; remove
13748 unnecessary parameter from calls to introduce_cast_before_cand.
13749 (replace_profitable_candidates): Remove unnecessary parameters
13750 from calls to replace_one_candidate.
13752 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13754 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13755 phi def as possibly hiding a basis for a CAND_ADD whose operands
13756 have been commuted in the analysis.
13757 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13759 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13761 * config/aarch64/aarch64.md
13762 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13763 shift value between 0-4.
13765 2013-05-07 Richard Biener <rguenther@suse.de>
13767 * double-int.h (rshift): New overload.
13768 * double-int.c (rshift): New function.
13769 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13770 (create_reference_ops_from_ref): Remove.
13771 (vn_reference_insert): Use shared ops for constructing the
13772 reference and copy it.
13774 2013-05-07 Richard Biener <rguenther@suse.de>
13776 PR middle-end/57190
13777 * tree-eh.c (sink_clobbers): Properly propagate
13778 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13780 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13782 PR tree-optimization/57149
13783 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13784 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13785 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13786 uninit_undefined_value_p instead of ssa_undefined_value_p.
13789 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13790 for modifier == EXPAND_INITIALIZER.
13792 2013-05-07 Anton Blanchard <anton@samba.org>
13794 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13795 for powerpc64 little endian.
13796 * configure: Regenerate.
13798 2013-05-06 Graham Stott <grahams@btinternet.com>
13800 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13801 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13802 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13803 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13805 2013-05-06 Graham Stott <grahams@btinternet.com>
13807 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13808 codes which allow non-lvalues.
13810 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13812 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13813 components are all 1s.
13814 (integer_minus_onep): New function.
13815 * tree.h (integer_minus_onep): Declare it.
13816 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13817 integer_minus_onep instead of integer_all_onesp.
13819 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13822 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13823 variations of these patterns.
13825 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13827 * config/i386/i386.md (isa): Add x64_sse4 member.
13828 (enabled): Handle x64_sse4.
13829 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13830 instruction for 64bit SSE4_1 targets. Update insn attributes.
13831 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13832 instruction for SSE4_1 targets. Update insn attributes.
13833 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13834 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13835 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13837 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13838 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13839 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13841 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13844 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13846 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13848 * final.c (do_assembler_dialects): Don't handle curly braces and
13849 vertical bar escaped by % as dialect delimiters.
13850 (output_asm_insn): Print curly braces and vertical bar if escaped
13851 by % and ASSEMBLER_DIALECT defined.
13852 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
13853 * doc/tm.texi: Regenerated.
13855 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13857 * config/mips/mips.c: Include tree-pass.h.
13858 (mips_reorg): Split in pre- and post-dbr_schedule parts.
13859 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
13860 (pass_mips_machine_reorg2): New machine specific pass.
13861 (insert_pass_mips_machine_reorg2): New pass plugin definition.
13862 (mips_option_override): Register the new pass.
13863 * rtl.h (cleanup_barriers): Remove prototype.
13864 (dbr_schedule): Likewise.
13865 * jump.c (cleanup_barriers): Make static.
13866 * reorg.c (dbr_schedule): Likewise.
13868 2013-05-06 Richard Biener <rguenther@suse.de>
13870 PR tree-optimization/57185
13871 * tree-parloops.c (add_field_for_reduction): Handle anonymous
13872 SSA names properly.
13874 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13877 * config/i386/i386.c (add_parameter_dependencies): Add dependence
13878 between "first_arg" and "insn", not "last" and "insn".
13880 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13882 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
13883 (find_candidates_in_block): Re-enable slsr_process_phi.
13884 (create_phi_basis): Fix double counting of candidate adjustment.
13886 2013-05-06 Richard Biener <rguenther@suse.de>
13888 PR middle-end/57147
13889 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
13890 the edge is also fallthru, preserve it and just clear the
13892 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
13893 also complex, preserve that and just clear the fallthru flag.
13894 * tree-inline.c (update_ssa_across_abnormal_edges): Also
13895 update virtual operands.
13897 2013-05-06 Alan Modra <amodra@gmail.com>
13899 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
13900 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
13901 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
13902 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
13903 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13904 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
13906 2013-05-06 Alan Modra <amodra@gmail.com>
13908 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
13909 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
13910 (DEFAULT_ASM_ENDIAN): Define.
13911 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13912 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
13913 Update -K PIC clause from sysv4.h.
13914 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
13915 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
13917 2013-05-06 Alan Modra <amodra@gmail.com>
13919 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
13920 twice for little-endian.
13921 (ashrdi3_no_power, ashrdi3): Support little-endian.
13923 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13926 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
13927 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
13929 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
13930 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
13931 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
13934 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
13935 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13937 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
13938 * config/pa/pa.opt: Make mbig-switch a no-op.
13939 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
13940 (CASE_VECTOR_MODE): Always return SImode.
13941 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
13942 for the !TARGET_BIG_SWITCH case.
13943 * config/pa/pa-linux.h: Likewise.
13944 * config/pa/pa-openbsd.h: Likewise.
13945 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
13946 * config/pa/pa.md (short_jump): Remove define_insn.
13947 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
13948 (casesi0): Remove define_insn.
13949 (type): Remove btable_branch.
13950 (pa_combine_type): Likewise.
13951 (in_nullified_branch_delay): Likewise.
13952 (in_call_delay): Likewise.
13953 (define_delay): Likewise.
13954 (define_insn_reservation "Z3"): Likewise.
13955 (define_insn_reservation "Z4"): Likewise.
13956 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
13957 (pa_adjust_insn_length): Remove adjustment for btable branches.
13958 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
13961 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
13963 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13964 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
13965 Add m->r,x alternatives.
13966 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
13967 splitters using SWI48x mode iterator.
13968 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
13969 TARGET_64BIT. Add m->x alternative.
13970 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
13971 Add o->x alternative. Enable for TARGET_SSE.
13972 (sse_storeq): Remove expander.
13973 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
13974 with memory input operand.
13975 (*vec_extractv2di_1 splitter): New.
13976 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
13977 * config/i386/i386.md (ssevecmodelower): New mode attribute.
13979 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13981 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
13982 (INT_LOWPART): Delete.
13983 (extract_MB): Adjust.
13984 (extract_ME): Adjust.
13985 (print_operand): Adjust.
13987 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13989 * config/rs6000/predicates.md (reg_or_add_cint_operand,
13990 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13991 (reg_or_logical_cint_operand, easy_fp_constant,
13992 logical_const_operand): Delete "CONST_DOUBLE" case.
13993 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
13994 "HOST_BITS_PER_WIDE_INT == 64" test.
13995 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
13996 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
13997 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
13998 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14000 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14001 CONST_DOUBLE DImode/VOIDmode case.
14002 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14003 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14004 CONST_DOUBLE VOIDmode case.
14005 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14006 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14007 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14008 Delete CONST_DOUBLE case.
14009 (splitters for mov FMOVE64 const_double): Delete
14010 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14011 "HOST_BITS_PER_WIDE_INT >= 64" test.
14012 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14014 (mov DI const_double): Delete.
14016 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14018 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14019 on op shows all bits zero in mode of a lowpart subreg, return zero.
14021 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14024 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14025 to save TFmode registers and DImode to save TImode registers for
14026 caller save operations.
14027 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14028 mark being partially clobbered since they only use the first
14031 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14032 and TDmode only use the upper 64-bits of each VSX register.
14034 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14036 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14037 (find_candidates_in_block): Disable slsr_process_phi.
14039 2013-05-03 Guozhi Wei <carrot@google.com>
14041 * coverage.c (coverage_obj_init): Move the construction of gcov
14043 (build_init_ctor): ... here.
14045 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14047 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14048 (slsr_cand_d): Redefine def_phi.
14049 (stride_status, phi_adjust_status, count_phis_status): New enums.
14050 (find_phi_def): New.
14051 (find_basis_for_base_expr): New.
14052 (find_basis_for_candidate): Handle hidden bases.
14053 (alloc_cand_and_find_basis): Handle phi candidates.
14054 (slsr_process_phi): New.
14055 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14056 (create_mul_imm_cand): Likewise.
14057 (create_add_ssa_cand): Exclude phi base candidates.
14058 (create_add_imm_cand): Likewise.
14059 (slsr_process_cast): Likewise.
14060 (slsr_process_copy): Likewise.
14061 (find_candidates_in_block): Handle phi candidates.
14062 (dump_candidate): Likewise.
14063 (unconditional_cands): Delete.
14064 (unconditional_cands_with_known_stride_p): Delete.
14065 (phi_dependent_cand_p): New.
14066 (cand_increment): Handle phi-dependent candidates.
14067 (replace_dependent): Delete.
14068 (replace_mult_candidate): New.
14069 (replace_unconditional_candidate): New.
14070 (incr_vec_index): Move to avoid forward reference.
14071 (create_add_on_incoming_edge): New.
14072 (create_phi_basis): New.
14073 (replace_dependents): Delete.
14074 (replace_conditional_candidate): New.
14075 (phi_add_costs): New.
14076 (replace_uncond_cands_and_profitable_phis): New.
14077 (record_increment): Handle phi adjustments.
14078 (record_phi_increments): New.
14079 (record_increments): Handle phi adjustments.
14080 (phi_incr_cost): New.
14081 (lowest_cost_path): Handle phis.
14082 (total_savings): Likewise.
14083 (analyze_increments): Likewise.
14084 (ncd_with_phi): New.
14085 (ncd_of_cand_and_phis): New.
14086 (nearest_common_dominator_for_cands): Handle phi increments.
14087 (all_phi_incrs_profitable): New.
14088 (replace_profitable_candidates): Handle phi-dependent candidates.
14089 (analyze_candidates_and_replace): Likewise.
14091 2013-05-03 Teresa Johnson <tejohnson@google.com>
14094 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14095 do not exceed REG_BR_PROB_BASE.
14097 2013-05-03 Jeff Law <law@redhat.com>
14099 PR tree-optimization/57144
14100 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14101 operand of the condition will bit into the new type when eliminating
14102 a cast feeding a condition.
14104 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14106 PR rtl-optimization/57130
14107 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14108 of COMPARE as in_code to the recursive call if needed.
14110 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14112 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14113 (enabled): Handle new members.
14114 * config/i386/sse.md (*vec_concatv2si): Merge from
14115 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14116 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14118 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14120 PR tree-optimization/57027
14121 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14122 for fnms opportunity, check we got the prerequisite kind
14123 of tree / gimple before using accessor functions.
14125 2013-05-03 Richard Biener <rguenther@suse.de>
14127 * double-int.h (lshift): New overload without precision
14128 and arith argument.
14129 (operator *=, operator +=, operator -=): Move ...
14130 * double-int.c (operator *=, operator +=, operator -=): ... here
14131 and implement more efficiently.
14132 (mul_double_with_sign): Remove.
14133 (lshift_double): Adjust to take unsinged shift argument, push
14134 dispatching code to callers.
14135 (mul_double_wide_with_sign): Add early out for callers that
14136 are not interested in high parts or overflow.
14137 (lshift): New function.
14138 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14139 dispatch code here.
14140 (lrotate, rrotate): Use logical shifts.
14141 * expr.c (get_inner_reference): Use lshift.
14142 * fixed-value.c (do_fixed_divide): Likewise.
14143 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14144 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14145 (indirect_refs_may_alias_p): Likewise.
14146 (stmt_kills_ref_p_1): Likewise.
14148 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14150 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14152 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14154 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14155 scalar form of FABD instruction.
14157 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14159 * lra-constraints.c (process_alt_operands): Add checking alt
14160 number to choose the best alternative.
14162 2013-05-02 Richard Biener <rguenther@suse.de>
14164 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14165 bitmap and its handling.
14166 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14168 2013-05-02 Richard Biener <rguenther@suse.de>
14170 PR middle-end/57140
14171 * tree-inline.c (copy_loops): Properly handle removed loops.
14172 (copy_cfg_body): Mark destination loops for fixup if source
14173 loops needed fixup.
14175 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14178 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14179 generating simple_return for naked functions.
14181 2013-05-02 Martin Jambor <mjambor@suse.cz>
14183 PR middle-end/56988
14184 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14185 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14187 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14188 ipa_agg_replacement_value structures.
14189 (known_aggs_to_agg_replacement_list): Likewise.
14190 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14191 (read_agg_replacement_chain): Likewise.
14192 (ipcp_transform_function): Also check that by_ref flags match.
14194 2013-05-02 Richard Biener <rguenther@suse.de>
14196 * graphds.h (struct graph): Add obstack member.
14197 * graphds.c (new_graph): Initialize obstack and allocate
14199 (add_edge): Allocate edge from the obstack.
14200 (free_graph): Free the obstack instead of all edges and vertices.
14202 2013-05-02 Teresa Johnson <tejohnson@google.com>
14204 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14206 * cfg.c (update_bb_profile_for_threading): Ditto.
14207 * tree-inline.c (copy_bb): Ditto.
14208 (copy_edges_for_bb): Ditto.
14209 (initialize_cfun): Ditto.
14210 (copy_cfg_body): Ditto.
14211 (expand_call_inline): Ditto.
14212 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14213 (estimate_node_size_and_time): Ditto.
14214 (inline_merge_summary): Ditto.
14215 * cgraphclones.c (cgraph_clone_edge): Ditto.
14216 (cgraph_clone_node): Ditto.
14217 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14218 (compute_trg_info): Ditto.
14220 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14222 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14223 S reg when fp attribute set.
14224 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14226 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14228 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14230 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14231 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14232 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14234 2013-05-02 Richard Biener <rguenther@suse.de>
14236 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14237 (struct instantiate_cache_entry): New type.
14238 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14239 (struct instantiate_cache_type): New type.
14240 (set_instantiated_value, get_instantiated_value): Remove.
14241 (get_instantiated_value_entry): New function.
14242 (instantiate_scev_name): Use the new cache and adjust.
14243 (instantiate_scev_poly): Adjust.
14244 (instantiate_scev_binary): Likewise.
14245 (instantiate_array_ref): Likewise.
14246 (instantiate_scev_convert): Likewise.
14247 (instantiate_scev_not): Likewise.
14248 (instantiate_scev_3): Likewise.
14249 (instantiate_scev_2): Likewise.
14250 (instantiate_scev_r): Likewise.
14251 (instantiate_scev): Likewise.
14252 (resolve_mixers): Likewise.
14254 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14257 * lra-constraints.c (best_small_class_operands_num): Remove.
14258 (process_alt_operands): Remove small_class_operands_num. Take
14259 small classes operands into losers and only if the operand is not
14260 matched. Modify debugging output.
14261 (curr_insn_transform): Remove best_small_class_operands_num.
14264 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14266 * config/aarch64/aarch64-builtins.c
14267 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14268 * config/aarch64/aarch64-simd-builtins.def
14269 (reduc_splus_): Add new modes.
14270 (reduc_uplus_): New.
14271 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14272 (reduc_uplus_v4sf): Likewise.
14273 (reduc_splus_v4sf): Likewise.
14274 (aarch64_addv<mode>): Likewise.
14275 (reduc_uplus_<mode>): Likewise.
14276 (reduc_splus_<mode>): Likewise.
14277 (aarch64_addvv2di): Likewise.
14278 (reduc_uplus_v2di): Likewise.
14279 (reduc_splus_v2di): Likewise.
14280 (aarch64_addvv2si): Likewise.
14281 (reduc_uplus_v2si): Likewise.
14282 (reduc_splus_v2si): Likewise.
14283 (reduc_<sur>plus_<mode>): New.
14284 (reduc_<sur>plus_v2di): Likewise.
14285 (reduc_<sur>plus_v2si): Likewise.
14286 (reduc_<sur>plus_v4sf): Likewise.
14287 (aarch64_addpv4sf): Likewise.
14288 * config/aarch64/arm_neon.h
14289 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14290 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14291 add UNSPEC_SADDV, UNSPEC_UADDV.
14293 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14295 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14297 * config/aarch64/arm_neon.h
14298 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14300 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14302 * config/aarch64/aarch64-builtins
14303 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14305 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14307 * config/aarch64/aarch64-simd-builtins.def
14308 (reduc_smax_): New.
14309 (reduc_smin_): Likewise.
14310 (reduc_umax_): Likewise.
14311 (reduc_umin_): Likewise.
14312 (reduc_smax_nan_): Likewise.
14313 (reduc_smin_nan_): Likewise.
14316 (smax): Update for V2SF, V4SF and V2DF modes.
14319 (smin_nan): Likewise.
14320 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14321 (<su><maxmin><mode>3): ...This, refactor.
14322 (s<maxmin><mode>3): New.
14323 (<maxmin_uns><mode>3): Likewise.
14324 (reduc_<maxmin_uns>_<mode>): Refactor.
14325 (reduc_<maxmin_uns>_v4sf): Likewise.
14326 (reduc_<maxmin_uns>_v2si): Likewise.
14327 (aarch64_<fmaxmin><mode>: Remove.
14328 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14330 (vmin<q>_f<32,64>): Likewise.
14331 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14333 (su): Add mappings for smax, smin, umax, umin.
14335 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14336 (FMAXMIN): Rename as...
14337 (FMAXMIN_UNS): ...This.
14339 (fmaxminv): Likewise.
14340 (fmaxmin): Likewise.
14342 (maxmin_uns_op): Likewise.
14344 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14346 * config/aarch64/arm_neon.h
14347 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14348 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14349 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14351 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14353 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14354 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14356 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14358 * config/aarch64/aarch64-simd.md
14359 (vcond<mode>_internal): Handle special cases for constant masks.
14360 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14361 (vcondu<mode><mode>): Likewise.
14362 (vcond<v_cmp_result><mode>): New.
14364 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14366 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14367 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14368 * config/aarch64/aarch64-simd-builtins.def
14369 (cmeq): Update to BUILTIN_VALLDI.
14374 * config/aarch64/arm_neon.h
14375 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14376 to builtins or C as appropriate.
14378 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14380 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14382 (cmhi): Rename to...
14384 * config/aarch64/aarch64-simd.md
14385 (simd_mode): Add SF.
14386 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14387 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14388 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14389 (cstore<mode>_neg): ...This.
14390 * config/aarch64/iterators.md
14392 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14393 (COMPARISONS): New.
14394 (UCOMPARISONS): Likewise.
14395 (optab): Add missing comparisons.
14401 (VCMP_S): Likewise.
14402 (VCMP_U): Likewise.
14403 (V_cmp_result): Add DF, SF modes.
14404 (v_cmp_result): Likewise.
14406 (vmtype): Likewise.
14407 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14409 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14411 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14412 define_insn to define_insn_and_split.
14413 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14414 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14415 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14416 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14417 (thumb2_negscc): Likewise.
14419 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14421 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14423 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14425 * config/arm/thumb2.md: Remove trailing whitespaces.
14427 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14429 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14430 Use gen_int_mode rather than GEN_INT.
14432 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14434 * value-prof.c (stream_in_histogram_value): Remove the strayed
14437 2013-04-30 Richard Biener <rguenther@suse.de>
14439 PR middle-end/57122
14440 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14442 2013-04-30 Richard Biener <rguenther@suse.de>
14444 PR middle-end/57107
14445 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14447 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14449 PR rtl-optimization/56957
14450 PR rtl-optimization/57105
14451 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14452 variable. Use just INSN_UID for determining whether an insn
14453 should be only disconnected from the insn stream.
14454 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14456 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14458 PR tree-optimization/57104
14459 * tsan.c (instrument_expr): Don't instrument accesses to
14460 DECL_HARD_REGISTER VAR_DECLs.
14462 2013-04-30 Richard Biener <rguenther@suse.de>
14464 * function.h (loops_for_fn): New inline function.
14465 (set_loops_for_fn): Likewise.
14466 * cfgloop.h (place_new_loop): Add struct function parameter.
14467 (get_loop): Likewise.
14468 (get_loops): Likewise.
14469 (number_of_loops): Likewise.
14470 (fel_next): Adjust.
14471 (fel_init): Likewise.
14472 * cfg.c (get_loop_copy): Adjust.
14473 * cfgloop.c (flow_loops_dump): Likewise.
14474 (record_loop_exits): Likewise.
14475 (verify_loop_structure): Likewise.
14476 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14477 (estimate_reg_pressure_cost): Likewise.
14478 (mark_loop_exit_edges): Likewise.
14479 * cfgloopmanip.c (place_new_loop): Likewise.
14480 (add_loop): Likewise.
14481 (duplicate_loop): Likewise.
14482 * graph.c (draw_cfg_nodes): Likewise.
14483 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14484 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14485 (extract_affine_chrec): Likewise.
14486 (build_scop_iteration_domain): Likewise.
14487 * graphite.c (graphite_initialize): Likewise.
14488 * ira-build.c (create_loop_tree_nodes): Likewise.
14489 (more_one_region_p): Likewise.
14490 (rebuild_regno_allocno_maps): Likewise.
14491 (mark_loops_for_removal): Likewise.
14492 (mark_all_loops_for_removal): Likewise.
14493 (remove_unnecessary_regions): Likewise.
14494 (ira_build): Likewise.
14495 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14496 * loop-init.c (fix_loop_structure): Likewise.
14497 (gate_rtl_move_loop_invariants): Likewise.
14498 (gate_rtl_unswitch): Likewise.
14499 (gate_rtl_unroll_and_peel_loops): Likewise.
14500 (rtl_doloop): Likewise.
14501 * lto-streamer-in.c (input_cfg): Likewise.
14502 * lto-streamer-out.c (output_cfg): Likewise.
14503 * modulo-sched.c (sms_schedule): Likewise.
14504 * predict.c (tree_estimate_probability): Likewise.
14505 (tree_estimate_probability_driver): Likewise.
14506 (estimate_loops): Likewise.
14507 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14508 (move_sese_region_to_fn): Likewise.
14509 (debug_loop_num): Likewise.
14510 * tree-chrec.c (chrec_evaluate): Likewise.
14511 (hide_evolution_in_other_loops_than_loop): Likewise.
14512 (chrec_component_in_loop_num): Likewise.
14513 (reset_evolution_in_loop): Likewise.
14514 (evolution_function_is_invariant_rec_p): Likewise.
14515 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14516 * tree-inline.c (copy_loops): Likewise.
14517 (copy_cfg_body): Likewise.
14518 (tree_function_versioning): Likewise.
14519 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14520 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14522 (add_to_evolution_1): Likewise.
14523 (scev_const_prop): Likewise.
14524 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14525 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14526 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14527 (tree_ssa_lim_initialize): Likewise.
14528 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14529 (verify_loop_closed_ssa): Likewise.
14530 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14531 (tree_ssa_loop_im): Likewise.
14532 (tree_ssa_loop_unswitch): Likewise.
14533 (tree_vectorize): Likewise.
14534 (check_data_deps): Likewise.
14535 (tree_ssa_loop_ivcanon): Likewise.
14536 (tree_ssa_loop_bounds): Likewise.
14537 (tree_complete_unroll): Likewise.
14538 (tree_complete_unroll_inner): Likewise.
14539 (tree_parallelize_loops): Likewise.
14540 (tree_ssa_loop_prefetch): Likewise.
14541 (tree_ssa_loop_ivopts): Likewise.
14542 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14543 * tree-vectorizer.c (vectorize_loops): Likewise.
14545 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14547 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14548 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14549 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14550 with EABI_LINK_SPEC.
14552 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14555 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14558 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14561 * lra-constraints.c (process_alt_operands): Discourage a bit more
14562 using memory for pseudos. Print cost dump for alternatives.
14563 Modify cost values for conflicts with early clobbers.
14564 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14566 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14569 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14571 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14573 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14574 from/to S register.
14575 (movdi_aarch64): Support LDR/STR from/to D register.
14577 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14579 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14580 or higher by default.
14582 2013-04-29 Richard Biener <rguenther@suse.de>
14584 PR middle-end/57075
14585 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14586 even if not adding abnormal edges for calls that can make
14589 2013-04-29 Richard Biener <rguenther@suse.de>
14591 PR middle-end/57103
14592 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14594 (move_stmt_r): Remove redundant checking.
14596 2013-04-29 Teresa Johnson <tejohnson@google.com>
14599 * basic-block.h (apply_scale): New function.
14600 (apply_probability): Use apply_scale.
14601 * gimple-streamer-in.c (input_bb): Ditto.
14602 * lto-streamer-in.c (input_cfg): Ditto.
14603 * lto-cgraph.c (merge_profile_summaries): Ditto.
14604 * tree-optimize.c (execute_fixup_cfg): Ditto.
14605 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14606 (copy_edges_for_bb): Ditto.
14607 (copy_cfg_body): Ditto.
14609 2013-04-29 Tom de Vries <tom@codesourcery.com>
14611 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14612 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14613 (tail_merge_optimize): Handle current_loops == NULL.
14615 2013-04-26 Jeff Law <law@redhat.com>
14617 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14618 (simplify_cond_using_ranges): Generalize code to simplify
14619 COND_EXPRs where one argument is a constant and the other
14620 is an SSA_NAME created by an integral type conversion.
14622 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14624 * config/arm/arm.md (store_minmaxsi): Use only when
14625 optimize_insn_for_size_p.
14627 2013-04-29 Christian Bruel <christian.bruel@st.com>
14630 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14632 2013-04-29 Richard Biener <rguenther@suse.de>
14634 PR middle-end/57089
14635 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14636 loop tree make sure to schedule a fixup for the child as well.
14637 (expand_omp_for_generic): Properly add loops.
14638 (expand_omp_for_static_nochunk): Likewise.
14639 (expand_omp_for_static_chunk): Likewise.
14640 (expand_omp_for): For the degenerate case fixup loops.
14641 (expand_omp_sections): Fix default bb placement in loops.
14642 (expand_omp_atomic_pipeline): Properly add loops.
14644 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14646 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14648 2013-04-29 Tom de Vries <tom@codesourcery.com>
14650 * tree-ssa-tail-merge.c: Update header comment.
14652 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14654 * config/aarch64/arm_neon.h
14655 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14656 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14657 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14658 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14659 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14660 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14661 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14662 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14664 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14666 * config/aarch64/aarch64-simd.md
14667 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14668 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14669 fix_trunc, fixuns_trunc.
14670 (ftrunc<VDQF:mode>2): New.
14671 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14672 (fix_trunc_optab): New.
14674 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14676 * config/aarch64/aarch64-builtins.c
14677 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14678 iceilf, lround, iroundf.
14680 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14683 * config/i386/i386.h (enum ix86_tune_indices)
14684 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14685 New, split from X86_TUNE_INTER_UNIT_MOVES.
14686 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14687 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14688 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14689 (TARGET_INTER_UNIT_MOVES): Remove.
14690 * config/i386/i386.c (initial_ix86_tune_features): Update.
14691 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14692 (ix86_expand_convert_uns_didf_sse): Use
14693 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14694 (ix86_expand_vector_init_one_nonzero): Ditto.
14695 (ix86_expand_vector_init_interleave): Ditto.
14696 (inline_secondary_memory_needed): Return true for moves from SSE class
14697 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14698 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14699 * config/i386/constraints.md (Yi, Ym): Depend on
14700 TARGET_INTER_UNIT_MOVES_TO_VEC.
14701 (Yj, Yn): New constraints.
14702 * config/i386/i386.md (*movdi_internal): Change constraints of
14703 operand 1 from Yi to Yj and from Ym to Yn.
14704 (*movsi_internal): Ditto.
14705 (*movdf_internal): Ditto.
14706 (*movsf_internal): Ditto.
14707 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14708 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14709 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14710 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14711 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14712 * config/i386/sse.md (movdi_to_sse): Ditto.
14713 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14714 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14715 TARGET_INTER_UNIT_MOVES.
14716 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14717 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14718 instead of TARGET_INTER_UNIT_MOVES.
14719 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14720 operand 1 from Yi to Yj and from Ym to Yn.
14722 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14724 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14725 (float_truncate_hi_): Likewise.
14726 (float_extend_lo_): Likewise.
14727 (float_truncate_lo_): Likewise.
14728 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14729 (aarch64_float_extend_lo_v2df): Likewise.
14730 (vec_unpacks_hi_v4sf): Likewise.
14731 (aarch64_float_truncate_lo_v2sf): Likewise.
14732 (aarch64_float_truncate_hi_v4sf): Likewise.
14733 (vec_pack_trunc_v2df): Likewise.
14734 (vec_pack_trunc_df): Likewise.
14736 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14738 * config/aarch64/aarch64-builtins.c
14739 (aarch64_fold_builtin): Fold float conversions.
14740 * config/aarch64/aarch64-simd-builtins.def
14741 (floatv2si, floatv4si, floatv2di): New.
14742 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14743 * config/aarch64/aarch64-simd.md
14744 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14745 * config/aarch64/iterators.md (FLOATUORS): New.
14746 (optab): Add float, floatuns.
14747 (su_optab): Likewise.
14749 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14751 * config/aarch64/aarch64-builtins.c
14752 (aarch64_builtin_vectorized_function): Use new names for
14754 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14755 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14756 (fcvtzu): Split as...
14757 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14758 (fcvtas): Split as...
14759 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14760 (fcvtau): Split as...
14761 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14762 (fcvtps): Split as...
14763 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14764 (fcvtpu): Split as...
14765 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14766 (fcvtms): Split as...
14767 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14768 (fcvtmu): Split as...
14769 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14770 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14771 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14772 (lfrintnusf, lfrintnudf): Likewise.
14773 * config/aarch64/aarch64-simd.md
14774 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14776 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14777 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14778 (fcvt_pattern): Likewise.
14780 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14782 * config/aarch64/aarch64-simd.md
14783 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14784 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14786 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14788 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14789 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14790 (vrnd<a,m,n,p>_f32): Implement using builtins.
14791 (vrnd<i,x><q>_f<32, 64>): New.
14793 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14795 * config/aarch64/aarch64-builtins.c
14796 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14797 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14798 (frintz): Rename to...
14800 (frintp): Rename to...
14802 (frintm): Rename to...
14804 (frinti): Rename to...
14805 (nearbyint): ...this.
14806 (frintx): Rename to...
14808 (frinta): Rename to...
14810 * config/aarch64/aarch64-simd.md
14811 (aarch64_frint<frint_suffix><mode>): Delete.
14812 (<frint_pattern><mode>2): Convert to insn.
14813 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14814 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14815 (frint_pattern): Likewise.
14816 (frint_suffix): Likewise.
14818 2013-04-29 Richard Biener <rguenther@suse.de>
14820 PR tree-optimization/57081
14821 * loop-init.c: Include tree-flow.h.
14822 (loop_optimizer_finalize): Free number of iteration estimates.
14823 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14825 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14827 PR tree-optimization/57083
14828 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14829 non-singleton shift count range, zero extend low_bound for uns case.
14831 * config/i386/predicates.md (general_vector_operand): New predicate.
14832 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14833 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14834 if they aren't nonimmediate operands. If their original values
14835 satisfy const_vector_equal_evenodd_p, don't shift them.
14836 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14837 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14838 if not nonimmediate_operand.
14839 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14840 instead of register_operand.
14841 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14843 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14845 * stor-layout.c (finalize_size_functions): Allocate a structure and
14846 reset cfun before dumping the functions.
14848 2013-04-27 Jakub Jelinek <jakub@redhat.com>
14850 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
14853 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
14854 use xop_pmacsdqh if uns_p.
14855 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
14856 the immediate rotate count.
14858 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
14860 * rtl.h (struct rtx_def): Add comment for field jump.
14861 (LRA_SUBREG_P): New macro.
14862 * recog.c (register_operand): Check LRA_SUBREG_P.
14863 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14865 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14866 (lra_final_code_change): Skip subreg change for operators.
14867 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14868 if there are no operand changes.
14869 * lra-constraints.c (curr_insn_set): New.
14870 (match_reload): Set LRA_SUBREG_P.
14871 (emit_spill_move): Ditto.
14872 (check_and_process_move): Use curr_insn_set. Process only single
14873 set insns. Don't initialize sec_mem_p and change_p.
14874 (simplify_operand_subreg): Use LRA_SUBREG_P.
14875 (reg_in_class_p): New function.
14876 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14877 of #ifdef. Add code to remove cycling.
14878 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14879 non-null disp. Reload inner instead of disp when base and index
14880 are null. Try to put lo_sum into register.
14881 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14882 (check_and_process_move): Move code for move cost check to
14883 simple_move_p. Remove equiv_substitution.
14884 (simple_move_p): New function.
14885 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14886 curr_insn_set. Call check_and_process_move only for single set
14887 insns. Use the new function. Move call of check_and_process_move
14888 after operand equiv substitution and address process.
14890 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14893 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
14894 with nonlocal goto receivers or returns twice calls, ignore
14895 unininitialized values from abnormal edges to nl goto receiver
14896 or returns twice call.
14898 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14900 PR tree-optimization/57051
14901 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
14902 and VEC_RSHIFT_EXPR if shift count is a multiple of element
14905 2013-04-26 Richard Biener <rguenther@suse.de>
14907 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
14908 (expand_omp_taskreg): Likewise. Mark loops for fixup.
14909 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
14910 (fixup_loop_arrays_after_move): New function.
14911 (move_sese_region_to_fn): Properly outline the loop tree parts
14912 of the SESE region.
14914 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
14916 * config/i386/i386.md (type, unit): Fix long lines.
14918 2013-04-26 Richard Biener <rguenther@suse.de>
14920 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
14921 (lto-streamer-out.o): Likewise.
14922 * cfgloop.c (init_loops_structure): Export, add struct function
14923 argument and adjust.
14924 (flow_loops_find): Adjust.
14925 * cfgloop.h (enum loop_estimation): Add EST_LAST.
14926 (init_loops_structure): Declare.
14927 * lto-streamer-in.c: Include cfgloop.h.
14928 (input_cfg): Input the loop tree.
14929 * lto-streamer-out.c: Include cfgloop.h.
14930 (output_cfg): Output the loop tree.
14931 (output_struct_function_base): Do not drop PROP_loops.
14933 2013-03-26 Richard Biener <rguenther@suse.de>
14935 * tree-cfg.c (execute_build_cfg): Build the loop tree.
14936 (pass_build_cfg): Provide PROP_loops.
14937 (move_sese_region_to_fn): Remove loops that are outlined into fn
14939 * tree-inline.c: Include cfgloop.h.
14940 (initialize_cfun): Do not drop PROP_loops.
14941 (copy_loops): New function.
14942 (copy_cfg_body): Copy loop structure.
14943 (tree_function_versioning): Initialize destination loop tree.
14944 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
14945 (pass_parallelize_loops): Do IL verification.
14946 * loop-init.c (loop_optimizer_init): Fixup loops if required.
14947 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
14948 the CFG make sure we fixup loops as well.
14949 * tree-ssa-tail-merge.c: Include cfgloop.h.
14950 (replace_block_by): When merging loop latches mark loops for fixup.
14951 * lto-streamer-out.c (output_struct_function_base): Drop
14952 PROP_loops for now.
14953 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
14954 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
14955 * ipa-split.c: Include cfgloop.h.
14956 (split_function): Add the new return block to the loop tree root.
14957 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
14958 whether we have removed the forwarder block.
14959 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
14960 * cfgloop.h (place_new_loop): Declare.
14961 * cfgloopmanip.c (place_new_loop): Export.
14962 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
14963 (tree-switch-conversion.o): Likewise.
14964 (tree-complex.o): Likewise.
14965 (tree-inline.o): Likewise.
14966 (tree-ssa-tailmerge.o): Likewise.
14967 (ipa-split.o): Likewise.
14968 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
14969 (tree-ssa-copy.o): Likewise.
14970 * tree-switch-conversion.c: Include cfgloop.h
14971 (process_switch): If we emit a bit-test cascade, schedule loops
14973 * tree-complex.c: Include cfgloop.h.
14974 (expand_complex_div_wide): Properly add new basic-blocks to loops.
14975 * asan.c: Include cfgloop.h.
14976 (create_cond_insert_point): Properly add new basic-blocks to
14977 loops, schedule loop fixup.
14978 * cfgloop.c (verify_loop_structure): Check that looks are not
14980 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
14982 (expand_omp_for_generic): Likewise.
14983 (expand_omp_sections): Likewise.
14984 (expand_omp_atomic_pipeline): Schedule loops for fixup.
14985 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
14986 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
14987 is initialized, not when loops are present.
14988 * tree-parloops.c (parallelize_loops): Remove checking here.
14989 * passes.c (init_optimization_passes): Schedule a copy-propagation
14990 pass before complete unrolling of inner loops.
14992 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14994 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
14995 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
14996 (colorize_init): Add argument to _WIN32 version.
14997 * toplev.c: Include diagnostic-color.h.
14998 (process_options): Default to -fdiagnostics-color=auto if
14999 GCC_COLORS env var is in the environment.
15000 * common.opt (fdiagnostics-color=): Add Var and Init.
15001 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15002 env var is in the environment, the default is auto rather than never.
15004 * diagnostic.h (file_name_as_prefix): Add context argument.
15005 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15006 the string as locus.
15007 * langhooks.c (lhd_print_error_function): Adjust caller.
15009 2013-04-25 Lawrence Crowl <crowl@google.com>
15011 * var-tracking.c (shared_hash_def::htab):
15012 Change type to hash_table. Update dependent calls and types.
15014 2013-04-25 Lawrence Crowl <crowl@google.com>
15016 * Makefile.in: Update as needed below.
15018 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15019 Move declaration to after the type's method definitons.
15021 * attribs.c (htab_t scoped_attributes::attribute_hash):
15022 Change type to hash_table. Update dependent calls and types.
15024 * bitmap.c (htab_t bitmap_desc_hash):
15025 Change type to hash_table. Update dependent calls and types.
15027 * cselib.c (htab_t cselib_hash_table):
15028 Change type to hash_table. Update dependent calls and types.
15030 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15031 (hash_string_slot_node): Move implementation into lto-streamer.h
15032 struct string_slot_hasher.
15033 (eq_string_slot_node): Likewise.
15035 * data-streamer-out.c: Update output_block::string_hash_table
15036 dependent calls and types.
15038 * dwarf2cfi.c (htab_t trace_index):
15039 Change type to hash_table. Update dependent calls and types.
15041 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15042 Change type to hash_table. Update dependent calls and types.
15043 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15044 (htab_t optimize_external_refs::map): Likewise.
15045 (htab_t output_comp_unit::extern_map): Likewise.
15046 (htab_t output_comdat_type_unit::extern_map): Likewise.
15047 (htab_t output_macinfo::macinfo_htab): Likewise.
15048 (htab_t optimize_location_lists::htab): Likewise.
15049 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15051 * except.c (htab_t ehspec_hash_type):
15052 Change type to hash_table. Update dependent calls and types.
15053 (assign_filter_values::ttypes): Likewise.
15054 (assign_filter_values::ehspec): Likewise.
15055 (sjlj_assign_call_site_values::ar_hash): Likewise.
15056 (convert_to_eh_region_ranges::ar_hash): Likewise.
15058 * gcse.c (htab_t pre_ldst_table):
15059 Change type to hash_table. Update dependent calls and types.
15061 * ggc-common.c (htab_t saving_htab):
15062 Change type to hash_table. Update dependent calls and types.
15063 (htab_t loc_hash): Likewise.
15064 (htab_t ptr_hash): Likewise.
15065 (call_count): Rename ggc_call_count.
15066 (call_alloc): Rename ggc_call_alloc.
15067 (loc_descriptor): Rename make_loc_descriptor.
15068 (add_statistics): Rename ggc_add_statistics.
15070 * ggc-common.c (saving_htab):
15071 Change type to hash_table. Update dependent calls and types.
15073 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15074 (push_gimplify_context): Likewise.
15075 (pop_gimplify_context): Likewise.
15076 (struct gimple_temp_hash_elt): Added.
15077 (struct gimplify_hasher): Likewise.
15078 (struct gimplify_ctx.temp_htab):
15079 Change type to hash_table. Update dependent calls and types.
15081 * gimple-fold.c: Include gimplify-ctx.h.
15083 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15084 Change type to hash_table. Update dependent calls and types.
15085 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15086 avoid potential global name collision.
15088 * gimplify.c: Include gimplify-ctx.h.
15089 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15090 (htab_t gimplify_ctx::temp_htab):
15091 Update dependent calls and types for new type hash_table.
15092 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15093 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15095 * gimplify-ctx.h: New.
15096 (struct gimple_temp_hash_elt): Move from gimplify.c.
15097 (class gimplify_hasher): New.
15098 (struct gimplify_ctx): Move from gimple.h.
15099 (htab_t gimplify_ctx::temp_htab):
15100 Change type to hash_table. Update dependent calls and types.
15102 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15103 (htab_t ivs_params::newivs_index):
15104 Change type to hash_table. Update dependent calls and types.
15105 (htab_t ivs_params::params_index): Likewise.
15106 (htab_t print_generated_program::params_index): Likewise.
15107 (htab_t gloog::newivs_index): Likewise.
15108 (htab_t gloog::params_index): Likewise.
15110 * graphite.c: Include graphite-htab.h.
15111 4htab_t graphite_transform_loops::bb_pbb_mapping):
15112 Change type to hash_table. Update dependent calls and types.
15114 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15115 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15116 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15118 * graphite-dependences.c: Include graphite-htab.h.
15119 (loop_is_parallel_p): Change hash table type of parameter.
15121 * graphite-htab.h: New.
15122 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15123 (extern find_pbb_via_hash): Move from graphite-poly.h.
15124 (extern loop_is_parallel_p): Move from graphite-poly.h.
15125 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15127 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15128 (extern loop_is_parallel_p): Move to graphite-htab.h.
15129 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15131 * haifa-sched.c (htab_t delay_htab):
15132 Change type to hash_table. Update dependent calls and types.
15133 (htab_t delay_htab_i2): Likewise.
15135 * ira-color.c (htab_t allocno_hard_regs_htab):
15136 Change type to hash_table. Update dependent calls and types.
15138 * ira-costs.c (htab_t cost_classes_htab):
15139 Change type to hash_table. Update dependent calls and types.
15141 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15142 Change type to hash_table. Update dependent calls and types.
15144 * loop-iv.c (htab_t bivs):
15145 Change type to hash_table. Update dependent calls and types.
15147 * loop-unroll.c (htab_t opt_info::insns_to_split):
15148 Change type to hash_table. Update dependent calls and types.
15149 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15151 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15152 (struct string_slot_hasher): New.
15153 (htab_t output_block::string_hash_table):
15154 Change type to hash_table. Update dependent calls and types.
15156 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15157 (htab_t file_name_hash_table):
15158 Change type to hash_table. Update dependent calls and types.
15160 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15163 * lto-streamer.c (htab_t tree_htab):
15164 Change type to hash_table. Update dependent calls and types.
15166 * omp-low.c: Include gimplify-ctx.h.
15168 * passes.c (htab_t name_to_pass_map):
15169 Change type to hash_table. Update dependent calls and types.
15170 (pass_traverse): Rename to passes_pass_traverse.
15172 * plugin.c (htab_t event_tab):
15173 Change type to hash_table. Update dependent calls and types.
15175 * postreload-gcse.c (htab_t expr_table):
15176 Change type to hash_table. Update dependent calls and types.
15177 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15179 * sese.c (debug_rename_map_1): Make extern.
15180 (htab_t copy_bb_and_scalar_dependences::rename_map):
15181 Change type to hash_table. Update dependent calls and types.
15183 * sese.h (extern debug_rename_map): Move to .c file.
15185 * store-motion.c (htab_t store_motion_mems_table):
15186 Change type to hash_table. Update dependent calls and types.
15188 * trans-mem.c (htab_t tm_new_mem_hash):
15189 Change type to hash_table. Update dependent calls and types.
15191 * tree-browser.c (htab_t TB_up_ht):
15192 Change type to hash_table. Update dependent calls and types.
15194 * tree-cfg.c (htab_t discriminator_per_locus):
15195 Change type to hash_table. Update dependent calls and types.
15197 * tree-complex.c: Include tree-hasher.h
15198 (htab_t complex_variable_components):
15199 Change type to hash_table. Update dependent calls and types.
15201 * tree-eh.c (htab_t finally_tree):
15202 Change type to hash_table. Update dependent calls and types.
15204 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15205 struct int_tree_hasher.
15206 (extern int_tree_map_eq): Likewise.
15207 (uid_decl_map_hash): Removed.
15208 (extern decl_tree_map_eq): Likewise.
15210 * tree-hasher.h: New.
15211 (struct int_tree_hasher): New.
15212 (typedef int_tree_htab_type): New.
15214 * tree-inline.c: Include gimplify-ctx.h.
15216 * tree-mudflap.c: Include gimplify-ctx.h.
15218 * tree-parloops.c: Include tree-hasher.h.
15219 (htab_t eliminate_local_variables_stmt::decl_address):
15220 Change type to hash_table. Update dependent calls and types.
15221 (htab_t separate_decls_in_region::decl_copies): Likewise.
15223 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15224 Change type to hash_table. Update dependent calls and types.
15226 * tree-sra.c (candidates):
15227 Change type to hash_table. Update dependent calls and types.
15229 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15231 (int_tree_map_hash): Likewise.
15233 * tree-ssa-dom.c (htab_t avail_exprs):
15234 Change type to hash_table. Update dependent calls and types.
15236 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15237 Change type to hash_table. Update dependent calls and types.
15239 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15240 Change type to hash_table. Update dependent calls and types.
15242 * tree-ssa-phiopt.c (seen_ssa_names):
15243 Change type to hash_table. Update dependent calls and types.
15245 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15246 Change type to hash_table. Update dependent calls and types.
15248 * tree-ssa-uncprop.c (equiv):
15249 Change type to hash_table. Update dependent calls and types.
15251 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15253 PR rtl-optimization/57003
15254 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15255 call note_stores with kill_clobbered_value callback again after
15256 killing regs_invalidated_by_call.
15258 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15260 * config/aarch64/aarch64-simd.md
15261 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15262 (aarch64_simd_bsl<mode>): Likewise.
15263 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15265 2013-04-25 Marek Polacek <polacek@redhat.com>
15267 PR tree-optimization/57066
15268 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15270 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15272 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15274 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15276 * config/aarch64/aarch64-builtins.c
15277 (aarch64_fold_builtin): New.
15278 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15279 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15280 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15281 * config/aarch64/arm_neon.h
15282 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15284 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15285 Tejas Belagod <tejas.belagod@arm.com>
15287 * config/aarch64/aarch64-builtins.c
15288 (aarch64_gimple_fold_builtin): New.
15289 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15290 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15291 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15292 (addvv4sf): Update.
15293 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15295 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15297 * config/aarch64/aarch64.md
15298 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15300 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15302 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15303 (*ngcsi_uxtw): New pattern.
15305 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15306 Julian Brown <julian@codesourcery.com>
15308 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15309 (TB_DREG): Add T_V4HF.
15310 (v4hf_UP): New macro.
15311 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15312 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15313 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15315 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15316 (arm_vector_mode_supported_p): Handle V4HF.
15317 (arm_mangle_map): Handle V4HFmode.
15318 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15319 * config/arm/arm_neon_builtins.def: Add entries for
15320 vcvtv4hfv4sf, vcvtv4sfv4hf.
15321 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15322 (neon_vcvtv4hfv4sf): Likewise.
15323 * config/arm/neon-gen.ml: Handle half-precision floating point
15325 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15326 * config/arm/arm_neon.h: Regenerate.
15327 * config/arm/neon.ml (type elts): Add F16.
15328 (type vectype): Add T_float16x4, T_floatHF.
15329 (type vecmode): Add V4HF.
15330 (type features): Add Requires_FP_bit feature.
15331 (elt_width): Handle F16.
15332 (elt_class): Likewise.
15333 (elt_of_class_width): Likewise.
15334 (mode_of_elt): Refactor.
15335 (type_for_elt): Handle F16, fix error messages.
15336 (vectype_size): Handle T_float16x4.
15337 (vcvt_sh): New function.
15338 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15339 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15340 (string_of_mode): Handle V4HF.
15341 * doc/arm-neon-intrinsics.texi: Regenerate.
15343 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15345 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15346 format specifier in 'X' case.
15348 2013-04-25 Alan Modra <amodra@gmail.com>
15351 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15352 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15353 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15354 Repeat for many other rotate/shift and mask patterns using subregs.
15355 Name lshiftrt insns.
15356 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15357 on WORDS_BIG_ENDIAN.
15359 2013-04-25 Alan Modra <amodra@gmail.com>
15361 * config.gcc: Support little-endian powerpc-linux targets.
15362 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15363 (LINK_OS_LINUX_SPEC): Define.
15364 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15365 Preserve MASK_LITTLE_ENDIAN.
15366 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15367 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15368 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15369 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15370 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15371 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15372 Correct fp word order for little-endian. Don't shift toc entries
15373 smaller than a word for little-endian.
15374 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15375 (bswapdi2 splits): Correct low-part subreg for little-endian.
15376 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15377 low/high where such is correct only for be.
15378 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15379 little-endian for -mcall-aixdesc.
15381 2013-04-25 Alan Modra <amodra@gmail.com>
15383 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15384 replace_equiv_address_nv.
15386 2013-04-25 Alan Modra <amodra@gmail.com>
15388 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15390 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15393 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15394 * rtl.h (struct rtx_def): ...
15396 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15398 PR rtl-optimizations/57046
15399 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15400 for multi-reg splits.
15402 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15404 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15406 2013-04-24 Sterling Augustine <saugustine@google.com>
15408 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15409 (comp_dir_string, debug_str_dwo_section): New.
15410 (DEBUG_STR_DWO_SECTION): Rename to ...
15411 (DEBUG_DWO_STR_SECTION): ... this.
15412 (DEBUG_NORM_STR_SECTION): Delete.
15413 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15414 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15415 (find_AT_string): Move most logic to ...
15416 (find_AT_string_in_table): ... here. New.
15417 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15418 add_skeleton_AT_string. Delete logic.
15419 (output_skeleton_debug_sections): Remove call to
15420 add_top_level_skeleton_die_attrs.
15421 (add_comp_dir_attribute): Move logic to comp_dir_string.
15422 (dwarf2out_init): Initialize debug_str_dwo_section.
15423 (output_indirect_string): Call find_string_form.
15424 (output_indirect_strings): Rewrite.
15425 (prune_unused_types): Empty skeleton_debug_str_hash.
15426 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15427 (dwarf2out_finish): Call output_indirect_strings.
15429 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15431 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15433 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15435 * rtl.h (struct rtx_def): Add comment for field jump.
15436 (LRA_SUBREG_P): New macro.
15437 * recog.c (register_operand): Check LRA_SUBREG_P.
15438 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15440 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15441 (lra_final_code_change): Skip subreg change for operators.
15442 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15443 if there are no operand changes.
15444 * lra-constraints.c (curr_insn_set): New.
15445 (match_reload): Set LRA_SUBREG_P.
15446 (emit_spill_move): Ditto.
15447 (check_and_process_move): Use curr_insn_set. Process only single
15448 set insns. Don't initialize sec_mem_p and change_p.
15449 (simplify_operand_subreg): Use LRA_SUBREG_P.
15450 (reg_in_class_p): New function.
15451 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15452 of #ifdef. Add code to remove cycling.
15453 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15454 non-null disp. Reload inner instead of disp when base and index
15455 are null. Try to put lo_sum into register.
15456 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15457 (check_and_process_move): Move code for move cost check to
15458 simple_move_p. Remove equiv_substitution.
15459 (simple_move_p): New function.
15460 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15461 curr_insn_set. Call check_and_process_move only for single set
15462 insns. Use the new function. Move call of check_and_process_move
15463 after operand equiv substitution and address process.
15465 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15467 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15468 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15469 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15471 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15473 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15475 2013-04-24 Marek Polacek <polacek@redhat.com>
15477 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15478 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15479 (select_loops_exit_conditions): Likewise.
15480 (number_of_iterations_for_all_loops): Likewise.
15481 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15482 (scev_analysis): Likewise.
15484 2013-04-02 Catherine Moore <clm@codesourcery.com>
15485 Chao-ying Fu <fu@mips.com>
15487 * config/mips/micromips.md (jraddiusp): New pattern.
15488 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15489 instruction if possible.
15491 2013-04-24 Alan Modra <amodra@gmail.com>
15493 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15495 2013-04-24 Julian Brown <julian@codesourcery.com>
15496 Chung-Lin Tang <cltang@codesourcery.com>
15498 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15499 dependency behavior in enumeration type DIE generation. Add TODO note
15500 to comments about future DW_FORM_sdata/udata re-work of related code.
15502 2013-04-23 Lawrence Crowl <crowl@google.com>
15504 * Makefile.in: Update as needed below.
15506 * hash-table.h (class hash_table):
15507 Correct many methods with parameter types compare_type to the correct
15508 value_type. (Correct code was unlikely to notice the change.)
15509 (hash_table::elements_with_deleted) New.
15510 (class hashtable::iterator): New.
15511 (hashtable::begin()): New.
15512 (hashtable::end()): New.
15513 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15515 * statistics.c (statistics_hashes):
15516 Change type to hash_table. Update dependent calls and types.
15518 * tree-into-ssa.c (var_infos):
15519 Change type to hash_table. Update dependent calls and types.
15521 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15522 Change type to hash_table. Update dependent calls and types.
15524 * tree-ssa-loop-im.c (struct mem_ref.refs):
15525 Change type to hash_table. Update dependent calls and types.
15527 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15528 Change type to hash_table. Update dependent calls and types.
15530 * tree-ssa-sccvn.c (vn_tables_s::nary):
15531 Change type to hash_table. Update dependent calls and types.
15532 (vn_tables_s::phis): Likewise.
15533 (vn_tables_s::references): Likewise.
15535 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15536 (vn_reference_eq): Update parameter and return types.
15538 * tree-ssa-structalias.c (pointer_equiv_class_table):
15539 Change type to hash_table. Update dependent calls and types.
15540 (location_equiv_class_table): Likewise.
15542 * tree-vect-data-refs.c: Consequential changes for making
15543 peeling a hash_table.
15545 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15546 (destroy_loop_vec_info): Dependent hash_table update.
15548 * tree-vectorizer.h (peeling_htab):
15549 Change type to hash_table. Update dependent calls and types.
15551 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15553 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15554 to check the register content is equal or not.
15555 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15556 to assign register content record.
15557 * lra-eliminations.c (update_reg_eliminate): Use
15558 lra_update_reg_val_offset to update register content offset.
15559 * lra-int.h (struct lra_reg): Add offset member.
15560 (lra_reg_val_equal_p): New static inline function.
15561 (lra_update_reg_val_offset): New static inline function.
15562 (lra_assign_reg_val): New static inline function.
15563 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15564 to assign register content record.
15565 (initialize_lra_reg_info_element): Initial offset to zero.
15567 2013-04-23 Catherine Moore <clm@codesourcery.com>
15569 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15570 operands. Record compression.
15572 2013-04-23 Xinliang David Li <davidxl@google.com>
15574 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15576 2013-04-23 Richard Biener <rguenther@suse.de>
15578 PR middle-end/57036
15579 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15580 parameter, only add abnormal goto edges from the copied body
15581 if the call could perform abnormal gotos.
15582 (copy_cfg_body): Adjust.
15584 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15586 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15588 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15590 * coretypes.h (gimple_stmt_iterator): Add struct to make
15593 2013-04-23 Richard Biener <rguenther@suse.de>
15595 PR tree-optimization/57026
15596 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15597 from SSA names occuring in abnormal PHI nodes.
15599 2013-04-22 Andi Kleen <ak@linux.intel.com>
15601 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15603 2013-04-22 Andi Kleen <ak@linux.intel.com>
15605 * lto/lto.c (print_lto_report_1): Declare early.
15606 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15608 2013-04-22 Andi Kleen <ak@linux.intel.com>
15610 * common.opt (-flto-report-wpa): Add.
15611 * doc/invoke.texi (-flto-report-wpa): Add.
15612 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15615 2013-04-22 Xinliang David Li <davidxl@google.com>
15617 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15618 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15619 * Makefile.in: New dependency
15621 David Daney <ddaney.cavm@gmail.com>
15623 * configure.ac (gcc_cv_as_micromips_support): Use the
15624 --fatal-warnings option.
15625 * configure: Regenerate.
15627 2013-04-22 Marek Polacek <polacek@redhat.com>
15630 * tsan.c (instrument_expr): Don't instrument expression
15631 in case its size is zero.
15633 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15637 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15639 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15641 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15643 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15644 (gimple_stmt_iterator): New typedef.
15645 * gimple.h (gimple_stmt_iterator): Rename to...
15646 (gimple_stmt_iterator_d): ... This.
15647 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15648 trees be valid for GIMPLE and GENERIC.
15649 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15650 * gimple-fold.c (gimple_fold_call): Call target hook
15651 gimple_fold_builtin.
15652 * hooks.c (hook_bool_gsiptr_false): New.
15653 * hooks.h (hook_bool_gsiptr_false): New.
15654 * target.def (fold_stmt): New.
15655 * doc/tm.texi: Regenerate.
15657 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15660 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15661 a set sp if no stack realignment.
15663 2013-04-22 Nick Clifton <nickc@redhat.com>
15665 * config.gcc (tilegx-linux): Extend extra_objs rather than
15667 (tilepro-linux): Likewise.
15669 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15671 * config/aarch64/aarch64-builtins.c
15673 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15674 (VAR<1-12>): Add MAP parameter.
15675 (BUILTIN_*): Likewise.
15676 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15677 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15678 (aarch64_ushl_n<mode>): Likewise.
15679 (aarch64_sshr_n<mode>): Likewise.
15680 (aarch64_ushr_n<mode>): Likewise.
15681 (aarch64_<maxmin><mode>): Likewise.
15682 (aarch64_sqrt<mode>): Likewise.
15683 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15684 (vshr<q>_n_*): Likewise.
15686 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15688 * config/aarch64/aarch64-builtins.c
15689 (aarch64_simd_builtin_type_mode): Handle SF types.
15691 (BUILTIN_GPF): Define.
15692 (aarch64_init_simd_builtins): Handle SF types.
15693 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15694 (frecps): Likewise.
15695 (frecpx): Likewise.
15696 * config/aarch64/aarch64-simd.md
15697 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15698 (aarch64_frecpe<mode>): New.
15699 (aarch64_frecps<mode>): Likewise.
15700 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15701 (v8type): Add frecp<esx>.
15702 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15703 (aarch64_frecps<mode>): Likewise.
15704 * config/aarch64/iterators.md (FRECP): New.
15705 (frecp_suffix): Likewise.
15706 * config/aarch64/arm_neon.h
15707 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15709 2013-04-22 Christian Bruel <christian.bruel@st.com>
15712 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15713 (REG_CLASS_NAMES): Idem.
15714 (REG_CLASS_CONTENTS): Idem.
15715 (REGCLASS_HAS_FP_REG): Idem.
15716 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15717 (sh_conditional_register_usage): Idem.
15719 2013-04-21 Jeff Law <law@redhat.com>
15721 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15722 (ssa_forward_propagate_and_combine): Use it.
15724 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15726 * lra.c: Update the flow chart diagram.
15728 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15730 PR rtl-optimization/56847
15731 * lra-constraints.c (process_alt_operands): Discourage alternative
15732 with non-matche doffsettable memory constraint fro memory with
15735 2013-04-19 Richard Biener <rguenther@suse.de>
15737 PR tree-optimization/56982
15738 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15740 * gimplify.c (gimplify_call_expr): Notice special calls.
15741 (gimplify_modify_expr): Likewise.
15742 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15743 abnormal control flow receivers.
15744 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15745 in the same way as cfun->has_nonlocal_labels.
15746 (gimple_purge_dead_abnormal_call_edges): Likewise.
15747 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15748 receivers start a basic-block.
15750 2013-04-19 Richard Biener <rguenther@suse.de>
15752 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15754 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15755 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15756 (SLP_TREE_LOAD_PERMUTATION): Add.
15757 (vect_transform_slp_perm_load): Adjust prototype.
15758 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15759 (vect_free_slp_instance): Likewise.
15760 (vect_create_new_slp_node): Likewise.
15761 (vect_supported_slp_permutation_p): Remove.
15762 (vect_slp_rearrange_stmts): Adjust.
15763 (vect_supported_load_permutation_p): Likewise. Inline
15764 vect_supported_slp_permutation_p here.
15765 (vect_analyze_slp_instance): Compute load permutations per
15766 slp node instead of per instance.
15767 (vect_get_slp_defs): Adjust.
15768 (vect_transform_slp_perm_load): Likewise.
15769 (vect_schedule_slp_instance): Remove redundant code.
15770 (vect_schedule_slp): Remove hack for PR56270, add it ...
15771 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15772 CSE loads for SLP. Adjust.
15774 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15776 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15777 spelling in two comments.
15779 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15782 * config/arm/arm.c (load_multiple_sequence): Require SP
15783 as base register for loads if SP is in the register list.
15785 2013-04-19 Martin Jambor <mjambor@suse.cz>
15787 PR tree-optimization/56718
15788 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15789 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15790 and made public. Adjusted all callers.
15791 (ipa_intraprocedural_devirtualization): New function.
15792 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15793 (ipa_intraprocedural_devirtualization): Likewise.
15794 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15796 2013-04-19 Richard Biener <rguenther@suse.de>
15798 PR tree-optimization/57000
15799 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15801 2013-04-19 Terry Guo <terry.guo@arm.com>
15803 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15805 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15806 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15807 (cortex_m4_fmacs): Use new reservations.
15808 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15810 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15812 PR rtl-optimization/56999
15813 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15815 (lra_coalesce): Remove split_origin_bitmap and related code.
15816 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15817 ranges if necessary.
15819 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15821 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15823 (ix86_expand_call): Remove clobbered_registers array and use
15824 x86_64_ms_sysv_extra_clobbered_registers instead.
15825 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15827 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15829 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15830 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15831 (*call_value_rex64_ms_sysv): Ditto.
15833 2013-04-18 Cary Coutant <ccoutant@google.com>
15835 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15836 parent instead of die_mark.
15838 2013-04-18 Diego Novillo <dnovillo@google.com>
15840 * gimple.c (create_gimple_tmp): New.
15841 (get_expr_type): New.
15842 (build_assign): New.
15843 (build_type_cast): New.
15844 * gimple.h (enum ssa_mode): Define.
15845 (gimple_seq_set_location): New.
15846 * asan.c (build_check_stmt): Change some gimple_build_* calls
15847 to use build_assign and build_type_cast.
15849 2013-04-18 Richard Biener <rguenther@suse.de>
15851 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
15852 handle negative step. Remove redundant checks.
15853 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
15854 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
15855 for negative step and grouped loads fail to vectorize.
15857 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15859 * emit-rtl.c (reset_insn_used_flags): New function.
15860 (reset_all_used_flags): Use it.
15861 (verify_insn_sharing): New function.
15862 (verify_rtl_sharing): Fix verification for SEQUENCEs.
15864 2013-04-18 Jakub Jelinek <jakub@redhat.com>
15866 PR tree-optimization/56984
15867 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
15868 and (x >> M) >= N don't register any assertion if N << M is the
15871 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15873 * lower-subreg.c (resolve_simple_move): If called self-recursive,
15874 do not delete_insn insns that have not yet been emitted, only
15875 unlink them with remove_insn.
15876 * df-scan.c (df_insn_delete): Revert r197492.
15878 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
15880 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
15881 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
15883 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15885 * config/arm/arm.md (movsicc_insn): Convert define_insn into
15886 define_insn_and_split.
15887 (and_scc,ior_scc,negscc): Likewise.
15888 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
15890 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15892 * config/arm/arm.c (use_return_insn): Return 0 for targets that
15893 can benefit from using a sequence of LDRD instructions in epilogue
15894 instead of a single LDM instruction.
15896 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
15899 * doc/extend.texi: Fix typo.
15901 2013-04-17 Richard Biener <rguenther@suse.de>
15903 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
15904 (vect_build_slp_tree): ... here.
15905 (vect_build_slp_tree_1): Compute which stmts of the SLP group
15906 match. Remove special-casing of mismatched complex loads.
15907 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
15908 re-try the match with swapped commutative operands.
15909 (vect_supported_load_permutation_p): Remove special-casing of
15910 mismatched complex loads.
15911 (vect_analyze_slp_instance): Adjust.
15913 2013-04-17 Richard Biener <rguenther@suse.de>
15915 PR rtl-optimization/56921
15916 * cfgloop.h (struct loop): Add simple_loop_desc member.
15917 (struct niter_desc): Mark with GTY(()).
15918 (simple_loop_desc): Do not use aux field but simple_loop_desc.
15919 * loop-iv.c (get_simple_loop_desc): Likewise.
15920 (free_simple_loop_desc): Likewise.
15923 2013-04-16 Richard Biener <rguenther@suse.de>
15925 PR rtl-optimization/56921
15926 * loop-init.c (pass_rtl_move_loop_invariants): Add
15927 TODO_do_not_ggc_collect to todo_flags_finish.
15928 (pass_rtl_unswitch): Same.
15929 (pass_rtl_unroll_and_peel_loops): Same.
15930 (pass_rtl_doloop): Same.
15932 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
15934 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
15935 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
15936 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
15938 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
15939 * tree-streamer.c (record_common_node): Adjust reference in comment.
15941 2013-04-17 Terry Guo <terry.guo@arm.com>
15943 * config/arm/cortex-m4.md: Add a new bypass.
15945 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15947 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
15949 (*subs_<optab><mode>_multp2): New pattern.
15950 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15951 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15953 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15955 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
15956 (*subs_mul_imm_<mode>): New pattern.
15958 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
15961 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
15962 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
15963 (vsx_movti_32bit): Change j->wa to O->wa.
15965 2013-04-16 Richard Biener <rguenther@suse.de>
15967 PR rtl-optimization/56921
15968 * loop-init.c (pass_rtl_move_loop_invariants): Add
15969 TODO_do_not_ggc_collect to todo_flags_finish.
15970 (pass_rtl_unswitch): Same.
15971 (pass_rtl_unroll_and_peel_loops): Same.
15972 (pass_rtl_doloop): Same.
15974 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
15976 * config/arm/arm.c (emit_multi_reg_push): New declaration
15977 for an existing function.
15978 (arm_emit_strd_push): New function.
15979 (arm_expand_prologue): Used here.
15980 (arm_emit_ldrd_pop): New function.
15981 (arm_expand_epilogue): Used here.
15982 (arm_get_frame_offsets): Update condition.
15983 (arm_emit_multi_reg_pop): Add a special case for load of a single
15984 register with writeback.
15986 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
15988 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
15991 2013-04-16 Richard Biener <rguenther@suse.de>
15993 PR tree-optimization/56756
15994 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
15995 (first_mem_ref_loc): New.
15996 (execute_sm): Place the load temporarily before a previous
15997 access instead of in the latch edge to ensure its SSA dependencies
15998 are defined at points dominating the load.
16000 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16002 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16003 correct fix by moving header and footer insn to the footer of
16004 the merged basic block. Clear BB_END of the merged-away block.
16006 PR middle-end/43631
16007 * emit-rtl.c (make_note_raw): New function.
16008 (link_insn_into_chain): New static inline function.
16009 (add_insn): Use it.
16010 (add_insn_before, add_insn_after): Factor insn chain linking code...
16011 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16012 using link_insn_into_chain.
16013 (note_outside_basic_block_p): New helper function for emit_note_after
16014 and emit_note_before.
16015 (emit_note_after): Use nobb variant of add_insn_after if the note
16016 should not be contained in a basic block.
16017 (emit_note_before): Use nobb variant of add_insn_before if the note
16018 should not be contained in a basic block.
16019 (emit_note_copy): Use make_note_raw.
16020 (emit_note): Likewise.
16021 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16022 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16023 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16024 the moved barrier the tail of the basic block it follows.
16025 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16027 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16029 PR tree-optimization/56962
16030 * gimple-ssa-strength-reduction.c (record_increment): Only set
16031 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16032 either rhs1 or rhs2 is equal to c->base_expr.
16034 2013-04-15 Richard Biener <rguenther@suse.de>
16036 PR tree-optimization/56933
16037 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16039 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16040 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16041 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16042 dependence check ...
16043 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16045 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16046 GROUP_READ_WRITE_DEPENDENCE.
16048 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16050 * emit-rtl.c (reset_all_used_flags): New function.
16051 (verify_rtl_sharing): Call reset_all_used_flags before and after
16052 performing the checks.
16054 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16056 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16057 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16058 * config/arm/constraints.md (De): New constraint.
16059 * config/arm/neon.md (anddi3_neon): Delete.
16060 (neon_vand<mode>): Expand to standard anddi3 pattern.
16061 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16062 Move earlier in the file.
16063 (neon_inv_logic_op2): Likewise.
16064 (arm_anddi_operand_neon): New predicate.
16066 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16068 * configure.ac (gcc_cv_ld_as_needed): Set
16069 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16070 Use -z ignore, -z record on *-*-solaris2*.
16071 (HAVE_LD_AS_NEEDED): Update comment.
16072 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16073 * configure: Regenerate.
16074 * config.in: Regenerate.
16075 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16076 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16077 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16078 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16079 equivalents. Fix markup.
16080 * doc/tm.texi: Regenerate.
16082 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16084 * config/i386/i386.opt: New option mstack-protector-guard=.
16085 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16086 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16087 TARGET_SSP_TLS_GUARD.
16088 * config/i386/i386.c (ix86_option_override_internal): Set
16089 ix86_stack_protector_guard.
16090 * config/i386/i386.md (stack_protect_set): Enable for
16091 TARGET_SSP_TLS_GUARD only.
16092 (stack_protect_set_<mode>): Ditto.
16093 (stack_protect_test): Ditto.
16094 (stack_protect_test_<mode>): Ditto.
16095 * doc/invoke.texi (i386 Option): Document.
16097 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16100 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16101 (S_MODES): Set H_MODE bit.
16102 (SF_MODES): Set only S_MODE and SF_MODE bits.
16103 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16104 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16105 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16106 <MODE_FLOAT>: Likewise.
16108 2013-04-15 Joey Ye <joey.ye@arm.com>
16110 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16112 2013-04-15 Joey Ye <joey.ye@arm.com>
16114 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16116 (thumb_far_jump_used_p): Count instruction size and set
16119 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16121 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16122 * resource.c (mark_target_live_regs): Reformat conditional block.
16124 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16126 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16127 notes, they are emitted only just before final.
16128 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16130 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16132 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16133 * cfgrtl.c (delete_insn): Call it here instead.
16134 * lra-spills.c (lra_final_code_change): Use delete_insn.
16135 * haifa-sched.c (sched_remove_insn): Likewise.
16136 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16137 returning to the nop pool.
16138 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16139 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16141 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16143 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16144 * doc/tm.texi: Regenerated.
16146 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16148 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16151 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16153 * df-core.c (df_find_def): Compare register numbers.
16154 (df_find_use): Likewise.
16156 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16159 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16160 lra_in_progress for return.
16162 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16164 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16165 define_insn into define_insn_and_split and emit movsicc patterns.
16167 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16169 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16171 2013-04-12 Richard Biener <rguenther@suse.de>
16173 * tree-pass.h (TODO_do_not_ggc_collect): New.
16174 * passes.c (execute_one_ipa_transform_pass): Honor
16175 TODO_do_not_ggc_collect.
16176 (execute_one_pass): Likewise.
16179 2013-04-10 Richard Biener <rguenther@suse.de>
16181 * passes.c (init_optimization_passes): Remove reload pass.
16182 * ira.c (do_reload): Merge into ...
16184 (rest_of_handle_reload): Remove.
16185 (pass_reload): Likewise.
16186 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16187 of reload for vzeroupper pass placement.
16189 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16191 PR tree-optimization/56918
16192 PR tree-optimization/56920
16193 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16194 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16195 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16196 use wide_mul_with_sign method.
16198 2013-04-12 Richard Biener <rguenther@suse.de>
16200 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16201 not be considered a gimple constant.
16203 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16205 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16206 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16209 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16210 Jakub Jelinek <jakub@redhat.com>
16212 * opts.c: Include diagnostic-color.h.
16213 (common_handle_option): Handle OPT_fdiagnostics_color_.
16214 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16215 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16216 (diagnostic-color.o): New.
16217 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16218 (diagnostic_color_rule): New enum.
16219 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16220 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16221 the location string.
16222 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16223 either NULL, or color kind.
16224 * diagnostic-color.c: New file.
16225 * diagnostic-color.h: New file.
16226 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16228 * doc/invoke.texi (-fdiagnostics-color): Document.
16229 * pretty-print.h (pp_show_color): Define.
16230 (struct pretty_print_info): Add show_color field.
16231 * diagnostic.c: Include diagnostic-color.h.
16232 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16233 macros. Colorize error:, warning: etc. strings and also the location
16235 (diagnostic_show_locus): Colorize the caret line.
16236 * pretty-print.c: Include diagnostic-color.h.
16237 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16238 inside of %< %> quotes or quoted through q format modifier.
16240 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16242 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16244 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16246 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16247 code in CC_NZ mode.
16248 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16251 2013-04-11 Marek Polacek <polacek@redhat.com>
16253 PR tree-optimization/48184
16254 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16256 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16258 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16259 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16260 (skip_simple_arithmetic): Tidy up.
16261 * tree.h (skip_simple_constant_arithmetic): Declare.
16263 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16265 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16267 2013-04-11 Richard Biener <rguenther@suse.de>
16269 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16270 generate vector constants.
16272 2013-04-11 Richard Biener <rguenther@suse.de>
16274 PR tree-optimization/56878
16275 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16276 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16278 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16279 Prefer to align the DR with the most invariant base address.
16281 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16283 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16286 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16288 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16289 floating-point vector comparisons against 0.
16291 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16293 PR tree-optimization/56899
16294 * fold-const.c (extract_muldiv_1): Apply distributive law
16295 only if TYPE_OVERFLOW_WRAPS (ctype).
16297 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16300 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16301 loading has side effect.
16303 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16305 * config/sparc/sparc.c: Include tree-pass.h.
16306 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16307 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16308 head of file. Change return type. Split off gate function.
16309 (sparc_gate_work_around_errata): New function.
16310 (pass_work_around_errata): New pass definition.
16311 (insert_pass_work_around_errata) New pass insert definition to
16312 insert pass_work_around_errata just after delayed-branch scheduling.
16313 (sparc_option_override): Insert the pass.
16314 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16316 2013-04-10 David S. Miller <davem@davemloft.net>
16318 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16319 or -mcpu=hypersparc.
16321 * target.def (cstore_mode): New hook.
16322 * target.h: Include insn-codes.h
16323 * targhooks.c: Likewise.
16324 (default_cstore_mode): New function.
16325 * targhooks.h: Declare it.
16326 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16327 * doc/tm.texi: Rebuild.
16328 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16329 target hook, rather than inspecting the insn_data.
16330 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16331 (TARGET_CSTORE_MODE): Redefine.
16332 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16334 * config/sparc/predicates.md (cstore_result_operand): New special
16336 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16337 Use it for operand 0.
16338 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16339 (*snesi_special): Likewise.
16340 (*snesi_zero): Likewise.
16341 (*seqsi_zero): Likewise.
16342 (*sltu_insn): Likewise.
16343 (*sgeu_insn): Likewise.
16344 (*seqdi_special): Make operand 0 and comparison operation be of
16346 (*snedi_special): Likewise.
16347 (*snedi_special_vis3): Likewise.
16348 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16349 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16350 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16351 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16352 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16353 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16354 (*sltu_extend_sp64): Likewise.
16355 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16356 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16357 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16358 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16359 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16361 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16363 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16364 (aarch64_start_file): Use the new function.
16366 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16367 Jason Merrill <jason@redhat.com>
16369 * common.opt: Add -gdwarf.
16370 * opts.c (common_handle_option): Handle it.
16371 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16373 2013-04-10 Richard Biener <rguenther@suse.de>
16375 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16376 (execute_one_ipa_transform_pass): But unconditionally here.
16377 (execute_one_pass): And here.
16378 (init_optimization_passes): Remove reload pass.
16379 * tree-pass.h (TODO_ggc_collect): Remove.
16380 (pass_reload): Likewise.
16381 * ira.c (do_reload): Merge into ...
16383 (rest_of_handle_reload): Remove.
16384 (pass_reload): Likewise.
16385 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16386 of reload for vzeroupper pass placement.
16387 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16388 and todo_flags_finish of all passes.
16390 2013-04-10 Richard Biener <rguenther@suse.de>
16392 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16393 first_const_oprnd field, rename first_def_type to first_op_type.
16394 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16395 (vect_get_and_check_slp_defs): Always use the type of the
16396 operand. Allow mixed vect_external_def, vect_constant_def types.
16397 (vect_get_constant_vectors): Handle mixed vect_external_def,
16398 vect_constant_def types.
16400 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16402 PR tree-optimization/55524
16403 * tree-ssa-math-opts.c
16404 (convert_mult_to_fma): Don't use an fms construct
16405 when we don't have an fms operation, but fnma, and it looks
16406 likely that we'll be able to use the latter.
16408 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16410 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16412 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16413 inline fail caused by overwritable functions.
16415 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16417 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16418 unnecessary bits in the constant power of two case.
16420 2013-04-10 Richard Biener <rguenther@suse.de>
16422 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16423 broken code swapping operands.
16424 (vect_build_slp_tree): Do not compute load permutations here.
16425 (vect_analyze_slp_instance): Compute load permutations here,
16426 after building the SLP tree.
16428 2013-04-09 Christian Bruel <christian.bruel@st.com>
16430 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16431 of next/prev_real_insn.
16433 2013-04-09 Jan Hubicka <jh@suse.cz>
16435 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16436 Drop aliased parameter.
16437 (function_and_variable_visibility): Do not handle alias pairs.
16438 * cgraph.c (varpool_externally_visible_p): Update prototype.
16439 * varpool.c (varpool_add_new_variable): Update.
16441 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16443 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16445 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16447 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16449 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16451 2013-04-09 Marek Polacek <polacek@redhat.com>
16453 PR tree-optimization/48762
16454 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16456 2013-04-09 Richard Biener <rguenther@suse.de>
16458 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16460 (vect_build_slp_tree): Likewise.
16461 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16462 calculating the cost of a SLP instance.
16463 (vect_analyze_slp_instance): Use it from here, after building
16466 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16468 PR middle-end/56883
16469 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16470 expand_omp_for_static_chunk): Use simple_p = true in
16471 force_gimple_operand_gsi calls when assigning to addressable decls.
16473 2013-04-09 Jeff Law <law@redhat.com>
16475 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16476 when the boolean was created by converting a wider object which
16477 had a boolean range.
16479 2013-04-09 Richard Biener <rguenther@suse.de>
16481 * tree-vectorizer.h (slp_void_p): Remove.
16482 (slp_tree): Typedef before _slp_tree declaration.
16483 (struct _slp_tree): Use a vector of slp_tree as children.
16484 (vect_get_place_in_interleaving_chain): Remove.
16485 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16487 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16489 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16490 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16491 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16492 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16493 Use slp_node instead of slp_void_p and adjust.
16495 2013-04-09 Richard Biener <rguenther@suse.de>
16497 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16498 work that is not necessary.
16500 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16502 PR tree-optimization/56854
16503 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16504 forward into clobber stmts if it would change MEM_REF lhs into
16507 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16509 * tree.c (type_hash_lookup, type_hash_add): Make static.
16510 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16512 2013-04-09 Richard Biener <rguenther@suse.de>
16514 * tree.h (unsave_expr_now): Remove.
16515 * tree-inline.c (mark_local_for_remap_r): Remove.
16516 (unsave_expr_1): Likewise.
16517 (unsave_r): Likewise.
16518 (unsave_expr_now): Likewise.
16519 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16520 (propagate_tree_value): Likewise.
16522 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16524 * doc/rtl.texi (sequence): Rewrite documentation to match the
16525 current use of SEQUENCE rtl objects.
16526 * rtl.def (SEQUENCE): Likewise.
16528 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16529 Update documentation.
16530 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16531 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16533 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16535 2013-04-08 Teresa Johnson <tejohnson@google.com>
16537 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16538 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16540 (estimate_edge_size_and_time): Add comment to suggest using rounding
16542 (estimate_node_size_and_time): Ditto.
16543 (remap_edge_change_prob): Use helper rounding divide methods.
16544 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16545 (gimple_mod_pow2_value_transform): Ditto.
16546 (gimple_mod_subtract_transform): Ditto.
16547 (gimple_ic_transform): Ditto.
16548 (gimple_stringops_transform): Ditto.
16549 * stmt.c (conditional_probability): Ditto.
16550 (emit_case_dispatch_table): Ditto.
16551 * lto-cgraph.c (merge_profile_summaries): Ditto.
16552 * tree-optimize.c (execute_fixup_cfg): Ditto.
16553 * cfgcleanup.c (try_forward_edges): Ditto.
16554 * cfgloopmanip.c (scale_loop_profile): Ditto.
16556 (duplicate_loop_to_header_edge): Ditto.
16557 (lv_adjust_loop_entry_edge): Ditto.
16558 * tree-vect-loop.c (vect_transform_loop): Ditto.
16559 * profile.c (compute_branch_probabilities): Ditto.
16560 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16561 * lto-streamer-in.c (input_cfg): Ditto.
16562 * gimple-streamer-in.c (input_bb): Ditto.
16563 * ipa-cp.c (update_profiling_info): Ditto.
16564 (update_specialized_profile): Ditto.
16565 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16566 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16568 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16569 (compute_trg_info): Ditto.
16570 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16571 (purge_dead_edges): Ditto.
16572 * loop-unswitch.c (unswitch_loop): Ditto.
16573 * cgraphclones.c (cgraph_clone_edge): Ditto.
16574 (cgraph_clone_node): Ditto.
16575 * tree-inline.c (copy_bb): Ditto.
16576 (copy_edges_for_bb): Ditto.
16577 (initialize_cfun): Ditto.
16578 (copy_cfg_body): Ditto.
16579 (expand_call_inline): Ditto.
16581 2013-04-08 Kai Tietz <ktietz@redhat.com>
16583 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16584 TARGET_CYGWIN64 by TARGET_64BIT.
16586 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16588 * config/epiphany/epiphany.md (GPR_1): New constant.
16589 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16590 * config/epiphany/epiphany.c (gen_compare_reg):
16591 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16592 is already in place.
16593 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16594 Don't require being called during rtl expansion; If y operlaps r0,
16596 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16597 (epiphany_expand_epilogue): Likewise.
16599 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16600 Don't use CC_FPmode for ORDERED / UNORDERED.
16601 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16603 * config/epiphany/constraints.md (CnL): New constraint.
16604 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16605 * config/epiphany/predicates.md (add_operand): Allow 1024.
16607 * config/epiphany/epiphany.md (logical_op): New code iterator.
16608 (op_mnc): New code attribute.
16609 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16610 (mov_f+1, mov_f+2): New peephole2 patterns.
16612 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16613 (cstoresi4): Also allow re-use of zero result when doing a NE
16614 comparison to a non-zero operand.
16615 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16617 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16618 Use gen_addsi3_i / gen_subsi3_i.
16620 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16624 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16625 contain anything but clobbers, at most one __builtin_stack_restore,
16626 optionally debug stmts and final resx, and if it has at least one
16627 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16628 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16629 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16630 which isn't defaut definition, remove them.
16631 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16632 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16633 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16634 with MEM_REF LHS with SSA_NAME address.
16636 2013-04-08 Jeff Law <law@redhat.com>
16638 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16640 2013-04-08 Richard Biener <rguenther@suse.de>
16642 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16644 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16645 determined vector type.
16646 (vect_analyze_data_refs): Likewise.
16647 (vect_get_new_vect_var): Adjust.
16648 (vect_create_destination_var): Preserve SSA name versions.
16649 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16650 not dump anything here.
16652 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16654 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16655 Add member lr_slot_known.
16656 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16658 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16659 Remove code that sets lr_slot_offset according to what a previous
16660 version of epiphany_emit_save_restore used to do.
16661 (epiphany_emit_save_restore): When doing an lr save or restore,
16662 set/verify lr_slot_known and lr_slot_offset.
16664 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16667 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16670 2013-04-08 Richard Biener <rguenther@suse.de>
16672 * alias.c (find_base_term): Fix thinko in previous change.
16674 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16676 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16677 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16678 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16679 if possible to compute val.
16680 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16681 For QImode integers don't require anything about precision. Use
16682 const_with_all_bytes_same to find out if the constant doesn't have
16683 repeated bytes in it.
16685 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16687 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16690 2013-04-08 Marek Polacek <polacek@redhat.com>
16692 PR rtl-optimization/48182
16693 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16696 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16699 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16700 nuses, make sure we have a label.
16702 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16705 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16706 (rs6000_emit_swdiv_low_precision): Remove.
16707 (rs6000_emit_swdiv): Rewrite to handle between one and four
16708 iterations of Newton-Raphson generally; modify required number of
16709 iterations for some cases.
16710 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16712 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16714 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16715 set-but-unused variable.
16717 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16718 basic blocks of released function bodies garbage-collectable.
16720 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16721 (struct rtl_opt_pass): Add TODO_df_finish.
16723 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16725 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16727 * config/arm/constraints.md (q): New constraint.
16728 * config/arm/ldrdstrd.md: New file.
16729 * config/arm/arm.md (ldrdstrd.md) New include.
16730 (arm_movdi): Use "q" instead of "r" constraint
16731 for double-word memory access.
16732 (movdf_soft_insn): Likewise.
16733 * config/arm/vfp.md (movdi_vfp): Likewise.
16734 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16735 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16736 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16737 (mem_ok_for_ldrd_strd): Likewise.
16738 (output_move_double): Update assertion.
16740 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16742 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16744 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16746 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16747 define_insn_and_split.
16748 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16750 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16752 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16753 define_insn_and_split.
16754 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16755 (shiftsi3_compare): New pattern.
16756 (rrx): New pattern.
16757 * config/arm/unspecs.md (UNSPEC_RRX): New.
16759 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16761 * config/arm/arm.md (negdi_extendsidi): New pattern.
16762 (negdi_zero_extendsidi): Likewise.
16764 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16766 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16767 define_insn_and_split.
16768 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16769 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16771 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16773 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16774 define_insn_and_split.
16775 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16776 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16778 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16780 * config/arm/arm.md (subsi3_carryin): New pattern.
16781 (subsi3_carryin_const): Likewise.
16782 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16783 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16785 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16787 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16789 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16791 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16792 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16794 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16796 * config/arm/arm.c (arm_expand_builtin): Change fcode
16797 type to unsigned int.
16799 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16801 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16803 2013-04-04 Ian Lance Taylor <iant@google.com>
16805 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16808 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16810 PR middle-end/56729
16811 * df-scan.c (df_insn_delete): Disable failing assert.
16813 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16815 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16816 New function prototype.
16817 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16818 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16819 (arm_builtin_vectorized_function): New function.
16821 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16823 * config/arm/arm_neon_builtins.def: New file.
16824 * config/arm/arm.c (neon_builtin_data): Move contents to
16825 arm_neon_builtins.def.
16826 (enum arm_builtins): Include neon builtin definitions.
16827 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16828 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16830 2013-04-04 Marek Polacek <polacek@redhat.com>
16832 PR tree-optimization/48186
16833 * predict.c (maybe_hot_frequency_p): Return false if
16834 HOT_BB_FREQUENCY_FRACTION is 0.
16835 (cgraph_maybe_hot_edge_p): Likewise.
16837 2013-04-04 Richard Biener <rguenther@suse.de>
16839 PR tree-optimization/56826
16840 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16843 2013-04-04 Richard Biener <rguenther@suse.de>
16845 PR tree-optimization/56213
16846 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16847 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16849 2013-04-04 Richard Biener <rguenther@suse.de>
16851 PR tree-optimization/56837
16852 * tree-loop-distribution.c (classify_partition): For non-zero
16853 values require that the value has the same precision as its
16854 mode to be useful as memset value.
16856 2013-04-03 Nick Clifton <nickc@redhat.com>
16858 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
16859 (fmssf4): Use fmsf.s on E3V5 architectures.
16860 (fnmasf4): Use fnmaf.s on E3V5 architectures.
16861 (fnmssf4): Use fnmsf.s on E3V5 architectures.
16863 2013-04-03 Jeff Law <law@redhat.com>
16865 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
16866 (lra-eliminations.o): Likewise.
16868 2013-04-03 Teresa Johnson <tejohnson@google.com>
16870 * gcov-io.c (compute_working_sets): Moved most of body of old
16871 compute_working_sets here from profile.c.
16872 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
16873 (gcov_working_set_t): Moved typedef here from basic-block.h
16874 (compute_working_set): Declare.
16875 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
16876 (get_working_sets): Renamed from compute_working_set,
16877 replace most of body with call to new compute_working_sets.
16878 (get_exec_counts): Replace call to compute_working_sets
16879 to get_working_sets.
16880 * profile.h (get_working_sets): Renamed from compute_working_set.
16881 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
16882 to get_working_sets.
16883 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
16884 * gcov-dump.c (dump_working_sets): New function.
16886 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
16888 * hwint.c (sext_hwi, zext_hwi): New functions.
16889 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
16890 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
16891 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
16892 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
16893 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
16894 (sext_hwi, zext_hwi): New functions.
16896 2013-04-03 Jeff Law <law@redhat.com>
16898 PR tree-optimization/56799
16899 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
16900 back test for widening conversion erroneously dropped in prior change.
16902 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16905 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
16906 instead of next_real_insn.
16908 2013-04-03 Marek Polacek <polacek@redhat.com>
16911 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
16913 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16916 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
16918 (thumb1_output_casesi): Likewise.
16919 (thumb2_output_casesi): Likewise.
16921 2013-04-03 Richard Biener <rguenther@suse.de>
16923 PR tree-optimization/56817
16924 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
16926 (tree_unroll_loops_completely_1): ... new function to manually
16927 walk the loop tree, properly defering outer loops of unrolled
16928 loops to later iterations.
16930 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16932 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
16933 (vectorizable_load): Likewise.
16934 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
16935 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
16937 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16939 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
16942 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16944 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
16946 2013-04-03 Bin Cheng <bin.cheng@arm.com>
16948 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
16950 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16952 PR tree-optimization/56790
16953 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
16956 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16958 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
16960 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
16961 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
16964 2013-04-03 Jakub Jelinek <jakub@redhat.com>
16967 * tree.h (force_folding_builtin_constant_p): New decl.
16968 * builtins.c (force_folding_builtin_constant_p): New variable.
16969 (fold_builtin_constant_p): Fold immediately also if
16970 force_folding_builtin_constant_p.
16972 2013-04-03 Richard Biener <rguenther@suse.de>
16974 PR tree-optimization/56812
16975 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
16976 DRs of the same interleaving chain are independent.
16978 2013-04-02 Jason Merrill <jason@redhat.com>
16980 * gdbinit.in (pbb): Use debug fn.
16982 2013-04-02 Lawrence Crowl <crowl@google.com>
16984 * sese.h (struct ivtype_map_elt_s): Remove unused.
16985 (extern debug_ivtype_map): Remove unused.
16986 (extern eq_ivtype_map_elts): Remove unused.
16987 * sese.c (debug_ivtype_map): Removed unused.
16988 (debug_ivtype_map_1): Removed unused.
16989 (debug_ivtype_elt): Remove unused.
16990 (eq_ivtype_map_elts): Remove unused.
16993 2013-04-02 Kai Tietz <ktietz@redhat.com>
16996 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
16997 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
16998 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17000 (legitimize_pe_coff_symbol): Likewise.
17001 (is_imported_p): New helper-function.
17002 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17003 for Windows x64 targets.
17004 (ix86_expand_prologue): Optimize for pe-coff targets.
17005 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17006 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17007 medium/large code-model.
17008 (legitimize_pic_address): Likewise.
17009 (legitimize_tls_address): Likewise.
17010 (ix86_expand_call): Likewise.
17011 (x86_output_mi_thunk): Likewise.
17012 (get_dllimport_decl): Add new beimport argument.
17013 (construct_plt_address): Don't assert for x64 pe-coff targets.
17014 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17016 (SYMBOL_FLAG_STUBVAR): New macro.
17017 (SYMBOL_REF_STUBVAR_P): Likewise.
17018 * config/i386/winnt.c (stub_list): New structure.
17019 (stub_head): New local variable.
17020 (i386_pe_record_stub): New function.
17021 (i386_pe_file_end): Emit refptr-stubs.
17023 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17025 PR rtl-optimization/56745
17026 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17027 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17030 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17031 and both of them are MEM_REFs, just compare first argument for
17032 equality and attempt to deal even with differing offsets.
17035 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17036 of gimple_clobber_p to be MEM_REF.
17037 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17038 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17039 after gimplification.
17040 * asan.c (get_mem_ref_of_assignment): Don't instrument
17041 gimple_clobber_p stmts.
17042 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17043 gimple_clobber_p stmt if they have MEM_REF lhs and
17044 are dead because of another gimple_clobber_p stmt.
17045 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17046 gimple_clobber_p stmts like debug stmts.
17047 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17048 that refer to unused VAR_DECLs or uninitialized values.
17049 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17050 gimple_clobber_p stmts if they refer to removed parameters.
17051 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17054 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17056 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17057 using SWI48 mode attribute.
17059 2013-04-02 Wei Mi <wmi@google.com>
17061 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17062 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17063 *<rotate_insn><mode>3_mask in i386.md.
17065 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17067 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17069 2013-04-02 Richard Biener <rguenther@suse.de>
17071 PR tree-optimization/56778
17072 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17073 Runtime alias tests are not supported for gather loads.
17074 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17075 stmts referenced from SSA operands before updating SSA form.
17077 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17078 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17080 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17081 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17082 * config/arm/cortex-a53.md: New file.
17083 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17084 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17085 * config/arm/arm.c (arm_issue_rate): Likewise.
17086 * config/arm/arm-tune.md: Regenerate
17087 * config/arm/arm-tables.opt: Regenerate.
17088 * config/arm/arm-cores.def: Add cortex-a53.
17090 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17092 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17095 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17097 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17098 scalar load/store operations using B/H registers.
17099 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17101 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17103 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17105 * config/aarch64/aarch64.c
17106 (aarch64_simd_scalar_immediate_valid_for_move): New.
17107 * config/aarch64/aarch64-protos.h
17108 (aarch64_simd_scalar_immediate_valid_for_move): New.
17109 * config/aarch64/constraints.md (Dh, Dq): New.
17110 * config/aarch64/iterators.md (hq): New.
17112 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17114 * reorg.c (get_branch_condition): Deal with conditional returns.
17115 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17117 2013-04-01 Wei Mi <wmi@google.com>
17119 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17120 Truncate operand 2 using %b asm operand modifier.
17121 (*<shift_insn><mode>3_mask): Ditto.
17122 (*<rotate_insn><mode>3_mask): Ditto.
17124 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17126 PR middle-end/56798
17127 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17129 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17131 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17133 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17135 2013-03-30 Lawrence Crowl <crowl@google.com>
17137 * dse.c (clear_alias_sets): Remove never set.
17138 (disqualified_clear_alias_sets): Remove never set.
17139 (clear_alias_mode_pool): Remove never set.
17140 (dse_step0): Remove condition that is never true.
17141 (canon_address): Remove condition that is never true.
17142 (dse_step7): Remove condition that is never true.
17143 (rest_of_handle_dse): Remove condition that is never true.
17144 (rest_of_handle_dse::did_global): Remove never read from above.
17145 (dse_step2_spill): Remove never called from above.
17146 (dse_step5_spill): Remove never called from above.
17148 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17150 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17151 JUMP_TABLE_DATA changes.
17152 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17153 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17154 (Insns) <jump_table_data>: New entry.
17155 * doc/tm.texi: Regenerate.
17157 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17159 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17160 for table jump at the end of a basic block using tablejump_p.
17161 * targhooks.c (default_invalid_within_doloop): Likewise.
17162 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17163 target hook implementation that is identical to the default hook.
17164 (rs6000_invalid_within_doloop): Remove.
17166 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17167 unused variable from tablejump_p call.
17169 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17170 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17171 (INSN_DELETED_P): Likewise.
17172 (emit_jump_table_data): New prototype.
17173 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17174 after 4th as unused.
17175 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17176 * sched-vis.c (print_insn): Likewise.
17177 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17178 insn for compatibility with back ends that use next_active_insn to
17179 identify jump table data.
17180 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17181 (remove_insn): Likewise.
17182 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17184 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17185 (emit_jump_table_data): New function.
17187 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17188 basic block, a JUMP_TABLE_DATA never is.
17189 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17190 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17191 off from code handling real insns.
17192 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17193 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17194 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17195 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17196 is not a NONDEBUG_INSN_P.
17197 * ira-costs.c (scan_one_insn): Likewise.
17198 * jump.c (mark_all_labels): Likewise.
17199 (mark_jump_label_1): Likewise.
17200 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17201 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17203 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17204 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17205 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17206 JUMP_TABLE_DATA_P insns.
17207 (calculate_elim_costs_all_insns): Likewise.
17208 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17209 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17210 (delete_output_reload): Code style fixups.
17211 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17212 insn flags on this non-insn.
17213 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17214 as scheduling barriers, for pre-change compatibility.
17215 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17216 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17218 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17219 redundant JUMP_TABLE_DATA_P test.
17220 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17221 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17222 (frv_for_each_packet): Likewise.
17223 * config/i386/i386.c (min_insn_size): Likewise.
17224 (ix86_avoid_jump_mispredicts): Likewise.
17225 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17226 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17227 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17228 (mips16_insn_length): Robustify.
17229 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17230 (mips16_split_long_branches): Likewise.
17231 * config/pa/pa.c (pa_combine_instructions): Likewise.
17232 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17233 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17234 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17235 as contributing to pool range lengths.
17236 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17237 Remove redundant JUMP_TABLE_DATA_P test.
17238 (sh_loop_align): Likewise.
17239 (split_branches): Likewise.
17240 (sh_insn_length_adjustment): Likewise.
17241 * config/spu/spu.c (get_branch_target): Likewise.
17243 2013-03-29 Jan Hubicka <jh@suse.cz>
17245 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17246 gcov streaming; stream hot bb threshold to ltrans.
17247 * predict.c (get_hot_bb_threshold): Break out from ....
17248 (maybe_hot_count_p): ... here.
17249 (set_hot_bb_threshold): New function.
17250 * lto-section-in.c (lto_section_name): Add profile.
17251 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17252 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17253 and data-streamer.h
17254 (histogram_entry): New structure.
17255 (histogram, histogram_pool): New global vars.
17256 (histogram_hash): New structure.
17257 (histogram_hash::hash): New method.
17258 (histogram_hash::equal): Likewise.
17259 (account_time_size): New function.
17260 (cmp_counts): New function.
17261 (dump_histogram): New function.
17262 (ipa_profile_generate_summary): New function.
17263 (ipa_profile_write_summary): New function.
17264 (ipa_profile_read_summary): New function.
17265 (ipa_profile): Decide on threshold.
17266 (pass_ipa_profile): Add ipa_profile_write_summary and
17267 ipa_profile_read_summary.
17268 * Makefile.in (ipa.o): Update dependencies.
17269 * lto-streamer.h (LTO_section_ipa_profile): New section.
17271 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17273 * tree.h (VAR_P): New.
17275 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17278 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17280 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17282 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17283 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17284 (control_flow_insn_p): Likewise.
17285 * cfgrtl.c (duplicate_insn_chain): Likewise.
17286 * final.c (get_attr_length_1): Likewise.
17287 (shorten_branches): Likewise.
17288 (final_scan_insn): Likewise.
17289 * function.c (instantiate_virtual_regs): Likewise.
17290 * gcse.c (insert_insn_end_basic_block): Likewise.
17291 * ira-costs.c (scan_one_insn): Likewise.
17292 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17293 * lra.c (check_rtl): Likewise.
17294 * reload1.c (elimination_costs_in_insn): Likewise.
17295 * reorg.c (follow_jumps): Likewise.
17297 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17298 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17299 (thumb_far_jump_used_p): Likewise.
17300 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17301 (workaround_speculation): Likewise.
17302 (add_sched_insns_for_speculation): Likewise.
17303 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17304 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17305 (frv_for_each_packet): Likewise.
17306 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17307 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17308 (final_emit_insn_group_barriers): Likewise.
17309 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17310 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17311 (mips16_insn_length): Likewise.
17312 * config/pa/pa.c (pa_reorg): Likewise.
17313 (pa_combine_instructions): Likewise.
17314 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17315 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17316 (sh_reorg): Likewise.
17317 (split_branches): Likewise.
17318 * config/spu/spu.c (get_branch_target): Likewise.
17320 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17323 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17325 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17326 Fix declaration name.
17328 2013-03-28 Lawrence Crowl <crowl@google.com>
17330 * graphds.h (struct graph.indicies): Remove unused.
17331 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17332 (SCOP_ORIGINAL_PDDRS): Remove unused.
17333 * sese.h (extern insert_loop_close_phis): Removed unused.
17334 (extern insert_guard_phis): Removed unused.
17335 (extern ivtype_map_elt_info): Removed unused.
17336 (new_ivtype_map_elt): Removed unused.
17337 * sese.c (ivtype_map_elt_info): Removed unused.
17339 2013-03-28 Lawrence Crowl <crowl@google.com>
17341 * Makefile.in: Add several missing include dependences.
17343 (test-dump.o): New. This object is not added to any executable,
17344 but is present for ad-hoc testing.
17346 (debug (const bitmap_head_def &)): New.
17347 (debug (const bitmap_head_def *)): New.
17349 (extern debug (const bitmap_head_def &)): New.
17350 (extern debug (const bitmap_head_def *)): New.
17352 (debug (edge_def &)): New.
17353 (debug (edge_def *)): New.
17355 (debug (basic_block_def &)): New.
17356 (debug (basic_block_def *)): New.
17358 (dump_node (const_tree, int, FILE *)): Correct source file.
17360 (debug (die_struct &)): New.
17361 (debug (die_struct *)): New.
17363 (extern debug (die_struct &)): New.
17364 (extern debug (die_struct *)): New.
17365 * gimple-pretty-print.c
17366 (debug (gimple_statement_d &)): New.
17367 (debug (gimple_statement_d *)): New.
17368 * gimple-pretty-print.h
17369 (extern debug (gimple_statement_d &)): New.
17370 (extern debug (gimple_statement_d *)): New.
17372 (debug (ira_allocno_copy &)): New.
17373 (debug (ira_allocno_copy *)): New.
17374 (debug (ira_allocno &)): New.
17375 (debug (ira_allocno *)): New.
17377 (extern debug (ira_allocno_copy &)): New.
17378 (extern debug (ira_allocno_copy *)): New.
17379 (extern debug (ira_allocno &)): New.
17380 (extern debug (ira_allocno *)): New.
17382 (debug (live_range &)): New.
17383 (debug (live_range *)): New.
17385 (debug (lra_live_range &)): New.
17386 (debug (lra_live_range *)): New.
17388 (debug (lra_live_range &)): New.
17389 (debug (lra_live_range *)): New.
17391 (debug (omega_pb_d &)): New.
17392 (debug (omega_pb_d *)): New.
17394 (extern debug (omega_pb_d &)): New.
17395 (extern debug (omega_pb_d *)): New.
17397 (debug (const rtx_def &)): New.
17398 (debug (const rtx_def *)): New.
17400 (debug_tree (tree): Move within file.
17401 (debug_raw (const tree_node &)): New.
17402 (debug_raw (const tree_node *)): New.
17403 (dump_tree_via_hooks (const tree_node *, int)): New.
17404 (debug (const tree_node &)): New.
17405 (debug (const tree_node *)): New.
17406 (debug_verbose (const tree_node &)): New.
17407 (debug_verbose (const tree_node *)): New.
17408 (debug_head (const tree_node &)): New.
17409 (debug_head (const tree_node *)): New.
17410 (debug_body (const tree_node &)): New.
17411 (debug_body (const tree_node *)): New.
17412 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17413 (debug (vec<tree, va_gc> &)): New.
17414 (debug (vec<tree, va_gc> *)): New.
17416 (extern debug (const rtx_def &)): New.
17417 (extern debug (const rtx_def *)): New.
17419 (debug_raw (simple_bitmap_def &)): New.
17420 (debug_raw (simple_bitmap_def *)): New.
17421 (debug (simple_bitmap_def &)): New.
17422 (debug (simple_bitmap_def *)): New.
17424 (extern debug (simple_bitmap_def &)): New.
17425 (extern debug (simple_bitmap_def *)): New.
17426 (extern debug_raw (simple_bitmap_def &)): New.
17427 (extern debug_raw (simple_bitmap_def *)): New.
17429 (debug (vinsn_def &)): New.
17430 (debug (vinsn_def *)): New.
17431 (debug_verbose (vinsn_def &)): New.
17432 (debug_verbose (vinsn_def *)): New.
17433 (debug (expr_def &)): New.
17434 (debug (expr_def *)): New.
17435 (debug_verbose (expr_def &)): New.
17436 (debug_verbose (expr_def *)): New.
17437 (debug (vec<rtx> &)): New.
17438 (debug (vec<rtx> *)): New.
17440 (extern debug (vinsn_def &)): New.
17441 (extern debug (vinsn_def *)): New.
17442 (extern debug_verbose (vinsn_def &)): New.
17443 (extern debug_verbose (vinsn_def *)): New.
17444 (extern debug (expr_def &)): New.
17445 (extern debug (expr_def *)): New.
17446 (extern debug_verbose (expr_def &)): New.
17447 (extern debug_verbose (expr_def *)): New.
17448 (extern debug (vec<rtx> &)): New.
17449 (extern debug (vec<rtx> *)): New.
17451 (_list_iter_cond_expr): Make inline instead of static.
17453 (debug (sreal &)): New.
17454 (debug (sreal *)): New.
17456 (extern debug (sreal &)): New.
17457 (extern debug (sreal *)): New.
17459 (extern debug_raw (const tree_node &)): New.
17460 (extern debug_raw (const tree_node *)): New.
17461 (extern debug (const tree_node &)): New.
17462 (extern debug (const tree_node *)): New.
17463 (extern debug_verbose (const tree_node &)): New.
17464 (extern debug_verbose (const tree_node *)): New.
17465 (extern debug_head (const tree_node &)): New.
17466 (extern debug_head (const tree_node *)): New.
17467 (extern debug_body (const tree_node &)): New.
17468 (extern debug_body (const tree_node *)): New.
17469 (extern debug (vec<tree, va_gc> &)): New.
17470 (extern debug (vec<tree, va_gc> *)): New.
17472 (debug (struct loop &)): New.
17473 (debug (struct loop *)): New.
17474 (debug_verbose (struct loop &)): New.
17475 (debug_verbose (struct loop *)): New.
17476 * tree-dump.c: Add header dependence.
17478 (extern debug (struct loop &)): New.
17479 (extern debug (struct loop *)): New.
17480 (extern debug_verbose (struct loop &)): New.
17481 (extern debug_verbose (struct loop *)): New.
17483 (debug (data_reference &)): New.
17484 (debug (data_reference *)): New.
17485 (debug (vec<data_reference_p> &)): New.
17486 (debug (vec<data_reference_p> *)): New.
17487 (debug (vec<ddr_p> &)): New.
17488 (debug (vec<ddr_p> *)): New.
17490 (extern debug (data_reference &)): New.
17491 (extern debug (data_reference *)): New.
17492 (extern debug (vec<data_reference_p> &)): New.
17493 (extern debug (vec<data_reference_p> *)): New.
17494 (extern debug (vec<ddr_p> &)): New.
17495 (extern debug (vec<ddr_p> *)): New.
17497 (debug (pt_solution &)): New.
17498 (debug (pt_solution *)): New.
17500 (extern debug (pt_solution &)): New.
17501 (extern debug (pt_solution *)): New.
17503 (debug (_var_map &)): New.
17504 (debug (_var_map *)): New.
17505 (debug (tree_live_info_d &)): New.
17506 (debug (tree_live_info_d *)): New.
17508 (extern debug (_var_map &)): New.
17509 (extern debug (_var_map *)): New.
17510 (extern debug (tree_live_info_d &)): New.
17511 (extern debug (tree_live_info_d *)): New.
17513 2013-03-28 Jan Hubicka <jh@suse.cz>
17515 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17517 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17519 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17520 record only when desired or required.
17522 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17524 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17525 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17527 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17529 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17530 (*andsi3_compare0_uxtw): New pattern.
17531 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17532 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17534 2013-03-28 Jan Hubicka <jh@suse.cz>
17536 * data-streamer-in.c (streamer_read_gcov_count): New function.
17537 * gimple-streamer-out.c: Include value-prof.h.
17538 (output_gimple_stmt): Output histogram.
17539 (output_bb): Use streamer_write_gcov_count.
17540 * value-prof.c: Include data-streamer.h
17541 (dump_histogram_value): Add HIST_TYPE_MAX.
17542 (stream_out_histogram_value): New function.
17543 (stream_in_histogram_value): New function.
17544 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17545 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17546 * data-streamer-out.c (streamer_write_gcov_count): New function.
17547 (streamer_write_gcov_count_stream): New function.
17548 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17549 (lto_output_node): Likewise.
17550 (input_node, input_edge): Likewise.
17551 * lto-streamer-out.c (output_cfg): Update streaming.
17552 * lto-streamer-in.c (input_cfg): Likewise.
17553 * data-streamer.h (streamer_write_gcov_count,
17554 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17555 * gimple-streamer-in.c: Include value-prof.h
17556 (input_gimple_stmt): Input histograms.
17557 (input_bb): Update profile streaming.
17559 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17561 * genmodes.c (emit_max_int): New function.
17562 (emit_insn_modes_h): Added call to emit_max_function.
17563 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17565 * machmode.def: Fixed comment.
17567 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17569 * combine.c (try_combine): Removed useless assert.
17570 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17572 2013-03-28 Marek Polacek <polacek@redhat.com>
17573 Richard Biener <rguenther@suse.de>
17575 PR tree-optimization/56695
17576 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17577 build signed result of a vector comparison.
17578 * tree-cfg.c (verify_gimple_comparison): Check that a result
17579 of a vector comparison has signed type.
17581 2013-03-28 Richard Biener <rguenther@suse.de>
17583 PR tree-optimization/37021
17584 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17585 do not restrict gaps between groups.
17586 * tree-vect-stmts.c (vectorizable_load): Properly account for
17587 a gap between groups.
17589 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17591 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17592 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17595 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17597 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17598 * doc/extend.texi (Named Address Spaces): Ditto.
17599 (Variable Attributes): Ditto.
17601 2013-03-27 Kai Tietz <ktietz@redhat.com>
17603 * config.build: Add support for cygwin x64 target.
17604 * config.gcc: Likewise.
17605 * config.host: Likewise.
17606 * configure.ac: Likewise
17607 * configure: Regenerated.
17609 2013-03-27 Kai Tietz <ktietz@redhat.com>
17611 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17612 * config/i386/t-cygwin-w64: New file.
17613 * config/i386/cygwin-w64.h: New file.
17614 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17615 and add support for x64-cygwin target.
17616 (CPP_SPEC): Likewise.
17617 (CXX_WRAP_SPEC_LIST): Undefine before define.
17618 (LIBGCJ_SONAME): Use 15 as version.
17620 2013-03-27 Richard Biener <rguenther@suse.de>
17622 PR tree-optimization/56716
17623 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17624 dumping for ref nodes.
17626 2013-03-27 Martin Jambor <mjambor@suse.cz>
17628 PR tree-optimization/55334
17629 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17630 restricted pointers to arrays.
17632 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17634 * Makefile.in (.SUFFIXES): Add .cc.
17635 (.c.o): Apply same recipe for implicit rule .cc.o.
17637 2013-03-27 Richard Biener <rguenther@suse.de>
17639 PR tree-optimization/37021
17640 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17641 REALPART/IMAGPART_EXPRs around the supported refs.
17642 * tree-ssa-structalias.c (find_func_aliases): Assume that
17643 floating-point values are not used to transfer pointers.
17645 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17647 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17648 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17649 * doc/tm.texi: Regenerate.
17650 * targhooks.h (default_has_ifunc_p): New.
17651 * targhooks.c (default_has_ifunc_p): Ditto.
17652 * config/linux-protos.h: New file.
17653 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17654 hook for linux which disables support of indirect functions in android.
17655 * config/linux-android.c: New file.
17656 * config/t-linux-android.c: Ditto.
17657 * config.gcc: Added new object file linux-android.o.
17658 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17659 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17660 * varasm.c (do_assemble_alias): Likewise.
17661 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17662 doesn't support indirect functions.
17663 * configure: Regenerate.
17665 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17668 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17669 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17672 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17674 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17676 2013-03-27 Terry Guo <terry.guo@arm.com>
17678 * config/arm/arm-cores.def: Added core cortex-r7.
17679 * config/arm/arm-tune.md: Regenerated.
17680 * config/arm/arm-tables.opt: Regenerated.
17681 * doc/invoke.texi: Added entry for core cortex-r7.
17683 2013-03-27 Walter Lee <walt@tilera.com>
17685 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17686 double-decrement of next_scratch_regno.
17688 2013-03-27 Walter Lee <walt@tilera.com>
17690 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17692 (insn_v1mulus): Ditto.
17693 (insn_v2muls): Ditto.
17695 2013-03-27 Walter Lee <walt@tilera.com>
17697 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17698 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17700 2013-03-27 Walter Lee <walt@tilera.com>
17702 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17703 (*sibcall_value): Ditto.
17705 2013-03-27 Walter Lee <walt@tilera.com>
17707 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17708 (insn_mnz_v8qi): ... this ...
17709 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17711 (insn_v<n>mnz): Replaced by ...
17712 (insn_v1mnz): ... this ...
17713 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17715 (insn_mz_<mode>): Replaced by ...
17716 (insn_mz_v8qi): ... this ...
17717 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17719 (insn_v<n>mz): Replaced by ...
17720 (insn_v1mz): ... this ...
17721 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17724 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17726 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17728 2013-03-26 Roland McGrath <mcgrathr@google.com>
17730 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17731 than fprintf with a non-constant, non-format string.
17733 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17735 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17736 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17737 operand 0 predicate.
17738 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17739 attribute. Use general_x64nomem_operand as operand 1 predicate.
17740 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17741 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17742 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17743 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17744 (mov<mode>_insv_1): Remove expander. Merge insn with
17745 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17746 Use general_x64nomem_operand as operand 1 predicate.
17747 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17748 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17749 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17750 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17751 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17752 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17753 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17754 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17755 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17756 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17757 (general_x64nomem_operand): Ditto.
17759 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17761 * config/rtems.opt: Add -pthread option.
17763 2013-03-26 Richard Biener <rguenther@suse.de>
17765 * alias.c (find_base_term): Avoid redundant and not used recursion.
17766 (base_alias_check): Get the initial base term from the caller.
17767 (true_dependence_1): Compute and pass base terms to base_alias_check.
17768 (write_dependence_p): Likewise.
17769 (may_alias_p): Likewise.
17771 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17773 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17774 PC-relative load in SI modes and above only.
17776 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17778 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17779 * config/arm/iwmmxt.md (WCGR0): Update.
17780 (WCGR1, WCGR2, WCGR3): Likewise.
17782 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17784 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17785 Use x64 and nox64 isa attributes.
17787 2013-03-26 Richard Biener <rguenther@suse.de>
17789 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17790 alignment computations and rely on get_object_alignment_1
17791 for the !TYPE_P case.
17792 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17794 2013-03-26 Walter Lee <walt@tilera.com>
17796 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17797 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17799 2013-03-25 Jeff Law <law@redhat.com>
17801 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17802 check for INTEGRAL_TYPE_P that was missing due to checking in
17803 wrong version of prior patch.
17805 2013-03-25 Walter Lee <walt@tilera.com>
17807 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17808 TILEGX_INSN_SHUFFLEBYTES1.
17809 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17811 (tilegx_builtins): Ditto.
17812 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17814 2013-03-25 Walter Lee <walt@tilera.com>
17816 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17817 (floatunssisf2): New pattern.
17818 (floatsidf2): New pattern.
17819 (floatunssidf2): New pattern.
17821 2013-03-25 Walter Lee <walt@tilera.com>
17823 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17824 tests for constraint J, K, N, P.
17826 2013-03-25 Walter Lee <walt@tilera.com>
17828 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17829 Use indirect/pcrel encoding.
17830 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17833 2013-03-25 Steve Ellcey <sellcey@mips.com>
17835 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17836 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17837 * config/mips/mips.c (mips_option_override): Set IMADD default.
17838 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17839 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17840 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17841 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17842 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17844 2013-03-25 Jeff Law <law@redhat.com>
17846 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17847 slightly to avoid creating and folding useless trees. Simplify
17848 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
17850 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
17852 * config/i386/i386.md (*zero_extendsidi2): Merge with
17853 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
17854 * config/i386/predicates.md (x86_64_zext_operand): Rename from
17855 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
17856 targets. Clarify comment.
17858 2013-03-25 Martin Jambor <mjambor@suse.cz>
17860 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
17861 pass-through jump functions differently.
17862 (ipa_read_jump_function): Likewise. Also use setter functions to set
17865 2013-03-25 Martin Jambor <mjambor@suse.cz>
17867 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
17868 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
17870 (ipa_get_indirect_edge_target): New function.
17871 (devirtualization_time_bonus): New parameter known_aggs, pass it to
17872 ipa_get_indirect_edge_target. Update all callers.
17873 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
17874 ipa_get_indirect_edge_target_1 instead of calling
17875 ipa_get_indirect_edge_target.
17876 (create_specialized_node): Pass aggvlas to
17877 ipcp_discover_new_direct_edges.
17879 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17881 * config/arm/arm.md (f_sels, f_seld): New types.
17882 (*cmov<mode>): New pattern.
17883 * config/arm/predicates.md (arm_vsel_comparison_operator): New
17886 2013-03-25 Kai Tietz <ktietz@redhat.com>
17888 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
17889 POSIX-printf for mingw-hosted builds.
17891 2013-03-25 Richard Biener <rguenther@suse.de>
17893 PR middle-end/56694
17894 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
17895 must-not-throw stmt location.
17897 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17899 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
17900 Emit load-acquire versions when acq is true.
17901 (arm_emit_store_exclusive): Add rel parameter.
17902 Emit store-release versions when rel is true.
17903 (arm_split_compare_and_swap): Use acquire-release instructions
17905 of barriers when appropriate.
17906 (arm_split_atomic_op): Likewise.
17907 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
17908 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
17909 (VUNSPEC_SLX): Likewise.
17910 (VUNSPEC_LDA): Likewise.
17911 (VUNSPEC_STL): Likewise.
17912 * config/arm/sync.md (atomic_load<mode>): New pattern.
17913 (atomic_store<mode>): Likewise.
17914 (arm_load_acquire_exclusive<mode>): Likewise.
17915 (arm_load_acquire_exclusivesi): Likewise.
17916 (arm_load_acquire_exclusivedi): Likewise.
17917 (arm_store_release_exclusive<mode>): Likewise.
17919 2013-03-25 Catherine Moore <clm@codesourcery.com>
17921 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
17922 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
17923 * config/mip/predicates.md (lwsp_swsp_operand,
17924 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
17925 sb16_operand, db4_operand, db7_operand, ib3_operand,
17926 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
17927 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
17928 andi16_operand): New predicates.
17929 * config/mips/mips.md (compression): New attribute.
17930 (enabled): New attribute.
17931 (length): Consider compression in computing length.
17932 (shift_compression): New code attribute.
17933 (*add<mode>3): New operands. Record compression.
17934 (sub<mode>3): Likewise.
17935 (one_cmpl<mode>2): Likewise.
17936 (*and<mode>3): Likewise.
17937 (*ior<mode>3): Likewise.
17938 (unnamed pattern for xor): Likewise.
17939 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
17940 (*<optab><mode>3): Likewise.
17941 (*mov<mode>_internal: Likewise.
17942 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
17943 (mips_unsigned_immediate_p): New.
17944 (umips_lwsp_swsp_address_p): New.
17945 (m16_based_address_p): New.
17946 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
17947 (mips_unsigned_immediate_p): New prototype.
17948 (lwsp_swsp_address_p): New prototype.
17949 (m16_based_address_p): New prototype.
17950 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
17951 (mips_signed_immediate_p): New function.
17952 (m16_based_address_p): New function.
17953 (lwsp_swsp_address_p): New function.
17954 (mips_print_operand_punctuation): Recognize short delay slot insns
17955 for microMIPS.add<mode>3"
17957 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17960 * config/arm/iterators.md (v_cmp_result): New mode attribute.
17961 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
17963 2013-03-25 Richard Biener <rguenther@suse.de>
17965 PR tree-optimization/56689
17966 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
17969 2013-03-25 Richard Biener <rguenther@suse.de>
17971 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
17973 (memory_references): Likewise.
17974 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
17975 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
17976 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
17977 (gather_mem_refs_in_loops): Fold into ...
17978 (analyze_memory_references): ... this. Move initialization
17979 to tree_ssa_lim_initialize.
17980 (fill_always_executed_in): Rename to ...
17981 (fill_always_executed_in_1): ... this.
17982 (fill_always_executed_in): Move contains_call computation to
17983 this new function from ...
17984 (tree_ssa_lim_initialize): ... here.
17985 (tree_ssa_lim): Call fill_always_executed_in.
17987 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17989 * postreload.c (reload_combine): Fix code detecting returns.
17991 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17993 * function.c (emit_use_return_register_into_block): On cc0 targets,
17994 do not emit the sequence between cc0 setter and user.
17996 2013-03-25 Kai Tietz <ktietz@redhat.com>
17998 * config/i386/predicates.md (local_symbolic_operand): Interpret
17999 dll-imported symbols as none-local.
18001 2013-03-25 Richard Biener <rguenther@suse.de>
18003 * tree-ssa-loop-im.c (struct depend): Remove.
18004 (struct lim_aux_data): Make depends a vec of gimples.
18005 (free_lim_aux_data): Adjust.
18006 (add_dependency): Likewise.
18007 (set_level): Likewise.
18009 2013-03-25 Richard Biener <rguenther@suse.de>
18011 PR middle-end/56434
18012 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18013 the pointer returned by calls with ECF_MALLOC set.
18015 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18017 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18019 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18021 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18022 using MMXMODE mode iterator.
18023 (*move<mode>_internal): Merge with *movv2sf_internal and
18024 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18026 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18028 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18029 (record_last_mem_set_info): Likewise.
18031 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18032 of XNEWVEC followed by memset.
18033 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18035 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18037 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18038 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18039 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18040 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18041 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18042 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18043 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18044 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18045 BARRIER_P instead of GET_CODE.
18047 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18049 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18050 inaccuracy in the probing code.
18052 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18053 (ctrapdi4): Likewise.
18055 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18057 * calls.c (expand_call): Add missing guard to code handling return
18058 of non-BLKmode structures in MSB.
18059 * function.c (expand_function_end): Likewise.
18061 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18063 * combine.c (try_combine): Adjust comment. Do not add the set of
18064 insn #0 if the destination indirectly is set or dies in insn #2.
18065 Tidy up code to distribute a new note.
18067 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18069 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18070 also for alternatives 16 and 17.
18072 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18074 * config/i386/sse.md (*mov<mode>_internal): Merge with
18075 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18076 Emit insn template depending on type attribute. Use
18077 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18078 movd instead of movq mnemonic for interunit moves. Rewrite mode
18079 attribute calculation. Remove unit attribute calculation.
18080 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18081 Set prefix_data16 attribute for DImode ssemov types.
18082 Use Ym instead of y for SSE-MMX conversion alternatives.
18083 Reorder operand constraints.
18085 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18087 * df.h (df_insn_delete): Adjust prototype.
18088 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18089 and let it decide whether mark the basic block dirty.
18090 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18091 * df-scan.c (df_insn_info_delete): New helper function, split
18092 off from df_insn_delete.
18093 (df_scan_free_bb_info): Use it.
18094 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18096 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18097 that the insn is actually an insn and it has a non-NULL basic block.
18098 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18100 2013-03-22 Richard Biener <rguenther@suse.de>
18102 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18104 (mem_ref_alloc): Do not allocate them.
18105 (refs_independent_p): Do not query or maintain a cache.
18107 2013-03-22 Richard Biener <rguenther@suse.de>
18109 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18110 (gather_mem_refs_in_loops): Do not compute it.
18111 (analyze_memory_references): Do not allocate it.
18112 (tree_ssa_lim_finalize): Do not free it.
18113 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18115 2013-03-22 Richard Biener <rguenther@suse.de>
18117 * is-a.h (as_a): Use gcc_checking_assert.
18119 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18121 * config/aarch64/aarch64.c (aarch64_print_operand): New
18122 format specifier for printing a constant in hex.
18123 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18124 format specifier for printing second operand.
18126 2013-03-22 Richard Biener <rguenther@suse.de>
18128 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18130 (gather_mem_refs_in_loops): Perform store accumulation here.
18131 (create_vop_ref_mapping_loop): Remove.
18132 (create_vop_ref_mapping): Likewise.
18133 (analyze_memory_references): Initialize refs_stored_in_loop.
18134 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18135 (record_indep_loop): Remove.
18136 (record_dep_loop): New function.
18137 (ref_indep_loop_p_1): Adjust to only walk over references
18138 in the loop, not its subloops.
18139 (ref_indep_loop_p): Rename to ...
18140 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18141 maintaining a more fine-grained cache.
18142 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18143 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18145 2013-03-22 Richard Biener <rguenther@suse.de>
18147 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18148 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18149 aggregate mem_ref_loc.
18150 (free_mem_ref_locs): Inline into ...
18151 (memref_free): ... this and adjust.
18152 (mem_ref_alloc): Adjust.
18153 (mem_ref_locs_alloc): Remove.
18154 (record_mem_ref_loc): Adjust.
18155 (get_all_locs_in_loop): Rewrite into ...
18156 (for_all_locs_in_loop): ... this iterator.
18157 (rewrite_mem_ref_loc): New functor.
18158 (rewrite_mem_refs): Use for_all_locs_in_loop.
18159 (sm_set_flag_if_changed): New functor.
18160 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18161 (ref_always_accessed): New functor.
18162 (ref_always_accessed_p): Use for_all_locs_in_loop.
18164 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18166 * tree-pass.h (PROP_gimple_lvec): New.
18167 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18168 (init_optimization_passes): Move pass_lower_vector.
18169 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18171 (pass_lower_vector): Provide PROP_gimple_lvec.
18172 (pass_lower_vector_ssa): Likewise.
18173 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18175 2013-03-21 Mark Wielaard <mjw@redhat.com>
18177 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18179 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18181 * config/i386/i386.md (*movdi_internal): Disparage slightly
18182 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18183 conversion alternatives.
18185 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18187 PR middle-end/48087
18188 * diagnostic.def (DK_WERROR): New kind.
18189 * diagnostic.h (werrorcount): Define.
18190 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18191 promoted to DK_ERROR, increment DK_WERROR counter instead of
18193 * toplev.c (toplev_main): Call print_ignored_options even if
18194 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18195 even if just werrorcount is non-zero.
18198 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18200 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18201 (string_cst_pool_decl): New function.
18202 (optimize_one_addr_into_implicit_ptr): New function.
18203 (resolve_addr_in_expr): Optimize DWARF location expression
18204 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18205 which doesn't live in memory, but has DW_AT_location or
18206 DW_AT_const_value, or refers to a string literal, into
18207 DW_OP_GNU_implicit_pointer.
18208 (optimize_location_into_implicit_ptr): New function.
18209 (resolve_addr): If removing DW_AT_location of a variable because
18210 it was DW_OP_addr of address of the variable, but the variable doesn't
18211 live in memory, try to emit const value attribute for the initializer.
18213 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18215 * tree.h (VECTOR_TYPE_P): New macro.
18216 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18217 TYPE_MODE): Use it.
18218 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18219 VEC_COND_EXPR cannot be lvalues.
18220 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18222 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18224 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18225 Restrict the transformation to equal modes.
18227 2013-03-21 Richard Biener <rguenther@suse.de>
18229 PR tree-optimization/39326
18230 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18231 (MEM_ANALYZABLE): Adjust.
18232 (record_mem_ref_loc): Move bitmap ops ...
18233 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18234 unanalyzable refs, do not record locations for it.
18235 (analyze_memory_references): Allocate ref zero as shared
18237 (refs_independent_p): Do not test for unanalyzed mems here.
18238 (ref_indep_loop_p_1): Special-case disambiguation against
18239 the unanalyzed ref.
18240 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18242 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18244 * config/arm/arm-protos.h (tune_params): Add
18245 prefer_neon_for_64bits field.
18246 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18247 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18248 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18249 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18250 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18251 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18252 (arm_option_override): Handle -mneon-for-64bits new option.
18253 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18254 (prefer_neon_for_64bits): Declare new variable.
18255 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18256 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18257 (arch_enabled): Handle new arch types. Remove support for onlya8
18259 (one_cmpldi2): Use new arch names.
18260 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18261 * config/arm/arm.opt (mneon-for-64bits): Add option.
18262 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18263 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18264 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18266 * doc/invoke.texi (-mneon-for-64bits): Document.
18268 2013-03-21 Richard Biener <rguenther@suse.de>
18270 PR tree-optimization/39326
18271 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18272 (sort_bbs_in_loop_postorder_cmp): New function.
18273 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18275 2013-03-21 Richard Biener <rguenther@suse.de>
18277 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18278 (vect_insert_into_interleaving_chain): Likewise.
18279 (vect_drs_dependent_in_basic_block): Inline ...
18280 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18282 (vect_analyze_data_ref_dependence): ... here. Simplify.
18283 (vect_check_interleaving): Simplify.
18284 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18285 (vect_slp_analyze_data_ref_dependences): ... this new function.
18286 (dr_group_sort_cmp): New function.
18287 (vect_analyze_data_ref_accesses): Compute data-reference groups
18288 here instead of in vect_analyze_data_ref_dependence. Use
18289 a more efficient algorithm.
18290 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18291 vect_slp_analyze_data_ref_dependences. Call
18292 vect_analyze_data_ref_accesses earlier.
18293 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18294 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18295 (vect_slp_analyze_data_ref_dependences): New prototype.
18297 2013-03-21 Richard Biener <rguenther@suse.de>
18299 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18300 ref is stored in the loop.
18301 (find_refs_for_sm): Walk only over all stores.
18302 (store_motion_loop): Allocate from lim_bitmap_obstack.
18303 (store_motion): Likewise.
18305 2013-03-21 Richard Biener <rguenther@suse.de>
18307 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18308 Update virtual SSA form.
18310 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18312 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18313 * configure: Regenerate.
18314 * config.in: Regenerate.
18315 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18316 if !HAVE_LD_EH_FRAME_CIEV3.
18318 2013-03-21 Richard Biener <rguenther@suse.de>
18320 * tree-cfg.c (verify_expr_no_block): New function.
18321 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18322 nor DECL_VALUE_EXPR have locations with associated blocks.
18323 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18324 (clear_unused_block_pointer): Remove code dealing with
18325 blocks in DECL_DEBUG_EXPR locations.
18327 2013-03-21 Richard Biener <rguenther@suse.de>
18329 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18330 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18331 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18332 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18333 instead of DECL_DEBUG_EXPR_IS_FROM.
18334 * gimplify.c (gimplify_modify_expr): Likewise.
18335 * tree-cfg.c (verify_expr_location_1): Likewise.
18336 * tree-complex.c (create_one_component_var): Likewise.
18337 * tree-sra.c (create_access_replacement): Likewise.
18338 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18339 (clear_unused_block_pointer): Likewise.
18340 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18341 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18342 * var-tracking.c (var_debug_decl): Likewise.
18343 (track_expr_p): Likewise.
18344 * tree-inline.c (add_local_variables): Likewise. Set
18345 DECL_HAS_DEBUG_EXPR_P after copying it.
18346 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18347 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18349 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18352 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18353 * configure: Regenerate.
18354 * config.in: Regenerate.
18355 * config/i386/i386.md (*movdf_internal): Use
18356 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18357 movd instead of movq mnemonic for interunit moves.
18358 (*movdi_internal): Ditto.
18360 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18362 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18363 (abd<mode>_3): New pattern.
18364 (aba<mode>_3): New pattern.
18365 (fabd<mode>_3): New pattern.
18367 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18369 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18370 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18371 occurrence of REGISTER_PREFIX as its empty string.
18373 2013-03-20 Jeff Law <law@redhat.com>
18375 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18376 addititional equivalences for equality comparisons between an SSA_NAME
18377 and a constant where the SSA_NAME was set from a widening conversion.
18379 2013-03-20 Walter Lee <walt@tilera.com>
18381 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18383 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18385 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18386 depending on type attribute.
18387 (*movti_internal): Ditto.
18388 (*movtf_internal): Ditto.
18389 (*movxf_internal): Ditto.
18390 (*movdf_internal): Ditto.
18391 (*movsf_internal): Ditto.
18393 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18395 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18396 maybe_vex for sselog1 and ssemov types.
18397 (*movdi_internal): Reorder operand constraints.
18398 (*movsi_internal): Ditto. Set prefix attribute to
18399 maybe_vex for sselog1 and ssemov types.
18400 (*movtf_internal): Set prefix attribute to maybe_vex
18401 for sselog1 and ssemov types.
18402 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18403 DImode ssemov types. Reorder operand constraints.
18404 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18405 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18406 attribute for SImode ssemov types. Reorder operand constraints.
18408 2013-03-20 Martin Jambor <mjambor@suse.cz>
18410 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18411 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18413 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18415 * config/rs6000/predicates.md (indexed_address, update_address_mem
18416 update_indexed_address_mem): New predicates.
18417 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18418 attribute for load/store instructions.
18419 * config/rs6000/dfp.md (movsd_store): Likewise.
18420 (movsd_load): Likewise.
18421 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18422 (unnamed HI->DI extend define_insn): Likewise.
18423 (unnamed SI->DI extend define_insn): Likewise.
18424 (unnamed QI->SI extend define_insn): Likewise.
18425 (unnamed QI->HI extend define_insn): Likewise.
18426 (unnamed HI->SI extend define_insn): Likewise.
18427 (unnamed HI->SI extend define_insn): Likewise.
18428 (extendsfdf2_fpr): Likewise.
18429 (movsi_internal1): Likewise.
18430 (movsi_internal1_single): Likewise.
18431 (movhi_internal): Likewise.
18432 (movqi_internal): Likewise.
18433 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18434 attribute for load/store instructions.
18435 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18437 (mov<mode>_softfloat): Likewise.
18438 (mov<mode>_hardfloat32): Likewise.
18439 (mov<mode>_hardfloat64): Likewise.
18440 (mov<mode>_softfloat64): Likewise.
18441 (movdi_internal32): Likewise.
18442 (movdi_internal64): Likewise.
18443 (probe_stack_<mode>): Likewise.
18445 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18447 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18448 floating point, and decimal floating point to reload iterator.
18450 * config/rs6000/constraints.md (wl constraint): New constraints to
18451 return FLOAT_REGS if certain options are used to reduce the number
18452 of separate patterns that exist in the file.
18453 (wx constraint): Likewise.
18454 (wz constraint): Likewise.
18456 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18457 -mdebug=reg, print wg, wl, wx, and wz constraints.
18458 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18459 Initialize the reload functions for 64-bit binary/decimal floating
18461 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18462 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18463 create the buffer on the stack to overcome not having a 32-bit
18465 (rs6000_emit_move): Likewise.
18466 (rs6000_secondary_memory_needed_rtx): Likewise.
18467 (rs6000_alloc_sdmode_stack_slot): Likewise.
18468 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18469 via xxlxor, just like DFmode 0.0.
18471 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18472 define as 1 if we are running on a power7 or newer.
18473 (enum r6000_reg_class_enum): Add new constraints.
18475 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18476 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18477 with other moves by using conditional constraits (wg). Use LFIWZX
18478 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18479 (movsd splitter): Likewise.
18480 (movsd_hardfloat): Likewise.
18481 (movsd_softfloat): Likewise.
18483 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18484 binary and decimal floating point moves.
18485 (fmove_ok): New attributes to combine binary and decimal floating
18486 point moves, and to combine power6x (mfpgpr) moves along normal
18488 (real_value_to_target): Likewise.
18489 (f32_lr): Likewise.
18490 (f32_lm): Likewise.
18491 (f32_li): Likewise.
18492 (f32_sr): Likewise.
18493 (f32_sm): Likewise.
18494 (f32_si): Likewise.
18495 (movsf): Combine binary and decimal floating point moves. Combine
18496 power6x (mfpgpr) moves with other moves by using conditional
18497 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18498 (mov<mode> for SFmode/SDmode); Likewise.
18499 (SFmode/SDmode splitters): Likewise.
18500 (movsf_hardfloat): Likewise.
18501 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18502 (movsf_softfloat): Likewise.
18503 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18505 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18506 wx and wz constraints.
18508 * config/rs6000/constraints.md (wg constraint): New constraint to
18509 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18511 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18514 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18515 -mdebug=reg, print wg, wl, wx, and wz constraints.
18516 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18517 Initialize the reload functions for 64-bit binary/decimal floating
18519 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18520 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18521 create the buffer on the stack to overcome not having a 32-bit
18523 (rs6000_emit_move): Likewise.
18524 (rs6000_secondary_memory_needed_rtx): Likewise.
18525 (rs6000_alloc_sdmode_stack_slot): Likewise.
18526 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18527 via xxlxor, just like DFmode 0.0.
18529 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18530 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18531 with other moves by using conditional constraits (wg). Use LFIWZX
18532 and STFIWX for loading SDmode on power7.
18533 (movdd splitters): Likewise.
18534 (movdd_hardfloat32): Likewise.
18535 (movdd_softfloat32): Likewise.
18536 (movdd_hardfloat64_mfpgpr): Likewise.
18537 (movdd_hardfloat64): Likewise.
18538 (movdd_softfloat64): Likewise.
18540 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18541 64-bit binary and decimal floating point moves.
18542 (FMOVE64X): Likewise.
18543 (movdf): Combine 64-bit binary and decimal floating point moves.
18544 Combine power6x (mfpgpr) moves with other moves by using
18545 conditional constraits (wg).
18546 (mov<mode> for DFmode/DDmode): Likewise.
18547 (DFmode/DDmode splitters): Likewise.
18548 (movdf_hardfloat32): Likewise.
18549 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18550 (movdf_softfloat32): Likewise.
18551 (movdf_hardfloat64_mfpgpr): Likewise.
18552 (movdf_hardfloat64): Likewise.
18553 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18554 (movdf_softfloat64): Likewise.
18555 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18556 (reload_<mode>_load): Move to later in the file so they aren't in
18557 the middle of the floating point move insns.
18558 (reload_<mode>_store): Likewise.
18560 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18563 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18564 constraint if -mdebug=reg.
18565 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18566 Enable using dd reload support if needed.
18568 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18569 binary and decimal floating point moves in rs6000.md.
18570 (movtd_internal): Likewise.
18572 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18573 decimal floating point moves.
18575 (movtf_internal): Likewise.
18576 (mov<mode>_internal, TDmode/TFmode): Likewise.
18577 (movtf_softfloat): Likewise.
18578 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18580 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18581 movdi_internal64, using wg constraint for move direct operations.
18582 (movdi_internal64): Likewise.
18584 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18585 MODES_TIEABLE_P for selected modes. Print the numerical value of
18586 the various virtual registers. Use GPR/FPR first/last values,
18587 instead of hard coding the register numbers. Print which modes
18588 have reload functions registered.
18589 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18590 settings before/after setting cpu, target and subtarget settings.
18591 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18592 and for secondary reload failures in rs6000_secondary_reload_inner.
18593 (rs6000_secondary_reload_fail): Likewise.
18594 (rs6000_secondary_reload_inner): Likewise.
18596 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18597 macros for first/last GPR and FPR registers.
18598 (LAST_GPR_REGNO): Likewise.
18599 (FIRST_FPR_REGNO): Likewise.
18600 (LAST_FPR_REGNO): Likewise.
18602 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18603 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18604 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18605 (vcond<mode><mode>): Likewise.
18606 (vcondu<mode><mode>): Likewise.
18607 (vector_gtu<mode>): Likewise.
18608 (vector_gte<mode>): Likewise.
18609 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18610 to prevent the compiler from converting DImode operations to TImode.
18611 (ior<mode>3): Likewise.
18612 (and<mode>3): Likewise.
18613 (one_cmpl<mode>2): Likewise.
18614 (nor<mode>3): Likewise.
18615 (andc<mode>3): Likewise.
18617 * config/rs6000/constraints.md (wt constraint): New constraint
18618 that returns VSX_REGS if TImode is allowed in VSX registers.
18620 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18621 constant under VSX.
18623 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18624 similar to TImode, but it is restricted to being in the GPRs.
18626 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18627 TImode to occupy a single VSX register.
18629 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18630 -mvsx-timode for power7/power8.
18631 (power7 cpu): Likewise.
18632 (power8 cpu): Likewise.
18634 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18635 sure that TFmode/TDmode take up two registers if they are ever
18636 allowed in the upper VSX registers.
18637 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18639 (rs6000_init_hard_regno_mode_ok): Likewise.
18640 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18641 constraint. Print if LRA is turned on.
18642 (rs6000_option_override_internal): Give an error if -mvsx-timode
18643 and VSX is not enabled.
18644 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18645 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18646 to reg+offset addressing. Use PTImode when checking offset
18647 addresses for validity.
18648 (reg_offset_addressing_ok_p): Likewise.
18649 (rs6000_legitimate_offset_address_p): Likewise.
18650 (rs6000_legitimize_address): Likewise.
18651 (rs6000_legitimize_reload_address): Likewise.
18652 (rs6000_legitimate_address_p): Likewise.
18653 (rs6000_eliminate_indexed_memrefs): Likewise.
18654 (rs6000_emit_move): Likewise.
18655 (rs6000_secondary_reload): Likewise.
18656 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18657 reloads to fpr registers to continue to use reg+offset addressing,
18658 but 64-bit reloads to altivec registers need reg+reg addressing.
18659 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18660 it. Treat LO_SUM like a PLUS operation.
18661 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18662 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18663 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18664 registers to share a register with a smaller sized type, since VSX
18665 puts scalars in the upper 64-bits.
18666 (print_operand): Add support for PTImode.
18667 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18668 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18669 registers, but don't have arithmetic support.
18670 (rs6000_memory_move_cost): Add test for VSX.
18671 (rs6000_opt_masks): Add -mvsx-timode.
18673 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18676 (VSr): Use wt constraint for TImode.
18677 (VSv): Drop TImode support.
18678 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18679 (vsx_movti_64bit): Likewise.
18680 (vsx_movti_32bit): Likewise.
18681 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18682 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18683 one '?' on the appropriate output constraint. Do not allow TImode
18684 logical operations on 32-bit systems.
18685 (vsx_ior<mode>3): Likewise.
18686 (vsx_xor<mode>3): Likewise.
18687 (vsx_one_cmpl<mode>2): Likewise.
18688 (vsx_nor<mode>3): Likewise.
18689 (vsx_andc<mode>3): Likewise.
18690 (vsx_concat_<mode>): Likewise.
18691 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18693 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18694 OPTION_MASK_VSX_TIMODE.
18695 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18696 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18698 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18699 (TI2 iterator): New iterator for TImode, PTImode.
18700 (wd mode attribute): Add values for vector types.
18701 (movti_string): Replace TI move operations with operations for TImode
18702 and PTImode. Add support for TImode being allowed in VSX registers.
18703 (mov<mode>_string, TImode/PTImode): Likewise.
18704 (movti_ppc64): Likewise.
18705 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18706 (TI mode splitters): Likewise.
18708 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18711 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18713 PR tree-optimization/56355
18714 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18715 Also handle integers with undefined overflow.
18717 2013-03-20 Catherine Moore <clm@codesourcery.com>
18718 Maciej W. Rozycki <macro@codesourcery.com>
18719 Tom de Vries <tom@codesourcery.com>
18720 Nathan Sidwell <nathan@codesourcery.com>
18721 Iain Sandoe <iain@codesourcery.com>
18722 Nathan Froyd <froydnj@codesourcery.com>
18723 Chao-ying Fu <fu@mips.com>
18725 * doc/extend.texi: (micromips, nomicromips, nocompression):
18726 Document new function attributes.
18727 * doc/invoke.texi (minterlink-compressed, mmicromips,
18728 m14k, m14ke, m14kec): Document new options.
18729 (minterlink-mips16): Update documentation.
18730 * doc/md.texi (ZC, ZD): Document new constraints.
18731 * configure.ac (gcc_cv_as_micromips): Check if linker
18732 supports the .set micromips directive.
18733 * configure: Regenerate.
18734 * config.in: Regenerate.
18735 * config/mips/mips-tables.opt: Regenerate.
18736 * config/mips/micromips.md: New file.
18737 * constraints.md (ZC, ZD): New constraints.
18738 * config/mips/predicates.md (movep_src_register): New predicate.
18739 (movep_src_operand): New predicate.
18740 (non_volatile_mem_operand): New predicate.
18741 * config/mips/mips.md (multimem): New type.
18742 (length): Differentiate between 17-bit and 18-bit branch offsets.
18743 (MOVEP1, MOVEP2): New mode iterator.
18744 (mov_<load>l): Use ZC constraint.
18745 (mov_<load>r): Likewise.
18746 (mov_<store>l): Likewise.
18747 (mov_<store>r): Likewise.
18748 (*branch_equality<mode>_inverted): Add microMIPS support.
18749 (*branch_equality<mode>): Likewise.
18750 (*jump_absolute): Likewise.
18751 (indirect_jump_<mode>): Likewise.
18752 (tablejump_<mode>): Likewise.
18753 (<optab>_internal): Likewise.
18754 (sibcall_internal): Likewise.
18755 (sibcall_value_internal): Likewise.
18756 (prefetch): Use constraint ZD.
18757 * config/mips/mips.opt (minterlink-compressed): New option.
18758 (minterlink-mips16): Now an alias for minterlink-compressed.
18759 (mmicromips): New option.
18760 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18761 (compare_and_swap_12): Likewise.
18762 (sync_add<mode>): Likewise.
18763 (sync_<optab>_12): Likewise.
18764 (sync_old_<optab>_12): Likewise.
18765 (sync_new_<optab>_12): Likewise.
18766 (sync_nand_12): Likewise.
18767 (sync_old_nand_12): Likewise.
18768 (sync_new_nand_12): Likewise.
18769 (sync_sub<mode>): Likewise.
18770 (sync_old_add<mode>): Likewise.
18771 (sync_old_sub<mode>): Likewise.
18772 (sync_new_add<mode>): Likewise.
18773 (sync_new_sub<mode>): Likewise.
18774 (sync_<optab><mode>): Likewise.
18775 (sync_old_<optab><mode>): Likewise.
18776 (sync_new_<optab><mode>): Likewise.
18777 (sync_nand<mode>): Likewise.
18778 (sync_old_nand<mode>): Likewise.
18779 (sync_new_nand<mode>): Likewise.
18780 (sync_lock_test_and_set<mode>): Likewise.
18781 (test_and_set_12): Likewise.
18782 (atomic_compare_and_swap<mode>): Likewise.
18783 (atomic_exchange<mode>_llsc): Likewise.
18784 (atomic_fetch_add<mode>_llsc): Likewise.
18785 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18786 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18787 (umips_save_restore_pattern_p): Likewise.
18788 (umips_load_store_pair_p): Likewise.
18789 (umips_output_load_store_pair): Likewise.
18790 (umips_movep_target_p): Likewise.
18791 (umips_12bit_offset_address_p): Likewise.
18792 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18793 (mips_base_mips16): Rename this...
18794 (mips_base_compression_flags): ...to this. Update all uses.
18795 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18796 (mips_mips16_decl_p): Delete.
18797 (mips_nomips16_decl_p): Delete.
18798 (mips_get_compress_on_flags): New function.
18799 (mips_get_compress_off_flags): New function.
18800 (mips_get_compress_mode): New function.
18801 (mips_get_compress_on_name): New function.
18802 (mips_get_compress_off_name): New function.
18803 (mips_insert_attributes): Support multiple compression types.
18804 (mips_merge_decl_attributes): Likewise.
18805 (umips_12bit_offset_address_p): New function.
18806 (mips_start_function_definition): Emit .set micromips directive.
18807 (mips_call_may_need_jalx_p): New function.
18808 (mips_function_ok_for_sibcall): Add microMIPS support.
18809 (mips_print_operand_punctuation): Support short delay slots and
18811 (umips_swm_mask, umips_swm_encoding): New.
18812 (umips_build_save_restore): New function.
18813 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18814 (was_mips16_p): Remove.
18815 (old_compression_mode): New.
18816 (mips_set_compression_mode): New function.
18817 (mips_set_current_function): Add microMIPS support.
18818 (mips_option_override): Likewise.
18819 (umips_save_restore_pattern_p): New function.
18820 (umips_output_save_restore): New function.
18821 (umips_load_store_pair_p_1): New function.
18822 (umips_load_store_pair_p): New function.
18823 (umips_output_load_store_pair_1): New function.
18824 (umips_output_load_store_pair): New function.
18825 (umips_movep_target_p) New function.
18826 (mips_prepare_pch_save): Add microMIPS support.
18827 * config/mips/mips.h (TARGET_COMPRESSION): New.
18828 (TARGET_CPU_CPP_BUILTINS): Update macro
18829 to use new compression flags and to support microMIPS.
18830 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18831 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18832 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18833 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18834 (ASM_SPEC): Support mmicromips and mno-micromips.
18835 (M16STORE_REG_P): New macro.
18836 (MIPS_CALL): Support TARGET_MICROMIPS.
18837 (MICROMIPS_J): New macro.
18838 (mips_base_mips16): Rename this...
18839 (mips_base_compression_flags): ...to this.
18840 (UMIPS_12BIT_OFFSET_P): New macro.
18841 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18842 (MULTILIB_DIRNAMES): Likewise.
18843 2013-03-20 Richard Biener <rguenther@suse.de>
18845 PR tree-optimization/56661
18846 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18847 the result does not have to be distinct.
18849 2013-03-20 Richard Biener <rguenther@suse.de>
18851 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
18854 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18855 Steven Bosscher <steven@gcc.gnu.org>
18857 PR rtl-optimization/56605
18858 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
18860 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18863 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
18864 that require movd instead of movq.
18866 2013-03-20 Richard Biener <rguenther@suse.de>
18868 * tree-ssa-structalias.c (struct variable_info): Add pointer
18869 to the first field of an aggregate with sub-vars. Make
18870 this and the pointer to the next subfield its ID.
18871 (vi_next): New function.
18872 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
18873 storedanything_id, integer_id): Increment by one.
18874 (new_var_info, get_call_vi, lookup_call_clobber_vi,
18875 get_call_clobber_vi): Adjust.
18876 (solution_set_expand): Simplify and speedup.
18877 (solution_set_add): Inline into ...
18878 (set_union_with_increment): ... this. Adjust accordingly.
18879 (do_sd_constraint): Likewise.
18880 (do_ds_constraint): Likewise.
18881 (do_complex_constraint): Simplify.
18882 (build_pred_graph): Adjust.
18883 (solve_graph): Likewise. Simplify and speedup.
18884 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
18885 get_constraint_for_component_ref, get_constraint_for_1,
18886 first_vi_for_offset, first_or_preceding_vi_for_offset,
18887 create_function_info_for, create_variable_info_for_1,
18888 create_variable_info_for, intra_create_variable_infos): Adjust.
18889 (init_base_vars): Push NULL for ID zero.
18890 (compute_points_to_sets): Adjust.
18892 2013-03-20 Richard Biener <rguenther@suse.de>
18894 * cfgloop.c (verify_loop_structure): Streamline and avoid
18895 ICEing on corrupt loop tree.
18896 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
18899 2013-03-20 Richard Biener <rguenther@suse.de>
18901 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
18902 check whether an SSA update is needed.
18904 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
18906 * config/mips/constraints.md (T): Rename to...
18910 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
18911 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
18913 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18915 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
18916 (*subsi3_carryin_uxtw): Likewise.
18918 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18920 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
18921 (*rorsi3_insn_uxtw): Likewise.
18923 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18925 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
18926 (*extrsi5_insn_uxtw): Likewise.
18928 2013-03-19 Richard Biener <rguenther@suse.de>
18930 PR tree-optimization/56273
18931 * passes.c (init_optimization_passes): Move second VRP after DOM.
18933 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
18935 * config/i386/i386.md (*movti_internal): Merge from
18936 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
18937 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
18938 nox64 isa attributes.
18940 2013-03-18 Richard Biener <rguenther@suse.de>
18942 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
18944 (merge_node_constraints): Likewise.
18945 (build_succ_graph): Likewise.
18946 (valid_graph_edge): Inline into single caller.
18947 (unify_nodes): Likewise. Use bitmap_set_bit return value
18949 (scc_visit): Fix formatting and variable use.
18950 (do_sd_constraint): Use gcc_checking_assert.
18951 (do_ds_constraint): Likewise.
18952 (do_complex_constraint): Likewise.
18953 (condense_visit): Likewise. Cleanup.
18954 (dump_pred_graph): New function.
18955 (perform_var_substitution): Dump the pred-graph before
18956 variable substitution.
18957 (find_equivalent_node): Use gcc_checking_assert.
18958 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
18960 2013-03-18 Richard Biener <rguenther@suse.de>
18962 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
18963 Remove cond_expr_stmt_list argument and do not gimplify the
18965 (vect_loop_versioning): Adjust.
18966 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
18967 Cleanup to use less temporaries.
18968 (vect_create_data_ref_ptr): Cleanup.
18970 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18972 PR tree-optimization/56635
18973 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
18974 require types_compatible_p types.
18976 2013-03-18 Nick Clifton <nickc@redhat.com>
18978 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
18979 spurious backslash.
18981 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
18982 Add missing line to comment describing function.
18984 2013-03-18 Richard Biener <rguenther@suse.de>
18986 PR tree-optimization/56210
18987 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18988 Handle string / character search functions.
18989 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
18991 2013-03-18 Richard Biener <rguenther@suse.de>
18993 PR middle-end/56483
18994 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
18995 and implement properly.
18996 * gimple.h (gimple_cond_single_var_p): Remove.
18998 2013-03-18 Richard Biener <rguenther@suse.de>
19000 * tree-data-ref.h (find_data_references_in_loop): Declare.
19001 * tree-data-ref.c (get_references_in_stmt): Use a stack
19002 vector pre-allocated in the callers.
19003 (find_data_references_in_stmt): Adjust.
19004 (graphite_find_data_references_in_stmt): Likewise.
19005 (create_rdg_vertices): Likewise.
19006 (find_data_references_in_loop): Export.
19007 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19008 Compute dependences here...
19009 (vect_analyze_data_refs): ...not here. When we encounter
19010 a non-vectorizable data reference in basic-block vectorization
19011 truncate the data reference vector. Do not bother to
19012 fixup data-dependence information for gather loads.
19013 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19014 of data references, as reported.
19016 2013-03-18 Richard Biener <rguenther@suse.de>
19018 PR tree-optimization/3713
19019 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19020 has_constants and expr.
19021 (stmt_has_constants): Properly valueize SSA names when deciding
19022 whether the stmt has constants.
19024 2013-03-18 Richard Biener <rguenther@suse.de>
19026 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19027 whole function when there is nothing to do.
19028 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19029 * tree-vectorizer.c (vectorize_loops): Update virtual and
19030 loop-closed SSA once.
19031 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19033 2013-03-18 Richard Biener <rguenther@suse.de>
19035 PR middle-end/56113
19036 * domwalk.c (bb_postorder): New global static.
19037 (cmp_bb_postorder): New function.
19038 (walk_dominator_tree): Replace scheme imposing an order for
19039 visiting dominator sons by one sorting them at the time they
19040 are pushed on the stack.
19042 2013-03-18 Richard Biener <rguenther@suse.de>
19044 PR tree-optimization/39326
19045 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19046 (struct mem_ref): Replace mem member with ao_ref typed member.
19047 (MEM_ANALYZABLE): Adjust.
19048 (memref_eq): Likewise.
19049 (mem_ref_alloc): Likewise.
19050 (gather_mem_refs_stmt): Likewise.
19051 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19052 (execute_sm_if_changed_flag_set): Adjust.
19053 (execute_sm): Likewise.
19054 (ref_always_accessed_p): Likewise.
19055 (refs_independent_p): Likewise.
19056 (can_sm_ref_p): Likewise.
19058 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19061 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19062 return 1 even for !unsignedp.
19064 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19066 * config/i386/i386.md (isa): Add x64 and nox64.
19067 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19068 (*pushtf): Enable *roF alternative for x64 isa only.
19069 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19070 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19071 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19072 (*movtf_internal): Merge from *movtf_internal_rex64 and
19073 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19074 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19075 nox64 isa attributes.
19076 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19077 nox64 isa attributes.
19078 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19080 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19082 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19084 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19087 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19088 class names. Remove trailing comma after "ALL_REGS".
19090 2013-03-16 Jan Hubicka <jh@suse.cz>
19092 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19093 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19094 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19095 of cgraph_get_create_node.
19096 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19098 2013-03-16 Jason Merrill <jason@redhat.com>
19101 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19102 with DW_AT_default_value.
19104 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19106 * BASE-VER: Set to 4.9.0.
19108 2013-03-14 Andi Kleen <ak@linux.intel.com>
19111 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19112 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19113 Document _x* TSX intrinsics.
19115 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19116 David Holsgrove <david.holsgrove@xilinx.com>
19118 * configure.ac: Add MicroBlaze TLS support detection.
19119 * configure: Regenerate.
19120 * config/microblaze/microblaze-protos.h
19121 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19122 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19123 * config/microblaze/microblaze.c (microblaze_address_type): Add
19124 ADDRESS_TLS and tls_reloc address types.
19125 (microblaze_address_info): Add tls_reloc.
19126 (TARGET_HAVE_TLS): Define.
19127 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19128 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19129 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19130 load_tls_operand, microblaze_call_tls_get_addr,
19131 microblaze_legitimize_tls_address): New functions.
19132 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19133 (get_base_reg): Use microblaze_tls_symbol_p.
19134 (microblaze_classify_address): Handle TLS.
19135 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19136 label_mentioned_p and microblaze_tls_referenced_p.
19137 (microblaze_legitimize_address): Handle TLS.
19138 (microblaze_address_insns): Handle ADDRESS_TLS.
19139 (pic_address_needs_scratch): Handle TLS.
19140 (print_operand_address): Handle TLS.
19141 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19142 (microblaze_expand_move): Handle TLS.
19143 (microblaze_legitimate_constant_p): Check
19144 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19145 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19146 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19147 (PIC_OFFSET_TABLE_REGNUM): Set.
19148 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19149 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19150 (addsi3, movsi_internal2, movdf_internal): Update constraints
19151 * config/microblaze/predicates.md (arith_plus_operand): Define
19152 (move_operand): Redefine as move_src_operand,
19153 check microblaze_tls_referenced_p.
19155 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19157 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19158 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19160 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19162 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19165 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19167 PR tree-optimization/53265
19168 * common.opt (Waggressive-loop-optimizations): New option.
19169 * tree-ssa-loop-niter.c: Include tree-pass.h.
19170 (do_warn_aggressive_loop_optimizations): New function.
19171 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19172 if number_of_latch_executions returned constant.
19173 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19174 early. If number_of_latch_executions returned constant, set
19175 nb_iterations_upper_bound back to it.
19176 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19178 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19179 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19181 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19182 (MULTILIB_OSDIRNAMES): Set.
19183 * genmultilib: If defaultosdirname doesn't start with :: , set
19184 defaultosdirname2 instead, clear it and emit two . multilib_raw
19185 entries instead of just one.
19187 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19189 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19190 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19191 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19192 (SUBTARGET_OVERRIDE_OPTIONS): New.
19194 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19197 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19198 (musermode): Convert to Var(TARGET_USERMODE).
19199 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19200 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19201 * config/sh/sh.c (sh_option_override): Use
19202 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19203 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19205 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19207 (udivsi3_i4_single, divsi3_i4_single): Use
19208 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19210 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19212 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19215 2013-03-13 Richard Biener <rguenther@suse.de>
19217 PR tree-optimization/56608
19218 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19219 calls when vectorizing basic-blocks.
19221 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19224 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19227 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19229 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19231 2013-03-11 Jan Hubicka <jh@suse.cz>
19234 * lto-streamer-out.c (output_symbol_p): Skip references from
19235 constructors of external variables.
19237 2013-03-11 Jan Hubicka <jh@suse.cz>
19239 PR middle-end/56571
19240 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19242 * emit-rtl.c (verify_rtx_sharing): Likewise.
19243 (copy_insn_1): Likewise.
19244 * rtl.c (copy_rtx): Likewise.
19246 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19249 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19250 output_operand_lossage message.
19252 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19255 * arm.c (shift_op): Validate RTL pattern on the fly.
19256 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19259 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19262 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19263 new call patterns that clobber %fr12.
19264 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19265 split and postreload patterns.
19266 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19267 registers %fr12 and %fr12R as call used.
19269 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19271 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19272 (canon_address, record_store, replace_read, check_mem_read_rtx,
19273 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19274 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19275 rest_of_handle_dse): Likewise.
19277 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19279 PR middle-end/56524
19280 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19282 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19283 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19284 (save_optabs_if_changed): Replace with...
19285 (init_tree_optimization_optabs): ...this.
19286 * optabs.c (save_optabs_if_changed): Rename to...
19287 (init_tree_optimization_optabs): ...this. Take the optimization node
19288 as argument. Do nothing if the base optabs are already correct.
19289 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19290 to recompute optabs.
19291 * function.h (function): Remove optabs field.
19292 * function.c (invoke_set_current_function_hook): Call
19293 init_tree_optimization_optabs. Use the result to initialize
19296 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19298 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19299 if GTMA_HAS_NO_INSTRUMENTATION.
19300 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19301 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19302 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19303 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19304 GTMA_HAS_NO_INSTRUMENTATION.
19306 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19308 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19311 2013-03-08 Marek Polacek <polacek@redhat.com>
19312 Jakub Jelinek <jakub@redhat.com>
19314 PR tree-optimization/56478
19315 * predict.c (is_comparison_with_loop_invariant_p): Change the
19316 type of loop_step to tree.
19317 (predict_loops): Adjust.
19318 (predict_iv_comparison): Perform the computations on double_ints.
19320 2013-03-08 Richard Biener <rguenther@suse.de>
19322 PR tree-optimization/56570
19323 * tree-cfg.c (verify_expr_location_1): Verify locations for
19325 * tree-sra.c (create_access_replacement): Strip locations
19326 from DECL_DEBUG_EXPRs.
19328 2013-03-08 Richard Biener <rguenther@suse.de>
19330 * tree-inline.c (expand_call_inline): Do not associate
19331 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19332 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19334 2013-03-08 Richard Biener <rguenther@suse.de>
19336 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19337 or block changes with -Og. Fix for location / block encoding
19338 changes and PHI arguments with locations.
19340 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19342 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19344 (struct output_info): Likewise.
19345 (register_overhead): Remove bad gcc_assert.
19346 (bitmap_find_bit): If there is only a single bitmap element, do not
19347 count a miss as a search.
19348 (print_statistics): Update for counter type changes.
19349 (dump_bitmap_statistics): Likewise. Print headers such that they
19350 are properly lined up with the printed counters.
19352 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19354 PR tree-optimization/56559
19355 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19356 check that it has only a single use.
19358 2013-03-07 Richard Biener <rguenther@suse.de>
19360 * doc/invoke.texi (fwhole-program): Discourage use in combination
19363 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19365 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19367 PR tree-optimization/56539
19368 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19369 instead of GSI_CONTINUE_LINKING as last argument to
19370 force_gimple_operand_gsi. Adjust function comment.
19372 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19375 PR middle-end/56548
19376 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19377 promoted mode, convert the result back to the original mode.
19379 2013-03-06 Richard Biener <rguenther@suse.de>
19381 PR middle-end/56294
19382 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19383 (insert_updated_phi_nodes_compare_uids): New function.
19384 (update_ssa): Sort symbols_to_rename after UID before
19385 traversing it to insert PHI nodes.
19387 2013-03-06 Richard Biener <rguenther@suse.de>
19389 PR middle-end/50494
19390 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19391 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19394 2013-02-13 Richard Biener <rguenther@suse.de>
19397 * varasm.c (output_constant_def_1): Get the decl representing
19398 the constant as argument.
19399 (output_constant_def): Wrap output_constant_def_1.
19400 (make_decl_rtl): Use output_constant_def_1 with the decl
19401 representing the constant.
19402 (build_constant_desc): Optionally re-use a decl already
19403 representing the constant.
19404 (tree_output_constant_def): Adjust.
19406 2013-03-06 Joey Ye <joey.ye@arm.com>
19409 * gcc.c (convert_white_space): New function.
19410 (main): Handles white space in function name.
19412 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19415 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19416 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19417 to SH_DIV_CALL_TABLE for TARGET_SH2.
19418 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19420 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19421 call-table options.
19423 2013-03-05 Sterling Augustine <saugustine@google.com>
19424 Cary Coutant <ccoutant@google.com>
19427 * dwarf2out.c (resolve_addr): Don't call
19428 remove_loc_list_addr_table_entries a second time for the same
19431 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19434 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19435 (avoid_complex_debug_insns): New function.
19436 (expand_debug_locations): Call it.
19438 PR rtl-optimization/56484
19439 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19440 lifetimes of hard registers on small register class machines.
19442 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19444 * config/microblaze/microblaze-protos.h: Rename
19445 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19446 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19448 (microblaze_fast_interrupt_function_p): New function.
19449 (microblaze_is_interrupt_handler): Rename to
19450 microblaze_is_interrupt_variant and add fast_interrupt check.
19451 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19452 (save_restore_insns): Likewise.
19453 (compute_frame_size): Likewise.
19454 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19455 (microblaze_globalize_label): Likewise.
19456 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19457 * config/microblaze/microblaze.md: Use wrapper
19458 microblaze_is_interrupt_variant.
19460 2013-03-05 Kai Tietz <ktietz@redhat.com>
19462 * sdbout.c (sdbout_one_type): Switch to current function's section
19463 supporting cold/hot.
19465 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19467 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19470 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19472 PR middle-end/56461
19473 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19474 if VALGRIND_GET_VBITS is defined, temporarily make object
19475 memory all defined, and restore previous valgrind addressability
19476 and definability afterwards. Free this_object at the end.
19478 PR middle-end/56461
19479 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19480 right before calling lra_create_live_ranges, also call it
19481 when clearing live_p. Only call lra_clear_live_ranges
19482 at the end if live_p.
19484 PR middle-end/56461
19485 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19487 2013-03-05 Richard Biener <rguenther@suse.de>
19489 PR tree-optimization/56521
19490 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19493 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19496 * except.h (remove_unreachable_eh_regions): New prototype.
19497 * except.c (remove_eh_handler_splicer): New function, split out
19498 of remove_eh_handler.
19499 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19500 warning about running it on many EH regions one at a time.
19501 (remove_unreachable_eh_regions_worker): New function, walk the
19502 EH tree in depth-first order and remove non-marked regions.
19503 (remove_unreachable_eh_regions): New function.
19504 * tree-eh.c (mark_reachable_handlers): New function, split out
19505 from remove_unreachable_handlers.
19506 (remove_unreachable_handlers): Use mark_reachable_handlers and
19507 remove_unreachable_eh_regions.
19508 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19509 and remove_unreachable_eh_regions.
19511 2013-03-05 Richard Biener <rguenther@suse.de>
19513 PR middle-end/56525
19514 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19515 not freeing them until the end.
19517 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19519 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19521 2013-03-05 Richard Biener <rguenther@suse.de>
19523 PR tree-optimization/56270
19524 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19525 of loads after scheduling an SLP instance.
19527 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19529 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19531 (check_gcc_parallelize): Run guality.exp as a separate job from
19532 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19533 struct-layout-1.exp with stackalign.exp.
19535 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19537 PR middle-end/56461
19538 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19539 load_index sbitmap even if some bit in it isn't set.
19541 PR middle-end/56461
19542 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19543 (discover_iteration_bound_by_body_walk): Change queues to
19544 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19545 spelling in comment. Call safe_push on queues[bound_index] directly.
19546 Release queues[queue_index] in every iteration unconditionally.
19547 Release bounds vector.
19549 PR middle-end/56461
19550 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19551 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19552 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19555 2013-03-05 Richard Biener <rguenther@suse.de>
19558 * tree-inline.c (remap_blocks_to_null): New function.
19559 (expand_call_inline): When expanding a call stmt without
19560 an associated block inline remap all callee blocks to NULL.
19562 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19564 PR rtl-optimization/56494
19565 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19566 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19567 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19569 PR middle-end/56461
19570 * sel-sched-ir.c (free_sched_pools): Release
19571 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19572 if succs_info_pool.max_top isn't -1.
19575 * opts.c (opts_obstack, opts_concat): Moved to...
19576 * opts-common.c (opts_obstack, opts_concat): ... here.
19578 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19580 PR middle-end/56461
19581 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19583 2013-03-04 Martin Jambor <mjambor@suse.cz>
19585 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19586 all appropriate places.
19588 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19590 PR tree-optimization/56424
19591 * ipa-split.c (split_function): Do not set the RSO flag if result is
19592 not by reference and its type is a register type.
19594 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19596 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19597 (microblaze_legitimate_pic_operand): Likewise
19598 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19599 new function microblaze_legitimate_pic_operand
19600 * config/microblaze/microblaze-protos.h
19601 (microblaze_legitimate_pic_operand): Declare.
19603 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19605 * config/microblaze/predicates.md (call_insn_simple_operand):
19606 New predicate for supported rtx code types.
19607 * config/microblaze/microblaze.md (call_internal1): Use
19608 call_insn_simple_operand predicate.
19610 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19612 PR middle-end/56461
19613 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19614 partitions.ordered_remove.
19616 PR middle-end/56461
19617 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19618 vec_oprnds0.create (1) for modifier == NONE.
19620 PR middle-end/56461
19621 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19622 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19623 vec_oprnds1 right before pushing anything to it for
19626 PR middle-end/56461
19627 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19628 set nbbs to 0 instead of having separate code path.
19629 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19630 instead of false as last argument if returning NULL.
19632 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19634 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19635 the attribute is now called "target" instead of "option".
19636 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19637 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19638 attribute/pragma name for TARGET_OPTION_VALID_P and
19639 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19640 * doc/tm.texi: Regenerated.
19642 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19644 * config/microblaze/microblaze.c:
19645 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19646 * config/microblaze/microblaze.h: Add -mxl-reorder to
19648 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19649 instructions emitted if TARGET_REORDER.
19650 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19651 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19654 2013-03-01 Xinliang David Li <davidxl@google.com>
19656 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19659 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19661 PR middle-end/56461
19662 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19663 vector even when returning true. Fix up function comment formatting.
19665 PR middle-end/56461
19666 * ira-build.c (ira_loop_nodes_count): New variable.
19667 (create_loop_tree_nodes): Initialize it.
19668 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19670 PR middle-end/56461
19671 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19672 method on dr_chain and result_chain.
19673 * tree-vect-stmts.c (vectorizable_store): Only call
19674 result_chain.create if j == 0.
19676 PR middle-end/56461
19677 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19678 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19679 before overwriting it.
19681 2013-03-01 Tobias Burnus <burnus@net-b.de>
19683 * doc/extended.texi (C Extensions): Change order in @menu
19685 (Other MIPS Built-in Functions): Move last MIPS entry before
19686 "picoChip Built-in Functions".
19687 (SH Built-in Functions): Move after RX Built-in Functions.
19688 * doc/gcc.texi (Introduction): Change order in @menu
19690 * doc/md.texi (Constraints): Ditto.
19691 * gty.texi (Type Information): Ditto.
19692 (User-provided marking routines for template types): Make
19694 * doc/invoke.texi (AArch64 Options): Move before
19695 "Adapteva Epiphany Options".
19697 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19698 Jakub Jelinek <jakub@redhat.com>
19701 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19702 no_address_safety_analysis attribute.
19703 * doc/extend.texi (no_address_safety_attribute): Rename to
19704 no_sanitize_address attribute, mention no_address_safety_analysis
19705 attribute as deprecated alias.
19707 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19709 PR middle-end/56461
19710 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19711 type to vec<vec<tree> > *.
19712 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19713 to be vec<tree> instead of vec<tree> *, set vec_defs
19714 to vNULL and call vec_defs.create (number_of_vects), adjust other
19716 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19717 vectorizable_condition): Adjust vect_get_slp_defs callers.
19719 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19721 * config/aarch64/aarch64.c
19722 (aarch64_float_const_representable): Remove unused variable.
19724 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19726 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19728 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19730 * config/aarch64/aarch64-builtins.c
19731 (aarch64_init_simd_builtins): Make static.
19733 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19735 * config/aarch64/aarch64.c
19736 (aarch64_simd_make_constant): Make static.
19738 2013-02-28 Martin Jambor <mjambor@suse.cz>
19740 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19741 with no initialization to the RHS of debug statements.
19743 2013-02-28 Martin Jambor <mjambor@suse.cz>
19745 PR tree-optimization/56294
19746 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19748 (get_access_replacement): Do not call create_access_replacement.
19749 Assert a replacement exists.
19750 (get_repl_default_def_ssa_name): Create the replacement declaration
19753 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19755 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19756 final_end_function.
19758 2013-02-28 Marek Polacek <polacek@redhat.com>
19760 PR rtl-optimization/56466
19761 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19762 if we're changing a loop.
19763 (peel_loops_completely): Likewise.
19765 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19768 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19770 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19773 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19774 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19775 INTX_FTYPE_FX, FX_FTYPE_INTX.
19776 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19778 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19780 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19781 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19782 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19783 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19784 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19785 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19786 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19787 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19788 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19789 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19790 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19791 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19792 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19793 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19794 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19795 (avrxmega6): Increase max flash segments from 5 to 6.
19796 * config/avr/t-multilib: Regenerate.
19797 * config/avr/avr-tables.opt: Regenerate.
19798 * doc/avr-mmcu.texi: Regenerate.
19800 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19802 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19803 (avr_device_to_arch): Rename to avr_device_to_ld.
19804 (avr_device_to_as): New prototype.
19805 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19806 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19807 * config/avr/driver-avr.c (avr_device_to_as): New.
19808 (avr_device_to_arch): Rename to avr_device_to_ld.
19810 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19812 PR middle-end/56461
19813 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19814 method on dr_chain and result_chain.
19816 PR middle-end/56461
19817 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19818 pointer_set_destroy on not_executed_last_iteration.
19820 PR middle-end/56461
19821 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19823 PR middle-end/56461
19824 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19825 FOR_EACH_DEFINED_FUNCTION when freeing state.
19827 PR middle-end/56461
19828 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19830 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19833 PR middle-end/56461
19834 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19835 known_aggs[i].items and release known_aggs vector.
19837 PR middle-end/56461
19838 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19840 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19842 * config/microblaze/microblaze.c (microblaze_emit_compare):
19843 Use xor for EQ/NE comparisions.
19844 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19845 (cbranchsf4): Adjust operator to comparison_operator.
19847 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19849 PR middle-end/56461
19850 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
19852 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
19853 vec_safe_push, always update *slot.
19854 (redirect_edge_var_map_clear): Use vec_free.
19855 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
19856 (free_var_map_entry): Use vec_free.
19857 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
19858 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
19860 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
19862 PR middle-end/45472
19863 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
19864 when the may_trap_p bit of the exprs being merged differs.
19865 Reorder tests for speculativeness in the logical and operator.
19867 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19869 * incpath.c (add_standard_paths): Use reconcat instead of concat
19870 where appropriate and avoid leaking memory.
19872 * opts.h: Include obstack.h.
19873 (opts_concat): New prototype.
19874 (opts_obstack): New declaration.
19875 * opts.c (opts_concat): New function.
19876 (opts_obstack): New variable.
19877 (init_options_struct): Call gcc_init_obstack on opts_obstack.
19878 (finish_options): Use opts_concat instead of concat
19879 and XOBNEWVEC instead of XNEWVEC.
19880 * opts-common.c (generate_canonical_option, decode_cmdline_option,
19881 generate_option): Likewise.
19882 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
19883 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
19886 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
19887 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
19889 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19891 PR middle-end/56461
19892 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
19894 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19896 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
19897 (arm_block_move_unaligned_straight): Likewise.
19898 (arm_adjust_block_mem): Likewise.
19900 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19903 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
19904 temp, cond and label.
19905 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
19908 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
19909 * config/c6x/c6x.h (dbx_register_map): Update declaration.
19912 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
19913 of prologue/epilogue functions.
19916 * config/tilegx/tilegx.c (tilegx_expand_prologue):
19917 Remove unused variable cfa_offset.
19918 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
19921 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
19922 type promotion to unsigned.
19925 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
19926 for HOST_WIDE_INT of 32 bit / same size as int.
19927 (arm_block_move_unaligned_straight): Likewise.
19928 (arm_adjust_block_mem): Likewise.
19931 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
19934 2013-02-26 Marek Polacek <polacek@redhat.com>
19936 PR tree-optimization/56426
19937 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
19939 2013-02-26 Richard Biener <rguenther@suse.de>
19942 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
19943 unused variable loops.
19945 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19947 PR tree-optimization/56448
19948 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
19949 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
19950 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
19951 later operands of the references, or even first operand for
19952 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
19954 PR tree-optimization/56443
19955 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
19956 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
19957 to type_for_mode langhook.
19959 2013-02-25 Matt Turner <mattst88@gmail.com>
19961 * doc/invoke.texi: Document r4700.
19963 2013-02-25 Richard Biener <rguenther@suse.de>
19965 PR tree-optimization/56175
19966 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
19968 (simplify_bitwise_binary): ... here. Also guard the conversion
19969 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
19971 2013-02-25 Catherine Moore <clm@codesourcery.com>
19974 2013-02-24 Catherine Moore <clm@codesourcery.com>
19975 Maciej W. Rozycki <macro@codesourcery.com>
19976 Tom de Vries <tom@codesourcery.com>
19977 Nathan Sidwell <nathan@codesourcery.com>
19978 Iain Sandoe <iain@codesourcery.com>
19979 Nathan Froyd <froydnj@codesourcery.com>
19980 Chao-ying Fu <fu@mips.com>
19982 * doc/extend.texi: (micromips, nomicromips, nocompression):
19983 Document new function attributes.
19984 * doc/invoke.texi (minterlink-compressed, mmicromips,
19985 m14k, m14ke, m14kec): Document new options.
19986 (minterlink-mips16): Update documentation.
19987 * doc/md.texi (ZC, ZD): Document new constraints.
19988 * configure.ac (gcc_cv_as_micromips): Check if linker
19989 supports the .set micromips directive.
19990 * configure: Regenerate.
19991 * config.in: Regenerate.
19992 * config/mips/mips-tables.opt: Regenerate.
19993 * config/mips/micromips.md: New file.
19994 * constraints.md (ZC, AD): New constraints.
19995 * config/mips/predicates.md (movep_src_register): New predicate.
19996 (movep_src_operand): New predicate.
19997 (non_volatile_mem_operand): New predicate.
19998 * config/mips/mips.md (multimem): New type.
19999 (length): Differentiate between 17-bit and 18-bit branch offsets.
20000 (MOVEP1, MOVEP2): New mode iterator.
20001 (mov_<load>l): Use ZC constraint.
20002 (mov_<load>r): Likewise.
20003 (mov_<store>l): Likewise.
20004 (mov_<store>r): Likewise.
20005 (*branch_equality<mode>_inverted): Add microMIPS support.
20006 (*branch_equality<mode>): Likewise.
20007 (*jump_absolute): Likewise.
20008 (indirect_jump_<mode>): Likewise.
20009 (tablejump_<mode>): Likewise.
20010 (<optab>_internal): Likewise.
20011 (sibcall_internal): Likewise.
20012 (sibcall_value_internal): Likewise.
20013 (prefetch): Use constraint ZD.
20014 * config/mips/mips.opt (minterlink-compressed): New option.
20015 (minterlink-mips16): Now an alias for minterlink-compressed.
20016 (mmicromips): New option.
20017 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20018 (compare_and_swap_12): Likewise.
20019 (sync_add<mode>): Likewise.
20020 (sync_<optab>_12): Likewise.
20021 (sync_old_<optab>_12): Likewise.
20022 (sync_new_<optab>_12): Likewise.
20023 (sync_nand_12): Likewise.
20024 (sync_old_nand_12): Likewise.
20025 (sync_new_nand_12): Likewise.
20026 (sync_sub<mode>): Likewise.
20027 (sync_old_add<mode>): Likewise.
20028 (sync_old_sub<mode>): Likewise.
20029 (sync_new_add<mode>): Likewise.
20030 (sync_new_sub<mode>): Likewise.
20031 (sync_<optab><mode>): Likewise.
20032 (sync_old_<optab><mode>): Likewise.
20033 (sync_new_<optab><mode>): Likewise.
20034 (sync_nand<mode>): Likewise.
20035 (sync_old_nand<mode>): Likewise.
20036 (sync_new_nand<mode>): Likewise.
20037 (sync_lock_test_and_set<mode>): Likewise.
20038 (test_and_set_12): Likewise.
20039 (atomic_compare_and_swap<mode>): Likewise.
20040 (atomic_exchange<mode>_llsc): Likewise.
20041 (atomic_fetch_add<mode>_llsc): Likewise.
20042 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20043 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20044 (umips_save_restore_pattern_p): Likewise.
20045 (umips_load_store_pair_p): Likewise.
20046 (umips_output_load_store_pair): Likewise.
20047 (umips_movep_target_p): Likewise.
20048 (umips_12bit_offset_address_p): Likewise.
20049 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20050 (mips_base_mips16): Rename this...
20051 (mips_base_compression_flags): ...to this. Update all uses.
20052 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20053 (mips_mips16_decl_p): Delete.
20054 (mips_nomips16_decl_p): Delete.
20055 (mips_get_compress_on_flags): New function.
20056 (mips_get_compress_off_flags): New function.
20057 (mips_get_compress_mode): New function.
20058 (mips_get_compress_on_name): New function.
20059 (mips_get_compress_off_name): New function.
20060 (mips_insert_attributes): Support multiple compression types.
20061 (mips_merge_decl_attributes): Likewise.
20062 (umips_12bit_offset_address_p): New function.
20063 (mips_start_function_definition): Emit .set micromips directive.
20064 (mips_call_may_need_jalx_p): New function.
20065 (mips_function_ok_for_sibcall): Add microMIPS support.
20066 (mips_print_operand_punctuation): Support short delay slots and
20068 (umips_swm_mask, umips_swm_encoding): New.
20069 (umips_build_save_restore): New function.
20070 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20071 (was_mips16_p): Remove.
20072 (old_compression_mode): New.
20073 (mips_set_compression_mode): New function.
20074 (mips_set_current_function): Add microMIPS support.
20075 (mips_option_override): Likewise.
20076 (umips_save_restore_pattern_p): New function.
20077 (umips_output_save_restore): New function.
20078 (umips_load_store_pair_p_1): New function.
20079 (umips_load_store_pair_p): New function.
20080 (umips_output_load_store_pair_1): New function.
20081 (umips_output_load_store_pair): New function.
20082 (umips_movep_target_p) New function.
20083 (mips_prepare_pch_save): Add microMIPS support.
20084 * config/mips/mips.h (TARGET_COMPRESSION): New.
20085 (TARGET_CPU_CPP_BUILTINS): Update macro
20086 to use new compression flags and to support microMIPS.
20087 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20088 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20089 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20090 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20091 (ASM_SPEC): Support mmicromips and mno-micromips.
20092 (M16STORE_REG_P): New macro.
20093 (MIPS_CALL): Support TARGET_MICROMIPS.
20094 (MICROMIPS_J): New macro.
20095 (mips_base_mips16): Rename this...
20096 (mips_base_compression_flags): ...to this.
20097 (UMIPS_12BIT_OFFSET_P): New macro.
20098 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20099 (MULTILIB_DIRNAMES): Likewise.
20101 2013-02-25 Tom de Vries <tom@codesourcery.com>
20103 PR rtl-optimization/56131
20104 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20105 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20106 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20108 2013-02-25 Tobias Burnus <burnus@net-b.de>
20110 * doc/invoke.texi (-fsanitize=): Move from optimization
20111 to debugging options.
20113 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20115 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20117 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20118 Alexander Monakov <amonakov@ispras.ru>
20120 PR middle-end/56077
20121 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20122 flush pending lists also on non-jumps. Adjust comment.
20124 2013-02-24 Catherine Moore <clm@codesourcery.com>
20125 Maciej W. Rozycki <macro@codesourcery.com>
20126 Tom de Vries <tom@codesourcery.com>
20127 Nathan Sidwell <nathan@codesourcery.com>
20128 Iain Sandoe <iain@codesourcery.com>
20129 Nathan Froyd <froydnj@codesourcery.com>
20130 Chao-ying Fu <fu@mips.com>
20132 * doc/extend.texi: (micromips, nomicromips, nocompression):
20133 Document new function attributes.
20134 * doc/invoke.texi (minterlink-compressed, mmicromips,
20135 m14k, m14ke, m14kec): Document new options.
20136 (minterlink-mips16): Update documentation.
20137 * doc/md.texi (ZC, ZD): Document new constraints.
20138 * configure.ac (gcc_cv_as_micromips): Check if linker
20139 supports the .set micromips directive.
20140 * configure: Regenerate.
20141 * config.in: Regenerate.
20142 * config/mips/mips-tables.opt: Regenerate.
20143 * config/mips/micromips.md: New file.
20144 * constraints.md (ZC, AD): New constraints.
20145 * config/mips/predicates.md (movep_src_register): New predicate.
20146 (movep_src_operand): New predicate.
20147 (non_volatile_mem_operand): New predicate.
20148 * config/mips/mips.md (multimem): New type.
20149 (length): Differentiate between 17-bit and 18-bit branch offsets.
20150 (MOVEP1, MOVEP2): New mode iterator.
20151 (mov_<load>l): Use ZC constraint.
20152 (mov_<load>r): Likewise.
20153 (mov_<store>l): Likewise.
20154 (mov_<store>r): Likewise.
20155 (*branch_equality<mode>_inverted): Add microMIPS support.
20156 (*branch_equality<mode>): Likewise.
20157 (*jump_absolute): Likewise.
20158 (indirect_jump_<mode>): Likewise.
20159 (tablejump_<mode>): Likewise.
20160 (<optab>_internal): Likewise.
20161 (sibcall_internal): Likewise.
20162 (sibcall_value_internal): Likewise.
20163 (prefetch): Use constraint ZD.
20164 * config/mips/mips.opt (minterlink-compressed): New option.
20165 (minterlink-mips16): Now an alias for minterlink-compressed.
20166 (mmicromips): New option.
20167 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20168 (compare_and_swap_12): Likewise.
20169 (sync_add<mode>): Likewise.
20170 (sync_<optab>_12): Likewise.
20171 (sync_old_<optab>_12): Likewise.
20172 (sync_new_<optab>_12): Likewise.
20173 (sync_nand_12): Likewise.
20174 (sync_old_nand_12): Likewise.
20175 (sync_new_nand_12): Likewise.
20176 (sync_sub<mode>): Likewise.
20177 (sync_old_add<mode>): Likewise.
20178 (sync_old_sub<mode>): Likewise.
20179 (sync_new_add<mode>): Likewise.
20180 (sync_new_sub<mode>): Likewise.
20181 (sync_<optab><mode>): Likewise.
20182 (sync_old_<optab><mode>): Likewise.
20183 (sync_new_<optab><mode>): Likewise.
20184 (sync_nand<mode>): Likewise.
20185 (sync_old_nand<mode>): Likewise.
20186 (sync_new_nand<mode>): Likewise.
20187 (sync_lock_test_and_set<mode>): Likewise.
20188 (test_and_set_12): Likewise.
20189 (atomic_compare_and_swap<mode>): Likewise.
20190 (atomic_exchange<mode>_llsc): Likewise.
20191 (atomic_fetch_add<mode>_llsc): Likewise.
20192 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20193 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20194 (umips_save_restore_pattern_p): Likewise.
20195 (umips_load_store_pair_p): Likewise.
20196 (umips_output_load_store_pair): Likewise.
20197 (umips_movep_target_p): Likewise.
20198 (umips_12bit_offset_address_p): Likewise.
20199 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20200 (mips_base_mips16): Rename this...
20201 (mips_base_compression_flags): ...to this. Update all uses.
20202 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20203 (mips_mips16_decl_p): Delete.
20204 (mips_nomips16_decl_p): Delete.
20205 (mips_get_compress_on_flags): New function.
20206 (mips_get_compress_off_flags): New function.
20207 (mips_get_compress_mode): New function.
20208 (mips_get_compress_on_name): New function.
20209 (mips_get_compress_off_name): New function.
20210 (mips_insert_attributes): Support multiple compression types.
20211 (mips_merge_decl_attributes): Likewise.
20212 (umips_12bit_offset_address_p): New function.
20213 (mips_start_function_definition): Emit .set micromips directive.
20214 (mips_call_may_need_jalx_p): New function.
20215 (mips_function_ok_for_sibcall): Add microMIPS support.
20216 (mips_print_operand_punctuation): Support short delay slots and
20218 (umips_swm_mask, umips_swm_encoding): New.
20219 (umips_build_save_restore): New function.
20220 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20221 (was_mips16_p): Remove.
20222 (old_compression_mode): New.
20223 (mips_set_compression_mode): New function.
20224 (mips_set_current_function): Add microMIPS support.
20225 (mips_option_override): Likewise.
20226 (umips_save_restore_pattern_p): New function.
20227 (umips_output_save_restore): New function.
20228 (umips_load_store_pair_p_1): New function.
20229 (umips_load_store_pair_p): New function.
20230 (umips_output_load_store_pair_1): New function.
20231 (umips_output_load_store_pair): New function.
20232 (umips_movep_target_p) New function.
20233 (mips_prepare_pch_save): Add microMIPS support.
20234 * config/mips/mips.h (TARGET_COMPRESSION): New.
20235 (TARGET_CPU_CPP_BUILTINS): Update macro
20236 to use new compression flags and to support microMIPS.
20237 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20238 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20239 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20240 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20241 (ASM_SPEC): Support mmicromips and mno-micromips.
20242 (M16STORE_REG_P): New macro.
20243 (MIPS_CALL): Support TARGET_MICROMIPS.
20244 (MICROMIPS_J): New macro.
20245 (mips_base_mips16): Rename this...
20246 (mips_base_compression_flags): ...to this.
20247 (UMIPS_12BIT_OFFSET_P): New macro.
20248 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20249 (MULTILIB_DIRNAMES): Likewise.
20251 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20254 * target-globals.c (save_target_globals): For init_reg_sets and
20255 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20257 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20259 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20260 * config/aarch64/t-aarch64
20261 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20263 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20265 PR inline-asm/56148
20266 * lra-constraints.c (process_alt_operands): Reload operand
20267 conflicting with earlier clobber only if no more other conflicting
20270 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20273 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20274 if not linking a shared library.
20276 2013-02-22 Seth LaForge <sethml@google.com>
20278 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20280 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20282 * config/arm/arm.md (split for extendsidi): Update condition.
20283 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20284 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20285 (qhs_zextenddi_cstr): Likewise.
20287 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20289 PR middle-end/56420
20290 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20291 avoid signed wrapping.
20292 (expand_mult): Handle properly multiplication by
20293 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20294 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20295 in the compiler if coeff is HOST_WIDE_INT_MIN.
20296 (expand_divmod): Don't make ext_op1 static, change it's type to
20297 uhwi. Avoid undefined behavior in -INTVAL (op1).
20299 PR rtl-optimization/50339
20300 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20302 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20303 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20304 into splitting_ashiftrt field.
20305 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20307 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20310 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20312 PR middle-end/56108
20313 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20314 are sure to go irrevocable.
20316 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20318 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20319 scalars are valid operands.
20321 2013-02-21 Martin Jambor <mjambor@suse.cz>
20323 PR tree-optimization/56310
20324 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20325 only matching indices and non-negative final offsets.
20326 (intersect_aggregates_with_edge): Pass src_idx to
20327 agg_replacements_to_vector. Pass src_idx insstead of index to
20328 intersect_with_agg_replacements.
20330 2013-02-21 Martin Jambor <mjambor@suse.cz>
20332 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20333 instead of hard-wired defaults.
20335 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20337 * doc/invoke.texi (MIPS Options): Update documentation of the
20338 floating-point multiply-accumulate instruction restrictions.
20340 2013-02-21 Kostya Serebryany <kcc@google.com>
20342 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20343 asan_shadow_offset on x86_64 linux.
20345 2013-02-21 Richard Biener <rguenther@suse.de>
20347 PR tree-optimization/56415
20349 2013-02-11 Richard Biener <rguenther@suse.de>
20351 PR tree-optimization/56273
20352 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20355 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20358 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20361 PR inline-asm/56405
20362 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20363 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20365 2013-02-20 Jan Hubicka <jh@suse.cz>
20367 PR tree-optimization/56265
20368 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20369 when target is referenced for first time.
20371 2013-02-20 Richard Biener <rguenther@suse.de>
20373 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20374 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20375 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20376 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20377 not return anything.
20378 (rename_ssa_copies): Do not remove unused locals.
20379 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20380 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20381 * passes.c (execute_function_todo): Do not schedule unused locals
20382 removal if cleanup_tree_cfg did something.
20383 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20384 about the number of removed locals.
20386 2013-02-20 Richard Biener <rguenther@suse.de>
20388 PR tree-optimization/56398
20389 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20391 2013-02-20 Martin Jambor <mjambor@suse.cz>
20393 PR tree-optimization/55334
20394 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20395 restricted pointers to arrays.
20397 2013-02-20 Richard Biener <rguenther@suse.de>
20398 Jakub Jelinek <jakub@redhat.com>
20400 PR tree-optimization/56396
20401 * tree-ssa-ccp.c (n_const_val): New static variable.
20402 (get_value): Return NULL for SSA names we don't have a lattice
20404 (ccp_initialize): Initialize n_const_val.
20405 * tree-ssa-copy.c (n_copy_of): New static variable.
20406 (init_copy_prop): Initialize n_copy_of.
20407 (get_value): Return NULL_TREE for SSA names we don't have a
20410 2013-02-20 Martin Jambor <mjambor@suse.cz>
20412 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20414 2013-02-20 Richard Biener <rguenther@suse.de>
20416 * genpreds.c (write_lookup_constraint): Do not compare first
20417 letter of the constraint again.
20419 2013-02-20 Richard Biener <rguenther@suse.de>
20421 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20423 (get_use_iv_cost): Terminate hashtable walk when coming across
20426 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20428 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20429 reassociation for avx2 targets.
20431 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20433 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20434 Add version check for v8.10.a to enable microblaze_has_clz
20435 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20436 version and TARGET_PATTERN_COMPARE check
20437 * config/microblaze/microblaze.md: New clzsi2 instruction
20439 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20441 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20442 function before branching.
20444 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20446 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20448 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20450 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20452 PR middle-end/55889
20453 * sel-sched.c: Include ira.h.
20454 (implicit_clobber_conflict_p): New function.
20455 (moveup_expr): Use it.
20456 * Makefile.in (sel-sched.o): Depend on ira.h.
20458 2013-02-19 Richard Biener <rguenther@suse.de>
20460 PR tree-optimization/56384
20461 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20462 (vn_hash_type): Split out from ...
20463 (vn_hash_constant_with_type): ... here.
20464 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20465 (vn_phi_eq): Compare types from vn_phi_s structure.
20466 (vn_phi_lookup): Populate vn_phi_s type.
20467 (vn_phi_insert): Likewise.
20469 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20471 PR tree-optimization/56350
20472 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20473 if haven't found reduction or nested cycle operand, rather than
20474 asserting we must find it.
20476 PR tree-optimization/56381
20477 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20480 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20481 Jakub Jelinek <jakub@redhat.com>
20484 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20485 (swap_optab_enable): Same.
20486 (init_all_optabs): Use argument instead of global.
20487 * tree.h (struct tree_optimization_option): New field target_optabs.
20488 * expr.h (init_all_optabs): Add argument to prototype.
20489 (TREE_OPTIMIZATION_OPTABS): New.
20490 (save_optabs_if_changed): Protoize.
20491 * optabs.h: Declare this_fn_optabs.
20492 * optabs.c (save_optabs_if_changed): New.
20493 Declare this_fn_optabs.
20494 (init_optabs): Add argument to init_all_optabs() call.
20495 * function.c (invoke_set_current_function_hook): Handle per
20497 * function.h (struct function): New field optabs.
20498 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20499 optimization_current_node has changed.
20500 * target-globals.h (save_target_globals_default_opts): Protoize.
20501 * target-globals.c (save_target_globals_default_opts): New.
20503 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20506 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20507 registers %fr12 and %fr12R as call used.
20510 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20511 and HImode, require all displacements to be an integer multiple of
20513 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20514 only allow QImode and HImode when reload is in progress and strict is
20515 true. Likewise for symbolic addresses. Use base14_operand to check
20516 displacements in REG+BASE addresses.
20518 2013-02-18 Richard Biener <rguenther@suse.de>
20520 PR tree-optimization/56366
20521 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20522 handle sign-conversion of outer-loop initial induction value.
20524 2013-02-18 Richard Biener <rguenther@suse.de>
20526 PR middle-end/56349
20527 * cfghooks.c (merge_blocks): If we merge a latch into another
20528 block adjust references to it.
20529 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20530 (verify_loop_structure): Verify that a recorded latch is in fact
20533 2013-02-18 Richard Biener <rguenther@suse.de>
20535 PR tree-optimization/56321
20536 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20537 order SSA name release and virtual operand unlinking.
20539 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20541 * config/microblaze/microblaze.md (save_stack_block): Define.
20542 (restore_stack_block): Likewise.
20544 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20546 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20547 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20548 * config/microblaze/microblaze.c (microblaze_option_override):
20549 Bail out early for PIC modes when target does not support PIC.
20551 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20553 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20554 Replace with a microblaze version.
20555 (microblaze_trampoline_init): Adapt for microblaze.
20556 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20559 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20560 Dodji Seketeli <dodji@redhat.com>
20563 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20564 (instrument_mem_region_access): Do not forget to always put
20565 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20566 0) statement, even for cases where either 'base' or 'base + len'
20567 are not instrumented -- because they have been previously
20568 instrumented. Simplify the logic by putting all the statements
20569 instrument 'base + len' inside a sequence, and then insert that
20570 sequence right before the current insertion point. Then, to
20571 instrument 'base + len', just get an iterator on that statement.
20572 And do not forget to update the pointer to iterator the function
20573 received as argument.
20575 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20577 PR rtl-optimization/56348
20578 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20580 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20582 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20583 (clean_graph_dump_file): Pass base to start_graph_dump.
20585 2013-02-14 Richard Henderson <rth@redhat.com>
20588 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20590 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20592 * collect2-aix.h: Define F_LOADONLY.
20594 2013-02-14 Richard Biener <rguenther@suse.de>
20597 * varasm.c (output_constant_def_1): Get the decl representing
20598 the constant as argument.
20599 (output_constant_def): Wrap output_constant_def_1.
20600 (make_decl_rtl): Use output_constant_def_1 with the decl
20601 representing the constant.
20602 (build_constant_desc): Optionally re-use a decl already
20603 representing the constant.
20604 (tree_output_constant_def): Adjust.
20606 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20609 * asan.c (instrument_builtin_call): Really put the length of the
20610 second source argument into src1_len.
20612 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20614 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20615 argument. If it is false, don't create edge from then_bb to
20617 (insert_if_then_before_iter): Pass true to it.
20618 (build_check_stmt): Pass false to it.
20619 (transform_statements): Flush hash table only on extended basic
20620 block boundaries, rather than at the beginning of every bb.
20621 Don't flush hash table on nonfreeing_call_p calls.
20622 * tree-flow.h (nonfreeing_call_p): New prototype.
20623 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20625 2013-02-13 David S. Miller <davem@davemloft.net>
20627 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20629 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20632 * ira.c (max_regno_before_ira): Move from ...
20634 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20637 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20639 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20641 2013-02-13 Richard Biener <rguenther@suse.de>
20644 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20645 globals in MEM_REFs.
20647 2013-02-13 Richard Biener <rguenther@suse.de>
20649 * loop-init.c (loop_optimizer_init): Clear loop state when
20650 re-initializing preserved loops.
20651 * loop-unswitch.c (unswitch_single_loop): Return whether
20652 we unswitched the loop. Do not verify loop state here.
20653 (unswitch_loops): When we unswitched a loop discover new loops.
20655 2013-02-13 Kostya Serebryany <kcc@google.com>
20657 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20659 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20661 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20663 Avoid instrumenting duplicated memory access in the same basic block
20664 * Makefile.in (asan.o): Add new dependency on hash-table.h
20665 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20666 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20667 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20668 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20669 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20670 (get_mem_ref_of_assignment): New functions.
20671 (get_mem_refs_of_builtin_call): Extract from
20672 instrument_builtin_call and tweak a little bit to make it fit with
20674 (instrument_builtin_call): Use the new
20675 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20676 of is_gimple_builtin_call.
20677 (instrument_derefs, instrument_mem_region_access): Insert the
20678 instrumented memory reference into the hash table.
20679 (maybe_instrument_assignment): Renamed instrument_assignment into
20680 this, and change it to advance the iterator when instrumentation
20681 actually happened and return true in that case. This makes it
20682 homogeneous with maybe_instrument_assignment, and thus give a
20683 chance to callers to be more 'regular'.
20684 (transform_statements): Clear the memory reference hash table
20685 whenever we enter a new BB, when we cross a function call, or when
20686 we are done transforming statements. Use
20687 maybe_instrument_assignment instead of instrumentation. No more
20688 need to special case maybe_instrument_assignment and advance the
20689 iterator after calling it; it's now handled just like
20690 maybe_instrument_call. Update comment.
20692 2013-02-13 Richard Biener <rguenther@suse.de>
20694 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20695 Fix loop discovery code.
20697 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20699 PR inline-asm/56148
20700 * lra-constraints.c (process_alt_operands): Match early clobber
20701 operand with itself. Check conflicts with earlyclobber only if
20702 the operand is not reloaded. Prefer to reload conflicting operand
20703 if earlyclobber and matching operands are the same.
20705 2013-02-12 Richard Biener <rguenther@suse.de>
20708 * lto-streamer-out.c (write_symbol): Do not output symbols
20709 for hard register variables.
20711 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20714 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20715 (umulsidi3_insn, mulsidi3_insn): New insns.
20717 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20719 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20720 (struct tune_params): Add vec_costs field.
20721 * config/arm/arm.c (arm_builtin_vectorization_cost)
20722 (arm_add_stmt_cost): New functions.
20723 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20724 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20725 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20726 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20727 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20728 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20729 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20731 2013-02-12 Richard Biener <rguenther@suse.de>
20734 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20735 decls again if possible.
20737 2013-02-12 Richard Biener <rguenther@suse.de>
20739 PR middle-end/56288
20740 * tree-ssa.c (verify_ssa_name): Fix check, move
20741 SSA_NAME_IN_FREE_LIST check up.
20743 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20744 Steven Bosscher <steven@gcc.gnu.org>
20746 PR rtl-optimization/56151
20747 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20748 equal to op0 or op1, and last_insn pattern is CODE operation
20749 with MEM dest and one of the operands matches that MEM.
20751 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20753 * doc/extend.texi: Document Function Multiversioning and "default"
20754 parameter string to target attribute.
20755 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20756 target attribute parameter is "default".
20757 (ix86_compare_version_priority): Remove checks for target attribute.
20758 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20759 Remove check for target attribute equal to NULL. Add assert.
20760 (ix86_generate_version_dispatcher_body): Change error to sorry.
20762 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20763 Jack Howarth <howarth@bromo.med.uc.edu>
20764 Patrick Marlier <patrick.marlier@gmail.com>
20767 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20768 define ENDFILE_SPEC as TM_DESTRUCTOR.
20769 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20771 2013-02-11 Alexander Potapenko <glider@google.com>
20772 Jack Howarth <howarth@bromo.med.uc.edu>
20773 Jakub Jelinek <jakub@redhat.com>
20776 * config/darwin.c (cdtor_record): Rename ctor_record.
20777 (sort_cdtor_records): Rename sort_ctor_records.
20778 (finalize_dtors): New routine to sort destructors by
20779 priority before use in assemble_integer.
20780 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20782 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20784 PR rtl-optimization/56275
20785 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20786 offset is non-negative and less than cmode size before
20787 calling simplify_subreg.
20789 2013-02-11 Richard Biener <rguenther@suse.de>
20791 PR tree-optimization/56264
20792 * cfgloop.h (fix_loop_structure): Adjust prototype.
20793 * loop-init.c (fix_loop_structure): Return the number of
20794 newly discovered loops.
20795 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20796 are discovered, do a full loop-closed SSA rewrite.
20798 2013-02-11 Richard Biener <rguenther@suse.de>
20800 PR tree-optimization/56273
20801 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20803 (check_array_ref): Fix missing newline in dumps.
20804 (search_for_addr_array): Likewise.
20806 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20808 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20810 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20813 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20815 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20817 PR rtl-optimization/56246
20818 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20820 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20821 constraints are satisfied.
20823 2013-02-08 Jeff Law <law@redhat.com>
20826 * emit-rtl.c (reg_is_parm_p): New function.
20827 * regs.h (reg_is_parm_p): New prototype.
20828 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20829 callee-clobbered registers.
20831 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20834 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20835 If there is no implicit builtin declaration, just return NULL.
20837 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20839 * config/i386/sse.md (FMAMODEM): New mode iterator.
20840 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20841 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20843 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20845 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20846 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20847 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20849 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20851 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
20852 (microblaze*-*-elf): Likewise.
20853 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
20855 * config/microblaze/microblaze-c.c: Add builtin defines for
20856 _LITTLE_ENDIAN and _BIG_ENDIAN.
20857 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
20858 add to TARGET_DEFAULT flags.
20859 Expand ASM_SPEC and LINK_SPEC.
20860 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
20861 * config/microblaze/microblaze.md: Update extendsidi2 and
20862 movdi_internal instructions to use low-order / high-order reg
20864 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
20865 options and inversemask / mask of LITTLE_ENDIAN.
20866 * config/microblaze/t-microblaze: Expand multilib options to
20867 include mlittle-endian (le) and update exceptions patterns.
20869 2013-02-08 Jakub Jelinek <jakub@redhat.com>
20871 PR rtl-optimization/56195
20872 * lra-constraints.c (get_reload_reg): Don't reuse regs
20873 if they have smaller mode than requested, if they have
20874 wider mode than requested, try to return a SUBREG.
20876 PR tree-optimization/56250
20877 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
20878 if type is unsigned and code isn't MULT_EXPR.
20880 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20882 PR tree-optimization/56064
20883 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
20884 bits according to mode.
20885 * fixed-value.h (fixed_from_double_int)
20886 (const_fixed_from_double_int): Adjust comments.
20888 2013-02-08 Richard Biener <rguenther@suse.de>
20891 * lto-streamer.h (struct data_in): Remove current_file, current_line
20892 and current_col members.
20893 * lto-streamer-out.c (lto_output_location): Stream changed bits
20894 en-block for efficiency.
20895 * lto-streamer-in.c (clear_line_info): Remove.
20896 (lto_input_location): Cache current file, line and column
20897 globally via local statics. Read changed bits en-block.
20898 (input_function): Do not call clear_line_info.
20899 (lto_read_body): Likewise.
20900 (lto_input_toplevel_asms): Likewise.
20902 2013-02-08 Michael Matz <matz@suse.de>
20904 PR tree-optimization/52448
20905 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
20906 (nt_call_phase): New static.
20907 (add_or_mark_expr): Only mark accesses with newer phase than any
20909 (nonfreeing_call_p): New.
20910 (nt_init_block): Update nt_call_phase, mark blocks as visited.
20911 (nt_fini_block): Keep blocks marked as visited.
20912 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
20914 2013-02-08 Richard Biener <rguenther@suse.de>
20916 * ira.c (ira): Free broken dominator information.
20918 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20920 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
20922 2013-02-08 Marek Polacek <polacek@redhat.com>
20924 * cfgloop.c (verify_loop_structure): Add more checking of headers.
20926 2013-02-08 Richard Biener <rguenther@suse.de>
20928 PR middle-end/56181
20929 * cfgloop.h (flow_loops_find): Adjust.
20930 (bb_loop_header_p): Declare.
20931 * cfgloop.c (bb_loop_header_p): New function split out from ...
20932 (flow_loops_find): ... here. Adjust function signature,
20933 support incremental loop structure update.
20934 (verify_loop_structure): Cleanup. Verify a loop is a loop.
20935 * cfgloopmanip.c (fix_loop_structure): Move ...
20936 * loop-init.c (fix_loop_structure): ... here.
20937 (apply_loop_flags): Split out from ...
20938 (loop_optimizer_init): ... here.
20939 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
20940 in incremental mode, only remove dead loops here.
20942 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20945 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
20946 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
20947 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
20948 (*round<mode>3.libgcc): New insns for fixed-modes.
20949 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
20950 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
20951 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
20952 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
20953 implementations. Define to __builtin_avr_absFX,
20954 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
20955 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
20956 __builtin_avr_countlsFX, respectively.
20957 * config/avr/avr-c.c (target.h): Include it.
20958 (enum avr_builtin_id): New enum.
20959 (avr_resolve_overloaded_builtin): New static function.
20960 (avr_register_target_pragmas): Use it to set
20961 targetm.resolve_overloaded_builtin.
20962 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
20963 tree nodes used by DEF_BUILTIN.
20964 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
20965 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
20966 <AVR_BUILTIN_xxBITS>: Same.
20968 2013-02-08 Richard Biener <rguenther@suse.de>
20970 * cfgloop.c (verify_loop_structure): Properly handle
20971 a loop exiting to another loop header.
20972 * ira-int.h (ira_loops): Remove.
20973 * ira.c (ira_loops): Remove.
20974 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
20975 (do_reload): Use loop_optimizer_finalize.
20976 * ira-build.c (create_loop_tree_nodes): Use get_loops and
20977 number_of_loops to access the loop tree.
20978 (more_one_region_p): Likewise.
20979 (finish_loop_tree_nodes): Likewise.
20980 (rebuild_regno_allocno_maps): Likewise.
20981 (mark_loops_for_removal): Likewise.
20982 (mark_all_loops_for_removal): Likewise.
20983 (remove_unnecessary_regions): Likewise.
20984 (ira_build): Likewise.
20985 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20987 2013-02-08 Richard Biener <rguenther@suse.de>
20989 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
20990 * ipa-pure-const.c (analyze_function): Avoid calling
20991 mark_irreducible_loops twice.
20992 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
20994 2013-02-07 David S. Miller <davem@davemloft.net>
20996 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
20998 * var-tracking.c (vt_add_function_parameter): Test the presence of
20999 HAVE_window_save properly and do not remap argument registers when
21000 we have a leaf function.
21002 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21005 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21007 * config/i386/i386.c (ix86_print_operand): Ditto.
21009 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21011 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21013 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21015 PR rtl-optimization/56225
21016 * lra-constraints.c (process_alt_operands): Check that reload hard
21017 reg can hold value for strict_low_part.
21019 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21022 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21023 dwarf2out_end_function.
21024 (in_first_function_p, maybe_at_text_label_p,
21025 first_loclabel_num_not_at_text_label): New variables.
21026 (dwarf2out_var_location): In the first function find out
21027 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21028 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21031 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21033 PR rtl-optimization/56178
21034 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21035 SUBREG of a register. Tidy up related block of code.
21036 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21037 note if the source is a register or a SUBREG of a register.
21039 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21042 * config/rs6000/rs6000.md (ptrm): New mode attr.
21043 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21044 call_value_indirect_aix<pttrsize>,
21045 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21048 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21050 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21051 if -bnortl. Convert to strcmp and strncmp.
21053 2013-02-07 Alan Modra <amodra@gmail.com>
21056 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21057 addresses won't wrap when offsetting.
21058 (rs6000_secondary_reload): Provide secondary reloads needed for
21059 wrapping LO_SUM addresses.
21061 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21063 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21064 MACH, just __MACH__.
21066 2013-02-06 Richard Biener <rguenther@suse.de>
21068 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21069 instead of calling fix_loop_structure.
21071 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21073 PR middle-end/56217
21074 * omp-low.c (use_pointer_for_field): Return false if
21075 lower_send_shared_vars doesn't generate any copy-out code.
21077 2013-02-06 Tom de Vries <tom@codesourcery.com>
21079 PR rtl-optimization/56131
21080 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21081 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21082 of the label is NULL. Add comment.
21084 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21086 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21089 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21090 (STATIC_LIBTSAN_LIBS): Likewise.
21091 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21092 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21093 is defined, don't add anything else beyond that.
21094 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21095 (LINK_COMMAND_SPEC): Use them.
21097 PR tree-optimization/56205
21098 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21099 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21100 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21102 2013-02-05 Richard Biener <rguenther@suse.de>
21104 PR tree-optimization/53342
21105 PR tree-optimization/53185
21106 * tree-vectorizer.h (vect_check_strided_load): Remove.
21107 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21108 not disallow peeling for vectorized strided loads.
21109 (vect_check_strided_load): Make static and simplify.
21110 (vect_analyze_data_refs): Adjust.
21111 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21112 correctly when vectorizing strided loads.
21114 2013-02-05 Richard Biener <rguenther@suse.de>
21116 * doc/install.texi: Refer to ISL, not PPL.
21118 2013-02-05 Jan Hubicka <jh@suse.cz>
21120 PR tree-optimization/55789
21121 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21123 2013-02-05 Jan Hubicka <jh@suse.cz>
21125 PR tree-optimization/55789
21126 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21127 the dead call anyway.
21129 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21132 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21134 2013-02-04 Alexander Potapenko <glider@google.com>
21135 Jack Howarth <howarth@bromo.med.uc.edu>
21136 Jakub Jelinek <jakub@redhat.com>
21139 * config/darwin.c (sort_ctor_records): Stabilized qsort
21140 on constructor priority by using original position.
21141 (finalize_ctors): New routine to sort constructors by
21142 priority before use in assemble_integer.
21143 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21145 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21148 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21149 about visibility on artificial decls.
21150 * config/sol2.c (solaris_assemble_visibility): Likewise.
21152 2013-02-04 Kai Tietz <ktietz@redhat.com>
21155 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21156 argument and improve checking of return-argument types for 16-byte
21158 (ix86_function_value_1): Add additional valtype argument on call
21159 of function_value_64.
21160 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21161 handling infunction_value_64 function.
21163 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21165 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21167 2013-02-04 Richard Biener <rguenther@suse.de>
21169 PR tree-optimization/56188
21170 * tree-ssa-structalias.c (label_visit): Consider case with
21171 initially non-empty points-to set.
21172 (perform_var_substitution): Dump node mapping and clean up.
21174 2013-02-04 Richard Guenther <rguenther@suse.de>
21177 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21178 node prevail as last resort.
21179 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21180 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21182 2013-02-04 Richard Biener <rguenther@suse.de>
21184 PR tree-optimization/56113
21185 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21187 (equiv_class_lookup_or_add): ... this.
21188 (label_visit): Adjust and fix error in previous patch.
21189 (perform_var_substitution): Adjust.
21191 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21193 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21194 * config/sh/sh4-300.md: Likewise.
21195 * config/sh/sh4a.md: Likewise.
21196 * config/sh/constraints.md: Likewise.
21197 * config/sh/sh.md: Likewise.
21198 * config/sh/netbsd-elf.h: Likewise.
21199 * config/sh/predicates.md: Likewise.
21200 * config/sh/sh-protos.h: Likewise.
21201 * config/sh/ushmedia.h: Likewise.
21202 * config/sh/linux.h: Likewise.
21203 * config/sh/sh.c: Likewise.
21204 * config/sh/superh.h: Likewise.
21205 * config/sh/elf.h: Likewise.
21206 * config/sh/sh4.md: Likewise.
21207 * config/sh/sh.h: Likewise.
21209 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21211 * config/pa/constraints.md: Adjust unused letters. Change "T"
21212 constraint to match_test floating_point_store_memory_operand().
21213 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21214 (base14_operand): New.
21215 (floating_point_store_memory_operand): New.
21216 (integer_store_memory_operand): Revise to use base14_operand and
21217 reg_plus_base_memory_operand.
21218 (move_dest_operand): Allow symbolic_memory_operands.
21219 (symbolic_memory_operand): Check for LO_SOM.
21220 (symbolic_operand): Change default case to break.
21221 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21222 CONST_DOUBLE values to be reloaded by putting them into memory when
21223 the destination is a floating point register.
21224 (movdf): Remove code to handle CONST_DOUBLE.
21226 (reload_indf_r1): New.
21227 (reload_insf_r1): New.
21228 Consistently use "Q" and "T" constraints with integer and floating
21229 point move instructions, respectively.
21230 (movdi): Remove FAIL.
21231 Change predicate for source operand unamed DImode move from
21232 general_operand to move_src_operand.
21233 (umulsidi3): Change predicate for destination operand to
21235 Likewise for similar unamed patterns.
21236 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21237 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21238 (hppa_legitimize_address): Simplify mask calculation.
21239 (pa_emit_move_sequence): Revised handling of secondary reloads from
21240 REG+D addresses for floating point loads and stores. Directly handle
21241 loading CONST0_RTX (mode) to a floating point register.
21242 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21243 to floating point registers. Don't restrict secondary reloads to
21244 floating point registers to integer modes. Revise some comments and
21246 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21247 (pa_legitimate_address_p): New.
21248 (pa_legitimize_reload_address): New.
21249 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21250 (STRICT_REG_OK_FOR_BASE_P): New.
21251 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21252 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21254 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21255 Andrew Dixie <andrewd@gentrack.com>
21257 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21260 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21262 * expmed.c (extract_bit_field_1): Pass the full width of the
21263 structure to get_best_reg_extraction_insn.
21265 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21268 * configure.ac (use_cxa_atexit): Add AIX.
21269 * configure: Regenerate.
21271 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21273 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21276 * final.c (need_profile_function): New variable.
21277 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21278 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21279 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21280 notes, targetm.asm_out.function_prologue doesn't emit anything,
21281 HAVE_prologue and profiler should be emitted before prologue,
21282 set need_profile_function instead of emitting it.
21283 (final_scan_insn): If need_profile_function, emit
21284 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21285 NOTE_INSN_FUNCTION_BEG note.
21287 2013-02-01 Richard Henderson <rth@redhat.com>
21289 * config/rs6000/rs6000.md (smulditi3): New.
21292 * config/alpha/alpha.md (umulditi3): New.
21294 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21296 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21297 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21299 2013-02-01 Richard Biener <rguenther@suse.de>
21301 PR tree-optimization/56113
21302 * tree-ssa-structalias.c (label_visit): Reduce work for
21303 single-predecessor nodes.
21305 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21307 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21308 range isn't testing for zero.
21310 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21312 PR middle-end/56113
21313 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21315 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21316 Nick Clifton <nickc@redhat.com>
21318 * config/v850/constraints.md (Q): Define as a memory constraint.
21319 * config/v850/predicates.md (label_ref_operand): New predicate.
21320 (e3v5_shift_operand): New predicate.
21321 (ior_operator): New predicate.
21322 * config/v850/t-v850: Add e3v5 multilib.
21323 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21324 (v850_gen_movdi): Prototype.
21325 * config/v850/v850.c: Add support for e3v5 architecture.
21326 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21328 (construct_save_jarl): Add e3v5 long JARL support.
21329 (v850_adjust_insn_length): New function. Adjust length of call
21330 insns when using e3v5 instructions.
21331 (v850_gen_movdi): New function: Generate instructions to move a
21333 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21334 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21335 (TARGET_USE_FPU): Enable for e3v5.
21336 (CONST_OK_FOR_W): New macro.
21337 (ADJUST_INSN_LENGTH): Define.
21338 * config/v850/v850.md (UNSPEC_LOOP): Define.
21339 (attr cpu): Add v850e3v5.
21340 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21341 (movdi): New pattern.
21342 (movdi_internal): New pattern.
21343 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21344 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21345 (cstoresf4): Likewise.
21346 (cstoredf4): Likewise.
21347 (insv): New pattern.
21348 (rotlso3_a): New pattern.
21349 (rotlsi3_b): New pattern
21350 (rotlsi3_v850e3v5): New pattern.
21351 (doloop_begin): New pattern.
21352 (fix_loop_counter): New pattern.
21353 (doloop_end): New pattern.
21354 (branch_normal): Add e3v5 long branch support.
21355 (branch_invert): Likewise.
21356 (branch_z_normal): Likewise.
21357 (branch_z_invert): Likewise.
21358 (branch_nz_normal): Likewise.
21359 (branch_nz_invert): Likewise.
21360 (call_internal_short): Add e3v5 register-indirect JARL support.
21361 (call_internal_long): Likewise.
21362 (call_value_internal_short): Likewise.
21363 (call_value_internal_long): Likewise.
21364 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21365 (mloop): New option.
21366 * config.gcc: Add support for configuring v840e3v5 target.
21367 * doc/invoke.texi: Document new v850 specific command line options.
21369 2013-01-31 Paul Koning <ni1d@arrl.net>
21373 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21374 children if parent is a class.
21375 (prune_unused_types_prune): Don't add DW_AT_declaration.
21377 2013-01-31 Richard Biener <rguenther@suse.de>
21379 PR tree-optimization/56157
21380 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21381 match up operand with SLP child.
21383 2013-01-31 Jason Merrill <jason@redhat.com>
21386 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21387 parameters the first time.
21388 (gen_scheduled_generic_parms_dies): Check completeness here.
21390 2013-01-31 Richard Biener <rguenther@suse.de>
21392 PR middle-end/53073
21393 * common.opt (faggressive-loop-optimizations): New flag,
21394 enabled by default.
21395 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21396 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21397 infer_loop_bounds_from_undefined by it.
21399 2013-01-31 Richard Biener <rguenther@suse.de>
21401 PR tree-optimization/56150
21402 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21403 visit virtual operands.
21404 (find_uses_to_rename_bb): Likewise.
21406 2013-01-31 Richard Biener <rguenther@suse.de>
21408 PR tree-optimization/56150
21409 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21410 mixed store non-store stmts.
21412 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21415 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21416 LIBASAN_EARLY_SPEC is defined.
21417 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21418 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21420 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21423 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21424 invalid args instead of ICEing on it.
21425 (ix86_valid_target_attribute_tree): Return error_mark_node if
21426 ix86_valid_target_attribute_inner_p failed.
21427 (ix86_valid_target_attribute_p): Return false only if
21428 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21429 target("default") attribute.
21430 (sorted_attr_string): Change argument from const char * to tree,
21431 merge in all target attribute arguments rather than just one.
21432 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21433 instead of free. Avoid using strcat.
21434 (ix86_mangle_function_version_assembler_name): Mangle
21435 target("default") as if no target attribute is present. Adjust
21436 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21437 instead of xmalloc and XDELETEVEC instead of free.
21438 (ix86_function_versions): Don't return true if one of the decls
21439 doesn't have target attribute. If they don't and one of the decls
21440 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21441 sorted_attr_string caller. Use XDELETEVEC instead of free.
21442 (ix86_supports_function_versions): Remove.
21443 (make_name): Fix up formatting.
21444 (make_dispatcher_decl): Remove resolver_name and its initialization.
21445 Avoid leaking memory.
21446 (is_function_default_version): Return true if there is
21447 target("default") attribute rather than no target attribute at all.
21448 (make_resolver_func): Avoid leaking memory.
21449 (ix86_generate_version_dispatcher_body): Likewise.
21450 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21451 * target.def (supports_function_versions): Remove.
21452 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21453 * doc/tm.texi: Regenerated.
21455 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21457 PR rtl-optimization/56144
21458 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21459 for values with side effects.
21461 2013-01-30 Richard Biener <rguenther@suse.de>
21463 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21464 (sparseset_pop): Likewise.
21465 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21466 to be able to use quick_push in the worker loop.
21468 2013-01-30 Marek Polacek <polacek@redhat.com>
21470 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21472 2013-01-30 Richard Biener <rguenther@suse.de>
21475 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21477 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21479 PR tree-optimization/56064
21480 * fixed-value.c (fixed_from_double_int): New function.
21481 * fixed-value.h (fixed_from_double_int): New prototype.
21482 (const_fixed_from_double_int): New static inline function.
21483 * fold-const.c (native_interpret_fixed): New static function.
21484 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21485 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21486 (native_encode_fixed): New static function.
21487 (native_encode_expr) <FIXED_CST>: Use it.
21488 (native_interpret_int): Move double_int worker code to...
21489 * double-int.c (double_int::from_buffer): ...this new static method.
21490 * double-int.h (double_int::from_buffer): Prototype it.
21492 2013-01-30 Richard Biener <rguenther@suse.de>
21494 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21495 New pointer-map and obstack.
21496 (init_alias_vars): Allocate pointer-map and obstack.
21497 (delete_points_to_sets): Free them.
21498 (find_what_var_points_to): Cache result.
21499 (find_what_p_points_to): Adjust for changed interface of
21500 find_what_var_points_to.
21501 (compute_points_to_sets): Likewise.
21502 (ipa_pta_execute): Likewise.
21504 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21506 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21507 * configure: Regenerate.
21508 * config.in: Regenerate.
21509 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21510 #nobits/#progbits if supported.
21512 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21515 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21516 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21517 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21519 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21521 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21522 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21524 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21526 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21527 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21529 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21531 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21533 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21534 * config/arm/cortex-a7.md: New bypasses using
21535 arm_mac_accumulator_is_result.
21537 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21539 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21540 (cortex_a7_neon_mla): Likewise.
21541 (cortex_a7_fpfmad): New reservation.
21542 (cortex_a7_fpmacs): Use ffmas and update required units.
21543 (cortex_a7_fpmuld): Update required units and latency.
21544 (cortex_a7_fpmacd): Likewise.
21545 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21546 (cortex_a7_neon). Likewise.
21547 (bypass) Update participating units.
21549 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21551 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21552 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21554 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21555 (vfp_fmul): Use ffmad.
21556 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21557 (cortex_r4_fmacd): Use ffmad.
21558 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21559 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21560 (cortex_a9_fmacd): Use ffmad.
21561 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21562 (cortex_a8_vfp_macd): Use ffmad.
21563 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21564 (cortex_a5_fpmacd): Use ffmad.
21565 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21566 (cortex_a15_vfp_macd): Use ffmad.
21567 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21569 2013-01-29 Jason Merrill <jason@redhat.com>
21572 * varasm.c (default_assemble_visibility): Don't warn about
21573 visibility on artificial decls.
21575 2013-01-29 Richard Biener <rguenther@suse.de>
21577 PR tree-optimization/56113
21578 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21580 (label_visit): Free duplicate bitmaps and record the leader instead.
21581 (perform_var_substitution): Adjust.
21583 2013-01-29 Richard Biener <rguenther@suse.de>
21585 PR tree-optimization/55270
21586 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21587 the CFG, schedule loops for fixup.
21589 2013-01-29 Nick Clifton <nickc@redhat.com>
21591 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21594 2013-01-28 Leif Ekblad <leif@rdos.net>
21596 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21597 * config/i386/i386.h (TARGET_RDOS): New macro.
21598 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21599 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21600 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21601 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21602 DEFAULT_LARGE_SECTION_THRESHOLD.
21603 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21604 * config/i386/rdos.h: New file.
21605 * config/i386/rdos64.h: New file.
21607 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21610 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21613 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21615 PR rtl-optimization/56117
21616 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21617 call cselib_lookup_from_insn on the MEM before calling
21618 add_insn_mem_dependence.
21620 2013-01-28 Richard Biener <rguenther@suse.de>
21622 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21623 to a stmt that didn't have one.
21624 (copy_phis_for_bb): Likewise for PHI arguments.
21625 (copy_debug_stmt): Likewise for debug stmts.
21627 2013-01-28 Richard Biener <rguenther@suse.de>
21629 PR tree-optimization/56034
21630 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21631 (partition_builtin_p): Adjust.
21632 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21633 it is the last partition.
21634 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21635 up the vertex for the definition.
21636 (classify_partition): Classify whether a partition is a
21637 PKIND_REDUCTION, thus has uses outside of the loop.
21638 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21639 Merge all PKIND_REDUCTION partitions into the last partition.
21640 (tree_loop_distribution): Seed partitions from reductions as well.
21642 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21644 PR tree-optimization/56125
21645 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21646 pow(x,c) into sqrt(x) * powi(x, n/2) or
21647 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21648 optimizing for size.
21649 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21650 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21653 PR tree-optimization/56094
21654 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21655 to UNKNOWN_LOCATION while gimplifying expr.
21657 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21660 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21661 operand 0 in movabs insn template for -masm=intel asm alternative.
21662 (*movabs<mode>_2): Ditto for operand 1.
21664 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21667 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21670 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21672 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21675 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21677 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21678 Undef to avoid warning.
21680 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21682 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21683 * configure: Regenerate.
21685 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21687 PR tree-optimization/56098
21688 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21689 for stmts with volatile ops.
21690 (cond_store_replacement): Don't optimize if assign has volatile ops.
21691 (cond_if_else_store_replacement_1): Don't optimize if either
21692 then_assign or else_assign have volatile ops.
21693 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21696 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21698 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21700 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21702 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21703 missing ':' in asm example.
21705 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21707 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21708 entries into lane and laneq entries.
21709 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21710 Remove AdvSIMD scalar modes.
21711 (aarch64_sq<r>dmulh_laneq<mode>): New.
21712 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21714 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21715 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21716 * config/aarch64/iterators.md (VCOND): New.
21719 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21722 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21723 Add NULL LIBNAME argument to existing definitions.
21724 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21725 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21726 * config/avr/avr.c (DEF_BUILTIN): Same.
21727 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21728 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21729 implementation is available (DECL_ASSEMBLER_NAME is set).
21730 (avr_fold_absfx): New static function.
21731 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21732 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21733 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21734 AVR_BUILTIN_ABSLLK.
21735 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21736 (abshk, absk, abslk, absllk): Provide as static inline functions.
21738 2013-01-25 Marek Polacek <polacek@redhat.com>
21740 PR tree-optimization/56035
21741 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21743 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21745 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21746 (*movtf_internal_rex64): Add (!o,C) alternative
21747 (*movxf_internal_rex64): Ditto.
21748 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21750 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21752 * doc/invoke.texi: fix typo.
21753 * doc/objc.texi: fix typo.
21755 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21757 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21758 for the first two alternatives.
21760 2013-01-24 Diego Novillo <dnovillo@google.com>
21762 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21763 (ggc-zone.o): Remove.
21764 * configure.ac: Remove option --with-gc.
21765 * configure: Re-generate.
21766 * doc/install.texi: Remove documentation for --with-gc.
21767 * gengtype.c (write_enum_defn): Remove. Update all users.
21768 (write_Types_process_field): Remove generation of gt_e_* argument.
21769 (output_type_enum): Remove. Update all users.
21770 (write_enum_defn): Remove. Update all users.
21771 (enum alloc_zone): Remove. Update all users.
21772 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21773 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21774 Update all callers.
21775 (struct ptr_data): Remove field TYPE. Update all users.
21776 (gt_pch_note_object): Remove argument TYPE. Update all users.
21777 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21779 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21780 (struct alloc_zone): Remove.
21781 (ggc_internal_alloc_zone_stat): Remove.
21782 (ggc_internal_cleared_alloc_zone_stat): Remove.
21783 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21784 (ggc_pch_count_object): Remove last argument. Update all users.
21785 (ggc_pch_alloc_object): Remove last argument. Update all users.
21786 (struct alloc_zone): Remove.
21787 * ggc-zone.c: Remove.
21788 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21789 (struct alloc_zone): Remove.
21790 (ggc_alloc_typed_stat): Remove.
21791 (ggc_alloc_typed): Remove.
21792 (ggc_splay_alloc): Remove first argument.
21793 (rtl_zone): Remove. Update all users.
21794 (tree_zone): Remove. Update all users.
21795 (tree_id_zone): Remove. Update all users.
21796 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21797 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21798 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21799 * tree-ssanames.c: Remove references to zone allocator in comments.
21801 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21803 * config/avr/avr.c (avr_out_fract): Make register numbers that
21804 might be outside of source operand signed.
21806 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21808 * config/i386/constraints.md (Yf): New constraint.
21809 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21810 of f constraint to conditionaly disable x87 register preferences.
21811 (*movdf_internal): Ditto.
21812 (*movsf_internal): Ditto.
21814 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21816 PR inline-asm/55934
21817 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21818 that have operands with impossible constraints.
21819 Add a FIXME for a speed-up opportunity.
21820 * lra-constraints.c (process_alt_operands): Verify that a class
21821 selected from constraints on asms is valid for the operand mode.
21822 (curr_insn_transform): Remove incorrect comment.
21824 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21826 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21827 TOC operand is a valid symbol ref in the constant pool.
21829 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21831 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21833 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21836 * config/avr/stdfix.h: New file.
21837 * t-avr (stdfix-gcc.h): New rule to build it.
21838 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21840 2013-01-23 Kostya Serebryany <kcc@google.com>
21842 * config/darwin.h: remove dependency on
21843 CoreFoundation (asan on Mac OS).
21845 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21848 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21849 instead of cmpdi_operand for first comparison operand.
21850 Don't assert that comparison operands aren't both constants.
21852 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21854 * doc/install.texi (Downloading the Source): Update references to
21855 downloading separate components.
21857 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21859 * doc/extend.texi (__int128): Improve grammar.
21861 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
21864 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
21865 alternative to (o,r).
21866 (*movdi_internal_rex64): Remove (!o,n) alternative.
21867 (DImode immediate->memory splitter): Remove.
21868 (DImode immediate->memory peephole2): Remove.
21869 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
21870 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
21871 alternative to (!o,*r).
21872 (*movtf_internal_sse): New pattern.
21873 (*movxf_internal_rex64): New pattern.
21874 (*movxf_internal): Disable for TARGET_64BIT.
21875 (*movdf_internal_rex64): Remove (!o,F) alternative.
21877 2013-01-22 Jakub Jelinek <jakub@redhat.com>
21879 PR middle-end/56074
21880 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
21881 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
21882 * tree-vect-loop-manip.c (find_loop_location): Also ignore
21883 stmt locations where LOCATION_LOCUS of the stmt location is
21884 UNKNOWN_LOCATION or BUILTINS_LOCATION.
21887 * config/i386/i386.md (UNSPEC_STOS): New.
21888 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
21889 *strsetqi_1): Add UNSPEC_STOS.
21891 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
21894 * doc/invoke.texi: Remove left over -Wsynth example.
21896 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21898 PR tree-optimization/56051
21899 * fold-const.c (fold_binary_loc): Don't fold
21900 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
21901 a narrowing conversion, or widening conversion from signed
21904 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21906 PR rtl-optimization/56023
21907 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
21908 dependent on debug instruction.
21910 2013-01-21 Martin Jambor <mjambor@suse.cz>
21912 PR middle-end/56022
21913 * function.c (allocate_struct_function): Call
21914 invoke_set_current_function_hook earlier.
21916 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21918 * reload1.c (init_reload): Only initialize reload_obstack
21919 during the first call.
21921 2013-01-21 Marek Polacek <polacek@redhat.com>
21923 * cfgloop.c (verify_loop_structure): Fix up grammar.
21925 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
21927 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
21928 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
21930 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21933 * config/arm/marvell-pj4.md: Update copyright year.
21934 Fix up use of alu to alu_reg and simple_alu_imm.
21936 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21938 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
21940 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
21943 * lra-constraints.c (curr_insn_transform): Don't reuse original
21944 insn for secondary memory move when memory mode should be different.
21946 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21948 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
21949 atomic_storedi_1): New patterns.
21951 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21953 btver2 pipeline descriptions.
21954 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
21956 * config/i386/i386.md (btver2_decode): New type attributes.
21957 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
21959 * config/i386/btver2.md: New file describing btver2 pipelines.
21961 2013-01-19 Andrew Pinski <apinski@cavium.com>
21963 PR tree-optimization/52631
21964 * tree-ssa-sccvn (visit_use): Before looking up the original
21965 statement, try looking up the simplified expression.
21967 2013-01-19 Anthony Green <green@moxielogic.com>
21969 * config/moxie/moxie.c (moxie_expand_prologue): Set
21970 current_function_static_stack_size.
21972 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21974 PR tree-optimization/56029
21975 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
21976 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
21978 2013-01-18 Sharad Singhai <singhai@google.com>
21980 PR tree-optimization/55995
21981 * dumpfile.c (dump_loc): Print location only if available.
21982 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
21984 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
21987 * lra-constraints.c (curr_insn_transform): Reuse original insn for
21988 secondary memory move.
21989 (inherit_reload_reg): Use rclass instead of cl for
21990 check_secondary_memory_needed_p.
21992 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21994 PR middle-end/56015
21995 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
21996 the case where writing real complex part of target modifies op1.
21998 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22000 * config/aarch64/aarch64-simd.md
22001 (aarch64_vcond_internal<mode>): Handle unordered cases.
22002 * config/aarch64/iterators.md (v_cmp_result): New.
22004 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22005 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22007 * config/arm/marvell-pj4.md: New file.
22008 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22009 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22010 (generic_vfp): Likewise.
22011 * config/arm/arm-cores.def: Add marvell-pj4.
22012 * config/arm/arm-tune.md: Regenerate.
22013 * config/arm/arm-tables.opt: Regenerate.
22014 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22015 * doc/invoke.texi: Document marvell-pj4.
22017 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22019 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22021 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22026 * var-tracking.c (negative_power_of_two_p): New.
22027 (global_get_addr_cache, local_get_addr_cache): New.
22028 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22029 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22031 (vt_stack_offset_p): Remove.
22032 (vt_canon_true_dep): Always canonicalize loc's address.
22033 (clobber_overlapping_mems): Make sure we have a MEM.
22034 (local_get_addr_clear_given_value): New.
22035 (val_reset): Clear local cached entries.
22036 (compute_bb_dataflow): Create and release the local cache.
22037 Disable duplicate MEMs clobbering.
22038 (emit_notes_in_bb): Clobber MEMs likewise.
22039 (vt_emit_notes): Create and release the local cache.
22040 (vt_initialize, vt_finalize): Create and release the global
22041 cache, respectively.
22042 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22044 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22046 PR libmudflap/53359
22047 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22048 not found in the symtab.
22050 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22053 PR rtl-optimization/55547
22054 PR rtl-optimization/53827
22057 * alias.c (offset_overlap_p): New, factored out of...
22058 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22059 the conservative special case for symbolic constants. Don't
22060 adjust zero sizes on alignment.
22062 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22064 PR rtl-optimization/52573
22065 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22066 REG_UNUSED for the same register.
22068 2013-01-17 Richard Biener <rguenther@suse.de>
22069 Marek Polacek <polacek@redhat.com>
22071 PR rtl-optimization/55833
22072 * loop-unswitch.c (unswitch_loops): Move loop verification...
22073 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22074 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22075 Set it to true when we're removing a loop from hierarchy tree in
22076 an irreducible region.
22077 (fix_bb_placements): Adjust caller.
22078 (fix_loop_placements): Likewise.
22080 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22082 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22083 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22084 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22085 Remove ID. Adjust comments.
22086 * config/avr/avr-c.c (avr_builtin_name): Remove.
22087 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22088 * config/avr/avr.c (avr_tolower): New static function.
22089 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22090 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22091 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22094 2013-01-17 Jan Hubicka <jh@suse.cz>
22096 PR tree-optimization/55273
22097 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22099 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22102 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22103 store through atomic_store<mode>_1.
22104 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22106 2013-01-17 Martin Jambor <mjambor@suse.cz>
22108 PR tree-optimizations/55264
22109 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22110 for virtual methods.
22111 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22112 virtual methods before inlining is over.
22113 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22115 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22118 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22120 PR rtl-optimization/56005
22121 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22122 pending reads for prefetch.
22124 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22126 * config/aarch64/aarch64.md
22127 (*cstoresi_neg_uxtw): New pattern.
22128 (*cmovsi_insn_uxtw): New pattern.
22129 (*<optab>si3_uxtw): New pattern.
22130 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22131 (*<optab>si3_insn_uxtw): New pattern.
22132 (*bswapsi2_uxtw): New pattern.
22134 2013-01-16 Richard Biener <rguenther@suse.de>
22136 * tree-inline.c (tree_function_versioning): Remove set but
22137 never used variable.
22139 2013-01-16 Richard Biener <rguenther@suse.de>
22141 PR tree-optimization/55964
22142 * tree-flow.h (rename_variables_in_loop): Remove.
22143 (rename_variables_in_bb): Likewise.
22144 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22145 (copy_loop_before): Adjust and delete update-ssa status.
22146 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22147 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22148 (rename_variables_in_loop): Remove.
22149 (slpeel_update_phis_for_duplicate_loop): Likewise.
22150 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22151 use available cfg machinery instead of duplicating it.
22152 Update PHI nodes and perform poor-mans SSA update here.
22153 (slpeel_tree_peel_loop_to_edge): Adjust.
22155 2013-01-16 Richard Biener <rguenther@suse.de>
22157 PR tree-optimization/54767
22158 PR tree-optimization/53465
22159 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22160 (vrp_visit_phi_node): For PHI arguments coming via backedges
22161 drop all symbolical range information.
22162 (execute_vrp): Compute backedges.
22164 2013-01-16 Richard Biener <rguenther@suse.de>
22166 * doc/install.texi: Update CLooG and ISL requirements to
22169 2013-01-16 Christian Bruel <christian.bruel@st.com>
22172 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22173 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22174 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22176 2013-01-16 DJ Delorie <dj@redhat.com>
22178 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22179 (UNSPECV_SP_SWITCH_E): New.
22180 (sp_switch_1): Change to an unspec.
22181 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22184 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22186 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22187 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22188 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22189 (expand_mem_thread_fence): Ditto.
22190 (expand_mem_signal_fence): Ditto.
22191 (expand_atomic_load): Ditto.
22192 (expand_atomic_store): Ditto.
22194 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22196 PR rtl-optimization/55547
22197 PR rtl-optimization/53827
22200 * alias.c (memrefs_conflict_p): Set sizes to negative after
22203 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22206 * function.c (thread_prologue_and_epilogue_insns): Always
22207 add crtl->drap_reg to set_up_by_prologue.set, even if
22208 stack_realign_drap is false.
22210 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22212 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22213 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22214 *call): Fix indention.
22216 2013-01-15 Tom de Vries <tom@codesourcery.com>
22219 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22222 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22224 PR rtl-optimization/55153
22225 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22227 2013-01-15 Martin Jambor <mjambor@suse.cz>
22229 PR tree-optimization/55920
22230 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22231 accesses as grp_to_be_debug_replaced.
22233 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22235 PR tree-optimization/55920
22236 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22237 there is non-useless type conversion needed from debug rhs to lhs,
22238 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22240 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22241 Mikael Pettersson <mikpe@it.uu.se>
22244 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22246 (ASM_OUTPUT_CASE_LABEL): Remove.
22247 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22248 * final.c (shorten_branches): Update alignment of labels before
22249 jump tables if CASE_VECTOR_SHORTEN_MODE.
22251 2013-01-15 Richard Biener <rguenther@suse.de>
22254 * system.h: Do not include gmp.h for building host tools.
22256 2013-01-15 Richard Biener <rguenther@suse.de>
22258 PR middle-end/55882
22259 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22260 account for bitpos when computing alignment.
22262 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22264 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22265 (ix86_target_macros_internal): Likewise.
22267 * config/i386/i386.c (m_CORE2I7): Removed.
22268 (m_CORE_HASWELL): New macro.
22269 (m_CORE_ALL): Likewise.
22270 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22271 (initial_ix86_arch_features): Likewise.
22272 (processor_target_table): Initializations for Core avx2.
22273 (cpu_names): New names "core-avx2".
22274 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22275 PROCESSOR_CORE_HASWELL.
22276 (ix86_issue_rate): New case.
22277 (ia32_multipass_dfa_lookahead): Likewise.
22278 (ix86_sched_init_global): Likewise.
22280 * config/i386/i386.h (TARGET_HASWELL): New macro.
22281 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22282 (processor_type): New PROCESSOR_HASWELL.
22284 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22286 PR tree-optimization/55955
22287 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22288 *SHIFT_EXPR and *ROTATE_EXPR codes.
22290 PR tree-optimization/48766
22291 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22292 -ftrapv disable -fwrapv.
22294 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22297 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22298 etc. to 1 and not to __flash.
22299 Use LL suffix for __INT24_MAX__ with -mint8.
22300 Use ULL suffix for __UINT24_MAX__ with -mint8.
22302 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22304 * config/avr/avr-arch.h
22305 (struct base_arch_s): Use typedef avr_arch_t instead.
22306 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22307 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22308 * config/avr/avr.c: Same.
22309 * config/avr/avr-devices.c: Same.
22310 * config/avr/driver-avr.c: Same.
22311 * config/avr/gen-avr-mmcu-texi.c: Same.
22312 * config/avr/avr-mcus.def: Adjust comment.
22314 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22316 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22317 * config/aarch64/iterators.md (VALLDI): New.
22319 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22320 Andi Kleen <ak@linux.intel.com>
22323 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22324 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22327 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22329 * config/avr/avr-stdint.h: Remove trailing blanks.
22330 * config/avr/avr-log.h: Same.
22331 * config/avr/avr-arch.h: Same.
22332 * config/avr/avr-devices.c: Same.
22333 * config/avr/avr-dimode.md: Same.
22334 * config/avr/predicates.md: Same.
22335 * config/avr/avr-c.c: Same. And fix typo.
22337 * config/avr/avr-protos.h: Same. And:
22338 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22339 (init_cumulative_args): Rename to avr_init_cumulative_args.
22340 (expand_prologue): Rename to avr_expand_prologue.
22341 (expand_epilogue): Rename to avr_expand_epilogue.
22342 (adjust_insn_length): Rename to avr_adjust_insn_length.
22343 (notice_update_cc): Rename to avr_notice_update_cc.
22344 (final_prescan_insn): Rename to avr_final_prescan_insn.
22345 * config/avr/avr.c: Same.
22346 * config/avr/avr.h: Same.
22347 * config/avr/avr.md: Remove trailing blanks.
22348 (prologue): Use avr_expand_prologue.
22349 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22351 2013-01-14 Richard Biener <rguenther@suse.de>
22353 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22354 verify_location, collect_subblocks): New functions.
22355 (verify_gimple_in_cfg): Verify that locations only reference
22356 BLOCKs in the functions BLOCK tree.
22358 2013-01-14 Richard Biener <rguenther@suse.de>
22360 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22362 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22364 (insert_out_of_ssa_copy_on_edge): Likewise.
22365 (rewrite_close_phi_out_of_ssa): Likewise.
22366 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22368 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22369 propagated constants.
22370 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22373 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22375 * config/avr/avr-modes.def: Add GPL copyright notice.
22377 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22379 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22380 MEMMODEL_MASK to determine memory model.
22381 (atomic_store<mode>): Ditto from operands[2].
22382 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22384 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22387 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22388 (fold_gimple_assign): Don't call unshare_expr here.
22389 (fold_ctor_reference): Call unshare_expr.
22391 2013-01-13 Terry Guo <terry.guo@arm.com>
22393 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22394 * doc/fragments.texi: Document MULTILIB_REUSE.
22395 * gcc.c (multilib_reuse): New internal spec.
22396 (set_multilib_dir): Also search multilib from multilib_reuse.
22397 * genmultilib (tmpmultilib3): Refactor code.
22398 (tmpmultilib4): Ditto.
22399 (multilib_reuse): New multilib argument.
22401 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22403 * Makefile.in: Update copyright.
22405 2013-01-12 Tom de Vries <tom@codesourcery.com>
22407 PR middle-end/55890
22408 * calls.c (expand_call): Check if arg_nr is valid.
22410 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22412 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22413 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22414 documentation. Add missing '__' in front of
22415 __builtin_ia32_packssdw256.
22417 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22420 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22421 NO_REGS for larl operands.
22422 (s390_reload_larl_operand): Use s390_load_address instead of
22425 2013-01-11 Richard Biener <rguenther@suse.de>
22427 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22428 (verify_node_sharing): ... here.
22429 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22431 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22433 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22434 Substitute TREECHECKING.
22435 * configure: Regenerate.
22436 * Makefile.in (TREECHECKING): New.
22438 2013-01-11 Richard Guenther <rguenther@suse.de>
22440 PR tree-optimization/44061
22441 * tree-vrp.c (extract_range_basic): Compute zero as
22442 value-range for __builtin_constant_p of function parameters.
22444 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22446 Update copyright years.
22448 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22450 PR rtl-optimization/55672
22451 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22452 const to be eliminable.
22454 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22456 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22457 * configure: Regenerate.
22459 2013-01-10 Richard Biener <rguenther@suse.de>
22461 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22463 2013-01-10 Richard Biener <rguenther@suse.de>
22466 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22467 locations for virtual PHI arguments.
22468 (rewrite_update_phi_arguments): Likewise.
22470 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22472 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22475 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22477 PR tree-optimization/55921
22478 * tree-complex.c (expand_complex_asm): New function.
22479 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22481 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22484 * config/s390/s390.c (s390_symref_operand_p)
22485 (s390_loadrelative_operand_p): Merge the two functions.
22486 (s390_check_qrst_address, print_operand_address): Add parameters
22487 to s390_loadrelative_operand_p invokation.
22488 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22489 (s390_reload_larl_operand, s390_secondary_reload): Use
22490 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22491 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22493 2013-01-09 Mike Stump <mikestump@comcast.net>
22495 * dse.c (record_store): Remove unnecessary assert.
22497 2013-01-09 Jan Hubicka <jh@suse.cz>
22499 PR tree-optimization/55569
22500 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22501 * cfgloop.h (scale_loop_profile): Likewise.
22503 2013-01-09 Jan Hubicka <jh@suse.cz>
22506 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22508 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22510 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22512 PR middle-end/55114
22513 * expr.h (maybe_emit_group_store): Declare.
22514 * expr.c (maybe_emit_group_store): New function.
22515 * builtins.c (expand_builtin_int_roundingfn): Call it.
22516 (expand_builtin_int_roundingfn_2): Likewise.
22518 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22520 PR rtl-optimization/55829
22521 * lra-constraints.c (match_reload): Add code for absent output.
22522 (curr_insn_transform): Add code for reloads of matched inputs
22525 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22527 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22528 attribute of movddup insn to DF.
22529 (*vec_interleave_lowv2df): Ditto.
22530 (vec_dupv2df): Ditto.
22532 2013-01-09 Jan Hubicka <jh@suse.cz>
22534 PR tree-optimiation/55875
22535 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22536 EVERY_ITERATION parameter.
22537 (number_of_iterations_exit): Check if exit is executed every iteration.
22538 (idx_infer_loop_bounds): Similarly here.
22539 (n_of_executions_at_most): Simplify
22540 to only test for cases where statement is dominated by the
22541 particular bound; handle correctly the "postdominance" test.
22542 (scev_probably_wraps_p): Use max loop iterations info
22543 as a global bound first.
22545 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22546 Nick Clifton <nickc@redhat.com>
22548 * config/v850/v850.md (cbranchsf4): New pattern.
22549 (cstoresf4): New pattern.
22550 (cbranchdf4): New pattern.
22551 (cstoredf4): New pattern.
22552 (movsicc): Disallow floating point comparisons.
22553 (cmpsf_le_insn): Fix order of operators.
22554 (cmpsf_lt_insn): Likewise.
22555 (cmpsf_eq_insn): Likewise.
22556 (cmpdf_le_insn): Likewise.
22557 (cmpdf_lt_insn): Likewise.
22558 (cmpdf_eq_insn): Likewise.
22559 (cmpsf_ge_insn): Use LE comparison.
22560 (cmpdf_ge_insn): Likewise.
22561 (cmpsf_gt_insn): Use LT comparison.
22562 (cmpdf_gt_insn): Likewise.
22563 (cmpsf_ne_insn): Delete pattern.
22564 (cmpdf_ne_insn): Delete pattern.
22565 * config/v850/v850.c (v850_gen_float_compare): Use
22566 gen_cmpdf_eq_insn for NE comparison.
22567 (v850_float_z_comparison_operator)
22568 (v850_float_nz_comparison_operator): Move from here ...
22569 * config/v850/predicates.md: ... to here. Move GT and GE
22570 comparisons into v850_float_z_comparison_operator.
22571 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22573 (v850_float_nz_comparison_operator): Likewise.
22575 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22577 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22578 with calls to gen_insvsi/gen_insvdi.
22580 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22582 * config/i386/i386.c (initial_ix86_tune_features): Set up
22583 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22585 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22586 Jakub Jelinek <jakub@redhat.com>
22588 PR tree-optimization/48189
22589 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22590 If nitercst is 0, don't predict the exit edge.
22592 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22594 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22595 in asm_fprintf with reg_names.
22596 (aarch64_print_operand_address): Likewise.
22597 (aarch64_return_addr): Likewise.
22598 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22600 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22602 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22603 (INT_U6_BITS): Likewise.
22604 * config/pa/predicates.md (uint6_operand): New predicate.
22605 (shift5_operand, shift6_operand): Likewise.
22606 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22608 (lshrdi3): Use shift6_operand.
22609 (shrpsi4, shrpdi4): New insn patterns.
22610 (extzv): Delete expander.
22611 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22612 predicates in unamed zero extract patterns. Tighten common constraint.
22613 (extv): Delete expander.
22614 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22615 predicates in unamed sign extract patterns. Tighten common constraint.
22616 (insv): Delete expander.
22617 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22618 predicates in unamed insert patterns. Tighten common constraint.
22619 Change uint32_operand predicate to uint6_operand predicate in unamed
22620 DImode pattern to insert constant values of type 1...1xxxx.
22622 2013-01-04 Jan Hubicka <jh@suse.cz>
22624 PR tree-optimization/55823
22625 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22628 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22629 Uros Bizjak <ubizjak@gmail.com>
22631 PR rtl-optimization/55845
22632 * df-problems.c (can_move_insns_across): Stop scanning at
22633 volatile_insn_p source instruction or give up if
22634 across_from .. across_to range contains any volatile_insn_p
22637 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22639 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22640 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22642 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22643 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22645 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22648 * asan.c (asan_clear_shadow): New function.
22649 (asan_emit_stack_protection): Use it.
22651 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22653 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22654 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22655 with tab instead of space.
22657 2013-01-08 Nick Clifton <nickc@redhat.com>
22659 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22660 register bank 0 at the start of an interrupt handler.
22661 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22664 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22666 * config/aarch64/aarch64-simd.md
22667 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22668 (aarch64_simd_bsl): Likewise.
22669 (aarch64_vcond_internal<mode>): Likewise.
22670 (vcond<mode><mode>): Likewise.
22671 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22672 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22674 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22676 * config/aarch64/aarch64-builtins.c
22677 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22679 2013-01-08 Martin Jambor <mjambor@suse.cz>
22682 * tree-sra.c (analyze_access_subtree): Return true also after
22683 potentially creating a debug-only replacement.
22685 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22687 PR middle-end/55890
22688 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22690 PR tree-optimization/54120
22691 * tree-vrp.c (range_fits_type_p): Don't allow
22692 src_precision < precision from signed vr to unsigned_p
22693 if vr->min or vr->max is negative.
22694 (simplify_float_conversion_using_ranges): Test can_float_p
22695 against CODE_FOR_nothing.
22697 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22698 Richard Biener <rguenther@suse.de>
22700 PR middle-end/55851
22701 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22702 types instead of just INTEGER_TYPE types.
22704 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22706 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22709 2013-01-07 Steve Ellcey <sellcey@mips.com>
22712 * config/mips/mips.opt: Change mad to mmad to match documentation.
22714 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22717 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22718 .progmemx.data now.
22720 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22723 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22724 (avr_addrspace_t): Add .section_name field.
22725 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22727 (avr_addrspace): Same. Initialize .section_name. Remove last
22728 NULL entry. Put __memx into .progmemx.data.
22729 (progmem_section_prefix): Remove.
22730 (avr_asm_init_sections): No need to initialize progmem_section.
22731 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22732 section name prefix.
22733 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22734 retrieve the progmem section.
22735 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22736 boundary to run over avr_addrspace[].
22737 (avr_register_target_pragmas): Ditto.
22739 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22741 * varasm.c (output_constant_def_contents): For asan_protect_global
22742 protected strings, adjust DECL_ALIGN if needed, before testing for
22744 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22745 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22747 (output_object_block): For asan protected decls, emit asan padding
22748 after their contents.
22749 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22750 (asan_finish_file): Test it here instead.
22752 2013-01-07 Nick Clifton <nickc@redhat.com>
22753 Matthias Klose <doko@debian.org>
22754 Doug Kwan <dougkwan@google.com>
22755 H.J. Lu <hongjiu.lu@intel.com>
22758 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22760 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22762 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22764 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22766 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22768 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22771 * doc/install.texi (Cross-Compiler-Specific Options): Document
22774 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22776 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22777 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22778 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22779 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22780 vqmovun_high_s64): Fix source operand number and update copyright.
22782 2013-01-07 Richard Biener <rguenther@suse.de>
22784 PR middle-end/55890
22785 * gimple.h (gimple_call_builtin_p): New overload.
22786 * gimple.c (validate_call): New function.
22787 (gimple_call_builtin_p): Likewise.
22788 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22789 Use gimple_call_builtin_p.
22790 (find_func_clobbers): Likewise.
22791 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22792 (strlen_optimize_stmt): Likewise.
22794 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22796 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22797 (vld1q_dup_*): Likewise.
22798 (vld1_*): Likewise.
22799 (vld1q_*): Likewise.
22800 (vld1_lane_*): Likewise.
22801 (vld1q_lane_*): Likewise.
22803 2013-01-07 Richard Biener <rguenther@suse.de>
22805 * lto-streamer.h (LTO_minor_version): Bump to 2.
22807 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22809 * config/aarch64/aarch64-protos.h
22810 (aarch64_const_double_zero_rtx_p): Rename to...
22811 (aarch64_float_const_zero_rtx_p): ...this.
22812 (aarch64_float_const_representable_p): New.
22813 (aarch64_output_simd_mov_immediate): Likewise.
22814 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22815 move immediate case.
22816 * config/aarch64/aarch64.c
22817 (aarch64_const_double_zero_rtx_p): Rename to...
22818 (aarch64_float_const_zero_rtx_p): ...this.
22819 (aarch64_print_operand): Allow printing of new constants.
22820 (aarch64_valid_floating_const): New.
22821 (aarch64_legitimate_constant_p): Check for valid floating-point
22823 (aarch64_simd_valid_immediate): Likewise.
22824 (aarch64_vect_float_const_representable_p): New.
22825 (aarch64_float_const_representable_p): Likewise.
22826 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22827 (aarch64_output_simd_mov_immediate): New.
22828 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22829 (*movdf_aarch64): Likewise.
22830 * config/aarch64/constraints.md (Ufc): New.
22831 (Y): call aarch64_float_const_zero_rtx.
22832 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22834 2013-01-07 Richard Biener <rguenther@suse.de>
22836 PR tree-optimization/55888
22837 PR tree-optimization/55862
22838 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22839 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22840 not if it is contained therein.
22842 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22844 * config/avr/t-avr: Typo.
22846 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22849 * config/avr/t-avr: Don't automatically rebuild
22850 $(srcdir)/config/avr/t-multilib
22851 $(srcdir)/config/avr/avr-tables.opt
22852 $(srcdir)/doc/avr-mmcu.texi
22853 (avr-mcus): New phony target to build them on request.
22854 (s-avr-mlib, s-avr-mmcu-texi): Remove.
22855 * avr/avr-mcus.def: Adjust comments.
22857 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
22859 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
22861 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22863 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
22865 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22867 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
22869 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
22871 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
22872 to generate profiling.
22873 * config/rs6000/aix64.h (LIB_SPEC): Same.
22875 2013-01-04 Andrew Pinski <apinski@cavium.com>
22877 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
22879 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
22881 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
22883 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
22885 (ix86_expand_move): Ditto.
22886 (ix86_zero_extend_to_Pmode): Ditto.
22887 (ix86_expand_call): Ditto.
22888 (ix86_expand_special_args_builtin): Ditto.
22889 (ix86_expand_builtin): Ditto.
22891 2013-01-04 Richard Biener <rguenther@suse.de>
22893 PR tree-optimization/55862
22894 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
22895 translating them through PHI nodes.
22897 2013-01-04 Martin Jambor <mjambor@suse.cz>
22899 PR tree-optimization/55755
22900 * tree-sra.c (sra_modify_assign): Do not check that an access has no
22901 children when trying to avoid producing a VIEW_CONVERT_EXPR.
22903 2013-01-04 Marek Polacek <polacek@redhat.com>
22905 PR middle-end/55859
22906 * opts.c (default_options_optimization): Clarify error message.
22908 2013-01-04 Richard Biener <rguenther@suse.de>
22910 PR middle-end/55863
22911 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
22914 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22917 * config/pa/pa.md (movsi): Revert previous change.
22918 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
22921 2013-01-03 Richard Henderson <rth@redhat.com>
22923 * config/i386/i386.c (ix86_expand_move): Always assign to op1
22924 after eliminating TLS symbols.
22926 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
22929 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
22930 * graphite-poly.c (debug_gmp_value): Likewise.
22932 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
22935 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
22936 selected code model, define __code_mode_small__, __code_model_medium__,
22937 __code_model_large__, __code_model_32__ or __code_model_kernel__.
22938 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
22939 xchg temporary register with %k. Declare temporary register as
22941 [__x86_64__]: For medium and large code models, preserve %rbx register.
22943 2013-01-03 Richard Biener <rguenther@suse.de>
22945 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
22946 (dump_subscript): Adjust.
22947 (finalize_ddr_dependent): Do not dump redundant info.
22948 (analyze_siv_subscript): Adjust.
22949 (subscript_dependence_tester): Likewise.
22950 (compute_affine_dependence): Likewise.
22952 2013-01-03 Richard Biener <rguenther@suse.de>
22955 2013-01-03 Richard Biener <rguenther@suse.de>
22957 PR tree-optimization/55857
22958 * tree-vect-stmts.c (vectorizable_load): Do not setup
22959 re-alignment for invariant loads.
22961 2013-01-02 Richard Biener <rguenther@suse.de>
22963 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22964 invariant load do not generate a vector load from the scalar location.
22966 2013-01-03 Richard Biener <rguenther@suse.de>
22968 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
22969 for not vectorizing.
22970 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
22971 not build INDIRECT_REFs, call get_name once only.
22972 (vect_create_data_ref_ptr): Likewise. Dump base object kind
22973 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
22975 2013-01-03 Richard Biener <rguenther@suse.de>
22977 PR tree-optimization/55857
22978 * tree-vect-stmts.c (vectorizable_load): Do not setup
22979 re-alignment for invariant loads.
22981 2013-01-03 Richard Biener <rguenther@suse.de>
22984 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
22985 prefer a built-in decl.
22987 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22989 * gcc.c (process_command): Update copyright notice dates.
22990 * gcov.c (print_version): Likewise.
22991 * gcov-dump.c (print_version): Likewise.
22993 PR rtl-optimization/55838
22994 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
22995 iv0.step, iv1.step and step.
22997 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22998 Marc Glisse <marc.glisse@inria.fr>
23000 PR tree-optimization/55832
23001 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23002 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23003 integer_{one,zero}_node.
23005 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23008 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23009 * var-tracking.c (reverse_op): Don't add reverse ops to
23010 VALUEs that have already
23011 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23013 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23015 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23017 2013-01-02 Teresa Johnson <tejohnson@google.com>
23019 * dumpfile.c (dump_loc): Print filename with location.
23020 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23021 new location_t parameter to emit complete unroll message with
23022 new dump framework.
23023 (canonicalize_loop_induction_variables): Compute loops location
23024 and pass to try_unroll_loop_completely.
23025 * loop-unroll.c (report_unroll_peel): New function.
23026 (peel_loops_completely): Use new dump format with location
23027 for main dumpfile message, and invoke report_unroll_peel on success.
23028 (decide_unrolling_and_peeling): Ditto.
23029 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23030 by report_unroll_peel.
23031 (decide_peel_completely): Ditto.
23032 (decide_unroll_constant_iterations): Ditto.
23033 (decide_unroll_runtime_iterations): Ditto.
23034 (decide_peel_simple): Ditto.
23035 (decide_unroll_stupid): Ditto.
23036 * cfgloop.c (get_loop_location): New function.
23037 * cfgloop.h (get_loop_location): Declare.
23039 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23041 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23044 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23046 PR middle-end/55198
23047 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23048 BLKmode objects when EXPAND_MEMORY is specified.
23050 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23052 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23054 (fold_builtin_cpu): Do not share cpu model decls across statements.
23056 2013-01-02 Jason Merrill <jason@redhat.com>
23059 * tree.c (build_array_type_1): Revert earlier change.
23061 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23063 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23065 * config/aarch64/aarch64-tune.md: Re-generate.
23067 2013-01-02 Richard Biener <rguenther@suse.de>
23069 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23070 invariant load do not generate a vector load from the scalar location.
23072 2013-01-02 Richard Biener <rguenther@suse.de>
23075 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23076 * configure: Regenerate.
23078 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23080 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23081 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23082 (expand_builtin_int_roundingfn_2): Keep the original target around
23083 for the fallback case.
23085 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23087 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23088 to be clear for sign changes.
23090 2013-01-01 Jan Hubicka <jh@suse.cz>
23092 * ipa-inline-analysis.c: Fix formatting.
23094 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23096 PR tree-optimization/55831
23097 * tree-vect-loop.c (get_initial_def_for_induction): Use
23098 gsi_after_labels instead of gsi_start_bb.
23100 Copyright (C) 2013 Free Software Foundation, Inc.
23102 Copying and distribution of this file, with or without modification,
23103 are permitted in any medium without royalty provided the copyright
23104 notice and this notice are preserved.