1 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
5 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6 Vladimir Makarov <vmakarov@redhat.com>
8 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
9 from the least significant bit.
10 (strip_address_mutations): Add bitfield operations handling.
11 (must_be_index_p): Add shifting and rotate operations handling.
12 (set_address_base): Use must_be_base_p predicate.
13 (set_address_index):Use must_be_index_p predicate.
15 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
16 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
17 Sergey Lega <sergey.s.lega@intel.com>
18 Anna Tikhonova <anna.tikhonova@intel.com>
19 Ilya Tocar <ilya.tocar@intel.com>
20 Andrey Turetskiy <andrey.turetskiy@intel.com>
21 Ilya Verbin <ilya.verbin@intel.com>
22 Kirill Yukhin <kirill.yukhin@intel.com>
23 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
25 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
27 (ix86_expand_vector_move_misalign): Support new unaligned load and
28 stores and use new names.
29 (CODE_FOR_sse2_storedqu): Rename to ...
30 (CODE_FOR_sse2_storedquv16qi): ... this.
31 (CODE_FOR_sse2_loaddqu): Rename to ...
32 (CODE_FOR_sse2_loaddquv16qi): ... this.
33 (CODE_FOR_avx_loaddqu256): Rename to ...
34 (CODE_FOR_avx_loaddquv32qi): ... this.
35 (CODE_FOR_avx_storedqu256): Rename to ...
36 (CODE_FOR_avx_storedquv32qi): ... this.
37 * config/i386/i386.md (fpint_logic): New.
38 * config/i386/sse.md (VMOVE): Extend for AVX512.
42 (VI_UNALIGNED_LOADSTORE): Ditto.
43 (sse2_avx_avx512f): Ditto.
44 (sse2_avx2): Extend for AVX512.
47 (sse): Extend for AVX512.
50 (avxsizesuffix): Ditto.
51 (sseintvecmode): Ditto.
53 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
54 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
55 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
56 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
57 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
58 (<sse2_avx_avx512f>_storedqu<mode): ... this.
59 (<sse>_movnt<mode>): Replace constraint "x" with "v".
60 (STORENT_MODE): Extend for AVX512.
61 (*absneg<mode>2): Replace constraint "x" with "v".
63 (*ieee_smin<mode>3): Ditto.
64 (*ieee_smax<mode>3): Ditto.
65 (avx_cmp<mode>3): Replace VF with VF_128_256.
66 (*<sse>_maskcmp<mode>3_comm): Ditto.
67 (<sse>_maskcmp<mode>3): Ditto.
68 (<sse>_andnot<mode>3): Extend for AVX512.
69 (<code><mode>3, anylogic): Replace VF with VF_128_256.
70 (<code><mode>3, fpint_logic): New.
71 (*<code><mode>3): Extend for AVX512.
72 (avx512flogicsuff): New.
73 (avx512f_<logic><mode>): Ditto.
74 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
76 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
77 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
78 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
79 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
80 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
81 (xop_vpermil2<mode>3): Ditto.
82 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
83 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
84 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
85 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
87 2013-09-25 Tom Tromey <tromey@redhat.com>
89 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
90 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
91 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
92 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
93 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
94 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
95 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
96 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
97 (GRAPHITE_HTAB_H): Remove.
99 2013-09-25 Tom Tromey <tromey@redhat.com>
101 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
103 2013-09-25 Tom Tromey <tromey@redhat.com>
105 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
107 2013-09-25 Tom Tromey <tromey@redhat.com>
109 * config/i386/t-i386 (i386.o): Remove.
110 (i386-c.o): Use COMPILE and POSTCOMPILE.
112 2013-09-25 Tom Tromey <tromey@redhat.com>
114 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
116 2013-09-25 Tom Tromey <tromey@redhat.com>
118 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
119 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
120 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
121 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
122 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
123 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
124 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
125 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
126 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
127 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
128 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
129 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
130 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
131 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
132 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
133 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
134 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
135 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
136 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
137 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
138 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
139 (tree-ssa-pre.o, tree-ssa-sccvn.o)
140 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
141 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
142 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
143 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
144 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
145 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
146 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
147 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
148 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
149 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
150 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
151 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
152 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
153 (graphite-clast-to-gimple.o, graphite-dependences.o)
154 (graphite-interchange.o, graphite-poly.o)
155 (graphite-scop-detection.o, graphite-sese-to-poly.o)
156 (graphite-optimize-isl.o, tree-vect-loop.o)
157 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
158 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
159 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
160 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
161 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
162 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
163 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
164 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
165 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
166 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
167 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
168 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
169 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
170 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
171 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
172 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
173 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
174 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
175 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
176 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
177 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
178 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
179 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
180 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
181 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
182 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
183 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
184 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
185 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
186 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
187 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
188 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
189 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
190 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
191 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
192 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
193 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
194 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
195 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
196 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
197 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
198 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
199 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
200 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
201 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
202 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
203 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
204 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
205 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
206 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
207 (gcov-dump.o): Remove.
208 (default-c.o): Use COMPILE and POSTCOMPILE.
209 (CFLAGS-gcc.o): New variable.
210 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
212 2013-09-25 Tom Tromey <tromey@redhat.com>
214 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
215 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
216 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
217 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
218 (c-family/c-pragma.o, c-family/c-pretty-print.o)
219 (c-family/c-semantics.o, c-family/c-ada-spec.o)
220 (c-family/array-notation-common.o, c-family/stub-objc.o)
221 (c-family/c-ubsan.o): Remove.
223 2013-09-25 Tom Tromey <tromey@redhat.com>
225 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
227 2013-09-25 Tom Tromey <tromey@redhat.com>
229 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
230 to add -DENABLE_SHARED_LIBGCC.
231 (gcc.o): Don't use subshell.
233 2013-09-25 Tom Tromey <tromey@redhat.com>
235 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
236 * configure.ac: Don't invoke AM_PROG_CC_C_O.
237 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
238 * configure, config.in: Rebuild.
240 2013-09-25 Tom Tromey <tromey@redhat.com>
242 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
243 (COMPILE, POSTCOMPILE): New variables.
244 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
245 (DEPFILES): New variable.
247 * configure.ac: Add checks for dependency checking.
248 * configure, aclocal.m4: Regenerate.
250 2013-09-25 Tom Tromey <tromey@redhat.com>
252 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
253 ($(ALL_HOST_OBJS)): Move order-only dependency to end
256 2013-09-25 Tom Tromey <tromey@redhat.com>
258 * Makefile.in (generated_files): Add options.h,
259 target-hooks-def.h, insn-opinit.h,
260 common/common-target-hooks-def.h, pass-instances.def,
261 c-family/c-target-hooks-def.h.
263 2013-09-25 Jeff Law <law@redhat.com>
265 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
266 than foo[foo.length () - 1] to access last member in a vec.
267 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
269 2013-09-25 Richard Biener <rguenther@suse.de>
272 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
274 2013-09-25 Jan Hubicka <jh@suse.cz>
276 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
279 2013-09-25 Marek Polacek <polacek@redhat.com>
282 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
283 when determining the type name.
285 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
287 * config/sh/sh.md: Fix formatting.
289 2013-09-24 Xinliang David Li <davidxl@google.com>
291 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
292 Check max peel iterations parameter.
293 * param.def: New parameter.
294 * doc/invoke.texi: Document New parameter.
296 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
298 * gimple-pretty-print.c: Various whitespace tweaks.
299 * tree-core.h: Likewise.
300 * tree-pretty-print.c: Likewise.
301 * tree-ssa-alias.c: Likewise.
302 * tree-ssa-copy.c: Likewise.
303 * tree-ssanames.c: Likewise.
304 * tree-ssanames.h: Likewise.
305 * tree-vrp.c: Likewise.
307 2013-09-24 Alan Modra <amodra@gmail.com>
311 * stmt.c (expand_asm_operands): Call expand_expr with
312 EXPAND_MEMORY for output operands that disallow regs. Don't
313 use EXPAND_WRITE on inout operands.
315 2013-09-24 Richard Biener <rguenther@suse.de>
318 * tree.c (reference_alias_ptr_type): Move ...
319 * alias.c (reference_alias_ptr_type): ... here and implement
320 in terms of the new reference_alias_ptr_type_1.
321 (ref_all_alias_ptr_type_p): New helper.
322 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
323 use ref_all_alias_ptr_type_p.
324 (get_deref_alias_set): Add flag_strict_aliasing check here.
325 (reference_alias_ptr_type_1): New function, split out from ...
326 (get_alias_set): ... here.
327 (alias_ptr_types_compatible_p): New function.
328 * alias.h (reference_alias_ptr_type): Declare.
329 (alias_ptr_types_compatible_p): Likewise.
330 * tree.h (reference_alias_ptr_type): Remove.
331 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
332 to compare MEM_REF alias types.
334 2013-09-24 Richard Biener <rguenther@suse.de>
336 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
338 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
340 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
341 reload helper function arrays into a single array reg_addr.
342 (reload_fpr_gpr): Likewise.
343 (reload_gpr_vsx): Likewise.
344 (reload_vsx_gpr): Likewise.
345 (struct rs6000_reg_addr): Likewise.
346 (reg_addr): Likewise.
347 (rs6000_debug_reg_global): Change rs6000_vector_reload,
348 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
349 (rs6000_init_hard_regno_mode_ok): Likewise.
350 (rs6000_secondary_reload_direct_move): Likewise.
351 (rs6000_secondary_reload): Likewise.
353 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
354 constraints: wu, ww, and wy. Repurpose wv constraint added during
355 power8 changes. Put wg constraint in alphabetical order.
357 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
358 for future work to add ISA 2.07 VSX single precision support.
359 (-mvsx-scalar-double): Change default from -1 to 1, update
360 documentation comment.
361 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
362 (-mupper-regs-df): New debug switch to control whether DF values
363 can go in the traditional Altivec registers.
364 (-mupper-regs-sf): New debug switch to control whether SF values
365 can go in the traditional Altivec registers.
367 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
369 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
370 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
371 Add new constraints, wu/ww/wy. Repurpose wv constraint.
372 (rs6000_debug_legitimate_address_p): Print if we are running
373 before, during, or after reload.
374 (rs6000_secondary_reload): Add a comment.
375 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
377 * config/rs6000/constraints.md (wa constraint): Sort w<x>
378 constraints. Update documentation string.
379 (wd constraint): Likewise.
380 (wf constraint): Likewise.
381 (wg constraint): Likewise.
382 (wn constraint): Likewise.
383 (ws constraint): Likewise.
384 (wt constraint): Likewise.
385 (wx constraint): Likewise.
386 (wz constraint): Likewise.
387 (wu constraint): New constraint for ISA 2.07 SFmode scalar
389 (ww constraint): Likewise.
390 (wy constraint): Likewise.
391 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
392 the previous submissions.
393 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
395 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
397 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
398 (REG_BR_PROB): Say that the probability is stored in an int_list.
399 * reg-notes.def: Update commentary to mention INT_LIST.
400 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
402 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
403 * rtlanal.c (int_reg_note_p): New function.
404 (alloc_reg_note): Assert that the note does not have an int argument.
405 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
406 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
407 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
408 rather than an INSN_LIST. Handle INT_LIST.
409 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
410 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
411 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
412 Manipulate them as ints rather than rtxes.
413 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
414 * regmove.c (copy_src_to_dest): Likewise.
415 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
417 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
418 into the cases that need it.
419 * config/arm/arm.c (arm_unwind_emit): Likewise.
421 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
422 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
423 * loop-doloop.c (add_test, doloop_modify): Likewise.
424 * loop-unswitch.c (compare_and_jump_seq): Likewise.
425 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
426 * predict.c (combine_predictions_for_insn): Likewise.
427 * print-rtl.c (print_rtx): Handle INT_LIST.
428 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
429 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
430 * config/arm/arm.c (emit_unlikely_jump): Likewise.
431 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
432 (ix86_split_fp_branch, predict_jump): Likewise.
433 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
434 * config/sh/sh.c (expand_cbranchsi4): Likewise.
435 * config/spu/spu.c (ea_load_store_inline): Likewise.
437 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
438 value of a REG_BR_PROB note.
439 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
440 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
441 * emit-rtl.c (try_split): Likewise.
442 * predict.c (br_prob_note_reliable_p): Likewise.
443 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
444 * reorg.c (mostly_true_jump): Likewise.
445 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
446 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
447 * config/i386/i386.c (ix86_print_operand): Likewise.
448 * config/ia64/ia64.c (ia64_print_operand): Likewise.
449 * config/mmix/mmix.c (mmix_print_operand): Likewise.
450 * config/rs6000/rs6000.c (output_cbranch): Likewise.
451 * config/s390/s390.c (s390_expand_tbegin): Likewise.
452 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
453 * config/sparc/sparc.c (output_cbranch): Likewise.
454 * config/spu/spu.c (get_branch_target): Likewise.
455 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
456 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
458 2013-09-23 Jan Hubicka <jh@suse.cz>
460 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
462 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
463 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
465 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
466 (gimple_fold_call): Dump inconsistent devirtualizations; add
467 sanity check for type based devirtualizations.
468 * ipa-prop.c: Include ipa-utils.h
469 (ipa_intraprocedural_devirtualization): Add sanity check.
470 (try_make_edge_direct_virtual_call): Likewise.
472 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
474 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
475 assignment statements.
477 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
479 * gimple-pretty-print.c (dump_ssaname_info): New function.
480 (dump_gimple_phi): Call it.
481 (pp_gimple_stmt_1): Likewise.
482 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
483 (range_info_def): Declare.
484 * tree-pretty-print.c (pp_double_int): New function.
485 (dump_generic_node): Call it.
486 * tree-pretty-print.h (pp_double_int): Declare.
487 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
488 * tree-ssanames.h (range_info_def): New structure.
489 (value_range_type): Move definition here.
490 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
492 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
494 (set_range_info): New function.
495 (get_range_info): Likewise.
496 (duplicate_ssa_name_range_info): Likewise.
497 (duplicate_ssa_name_fn): Check pointer type and call
498 duplicate_ssa_name_range_info.
499 * tree-ssa-copy.c (fini_copy_prop): Likewise.
500 * tree-vrp.c (value_range_type): Remove definition, now in
502 (vrp_finalize): Call set_range_info to update value range of
504 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
505 (SSA_NAME_RANGE_INFO): New macro.
507 2013-09-23 Richard Biener <rguenther@suse.de>
509 PR tree-optimization/58464
510 * tree-ssa-pre.c (phi_trans_lookup): Remove.
511 (phi_trans_add): Change to add conditionally on being not
513 (phi_translate_1): Remove recursion detection here.
514 (phi_translate): Pre-seed the cache with NULL to catch
515 recursion here in a more generic way.
516 (bitmap_find_leader): Adjust comment.
517 (get_representative_for): Dump value-numbers.
518 (create_expression_by_pieces): Likewise.
519 (insert_into_preds_of_block): Likewise.
521 2013-09-23 Christian Bruel <christian.bruel@st.com>
524 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
525 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
527 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
530 2013-09-20 Renlin Li <renlin.li@arm.com>
532 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
533 (aarch64_expand_epilogue): Likewise.
534 (aarch64_legitimize_reload_address): Likewise.
536 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
538 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
539 NULL_TREE before pushing them onto the vector. Likewise for labels.
541 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
543 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
544 comparison operators when -fno-trapping-math is in effect.
545 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
546 comparison operators in TFmode and assert that unsupported operators
548 (ia64_print_operand): Likewise.
550 2013-09-21 Jan Hubicka <jh@suse.cz>
552 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
553 (sse_typeless_stores): Enable for core
554 (sse_load0_by_pxor): Likewise.
555 (four_jump_limit): Disable for core.
556 (pad_returns): Likewise.
557 (avoid_vector_decode): Likewise.
558 (fuse_cmp_and_branch): Enable for cores.
559 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
561 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
564 * config/pa/pa.c (pa_option_override): Disable auto increment and
565 decrement instructions until reload is completed.
567 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
568 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
569 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
571 2013-09-20 DJ Delorie <dj@redhat.com>
572 Nick Clifton <nickc@redhat.com>
574 * config/rl78/rl78.c: Various whitespace and comment tweaks.
575 (need_to_save): Save bank 0 on interrupts.
576 (characterize_address): Strip far address wrappers.
577 (rl78_as_legitimate_address): Likewise.
578 (transcode_memory_rtx): Likewise.
579 (rl78_peep_movhi_p): Disable this peephole after devirt.
580 (rl78_propogate_register_origins): Forget all origins when a
582 * config/rl78/rl78-virt.md: Various whitespace tweaks.
583 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
585 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
586 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
587 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
589 * config/rl78/constraints.md: Various whitespace and paren tweaks.
591 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
593 * config/pa/pa.md: In "scc" insn patterns, change output template to
594 handle const0_rtx in reg_or_0_operand operands.
596 2013-09-20 Martin Husemann <martin@NetBSD.org>
599 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
600 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
602 2013-09-20 Richard Biener <rguenther@suse.de>
605 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
606 remembering SSA name version and block index.
607 (new_scev_info_str): Adjust.
608 (hash_scev_info): Likewise. Also hash the block index.
609 (eq_scev_info): Adjust.
610 (find_var_scev_info): Likewise.
611 (struct instantiate_cache_entry): Remove.
612 (struct instantiate_cache_type): Use a htab to map name, block
614 (instantiate_cache_type::~instantiate_cache_type): Adjust.
615 (get_instantiated_value_entry): Likewise.
616 (hash_idx_scev_info, eq_idx_scev_info): New functions.
617 (instantiate_scev_name): Adjust.
619 2013-09-20 Jeff Law <law@redhat.com>
621 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
623 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
625 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
626 Call aarch64_simd_expand_args to update op[argc].
628 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
630 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
633 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
635 * gengtype.c (file_rules): Added rule for *.cc files.
636 (get_output_file_with_visibility): Give fatal message when no
639 2013-09-20 Renlin Li <renlin.li@arm.com>
641 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
642 (aarch64_expand_epilogue): Likewise.
643 (aarch64_legitimize_reload_address): Likewise.
645 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
648 * expr.c (expand_assignment): Remove misalignp code path.
650 2013-09-20 Marek Polacek <polacek@redhat.com>
653 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
654 TYPE_PRECISION. Add asserts.
656 2013-09-20 Richard Biener <rguenther@suse.de>
658 PR tree-optimization/58453
659 * tree-loop-distribution.c (distribute_loop): Apply the cost
660 model for -ftree-loop-distribute-patterns, too.
662 2013-09-20 Richard Biener <rguenther@suse.de>
665 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
666 make type comparison less strict.
668 2013-09-20 Alan Modra <amodra@gmail.com>
670 * configure: Regenerate.
671 * aclocal.m4: Regenerate.
673 2013-09-20 Marek Polacek <polacek@redhat.com>
676 * doc/extend.texi: Document that attribute used is meant to be used
677 on variables with static storage duration.
679 2013-09-19 Jakub Jelinek <jakub@redhat.com>
681 PR tree-optimization/58472
682 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
683 simd_lane_access set inv_p = false.
684 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
685 the simduid magic VAR_DECL.
687 2013-09-19 Jan Hubicka <jh@suse.cz>
689 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
691 2013-09-17 Jeff Law <law@redhat.com>
693 * tree-ssa-dom.c (record_temporary_equivalences): New function
694 split out of dom_opt_dom_walker::after_dom_children.
695 (dom_opt_dom_walker::thread_across_edge): Move common code
696 in here from dom_opt_dom_walker::after_dom_children.
697 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
699 2013-09-19 Jan Hubicka <jh@suse.cz>
701 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
702 (TARGET_GENERIC): Use PROCESOR_GENERIC
703 (enum processor_type): Unify generic32 and 64.
704 * i386.md (cpu): Likewise.
705 * x86-tune.def (use_leave): Enable for generic32.
706 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
707 * athlon.md: Change generic64 to generic in all occurences.
708 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
709 (ix86_target_macros_internal): Likewise.
710 * driver-i386.c (host_detect_local_cpu): Likewise.
711 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
712 (generic_memcpy, generic_memset, generic_cost): This one.
713 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
714 (m_GENERIC32, m_GENERIC64): Remove.
715 (m_GENERIC): Turn into one flag.
716 (processor_target): Unify generic tunnings.
717 (ix86_option_override_internal): Replace generic32/64 by generic.
718 (ix86_issue_rate): Likewise.
719 (ix86_adjust_cost): Likewise.
721 2013-09-19 Jan Hubicka <jh@suse.cz>
723 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
726 2013-09-19 Jakub Jelinek <jakub@redhat.com>
728 * omp-low.c (expand_omp_sections): Always pass len - 1 to
729 GOMP_sections_start, even if !exit_reachable.
731 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
733 * lra-constraints.c (need_for_all_save_p): Use macro
734 HARD_REGNO_CALL_PART_CLOBBERED.
735 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
736 set up pseudo conflict hard regs.
738 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
741 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
744 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
746 PR rtl-optimization/58438
747 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
748 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
749 from previous subpasses.
751 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
753 * arm.c (arm_get_frame_offsets): Validate architecture supports
754 LDRD/STRD before accepting the tuning preference.
755 (arm_expand_prologue): Likewise.
756 (arm_expand_epilogue): Likewise.
758 2013-09-18 Richard Biener <rguenther@suse.de>
760 PR tree-optimization/58417
761 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
762 have chrecs with symbols defined in the loop as operands.
763 (chrec_fold_multiply): Likewise.
764 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
765 parameters before folding binary operations.
766 (struct instantiate_cache_entry_hasher): Remove.
767 (struct instantiate_cache_type): Use a pointer-map.
768 (instantiate_cache_type::instantiate_cache_type): New function.
769 (instantiate_cache_type::get): Likewise.
770 (instantiate_cache_type::set): Likewise.
771 (instantiate_cache_type::~instantiate_cache_type): Adjust.
772 (get_instantiated_value_entry): Likewise.
773 (global_cache): New global.
774 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
775 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
776 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
777 (instantiate_scev_name): Adjust.
778 (instantiate_scev): Construct global instead of local cache.
779 (resolve_mixers): Likewise.
781 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
782 Paolo Carlini <paolo.carlini@oracle.com>
785 * doc/implement-cxx.texi: Fix references to the C++ standards.
787 2013-09-18 Jakub Jelinek <jakub@redhat.com>
789 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
790 * tree-vect-data-refs.c (vect_analyze_data_refs): For
791 simd_lane_access drs, update also DR_ALIGNED_TO.
793 2013-09-18 Marek Polacek <polacek@redhat.com>
796 * doc/extend.texi: Document no_sanitize_undefined attribute.
797 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
798 no_sanitize_undefined attribute.
800 2013-09-18 Nick Clifton <nickc@redhat.com>
802 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
803 (ASM_DECLARE_FUNCTION_NAME): Define.
805 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
807 * compare-elim.c (find_comparison_dom_walker): New class
808 (find_comparisons_in_bb): Rename to
809 find_comparison_dom_walker::before_dom_children
810 (find_comparisons): Adjust
811 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
813 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
814 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
815 (init_walk_dominator_tree): Remove declaration.
816 (fini_walk_dominator_tree): Remove declaration.
817 * fwprop.c (single_def_use_dom_walker): New class
818 (single_def_use_enter_block): Convert to
819 single_def_use_dom_walker::before_dom_children.
820 (single_def_use_leave_block): Convert to
821 single_def_use_dom_walker::after_dom_children.
822 (build_single_def_use_links): Adjust.
823 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
825 (find_candidates_in_block): Convert to
826 find_candidates_dom_walker::before_dom_children.
827 (execute_strength_reduction): Adjust.
828 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
829 (sese_dom_walker): New class.
830 (sese_dom_walker::sese_dom_walker): New constructor.
831 (sese_dom_walker::~sese_dom_walker): New destructor.
832 (build_sese_conditions_before): Convert to
833 sese_dom_walker::before_dom_children.
834 (build_sese_conditions_after): Convert to
835 sese_dom_walker::after_dom_children.
836 (build_poly_scop): Adjust
837 * tree-into-ssa.c (rewrite_dom_walker): New class
838 (rewrite_enter_block): Convert to
839 rewrite_dom_walker::before_dom_children.
840 (rewrite_leave_block): Convert to
841 rewrite_dom_walker::after_dom_children.
842 (rewrite_update_dom_walker): New class.
843 (rewrite_update_enter_block): Convert to
844 rewrite_update_dom_walker::before_dom_children.
845 (rewrite_update_leave_block): Convert to
846 rewrite_update_dom_walker::after_dom_children.
847 (rewrite_blocks, rewrite_into_ssa): Adjust.
848 (mark_def_dom_walker): New class.
849 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
850 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
851 (mark_def_sites_blocks): Convert to
852 mark_def_dom_walker::before_dom_children.
853 (mark_def_site_blocks): Remove.
854 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
855 (tree_ssa_dominator_optimize): Adjust.
856 (dom_thread_across_edge): Convert to method
857 dom_opt_dom_walker::thread_across_edge.
858 (dom_opt_enter_block): Convert to member function
859 dom_opt_dom_walker::before_dom_children.
860 (dom_opt_leave_block): Convert to member function
861 dom_opt_dom_walker::after_dom_children.
862 * tree-ssa-dse.c (dse_dom_walker): New class.
863 (dse_enter_block): Convert to member function
864 dse_dom_walker::before_dom_children.
865 (tree_ssa_dse): Adjust.
866 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
867 (determine_invariantness_stmt): Convert to method
868 invariantness_dom_walker::before_dom_children.
869 (determine_invariantness): Remove
870 (move_computations_dom_walker): New class.
871 (move_computations_stmt): Convert to method
872 move_computations_dom_walker::before_dom_children.
873 (move_computations, tree_ssa_lim): Adjust.
874 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
875 (nt_init_block): Make method
876 notrappping_dom_walker::before_dom_children.
877 (nt_fini_block): Make
878 method nontrapping_dom_walker::after_dom_children.
879 (get_non_trapping): Adjust.
880 * tree-ssa-pre.c (eliminate_dom_walker): New class.
881 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
882 (eliminate_leave_block): Make method.
883 eliminate_dom_walker::after_dom_children.
885 * tree-ssa-strlen.c (strlen_dom_walker): New class.
886 (strlen_enter_block): Make method
887 strlen_dom_walker::before_dom_children.
888 (strlen_leave_block): Make
889 method strlen_dom_walker::after_dom_children.
890 (tree_ssa_strlen): Adjust.
891 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
892 (tree_ssa_uncprop): Adjust.
893 (uncprop_leave_block): Make method
894 uncprop_dom_walker::after_dom_children.
895 (uncprop_leave_block): Make method
896 uncprop_dom_walker::before_dom_children.
898 2013-09-18 Bin Cheng <bin.cheng@arm.com>
900 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
901 before branch in same basic block. Check both src and dest of
904 2013-09-17 Nick Clifton <nickc@redhat.com>
906 * config/rl78/rl78-real.md (bf): New pattern.
908 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
909 (rl78_print_operand): Do not put a # before a %B.
910 * config/rl78/rl78.opt: Tweak doc strings.
912 2013-09-17 DJ Delorie <dj@redhat.com>
914 * config/rl78/constraints.md (Wcv): Allow up to $r31.
915 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
916 (rl78_option_override): Likewise, if -mallregs.
917 (is_virtual_register): Likewise.
918 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
919 (REGNO_OK_FOR_BASE_P): Likewise.
920 * config/rl78/rl78.opt (-mallregs): New.
922 2013-09-17 Nick Clifton <nickc@redhat.com>
924 * config/rl78/rl78.c (need_to_save): Change return type to bool.
925 For interrupt functions: save all call clobbered registers if the
926 interrupt handler is not a leaf function.
927 (rl78_expand_prologue): Always recompute the frame information.
928 For interrupt functions: only select bank 0 if one of the bank 0
929 registers is going to be psuhed.
931 2013-09-17 DJ Delorie <dj@redhat.com>
933 * config/rl78/constraints.md: For each W* constraint, rename to C*
934 and create a W* constraint that checks for an optional ES: prefix
936 * config/rl78/rl78.md (UNS_ES_ADDR): New.
937 (es_addr): New. Used to wrap far addresses.
938 * config/rl78/rl78-protos.h (rl78_es_addr): New.
940 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
941 wrapped far addresses.
942 (rl78_print_operand_1): Unwrap far addresses before processing.
943 (rl78_lo16): Wrap far addresses in unspecs.
946 (insn_ok_now): Check for not-yet-wrapped far addresses.
947 (transcode_memory_rtx): Properly re-wrap far addresses.
949 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
951 * config/sparc/t-rtems: Add leon3 multilibs.
953 2013-09-17 Cong Hou <congh@google.com>
955 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
956 when checking the dot production pattern. The type of rhs operand
957 of multiply is now checked correctly.
959 2013-09-17 Jeff Law <law@redhat.com>
961 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
962 edge implied equivalences into successor phis.
963 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
964 here from tree-ssa-threadedge.c.
965 (mark_threaded_blocks): When threading through a joiner, if both
966 successors of the joiner's clone reach the same block, verify the
967 PHI arguments are equal. If not, cancel the jump threading request.
968 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
969 tree-ssa-threadupdate.c
970 (thread_across_edge): Don't check PHI argument equality when
971 threading through joiner block here.
973 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
975 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
976 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
977 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
978 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
979 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
980 (has_undefined_value_p): New. Pass dependant parts of
981 ssa_undefined_value_p.
982 (uninit_undefined_value_p): Use has_undefined_value_p.
983 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
984 make_pass_early_warn_uninitialized): Move from tree-ssa.c
985 * tree-ssa.h: Adjust prototypes
987 2013-09-17 Jan Hubicka <jh@suse.cz>
990 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
991 * ipa-inline.c (can_inline_edge_p): Do not downgrade
992 FUNCTION_NOT_OPTIMIZED.
993 * ipa-inline-analysis.c (compute_inline_parameters): Function
994 not optimized is not inlinable unless it is alwaysinline.
995 (inline_analyze_function): Force calls in not optimized
996 function not inlinable.
998 2013-09-17 Jan Hubicka <jh@suse.cz>
1001 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1003 * ipa.c (function_and_variable_visibility): Likewise.
1004 * ipa-profile.c (ipa_profile): Likewise.
1006 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1009 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1010 attribute, and don't inline the resolver in this case.
1012 2013-09-17 Teresa Johnson <tejohnson@google.com>
1014 * coverage.c (get_coverage_counts): Add missing newline.
1016 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1018 PR tree-optimization/58088
1019 * fold-const.c (mask_with_trailing_zeros): New function.
1020 (fold_binary_loc): Make sure we don't recurse infinitely
1021 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1022 Use mask_with_trailing_zeros where appropriate.
1024 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1026 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1027 of 'prev' var to get better distance estimation.
1029 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1031 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1032 * tree-inline.c (is_parameter_of): New predicate.
1033 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1034 a parameter has been remapped.
1035 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1036 (optimize_inline_calls): Initialize transform_parameter.
1037 (copy_gimple_seq_and_replace_locals): Likewise.
1038 (tree_function_versioning): Likewise.
1039 (maybe_inline_call_in_expr): Likewise.
1041 2013-09-17 Nick Clifton <nickc@redhat.com>
1043 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1044 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1046 (is_attr_func): New function.
1047 (msp430_is_interrupt_func): New function.
1048 (is_naked_func): New function.
1049 (is_reentrant_func): New function.
1050 (is_critical_func): New function.
1051 (msp430_start_function): Add annotations for function attributes.
1052 (msp430_attr): New function.
1053 (msp430_attribute_table): New.
1054 (msp430_function_section): New function.
1055 (TARGET_ASM_FUNCTION_SECTION): Define.
1056 (msp430_builtin): New enum.
1057 (msp430_init_builtins): New function.
1058 (msp430_builtin_devl): New function.
1059 (msp430_expand_builtin): New function.
1060 (TARGET_INIT_BUILTINS): Define.
1061 (TARGET_EXPAND_BUILTINS): Define.
1062 (TARGET_BUILTIN_DECL): Define.
1063 (msp430_expand_prologue): Add support for naked, interrupt,
1064 critical and reentrant functions.
1065 (msp430_expand_epilogue): Likewise.
1066 (msp430_print_operand): Handle 'O' character.
1067 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1069 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1070 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1071 (pushm): Use a 'n' rather than an 'i' constraint.
1072 (msp_return): Add generation of the interrupt return instruction.
1073 (disable_interrupts): New pattern.
1074 (enable_interrupts): New pattern.
1075 (push_intr_state): New pattern.
1076 (pop_intr_state): New pattern.
1077 (bic_SR): New pattern.
1078 (bis_SR): New pattern.
1079 * doc/extend.texi: Document MSP430 function attributes and builtin
1082 2013-09-17 Richard Biener <rguenther@suse.de>
1084 PR tree-optimization/58432
1085 * tree-loop-distribution.c (tree_loop_distribution): Also
1086 scan PHIs for outside loop uses and seed a partition from them.
1088 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1090 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1091 (restructure_reference): Call backtrace_base_for_ref.
1093 2013-09-17 Alan Modra <amodra@gmail.com>
1096 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
1098 2013-09-16 DJ Delorie <dj@redhat.com>
1100 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1101 vregs location for RL78/G10.
1102 (rl78_expand_prologue): Avoid SEL on G10.
1103 (rl78_expand_epilogue): Likewise.
1104 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1105 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1106 __RL78_G10__ when appropriate.
1107 (ASM_SPEC): Pass -mg10 along to the assembler.
1108 * config/rl78/rl78.md (sel_rb): Disable for G10.
1109 * config/rl78/rl78.opt: Add -mg10 option.
1110 * config/rl78/t-rl78: Add -mg10 multilib.
1112 2013-09-16 Xinliang David Li <davidxl@google.com>
1114 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1115 * omp-low.c (omp_max_vf): Ditto.
1116 (expand_omp_simd): Ditto.
1117 * tree-vectorizer.c (vectorize_loops): Ditto.
1118 (gate_vect_slp): Ditto.
1119 (gate_increase_alignment): Ditto.
1120 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1121 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1122 (gate_tree_vectorize): Name change.
1123 (tree_vectorize): Ditto.
1124 (pass_vectorize::gate): Call new function.
1125 (pass_vectorize::execute): Ditto.
1126 opts.c: O3 default setting change.
1127 (finish_options): Check new flag.
1128 * doc/invoke.texi: Document new flags.
1129 * common.opt: New flags.
1131 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1133 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1134 (Condition Code): Don't reference it.
1135 * doc/tm.texi: Regenerate.
1137 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1140 * lra-constraints.c (undo_optional_reloads): Consider all optional
1141 reload even if it did not get a hard reg.
1143 2013-09-16 Teresa Johnson <tejohnson@google.com>
1145 * dumpfile.c (dump_loc): Remove newline emission.
1146 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1147 emission to dump_printf_loc calls where missing.
1148 (vect_mark_for_runtime_alias_test): Ditto.
1149 (vect_analyze_data_ref_dependence): Ditto.
1150 (vect_analyze_data_ref_dependences): Ditto.
1151 (vect_slp_analyze_data_ref_dependence): Ditto.
1152 (vect_slp_analyze_data_ref_dependences): Ditto.
1153 (vect_compute_data_ref_alignment): Ditto.
1154 (vect_update_misalignment_for_peel): Ditto.
1155 (vect_verify_datarefs_alignment): Ditto.
1156 (vector_alignment_reachable_p): Ditto.
1157 (vect_get_data_access_cost): Ditto.
1158 (vect_enhance_data_refs_alignment): Ditto.
1159 (vect_find_same_alignment_drs): Ditto.
1160 (vect_analyze_data_refs_alignment): Ditto.
1161 (vect_analyze_group_access): Ditto.
1162 (vect_analyze_data_ref_access): Ditto.
1163 (vect_analyze_data_ref_accesses): Ditto.
1164 (vect_prune_runtime_alias_test_list): Ditto.
1165 (vect_analyze_data_refs): Ditto.
1166 (vect_create_addr_base_for_vector_ref): Ditto.
1167 (vect_create_data_ref_ptr): Ditto.
1168 (vect_grouped_store_supported): Ditto.
1169 (vect_grouped_load_supported): Ditto.
1170 * value-prof.c (check_counter): Ditto.
1171 (check_ic_target): Ditto.
1172 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1173 (vect_recog_widen_mult_pattern): Ditto.
1174 (vect_recog_widen_sum_pattern): Ditto.
1175 (vect_recog_over_widening_pattern): Ditto.
1176 (vect_recog_widen_shift_pattern): Ditto.
1177 (vect_recog_rotate_pattern): Ditto.
1178 (vect_recog_vector_vector_shift_pattern): Ditto.
1179 (vect_recog_divmod_pattern): Ditto.
1180 (vect_recog_mixed_size_cond_pattern): Ditto.
1181 (vect_recog_bool_pattern): Ditto.
1182 (vect_pattern_recog_1): Ditto.
1183 (vect_pattern_recog): Ditto.
1184 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1185 (vect_is_simple_iv_evolution): Ditto.
1186 (vect_analyze_scalar_cycles_1): Ditto.
1187 (vect_get_loop_niters): Ditto.
1188 (vect_analyze_loop_1): Ditto.
1189 (vect_analyze_loop_form): Ditto.
1190 (vect_analyze_loop_operations): Ditto.
1191 (vect_analyze_loop_2): Ditto.
1192 (vect_analyze_loop): Ditto.
1193 (report_vect_op): Ditto.
1194 (vect_is_slp_reduction): Ditto.
1195 (vect_is_simple_reduction_1): Ditto.
1196 (vect_get_known_peeling_cost): Ditto.
1197 (vect_estimate_min_profitable_iters): Ditto.
1198 (vect_model_reduction_cost): Ditto.
1199 (vect_model_induction_cost): Ditto.
1200 (get_initial_def_for_induction): Ditto.
1201 (vect_create_epilog_for_reduction): Ditto.
1202 (vectorizable_reduction): Ditto.
1203 (vectorizable_induction): Ditto.
1204 (vectorizable_live_operation): Ditto.
1205 (vect_loop_kill_debug_uses): Ditto.
1206 (vect_transform_loop): Ditto.
1207 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1208 (vect_stmt_relevant_p): Ditto.
1209 (process_use): Ditto.
1210 (vect_mark_stmts_to_be_vectorized): Ditto.
1211 (vect_model_simple_cost): Ditto.
1212 (vect_model_promotion_demotion_cost): Ditto.
1213 (vect_model_store_cost): Ditto.
1214 (vect_get_store_cost): Ditto.
1215 (vect_model_load_cost): Ditto.
1216 (vect_get_load_cost): Ditto.
1217 (vect_init_vector_1): Ditto.
1218 (vect_get_vec_def_for_operand): Ditto.
1219 (vect_finish_stmt_generation): Ditto.
1220 (vectorizable_call): Ditto.
1221 (vectorizable_conversion): Ditto.
1222 (vectorizable_assignment): Ditto.
1223 (vectorizable_shift): Ditto.
1224 (vectorizable_operation): Ditto.
1225 (vectorizable_store): Ditto.
1226 (vectorizable_load): Ditto.
1227 (vectorizable_condition): Ditto.
1228 (vect_analyze_stmt): Ditto.
1229 (vect_transform_stmt): Ditto.
1230 (vect_is_simple_use): Ditto.
1231 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1232 (vect_can_advance_ivs_p): Ditto.
1233 (vect_update_ivs_after_vectorizer): Ditto.
1234 (vect_do_peeling_for_loop_bound): Ditto.
1235 (vect_gen_niters_for_prolog_loop): Ditto.
1236 (vect_update_inits_of_drs): Ditto.
1237 (vect_create_cond_for_alias_checks): Ditto.
1238 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1239 (vect_build_slp_tree_1): Ditto.
1240 (vect_supported_load_permutation_p): Ditto.
1241 (vect_analyze_slp_instance): Ditto.
1242 (vect_analyze_slp): Ditto.
1243 (vect_make_slp_decision): Ditto.
1244 (vect_detect_hybrid_slp): Ditto.
1245 (vect_bb_vectorization_profitable_p): Ditto.
1246 (vect_slp_analyze_bb_1): Ditto.
1247 (vect_update_slp_costs_according_to_vf): Ditto.
1248 (vect_get_mask_element): Ditto.
1249 (vect_transform_slp_perm_load): Ditto.
1250 (vect_schedule_slp_instance): Ditto.
1251 (vect_schedule_slp): Ditto.
1252 (vect_slp_transform_bb): Ditto.
1253 * profile.c (read_profile_edge_counts): Ditto.
1254 (compute_branch_probabilities): Ditto.
1255 * coverage.c (get_coverage_counts): Ditto.
1257 2013-09-16 Diego Novillo <dnovillo@google.com>
1259 * tree-core.h: Add missing comment lines from refactoring
1262 2013-09-16 Jan Hubicka <jh@suse.cz>
1264 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1265 abstract functions; for static functions check the presence
1268 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1270 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1271 * config/aarch64/aarch64-simd.md
1272 (aarch64_mla_elt<mode>): New.
1273 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1274 (aarch64_mls_elt<mode>): Likewise.
1275 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1276 (aarch64_fma4_elt<mode>): Likewise.
1277 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1278 (aarch64_fma4_elt_to_128v2df): Likewise.
1279 (aarch64_fma4_elt_to_64df): Likewise.
1280 (fnma<mode>4): Likewise.
1281 (aarch64_fnma4_elt<mode>): Likewise.
1282 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1283 (aarch64_fnma4_elt_to_128v2df): Likewise.
1284 (aarch64_fnma4_elt_to_64df): Likewise.
1285 * config/aarch64/iterators.md (VDQSF): New.
1286 * config/aarch64/arm_neon.h
1287 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1288 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1290 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1292 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1293 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1294 (aarch64_mul3_elt_to_128df): Likewise.
1295 (aarch64_mul3_elt_to_64v2df): Likewise.
1296 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1298 (VMUL_CHANGE_NLANES) Likewise.
1301 * config/aarch64/arm_neon.h
1302 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1304 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1306 * config/aarch64/arm_neon.h
1307 (vcvtx_high_f32_f64): Fix parameters.
1309 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1310 Uros Bizjak <ubizjak@gmail.com>
1312 * config/alpha.c: Include tree-ssanames.h.
1314 2013-09-16 Richard Biener <rguenther@suse.de>
1316 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1317 (dot_rdg_1): Handle control_dd.
1318 (create_edge_for_control_dependence): New function.
1319 (create_rdg_edges): Add control dependences if asked for.
1320 (build_rdg): Likewise.
1321 (generate_loops_for_partition): If there are not necessary
1322 control stmts remove all their dependencies.
1323 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1324 (distribute_loop): Pass on control dependences.
1325 (tree_loop_distribution): Compute control dependences and remove
1326 restriction on number of loop nodes.
1328 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1330 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1333 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1335 * cse.c (try_const_anchors): Punt on CC modes.
1337 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1339 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1341 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1344 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1347 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1350 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1352 (darwin_objc1_section): Likewise.
1353 (darwin_file_end): Emit Image Info section when required.
1355 2013-09-14 Jan Hubicka <jh@suse.cz>
1357 * tree-into-ssa.c (gate_into_ssa): New.
1358 (pass_data_build_ssa): Use it.
1359 * cgraph.h (expand_thunk): Update prototype.
1360 * cgraphunit.c (analyze_function): Expand thunks early.
1361 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1362 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1363 set lowered flag; do not add new function.
1364 (assemble_thunks_and_aliases): Update.
1365 * tree-ssa.c (gate_init_datastructures): New gate.
1366 (pass_data_init_datastructures): Use it.
1368 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1371 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1372 xmm register set described in the psABI.
1374 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1376 * dwarf2out.c (should_emit_struct_debug): Add check
1377 for type_decl variable is not NULL.
1379 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1381 * config.gcc: Use new winnt-c.c target hooks
1382 * config/t-winnt: New file
1383 * config/winnt-c.c: New file
1384 * doc/tm.texi.in: Document new hook
1385 * doc/tm.texi: Regenerated
1387 2013-09-13 Jan Hubicka <jh@suse.cz>
1390 * ipa-inline.c (check_callers): New function.
1391 (check_caller_edge): Remove.
1392 (want_inline_function_to_all_callers_p): Also permit alises that are
1394 (inline_to_all_callers): Terminate the walk when devirtualization
1395 introduce new calls.
1397 2013-09-13 Jan Hubicka <jh@suse.cz>
1399 * ipa-inline-analysis.c (struct growth_data): Add node.
1400 (do_estimate_growth_1): Fix detection of recursion.
1402 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1404 PR tree-optimization/58392
1405 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1406 to avoid shadowing of outer loop variable. If
1407 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1408 replace_by_duplicate_decl simduid of loops that have it set and
1409 set dest_cfun->has_simduid_loops and/or
1410 dest_cfun->has_force_vect_loops.
1411 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1412 instead of maybe_lookup_decl.
1413 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1414 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1415 call. Copy over force_vect and copy and remap simduid. Set
1416 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1417 (copy_cfg_body): Remove blocks_to_copy argument. Use
1418 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1419 caller. Don't set cfun->has_simduid_loops and/or
1420 cfun->has_force_vect_loops here.
1421 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1423 (expand_call_inline, tree_function_versioning): Adjust copy_body
1426 2013-09-13 Martin Jambor <mjambor@suse.cz>
1429 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1430 the assert if the edge was a speculative one.
1432 2013-09-13 Richard Biener <rguenther@suse.de>
1434 * tree-data-ref.h (known_dependences_p): Move here ...
1435 * tree-loop-distribution.c (known_dependences_p): ... from here.
1436 (dump_rdg_component, debug_rdg_component): Remove.
1438 (generate_loops_for_partition): Use gimple_uid instead of
1439 relying on matching stmt visit order.
1440 (rdg_build_partitions): Take starting stmt vector.
1441 (ldist_gen): Merge into ...
1442 (distribute_loop): ... this function. Do not compute starting
1444 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1446 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1448 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1449 Set type attribute correctly. Set predicable_short_it attribute.
1450 (cmpsi_shiftsi): Remove %? from output template.
1452 2013-09-13 Richard Biener <rguenther@suse.de>
1454 * tree-loop-distribution.c (struct rdg_component,
1455 rdg_defs_used_in_other_loops_p, free_rdg_components,
1456 rdg_build_components): Remove.
1457 (stmts_from_loop): Do not record virtual PHIs.
1458 (generate_loops_for_partition): Skip virtual PHIs.
1459 (build_rdg_partition_for_component): Rename to ...
1460 (build_rdg_partition_for_vertex): ... this and adjust.
1461 (rdg_build_partitions): Take a vector of starting vertices
1462 instead of components. Remove unnecessary leftover handling.
1463 (ldist_gen): Do not build components or record other stores.
1464 (distribute_loop): Do not distribute loops containing stmts
1467 2013-09-13 Christian Bruel <christian.bruel@st.com>
1470 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1472 2013-09-13 Kai Tietz <ktietz@redhat.com>
1474 * config.gcc: Separate cases for mingw and cygwin targets,
1475 and add 64-bit cygwin target case.
1477 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1478 dll-export inline-functions.
1479 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1482 2013-09-13 Jeff Law <law@redhat.com>
1486 2013-09-06 Jeff Law <law@redhat.com>
1488 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1489 edge implied equivalences into successor phis.
1491 2013-09-12 DJ Delorie <dj@redhat.com>
1493 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1496 2013-09-12 Brooks Moses <bmoses@google.com>
1499 * Makefile.in: Do not install driver binaries in $(target)/bin.
1501 2013-09-12 DJ Delorie <dj@redhat.com>
1503 * config/rl78/rl78.opt (mrelax): New.
1504 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1505 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1507 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1509 (rl78_expand_epilogue): Likewise.
1511 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1514 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1515 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1516 pointer elimination with using remove_reg_equal_offset_note.
1518 2013-09-12 DJ Delorie <dj@redhat.com>
1520 * config/msp430/: New port.
1521 * config.gcc (msp430): Added.
1522 * doc/invoke.texi: Document MSP430 options.
1523 * doc/install.texi: Document msp430-elf
1524 * doc/md.texi: Document msp430-elf
1525 * doc/contrib.texi: Document msp430-elf
1527 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1530 2013-09-12 Martin Jambor <mjambor@suse.cz>
1533 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1534 reference descriptor is NULL.
1535 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1538 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1540 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1541 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1542 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1543 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1544 make_temp_ssa_name): move to tree-ssanames.h
1545 * tree-ssa-alias.h: Move prototype.
1546 * tree-ssa.h: Include tree-ssanames.h.
1547 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1548 * tree-ssanames.h: New. Move items from tree-flow*.h
1549 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1551 2013-09-12 Richard Biener <rguenther@suse.de>
1553 PR tree-optimization/58404
1554 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1555 propagate non-invariant addresses into dereferences wrapped
1556 in component references.
1558 2013-09-12 Richard Biener <rguenther@suse.de>
1560 PR tree-optimization/58402
1561 * passes.def: Move pass_late_warn_uninitialized later.
1563 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1565 * tree-ssa.h: New. Move content from tree-flow.h and
1567 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1568 Move prototypes belonging to tree-ssa.c.
1569 * tree-flow-inline.h (redirect_edge_var_map_def,
1570 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1572 * gimple.h: Adjust prototypes.
1573 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1575 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1576 * tree.h: Move prototype to tree-ssa.h.
1577 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1578 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1579 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1580 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1581 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1582 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1583 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1584 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1585 graphite-blocking.c, graphite-clast-to-gimple.c,
1586 graphite-dependences.c, graphite-interchange.c,
1587 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1588 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1589 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1590 ipa-reference.c, ipa-split.c, ipa-utils.c,
1591 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1592 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1593 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1594 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1595 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1596 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1597 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1598 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1599 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1600 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1601 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1602 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1603 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1604 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1605 value-prof.c, var-tracking.c,
1606 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1608 2013-09-12 Richard Biener <rguenther@suse.de>
1610 PR tree-optimization/58396
1611 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1612 (build_rdg): Take a loop-nest parameter, fix memleaks.
1613 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1615 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1617 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1620 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1622 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1623 The FMA instruction names should have a 'v' prefix.
1625 2013-09-12 Richard Biener <rguenther@suse.de>
1627 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1628 (dot_rdg): Use popen instead of system in optional code.
1629 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1630 (already_processed_vertex_p): Adjust.
1631 (has_anti_or_output_dependence, predecessor_has_mem_write,
1632 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1633 rdg_flag_uses): Remove.
1634 (rdg_flag_vertex): Simplify.
1635 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1637 (build_rdg_partition_for_component): Process the first vertex
1638 of a component only.
1639 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1641 2013-09-12 Alan Modra <amodra@gmail.com>
1643 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1645 2013-09-11 DJ Delorie <dj@redhat.com>
1646 Nick Clifton <nickc@redhat.com>
1648 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1649 (rl78_stack_based_mem): New.
1650 * config/rl78/constraints.md (Iv08): New.
1659 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1661 (movsi): Change from expand to insn-and-split.
1662 (ashrsi3): Clobber AX.
1666 * config/rl78/rl78.md (CC_REG): Fix.
1667 (addsi3): Allow memory and immediate operands.
1668 (addsi3_internal): Split into...
1669 (addsi3_internal_virt): ...new, and ...
1670 (addsi3_internal_real): ...new.
1672 (subsi3_internal_virt): New.
1673 (subsi3_internal_real): New.
1674 (mulsi3): Add memory operand.
1675 (mulsi3_rl78): Likewise.
1676 (mulsi3_g13): Likewise.
1677 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1678 (cbranchqi4_real): Add more constraint options.
1679 (cbranchhi4_real): Expand pattern.
1680 (cbranchhi4_real_signed): New.
1681 (cbranchhi4_real_inverted): New.
1682 (cbranchsi4_real_lt): New.
1683 (cbranchsi4_real_ge): New.
1684 (cbranchsi4_real_signed): New.
1685 (cbranchsi4_real): New.
1687 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1689 (lshrsi3_virt): Likewise.
1690 (ashlsi3_virt): Likewise.
1691 (cbranchqi4_virt_signed): New.
1692 (cbranchhi4_virt_signed): New.
1693 (cbranchsi4_virt): New.
1694 * config/rl78/rl78.c: Whitespace fixes throughout.
1695 (move_elim_pass): New.
1696 (pass_data_rl78_move_elim): New.
1697 (pass_rl78_move_elim): New.
1698 (make_pass_rl78_move_elim): New.
1699 (rl78_devirt_info): Run devirt earlier.
1700 (rl78_move_elim_info): New.
1701 (rl78_asm_file_start): Register it.
1702 (rl78_split_movsi): New.
1703 (rl78_as_legitimate_address): Allow virtual base registers when
1705 (rl78_addr_space_convert): Remove spurious debug stuff.
1706 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1707 (rl78_print_operand): More cases for not printing '#'.
1708 (rl78_expand_compare): Remove most of the logic.
1709 (content_memory): New.
1710 (clear_content_memory): New.
1711 (get_content_index): New.
1712 (get_content_name): New.
1713 (display_content_memory): New.
1714 (update_content): New.
1715 (record_content): New.
1716 (already_contains): New.
1717 (insn_ok_now): Re-recog insns with virtual registers.
1718 (add_postponed_content_update): New.
1719 (process_postponed_content_update): New.
1720 (gen_and_emit_move): New.
1721 (transcode_memory_rtx): Record new location content.
1722 Use gen_and_emit_move.
1723 (force_into_acc): New.
1724 (move_to_acc): Use gen_and_emit_move.
1725 (move_from_acc): Likewise.
1726 (move_acc_to_reg): Likewise.
1727 (move_to_x): Likewise.
1728 (move_to_hl): Likewise.
1729 (move_to_de): Likewise.
1730 (rl78_alloc_physical_registers_op1): Record location content.
1731 (has_constraint): New.
1732 (rl78_alloc_physical_registers_op2): Record location content.
1734 (rl78_alloc_physical_registers_ro1): Likewise.
1735 (rl78_alloc_physical_registers_cmp): Likewise.
1736 (rl78_alloc_physical_registers_umul): Likewise.
1737 (rl78_alloc_address_registers_macax): New.
1738 (rl78_alloc_physical_registers): Initialize and set location
1739 content memory as needed.
1740 (rl78_reorg): Make sure split2 is called.
1741 (rl78_rtx_costs): New.
1743 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1745 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1746 for (not (neg ...)) and (neg (not ...)) cases.
1748 2013-09-11 Richard Biener <rguenther@suse.de>
1751 * passes.def: Split critical edges before late uninit warning passes.
1752 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1754 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1756 PR tree-optimization/58385
1757 * fold-const.c (build_range_check): If both low and high are NULL,
1758 use omit_one_operand_loc to preserve exp side-effects.
1760 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1762 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1764 2013-09-11 Richard Biener <rguenther@suse.de>
1766 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1767 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1768 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1769 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1770 stmts_from_loop, known_dependences_p, build_empty_rdg,
1771 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1772 * tree-loop-distribution.c: ... here.
1773 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1774 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1775 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1776 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1777 * tree-loop-distribution.c: ... here.
1778 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1779 (struct partition_s): Add loops member.
1780 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1781 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1782 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1784 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1785 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1786 Sergey Lega <sergey.s.lega@intel.com>
1787 Anna Tikhonova <anna.tikhonova@intel.com>
1788 Ilya Tocar <ilya.tocar@intel.com>
1789 Andrey Turetskiy <andrey.turetskiy@intel.com>
1790 Ilya Verbin <ilya.verbin@intel.com>
1791 Kirill Yukhin <kirill.yukhin@intel.com>
1792 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1794 * config/i386/constraints.md (k): New.
1796 * config/i386/i386.c (const regclass_map): Add new mask registers.
1797 (dbx_register_map): Ditto.
1798 (dbx64_register_map): Ditto.
1799 (svr4_dbx_register_map): Ditto.
1800 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1802 (ix86_preferred_reload_class): Disable constants for mask registers.
1803 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1804 (ix86_hard_regno_mode_ok): Support new mask registers.
1805 (x86_order_regs_for_local_alloc): Ditto.
1806 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1807 (FIXED_REGISTERS): Add new mask registers.
1808 (CALL_USED_REGISTERS): Ditto.
1809 (REG_ALLOC_ORDER): Ditto.
1810 (VALID_MASK_REG_MODE): New.
1811 (FIRST_MASK_REG): Ditto.
1812 (LAST_MASK_REG): Ditto.
1813 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1814 (MAYBE_MASK_CLASS_P): New.
1815 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1816 (REG_CLASS_CONTENTS): Ditto.
1817 (MASK_REGNO_P): New.
1818 (ANY_MASK_REG_P): Ditto.
1819 (HI_REGISTER_NAMES): Add new mask registers.
1820 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1821 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1823 (attribute "type"): Add mskmov, msklog.
1824 (attribute "length_immediate"): Support them.
1825 (attribute "memory"): Ditto.
1826 (attribute "prefix_0f"): Ditto.
1827 (*movhi_internal): Support new mask registers.
1828 (*movqi_internal): Ditto.
1829 (define_split): Split out clobber pattern is a logic
1830 insn on mask registers.
1831 (*k<logic><mode>): New.
1832 (*andhi_1): Extend to support mask regs.
1833 (*andqi_1): Extend to support mask regs.
1835 (define_split): Split and-not to and and not if operands
1837 (*<code><mode>_1): Separate HI mode to new pattern...
1838 (*<code>hi_1): This.
1839 (*<code>qi_1): Extend to support mask regs.
1841 (kortestzhi): Ditto.
1842 (kortestchi): Ditto.
1844 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1845 (*one_cmplhi2_1): ...Here, now with mask registers support.
1846 (*one_cmplqi2_1): Support new mask registers.
1847 (HI/QImode arithmetics splitter): Don't split if mask registers
1849 (HI/QImode not splitter): Ditto.
1850 * config/i386/predicated.md (mask_reg_operand): New.
1851 (general_reg_operand): Ditto.
1853 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1855 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1856 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1858 2013-09-10 Jeff Law <law@redhat.com>
1860 PR tree-optimization/58380
1861 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1862 of threading through a buried loop header.
1864 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1865 return value for single successor case.
1867 2013-09-10 Jan Hubicka <jh@suse.cz>
1869 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1871 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1874 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1875 support conditional execution.
1876 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1878 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1880 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1882 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1883 Check destination too to check move insn.
1884 (undo_optional_reloads): Add check that the original peudo did not
1885 changed its allocation and the optional reload was inherited on last
1886 inheritance pass. Break loop after deciding to keep optional reload.
1887 (lra_undo_inheritance): Add check that inherited pseudo still in
1890 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1892 * config/aarch64/aarch64.md (generic_sched): New.
1893 * config/aarch64/aarch64-generic.md (load): Make conditional
1894 on generic_sched attribute.
1895 (nonload): Likewise.
1897 2013-09-10 Jan Hubicka <jh@suse.cz>
1899 * lto-cgraph.c: Include ipa-utils.h.
1900 (compute_ltrans_boundary): Also add possible targets into the boundary.
1902 2013-09-10 Jan Hubicka <jh@suse.cz>
1904 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1905 VAR_DECL of vtable rather than full expression.
1907 2013-09-10 Jan Hubicka <jh@suse.cz>
1908 Paolo Carlini <paolo.carlini@oracle.com>
1910 * cgraphunit.c (analyze_functions): Save input_location, set it
1911 to UNKNOWN_LOCATION and restore it at the end.
1913 2013-09-10 Martin Jambor <mjambor@suse.cz>
1915 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1916 represented by a thunk.
1918 2013-09-10 Jeff Law <law@redhat.com>
1920 PR tree-optimization/58343
1921 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1922 jump threading requests through loop headers buried in the middle
1923 of a jump threading path.
1925 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1926 in return value/type.
1928 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1930 PR rtl-optimization/58365
1931 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1932 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1935 2013-09-10 Richard Biener <rguenther@suse.de>
1937 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1938 * tree-data-ref.c (create_rdg_vertices): Collect all data
1939 references, signal failure to the caller, use data-ref API.
1940 (build_rdg): Compute data references only once. Maintain lifetime
1941 of data references and data dependences from within RDG.
1942 (free_rdg): Free dependence relations.
1943 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1944 inventing extra dependences.
1945 (distribute_loop): Update for RDG API changes.
1947 2013-09-10 Kai Tietz <ktietz@redhat.com>
1949 * doc/invoke.texi (fms-extensions): Document changed
1950 behavior for ms-abi targets.
1951 * config/i386/i386.c (ix86_option_override_internal):
1952 Set default value of option -fms-extension for ms-abi targets.
1954 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1956 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
1958 2013-09-10 Alan Modra <amodra@gmail.com>
1961 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
1963 2013-09-10 Alan Modra <amodra@gmail.com>
1965 * config/rs6000/predicates.md (add_cint_operand): New.
1966 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
1967 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
1968 using add_cint_operand.
1969 (largetoc_high_plus_aix): Likewise.
1971 2013-09-09 Jakub Jelinek <jakub@redhat.com>
1973 PR tree-optimization/58364
1974 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
1975 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
1976 the current range can't be an unconditional true or false.
1978 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
1980 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
1982 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
1984 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
1986 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
1989 * doc/invoke.texi (-Wdelete-incomplete): Document it.
1991 2013-09-09 Ian Bolton <ian.bolton@arm.com>
1993 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
1994 NO_REGS for immediate that can't be moved directly into FP_REGS.
1996 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1998 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
1999 comparison with negated operand.
2000 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2003 2013-09-09 Richard Biener <rguenther@suse.de>
2006 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2007 of a subloop record all its block as affecting loop-closed SSA form.
2009 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2011 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2012 of an rtx/bitpos pair.
2013 (store_fixed_bit_field): Update accordingly.
2015 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2017 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2019 * builtins.c (expand_errno_check): Likewise.
2020 * dwarf2cfi.c (init_return_column_size): Likewise.
2021 * except.c (sjlj_mark_call_sites): Likewise.
2022 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2023 * lra-constraints.c (emit_inc): Likewise.
2024 * ree.c (combine_set_extension): Likewise.
2025 * regmove.c (fixup_match_2): Likewise.
2026 * reload1.c (inc_for_reload): Likewise.
2028 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2030 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2031 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2032 Use gen_int_mode with the mode of the associated simplify_* call.
2033 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2034 * expmed.c (expand_shift_1): Likewise.
2035 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2036 * loop-iv.c (iv_number_of_iterations): Likewise.
2037 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2038 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2040 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2042 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2043 of the associated expand_* call.
2044 (asan_emit_stack_protection): Likewise.
2045 * builtins.c (round_trampoline_addr): Likewise.
2046 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2047 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2048 (emit_store_flag): Likewise.
2049 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2050 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2052 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2053 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2054 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2055 * modulo-sched.c (generate_prolog_epilog): Likewise.
2056 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2057 (expand_ctz, expand_ffs, expand_unop): Likewise.
2059 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2061 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2062 of the associated gen_rtx_* call.
2063 * caller-save.c (init_caller_save): Likewise.
2064 * combine.c (find_split_point, make_extraction): Likewise.
2065 (make_compound_operation): Likewise.
2066 * dwarf2out.c (mem_loc_descriptor): Likewise.
2067 * explow.c (plus_constant, probe_stack_range): Likewise.
2068 * expmed.c (expand_mult_const): Likewise.
2069 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2070 * reload1.c (init_reload): Likewise.
2071 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2072 * var-tracking.c (adjust_mems): Likewise.
2073 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2074 rather than gen_rtx_fmt_ee.
2076 2013-09-09 Jan Hubicka <jh@suse.cz>
2079 * value-prof.c (gimple_ic): Copy also abnormal edges.
2081 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2083 * asan.c (asan_shadow_cst): Use gen_int_mode.
2085 2013-09-08 Jan Hubicka <jh@suse.cz>
2087 * ipa-profile.c: Add toplevel comment.
2088 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2089 (contains_hot_call_p): New function.
2090 (ipa_propagate_frequency): Set frequencies based on counts when
2092 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2093 profile; do not tamper with profile after inlining if it is read.
2095 2013-09-08 Jan Hubicka <jh@suse.cz>
2097 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2100 2013-09-08 Jan Hubicka <jh@suse.cz>
2102 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2105 2013-09-08 Jeff Law <law@redhat.com>
2108 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2111 2013-09-08 Andi Kleen <ak@linux.intel.com>
2113 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2115 2013-09-08 Jan Hubicka <jh@suse.cz>
2117 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2119 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2121 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2122 for non-debug insns.
2123 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2124 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2125 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2128 2013-09-08 Jan Hubicka <jh@suse.cz>
2130 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2131 targets and devirtualize to BUILT_IN_UNREACHABLE.
2132 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2133 * ipa.c (walk_polymorphic_call_targets): New function.
2134 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2135 functions; use the new timevar.
2136 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2137 was removed from the program.
2138 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2139 not consider it in the walk when its vtable is dead.
2140 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2142 (devirt_variable_node_removal_hook): New function.
2143 (possible_polymorphic_call_targets): Also register
2144 devirt_variable_node_removal_hook.
2145 (ipa_devirt): Do not do non-speculative devirtualization.
2146 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2148 2013-09-08 Jan Hubicka <jh@suse.cz>
2150 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2151 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2152 varpool_remove_variable_insertion_hook): Declare.
2153 * varpool.c (varpool_node_hook_list): New structure.
2154 (first_varpool_node_removal_hook,
2155 first_varpool_variable_insertion_hook): New variables.
2156 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2157 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2158 varpool_remove_variable_insertion_hook,
2159 varpool_call_variable_insertion_hooks): New functions.
2160 (varpool_remove_node): Use it.
2162 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2165 * diagnostic.c (diagnostic_build_prefix): When s.file is
2166 "<built-in>" don't output line and column numbers.
2168 2013-09-06 Jan Hubicka <jh@suse.cz>
2170 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2171 * lto-streamer-out.c: Stream thunks, too.
2172 * lto-streamer-in.c (input_function): Pop cfun here
2173 (lto_read_body): Instead of here.
2175 2013-09-06 Caroline Tice <cmtice@google.com>
2177 * doc/install.texi: Add documentation for the --enable-vtable-verify
2178 and the --disable-libvtv configure options.
2180 2013-09-06 Jeff Law <law@redhat.com>
2182 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2183 edge implied equivalences into successor phis.
2185 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2187 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2189 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2191 * resource.c (mark_target_live_regs): Compute resources taking
2192 into account if a call is predicated or not.
2194 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2196 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2197 the compiler in the function names.
2199 2013-09-06 Jan Hubicka <jh@suse.cz>
2202 * ipa-inline.c (has_caller_p): New function.
2203 (want_inline_function_to_all_callers_p): Use it.
2204 (sum_callers, inline_to_all_callers): Break out from ...
2205 (ipa_inline): ... here.
2207 2013-09-06 Jan Hubicka <jh@suse.cz>
2209 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2210 only when AVX is enabled.
2212 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2214 * config/aarch64/aarch64.md
2215 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2216 is fpsimd_<load/store>2.
2217 (load_pair<mode>): Likewise.
2218 (store_pair<mode>): Likewise.
2220 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2222 * config/arm/types.md (type): Add "mrs" type.
2223 * config/aarch64/aarch64.md
2224 (aarch64_load_tp_hard): Make type "mrs".
2226 (load_tp_hard): Make type "mrs".
2227 * config/arm/cortex-a15.md: Update with new attributes.
2228 * config/arm/cortex-a5.md: Update with new attributes.
2229 * config/arm/cortex-a53.md: Update with new attributes.
2230 * config/arm/cortex-a7.md: Update with new attributes.
2231 * config/arm/cortex-a8.md: Update with new attributes.
2232 * config/arm/cortex-a9.md: Update with new attributes.
2233 * config/arm/cortex-m4.md: Update with new attributes.
2234 * config/arm/cortex-r4.md: Update with new attributes.
2235 * config/arm/fa526.md: Update with new attributes.
2236 * config/arm/fa606te.md: Update with new attributes.
2237 * config/arm/fa626te.md: Update with new attributes.
2238 * config/arm/fa726te.md: Update with new attributes.
2240 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2242 * config/aarch64/aarch64.md
2243 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2244 (*movtf_aarch64): Likewise.
2246 (thumb1_movdi_insn): Use "multiple" for type where more than one
2247 instruction is used for a move.
2248 (*arm32_movhf): Likewise.
2249 (*thumb_movdf_insn): Likewise.
2251 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2253 * config/arm/types.md (type): Rename fcpys to fmov.
2255 (*arm_movsi_vfp): Rename type fcpys as fmov.
2256 (*thumb2_movsi_vfp): Likewise
2257 (*movhf_vfp_neon): Likewise
2258 (*movhf_vfp): Likewise
2259 (*movsf_vfp): Likewise
2260 (*thumb2_movsf_vfp): Likewise
2261 (*movsfcc_vfp): Likewise
2262 (*thumb2_movsfcc_vfp): Likewise
2263 * config/aarch64/aarch64-simd.md
2264 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2265 * config/aarch64/aarch64.md
2266 (*movsi_aarch64): Replace type mov_reg with fmovs.
2267 (*movdi_aarch64): Likewise
2268 (*movsf_aarch64): Likewise
2269 (*movdf_aarch64): Likewise
2271 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2272 * config/arm/iwmmxt.md
2273 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2274 * config/arm/arm1020e.md: Update with new attributes.
2275 * config/arm/cortex-a15-neon.md: Update with new attributes.
2276 * config/arm/cortex-a5.md: Update with new attributes.
2277 * config/arm/cortex-a53.md: Update with new attributes.
2278 * config/arm/cortex-a7.md: Update with new attributes.
2279 * config/arm/cortex-a8-neon.md: Update with new attributes.
2280 * config/arm/cortex-a9.md: Update with new attributes.
2281 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2282 * config/arm/cortex-r4f.md: Update with new attributes.
2283 * config/arm/marvell-pj4.md: Update with new attributes.
2284 * config/arm/vfp11.md: Update with new attributes.
2286 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2288 * config/aarch64/aarch64.md
2289 (*madd<mode>): Fix type attribute.
2290 (*maddsi_uxtw): Likewise.
2291 (*msub<mode>): Likewise.
2292 (*msubsi_uxtw): Likewise.
2293 (<su_optab>maddsidi4): Likewise.
2294 (<su_optab>msubsidi4): Likewise.
2296 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2298 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2299 * config/arm/arm.md (core_cycles): Remove fdiv.
2300 * config/arm/vfp.md:
2301 (*sqrtsf2_vfp): Update for attribute changes.
2302 (*sqrtdf2_vfp): Likewise.
2303 * config/aarch64/aarch64.md:
2304 (sqrt<mode>2): Update for attribute changes.
2305 * config/arm/arm1020e.md: Update with new attributes.
2306 * config/arm/cortex-a15-neon.md: Update with new attributes.
2307 * config/arm/cortex-a5.md: Update with new attributes.
2308 * config/arm/cortex-a53.md: Update with new attributes.
2309 * config/arm/cortex-a7.md: Update with new attributes.
2310 * config/arm/cortex-a8-neon.md: Update with new attributes.
2311 * config/arm/cortex-a9.md: Update with new attributes.
2312 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2313 * config/arm/cortex-r4f.md: Update with new attributes.
2314 * config/arm/marvell-pj4.md: Update with new attributes.
2315 * config/arm/vfp11.md: Update with new attributes.
2317 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2319 * config/arm/types.md
2320 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2321 * config/aarch64/aarch64.md
2322 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2324 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2325 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2326 (float<GPI:mode><GPF:mode>2): Likewise.
2328 (*truncsisf2_vfp): Update with new attributes.
2329 (*truncsidf2_vfp): Likewise.
2330 (fixuns_truncsfsi2): Likewise.
2331 (fixuns_truncdfsi2): Likewise.
2332 (*floatsisf2_vfp): Likewise.
2333 (*floatsidf2_vfp): Likewise.
2334 (floatunssisf2): Likewise.
2335 (floatunssidf2): Likewise.
2336 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2337 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2338 * config/arm/arm1020e.md: Update with new attributes.
2339 * config/arm/cortex-a15-neon.md: Update with new attributes.
2340 * config/arm/cortex-a5.md: Update with new attributes.
2341 * config/arm/cortex-a53.md: Update with new attributes.
2342 * config/arm/cortex-a7.md: Update with new attributes.
2343 * config/arm/cortex-a8-neon.md: Update with new attributes.
2344 * config/arm/cortex-a9.md: Update with new attributes.
2345 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2346 * config/arm/cortex-r4f.md: Update with new attributes.
2347 * config/arm/marvell-pj4.md: Update with new attributes.
2348 * config/arm/vfp11.md: Update with new attributes.
2350 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2352 * config/aarch64/arm_neon.h
2353 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2354 (vqtbx<1,2,3,4><q>_s8): Likewise.
2356 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2358 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2359 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2360 (add<mode>3): Add type attribute.
2361 (add<mode>3): Likewise.
2362 (usadd<mode>3): Likewise.
2363 (ssadd<mode>3): Likewise.
2364 (sub<mode>3): Likewise.
2365 (sub<mode>3): Likewise.
2366 (ussub<mode>3): Likewise.
2367 (sssub<mode>3): Likewise.
2368 (ssmulsa3): Likewise.
2369 (usmulusa3): Likewise.
2370 (arm_usatsihi): Likewise.
2372 (*movdi_vfp): Add types for all instructions.
2373 (*movdi_vfp_cortexa8): Likewise.
2374 (*movhf_vfp_neon): Likewise.
2375 (*movhf_vfp): Likewise.
2376 (*movdf_vfp): Likewise.
2377 (*thumb2_movdf_vfp): Likewise.
2378 (*thumb2_movdfcc_vfp): Likewise.
2379 * config/arm/arm.md: Add type attribute to all insn patterns.
2380 (*thumb1_adddi3): Add type attribute.
2381 (*arm_adddi3): Likewise.
2382 (*adddi_sesidi_di): Likewise.
2383 (*adddi_zesidi_di): Likewise.
2384 (*thumb1_addsi3): Likewise.
2385 (addsi3_compare0): Likewise.
2386 (*addsi3_compare0_scratch): Likewise.
2387 (*compare_negsi_si): Likewise.
2388 (cmpsi2_addneg): Likewise.
2389 (*addsi3_carryin_<optab>): Likewise.
2390 (*addsi3_carryin_alt2_<optab>): Likewise.
2391 (*addsi3_carryin_clobercc_<optab>): Likewise.
2392 (*subsi3_carryin): Likewise.
2393 (*subsi3_carryin_const): Likewise.
2394 (*subsi3_carryin_compare): Likewise.
2395 (*subsi3_carryin_compare_const): Likewise.
2396 (*arm_subdi3): Likewise.
2397 (*thumb_subdi3): Likewise.
2398 (*subdi_di_zesidi): Likewise.
2399 (*subdi_di_sesidi): Likewise.
2400 (*subdi_zesidi_di): Likewise.
2401 (*subdi_sesidi_di): Likewise.
2402 (*subdi_zesidi_ze): Likewise.
2403 (thumb1_subsi3_insn): Likewise.
2404 (*arm_subsi3_insn): Likewise.
2405 (*anddi3_insn): Likewise.
2406 (*anddi_zesidi_di): Likewise.
2407 (*anddi_sesdi_di): Likewise.
2408 (*ne_zeroextracts): Likewise.
2409 (*ne_zeroextracts): Likewise.
2410 (*ite_ne_zeroextr): Likewise.
2411 (*ite_ne_zeroextr): Likewise.
2412 (*anddi_notdi_di): Likewise.
2413 (*anddi_notzesidi): Likewise.
2414 (*anddi_notsesidi): Likewise.
2415 (andsi_notsi_si): Likewise.
2416 (thumb1_bicsi3): Likewise.
2417 (*iordi3_insn): Likewise.
2418 (*iordi_zesidi_di): Likewise.
2419 (*iordi_sesidi_di): Likewise.
2420 (*thumb1_iorsi3_insn): Likewise.
2421 (*xordi3_insn): Likewise.
2422 (*xordi_zesidi_di): Likewise.
2423 (*xordi_sesidi_di): Likewise.
2424 (*arm_xorsi3): Likewise.
2425 (*andsi_iorsi3_no): Likewise.
2426 (*smax_0): Likewise.
2427 (*smax_m1): Likewise.
2428 (*arm_smax_insn): Likewise.
2429 (*smin_0): Likewise.
2430 (*arm_smin_insn): Likewise.
2431 (*arm_umaxsi3): Likewise.
2432 (*arm_uminsi3): Likewise.
2433 (*minmax_arithsi): Likewise.
2434 (*minmax_arithsi_): Likewise.
2435 (*satsi_<SAT:code>): Likewise.
2436 (arm_ashldi3_1bit): Likewise.
2437 (arm_ashrdi3_1bit): Likewise.
2438 (arm_lshrdi3_1bit): Likewise.
2439 (*arm_negdi2): Likewise.
2440 (*thumb1_negdi2): Likewise.
2441 (*arm_negsi2): Likewise.
2442 (*thumb1_negsi2): Likewise.
2443 (*negdi_extendsid): Likewise.
2444 (*negdi_zero_extend): Likewise.
2445 (*arm_abssi2): Likewise.
2446 (*thumb1_abssi2): Likewise.
2447 (*arm_neg_abssi2): Likewise.
2448 (*thumb1_neg_abss): Likewise.
2449 (one_cmpldi2): Likewise.
2450 (extend<mode>di2): Likewise.
2451 (*compareqi_eq0): Likewise.
2452 (*arm_extendhisi2addsi): Likewise.
2453 (*arm_movdi): Likewise.
2454 (*thumb1_movdi_insn): Likewise.
2455 (*arm_movt): Likewise.
2456 (*thumb1_movsi_insn): Likewise.
2457 (pic_add_dot_plus_four): Likewise.
2458 (pic_add_dot_plus_eight): Likewise.
2459 (tls_load_dot_plus_eight): Likewise.
2460 (*thumb1_movhi_insn): Likewise.
2461 (*thumb1_movsf_insn): Likewise.
2462 (*movdf_soft_insn): Likewise.
2463 (*thumb_movdf_insn): Likewise.
2464 (cbranchsi4_insn): Likewise.
2465 (cbranchsi4_scratch): Likewise.
2466 (*negated_cbranchsi4): Likewise.
2467 (*tbit_cbranch): Likewise.
2468 (*tlobits_cbranch): Likewise.
2469 (*tstsi3_cbranch): Likewise.
2470 (*cbranchne_decr1): Likewise.
2471 (*addsi3_cbranch): Likewise.
2472 (*addsi3_cbranch_scratch): Likewise.
2473 (*arm_cmpdi_insn): Likewise.
2474 (*arm_cmpdi_unsig): Likewise.
2475 (*arm_cmpdi_zero): Likewise.
2476 (*thumb_cmpdi_zero): Likewise.
2477 (*deleted_compare): Likewise.
2478 (*mov_scc): Likewise.
2479 (*mov_negscc): Likewise.
2480 (*mov_notscc): Likewise.
2481 (*cstoresi_eq0_thumb1_insn): Likewise.
2482 (cstoresi_nltu_thumb1): Likewise.
2483 (cstoresi_ltu_thu): Likewise.
2484 (thumb1_addsi3_addgeu): Likewise.
2485 (*arm_jump): Likewise.
2486 (*thumb_jump): Likewise.
2487 (*check_arch2): Likewise.
2488 (arm_casesi_internal): Likewise.
2489 (thumb1_casesi_dispatch): Likewise.
2490 (*arm_indirect_jump): Likewise.
2491 (*thumb1_indirect_jump): Likewise.
2493 (*and_scc): Likewise.
2494 (*ior_scc): Likewise.
2495 (*compare_scc): Likewise.
2496 (*cond_move): Likewise.
2497 (*cond_arith): Likewise.
2498 (*cond_sub): Likewise.
2499 (*cmp_ite0): Likewise.
2500 (*cmp_ite1): Likewise.
2501 (*cmp_and): Likewise.
2502 (*cmp_ior): Likewise.
2503 (*ior_scc_scc): Likewise.
2504 (*ior_scc_scc_cmp): Likewise.
2505 (*and_scc_scc): Likewise.
2506 (*and_scc_scc_cmp): Likewise.
2507 (*and_scc_scc_nod): Likewise.
2508 (*negscc): Likewise.
2509 (movcond_addsi): Likewise.
2510 (movcond): Likewise.
2511 (*ifcompare_plus_move): Likewise.
2512 (*if_plus_move): Likewise.
2513 (*ifcompare_move_plus): Likewise.
2514 (*if_move_plus): Likewise.
2515 (*ifcompare_arith_arith): Likewise.
2516 (*if_arith_arith): Likewise.
2517 (*ifcompare_arith_move): Likewise.
2518 (*if_arith_move): Likewise.
2519 (*ifcompare_move_arith): Likewise.
2520 (*if_move_arith): Likewise.
2521 (*ifcompare_move_not): Likewise.
2522 (*if_move_not): Likewise.
2523 (*ifcompare_not_move): Likewise.
2524 (*if_not_move): Likewise.
2525 (*ifcompare_shift_move): Likewise.
2526 (*if_shift_move): Likewise.
2527 (*ifcompare_move_shift): Likewise.
2528 (*if_move_shift): Likewise.
2529 (*ifcompare_shift_shift): Likewise.
2530 (*ifcompare_not_arith): Likewise.
2531 (*ifcompare_arith_not): Likewise.
2532 (*if_arith_not): Likewise.
2533 (*ifcompare_neg_move): Likewise.
2534 (*if_neg_move): Likewise.
2535 (*ifcompare_move_neg): Likewise.
2536 (*if_move_neg): Likewise.
2537 (prologue_thumb1_interwork): Likewise.
2538 (*cond_move_not): Likewise.
2539 (*sign_extract_onebit): Likewise.
2540 (*not_signextract_onebit): Likewise.
2541 (stack_tie): Likewise.
2542 (align_4): Likewise.
2543 (align_8): Likewise.
2544 (consttable_end): Likewise.
2545 (consttable_1): Likewise.
2546 (consttable_2): Likewise.
2547 (consttable_4): Likewise.
2548 (consttable_8): Likewise.
2549 (consttable_16): Likewise.
2550 (*thumb1_tablejump): Likewise.
2551 (prefetch): Likewise.
2552 (force_register_use): Likewise.
2553 (thumb_eh_return): Likewise.
2554 (load_tp_hard): Likewise.
2555 (load_tp_soft): Likewise.
2556 (tlscall): Likewise.
2557 (*arm_movtas_ze): Likewise.
2558 (*arm_rev): Likewise.
2559 (*arm_revsh): Likewise.
2560 (*arm_rev16): Likewise.
2561 * config/arm/thumb2.md
2562 (*thumb2_smaxsi3): Likewise.
2563 (*thumb2_sminsi3): Likewise.
2564 (*thumb32_umaxsi3): Likewise.
2565 (*thumb2_uminsi3): Likewise.
2566 (*thumb2_negdi2): Likewise.
2567 (*thumb2_abssi2): Likewise.
2568 (*thumb2_neg_abss): Likewise.
2569 (*thumb2_movsi_insn): Likewise.
2570 (tls_load_dot_plus_four): Likewise.
2571 (*thumb2_movhi_insn): Likewise.
2572 (*thumb2_mov_scc): Likewise.
2573 (*thumb2_mov_negs): Likewise.
2574 (*thumb2_mov_negs): Likewise.
2575 (*thumb2_mov_nots): Likewise.
2576 (*thumb2_mov_nots): Likewise.
2577 (*thumb2_movsicc_): Likewise.
2578 (*thumb2_movsfcc_soft_insn): Likewise.
2579 (*thumb2_indirect_jump): Likewise.
2580 (*thumb2_and_scc): Likewise.
2581 (*thumb2_ior_scc): Likewise.
2582 (*thumb2_ior_scc_strict_it): Likewise.
2583 (*thumb2_cond_move): Likewise.
2584 (*thumb2_cond_arith): Likewise.
2585 (*thumb2_cond_ari): Likewise.
2586 (*thumb2_cond_sub): Likewise.
2587 (*thumb2_negscc): Likewise.
2588 (*thumb2_movcond): Likewise.
2589 (thumb2_casesi_internal): Likewise.
2590 (thumb2_casesi_internal_pic): Likewise.
2591 (*thumb2_alusi3_short): Likewise.
2592 (*thumb2_mov<mode>_shortim): Likewise.
2593 (*thumb2_addsi_short): Likewise.
2594 (*thumb2_subsi_short): Likewise.
2595 (thumb2_addsi3_compare0): Likewise.
2596 (*thumb2_cbz): Likewise.
2597 (*thumb2_cbnz): Likewise.
2598 (*thumb2_one_cmplsi2_short): Likewise.
2599 (*thumb2_negsi2_short): Likewise.
2600 (*orsi_notsi_si): Likewise.
2601 * config/arm/arm1020e.md: Update with new attributes.
2602 * config/arm/arm1026ejs.md: Update with new attributes.
2603 * config/arm/arm1136jfs.md: Update with new attributes.
2604 * config/arm/arm926ejs.md: Update with new attributes.
2605 * config/arm/cortex-a15.md: Update with new attributes.
2606 * config/arm/cortex-a5.md: Update with new attributes.
2607 * config/arm/cortex-a53.md: Update with new attributes.
2608 * config/arm/cortex-a7.md: Update with new attributes.
2609 * config/arm/cortex-a8.md: Update with new attributes.
2610 * config/arm/cortex-a9.md: Update with new attributes.
2611 * config/arm/cortex-m4.md: Update with new attributes.
2612 * config/arm/cortex-r4.md: Update with new attributes.
2613 * config/arm/fa526.md: Update with new attributes.
2614 * config/arm/fa606te.md: Update with new attributes.
2615 * config/arm/fa626te.md: Update with new attributes.
2616 * config/arm/fa726te.md: Update with new attributes.
2618 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2620 * config/aarch64/aarch64-simd.md
2621 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2622 <vwx> iterator to ensure correct register choice.
2623 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2624 (aarch64_sqdmull_n<mode>): Likewise.
2625 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2626 * config/aarch64/arm_neon.h
2627 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2628 (vml<as><q>_n_<su>16): Likewise.
2629 (vml<as>l_high_lane<q>_<su>16): Likewise.
2630 (vml<as>l_high_n_<su>16): Likewise.
2631 (vml<as>l_lane<q>_<su>16): Likewise.
2632 (vml<as>l_n_<su>16): Likewise.
2633 (vmul<q>_lane<q>_<su>16): Likewise.
2634 (vmul<q>_n_<su>16): Likewise.
2635 (vmull_lane<q>_<su>16): Likewise.
2636 (vmull_n_<su>16): Likewise.
2637 (vmull_high_lane<q>_<su>16): Likewise.
2638 (vmull_high_n_<su>16): Likewise.
2639 (vqrdmulh<q>_n_s16): Likewise.
2641 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2643 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2644 have the correct lane parameter.
2646 2013-09-06 Richard Biener <rguenther@suse.de>
2648 * cfganal.c (control_dependences::~control_dependences):
2649 Properly free all of the vector.
2651 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2654 * config/i386/i386.c (ix86_conditional_register_usage):
2655 Proper initialize extended SSE registers.
2657 2013-09-06 Jan Hubicka <jh@suse.cz>
2659 PR tree-optimization/58311
2660 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2662 2013-09-06 Jan Hubicka <jh@suse.cz>
2664 * Makefile.in (tree-sra.o): Update dependencies.
2665 * tree-sra.c: Include ipa-utils.h
2666 (scan_function): Use recursive_call_p.
2667 (has_caller_p): New function.
2668 (cgraph_for_node_and_aliases): Count also callers of aliases.
2670 2013-09-06 Jan Hubicka <jh@suse.cz>
2673 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2674 * tree-tailcall.c: Include ipa-utils.h.
2675 (find_tail_calls): Use it.
2676 * ipa-pure-const.c (check_call): Likewise.
2677 * ipa-utils.c (recursive_call_p): New function.
2678 * ipa-utils.h (recursive_call_p): Dclare.
2679 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2680 (symtab_semantically_equivalent_p): New function.
2681 * Makefile.in (tree-tailcall.o): Update dependencies.
2683 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2685 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2688 2013-09-06 Richard Biener <rguenther@suse.de>
2690 * lto-streamer.h (lto_global_var_decls): Remove.
2691 * Makefile.in (OBJS): Remove lto-symtab.o.
2692 (lto-symtab.o): Remove.
2693 (GTFILES): Remove lto-symtab.c
2694 * lto-symtab.c: Move to lto/
2696 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2698 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2699 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2700 (UNSPEC_FPINT_RINT): New constant definitions.
2701 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2702 definition with 2 attributes.
2703 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2704 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2707 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2709 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2710 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2711 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2712 attribute to "z196_alone".
2713 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2716 2013-09-06 Richard Biener <rguenther@suse.de>
2718 * basic-block.h (class control_dependences): New.
2719 * tree-ssa-dce.c (control_dependence_map): Remove.
2721 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2722 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2723 find_pdom, find_control_dependence, find_all_control_dependences):
2725 (mark_control_dependent_edges_necessary,
2726 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2727 tree_dce_done, perform_tree_ssa_dce): Adjust.
2728 * cfganal.c (set_control_dependence_map_bit,
2729 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2730 find_all_control_dependences): Move from tree-ssa-dce.c and
2731 implement as methods of control_dependences class.
2732 (control_dependences::control_dependences): New.
2733 (control_dependences::~control_dependences): Likewise.
2734 (control_dependences::get_edges_dependent_on): Likewise.
2735 (control_dependences::get_edge): Likewise.
2737 2013-09-04 Jan Hubicka <jh@suse.cz>
2739 * tree.c (types_same_for_odr): Drop overactive check.
2740 * ipa-devirt.c (hash_type_name): Likewise.
2742 2013-09-04 Jan Hubicka <jh@suse.cz>
2744 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2745 (analyze_functions): ... here.
2747 2013-09-04 Jan Hubicka <jh@suse.cz>
2750 * cgraphunit.c (analyze_functions): Clear AUX fields
2751 after processing; initialize assembler name has.
2753 2013-09-05 Jeff Law <law@redhat.com>
2755 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2756 from thread_around_empty_block. Record threading path into PATH.
2757 Recurse if threading through the initial block is successful.
2758 (thread_across_edge): Corresponding changes to slightly simplify.
2760 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2762 * config/aarch64/aarch64.md
2763 (type): Remove frecpe, frecps, frecpx.
2764 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2765 fix to be a TARGET_SIMD instruction.
2766 (aarch64_frecps): Remove.
2767 * config/aarch64/aarch64-simd.md
2768 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2769 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2771 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2772 Sofiane Naci <sofiane.naci@arm.com>
2774 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2775 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2776 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2777 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2778 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2779 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2780 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2781 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2782 into "clz, "rbit". Rename "shift" to "shift_imm".
2783 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2784 changes. Update for attribute changes all occurrences of arlo_* and
2786 * config/arm/arm-fixed.md: Update for attribute changes
2787 all occurrences of arlo_* types.
2788 * config/arm/thumb2.md: Update for attribute changes all occurrences
2790 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2791 (cortexa7_older_only): Likewise.
2792 (cortexa7_younger): Likewise.
2793 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2794 (1020alu_shift_op): Likewise.
2795 (1020alu_shift_reg_op): Likewise.
2796 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2797 (alu_shift_op): Likewise.
2798 (alu_shift_reg_op): Likewise.
2799 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2800 (11_alu_shift_op): Likewise.
2801 (11_alu_shift_reg_op): Likewise.
2802 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2803 (9_alu_shift_reg_op): Likewise.
2804 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2806 (cortex_a15_alu_shift): Likewise.
2807 (cortex_a15_alu_shift_reg): Likewise.
2808 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2810 (cortex_a5_alu_shift): Likewise.
2811 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2813 (cortex_a53_alu_shift): Likewise.
2814 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2816 (cortex_a7_alu_reg): Likewise.
2817 (cortex_a7_alu_shift): Likewise.
2818 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2820 (cortex_a8_alu_shift): Likewise.
2821 (cortex_a8_alu_shift_reg): Likewise.
2822 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2823 (cortex_a9_dp_shift): Likewise.
2824 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2826 * config/arm/cortex-r4.md
2827 (cortex_r4_alu): Update for attribute changes.
2828 (cortex_r4_mov): Likewise.
2829 (cortex_r4_alu_shift_reg): Likewise.
2830 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2831 (526_alu_shift_op): Likewise.
2832 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2833 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2834 (626te_alu_shift_op): Likewise.
2835 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2836 (726te_alu_shift_op): Likewise.
2837 (726te_alu_shift_reg_op): Likewise.
2838 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2839 (mp626_alu_shift_op): Likewise.
2840 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2841 (pj4_alu_conds): Likewise.
2842 (pj4_shift): Likewise.
2843 (pj4_shift_conds): Likewise.
2844 (pj4_alu_shift): Likewise.
2845 (pj4_alu_shift_conds): Likewise.
2846 * config/aarch64/aarch64.md: Update for attribute change
2847 all occurrences of arlo_* and shift* types.
2849 2013-09-05 Mike Stump <mikestump@comcast.net>
2851 * tree.h: Move documentation for tree_function_decl to tree-core.h
2852 with the declaration.
2854 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2857 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2858 looking for widest mode.
2860 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2862 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2864 2013-09-05 Richard Biener <rguenther@suse.de>
2866 PR tree-optimization/58137
2867 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2868 Do not create vectors of pointers.
2869 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2870 types for the components of the vector initializer.
2871 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2872 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2874 2013-09-05 Martin Jambor <mjambor@suse.cz>
2876 * ipa-prop.c (remove_described_reference): Accept missing references,
2877 return false if that hppens, otherwise return true.
2878 (cgraph_node_for_jfunc): New function.
2879 (try_decrement_rdesc_refcount): Likewise.
2880 (try_make_edge_direct_simple_call): Use them.
2881 (ipa_edge_removal_hook): Remove references from rdescs.
2882 (ipa_edge_duplication_hook): Clone rdescs and their references
2883 when the new edge has the same caller as the old one.
2884 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2885 reference before removing any edges.
2887 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2889 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2891 * thumb2.md (thumb2_storewb_parisi): New pattern.
2893 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2895 * config/aarch64/aarch64-option-extensions.def: Add
2896 AARCH64_OPT_EXTENSION of 'crc'.
2897 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2898 (AARCH64_ISA_CRC): Ditto.
2899 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2900 description of the CRC extension.
2902 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2904 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2905 * config/rs6000/linux.h: Ditto.
2906 * alpha/linux.h: Ditto.
2907 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2908 no_c99_libc_has_function.
2909 * config/c6x/uclinux-elf.h: Ditto.
2910 * config/lm32/uclinux-elf.h: Ditto.
2911 * config/m68k/uclinux.h: Ditto.
2912 * config/moxie/uclinux.h: Ditto.
2913 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2914 (crisv32-*-linux*, cris-*-linux*): Ditto.
2915 * config/bfin/bfin.c: Include "tm_p.h".
2917 2013-09-05 Richard Biener <rguenther@suse.de>
2919 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2920 check for a definition without a basic-block.
2922 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2923 Sofiane Naci <sofiane.naci@arm.com>
2925 * config/aarch64/aarch64.md
2926 (*movti_aarch64): Rename r_2_f and f_2_r.
2927 (*movsf_aarch64): Likewise.
2928 (*movdf_aarch64): Likewise.
2929 (*movtf_aarch64): Likewise.
2930 (aarch64_movdi_<mode>low): Likewise.
2931 (aarch64_movdi_<mode>high): Likewise.
2932 (aarch64_mov<mode>high_di): Likewise.
2933 (aarch64_mov<mode>low_di): Likewise.
2934 (aarch64_movtilow_tilow): Likewise.
2935 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2936 values to config/arm/types.md
2937 (attribute "conds"): Update for attribute change.
2938 (anddi3_insn): Likewise.
2939 (iordi3_insn): Likewise.
2940 (xordi3_insn): Likewise.
2941 (one_cmpldi2): Likewise.
2942 * config/arm/types.md (type): Add Neon types.
2943 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2944 use "type" attribute.
2945 (movmisalign<mode>_neon_store): Likewise.
2946 (movmisalign<mode>_neon_load): Likewise.
2947 (vec_set<mode>_internal): Likewise.
2948 (vec_setv2di_internal): Likewise.
2949 (vec_extract<mode>): Likewise.
2950 (vec_extractv2di): Likewise.
2951 (add<mode>3_neon): Likewise.
2952 (adddi3_neon): Likewise.
2953 (sub<mode>3_neon): Likewise.
2954 (subdi3_neon): Likewise.
2955 (mul<mode>3_neon): Likewise.
2956 (mul<mode>3add<mode>_neon): Likewise.
2957 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
2958 (fma<VCVTF:mode>4)): Likewise.
2959 (fma<VCVTF:mode>4_intrinsic): Likewise.
2960 (fmsub<VCVTF:mode>4)): Likewise.
2961 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2962 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2963 (ior<mode>3): Likewise.
2964 (and<mode>3): Likewise.
2965 (anddi3_neon): Likewise.
2966 (orn<mode>3_neon): Likewise.
2967 (orndi3_neon): Likewise.
2968 (bic<mode>3_neon): Likewise.
2969 (bicdi3_neon): Likewise.
2970 (xor<mode>3): Likewise.
2971 (one_cmpl<mode>2): Likewise.
2972 (abs<mode>2): Likewise.
2973 (neg<mode>2): Likewise.
2974 (umin<mode>3_neon): Likewise.
2975 (umax<mode>3_neon): Likewise.
2976 (smin<mode>3_neon): Likewise.
2977 (smax<mode>3_neon): Likewise.
2978 (vashl<mode>3): Likewise.
2979 (vashr<mode>3_imm): Likewise.
2980 (vlshr<mode>3_imm): Likewise.
2981 (ashl<mode>3_signed): Likewise.
2982 (ashl<mode>3_unsigned): Likewise.
2983 (neon_load_count): Likewise.
2984 (ashldi3_neon_noclobber): Likewise.
2985 (signed_shift_di3_neon): Likewise.
2986 (unsigned_shift_di3_neon): Likewise.
2987 (ashrdi3_neon_imm_noclobber): Likewise.
2988 (lshrdi3_neon_imm_noclobber): Likewise.
2989 (widen_ssum<mode>3): Likewise.
2990 (widen_usum<mode>3): Likewise.
2991 (quad_halves_<code>v4si): Likewise.
2992 (quad_halves_<code>v4sf): Likewise.
2993 (quad_halves_<code>v8hi): Likewise.
2994 (quad_halves_<code>v16qi): Likewise.
2995 (reduc_splus_v2di): Likewise.
2996 (neon_vpadd_internal<mode>): Likewise.
2997 (neon_vpsmin<mode>): Likewise.
2998 (neon_vpsmax<mode>): Likewise.
2999 (neon_vpumin<mode>): Likewise.
3000 (neon_vpumax<mode>): Likewise.
3001 (ss_add<mode>_neon): Likewise.
3002 (us_add<mode>_neon): Likewise.
3003 (ss_sub<mode>_neon): Likewise.
3004 (us_sub<mode>_neon): Likewise.
3005 (neon_vadd<mode>_unspec): Likewise.
3006 (neon_vaddl<mode>): Likewise.
3007 (neon_vaddw<mode>): Likewise.
3008 (neon_vhadd<mode>): Likewise.
3009 (neon_vqadd<mode>): Likewise.
3010 (neon_vaddhn<mode>): Likewise.
3011 (neon_vmul<mode>): Likewise.
3012 (neon_vmla<mode>): Likewise.
3013 (neon_vmlal<mode>): Likewise.
3014 (neon_vmls<mode>): Likewise.
3015 (neon_vmlsl<mode>): Likewise.
3016 (neon_vqdmulh<mode>): Likewise.
3017 (neon_vqdmlal<mode>): Likewise.
3018 (neon_vqdmlsl<mode>): Likewise.
3019 (neon_vmull<mode>): Likewise.
3020 (neon_vqdmull<mode>): Likewise.
3021 (neon_vsub<mode>_unspec): Likewise.
3022 (neon_vsubl<mode>): Likewise.
3023 (neon_vsubw<mode>): Likewise.
3024 (neon_vqsub<mode>): Likewise.
3025 (neon_vhsub<mode>): Likewise.
3026 (neon_vsubhn<mode>): Likewise.
3027 (neon_vceq<mode>): Likewise.
3028 (neon_vcge<mode>): Likewise.
3029 (neon_vcgeu<mode>): Likewise.
3030 (neon_vcgt<mode>): Likewise.
3031 (neon_vcgtu<mode>): Likewise.
3032 (neon_vcle<mode>): Likewise.
3033 (neon_vclt<mode>): Likewise.
3034 (neon_vcage<mode>): Likewise.
3035 (neon_vcagt<mode>): Likewise.
3036 (neon_vtst<mode>): Likewise.
3037 (neon_vabd<mode>): Likewise.
3038 (neon_vabdl<mode>): Likewise.
3039 (neon_vaba<mode>): Likewise.
3040 (neon_vabal<mode>): Likewise.
3041 (neon_vmax<mode>): Likewise.
3042 (neon_vmin<mode>): Likewise.
3043 (neon_vpaddl<mode>): Likewise.
3044 (neon_vpadal<mode>): Likewise.
3045 (neon_vpmax<mode>): Likewise.
3046 (neon_vpmin<mode>): Likewise.
3047 (neon_vrecps<mode>): Likewise.
3048 (neon_vrsqrts<mode>): Likewise.
3049 (neon_vqabs<mode>): Likewise.
3050 (neon_vqneg<mode>): Likewise.
3051 (neon_vcls<mode>): Likewise.
3052 (clz<mode>2): Likewise.
3053 (popcount<mode>2): Likewise.
3054 (neon_vrecpe): Likewise.
3055 (neon_vrsqrte): Likewise.
3056 (neon_vget_lane<mode>_sext_internal): Likewise.
3057 (neon_vget_lane<mode>_zext_internal): Likewise.
3058 (neon_vdup_n<mode>): Likewise.
3059 (neon_vdup_nv2di): Likewise.
3060 (neon_vdpu_lane<mode>_internal): Likewise.
3061 (neon_vswp<mode>): Likewise.
3062 (float<mode><V_cvtto>2): Likewise.
3063 (floatuns<mode><V_cvtto>2): Likewise.
3064 (fix_trunc<mode><V_cvtto>)2): Likewise
3065 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3066 (neon_vcvt<mode>): Likewise.
3067 (neon_vcvtv4sfv4hf): Likewise.
3068 (neon_vcvtv4hfv4sf): Likewise.
3069 (neon_vcvt_n<mode>): Likewise.
3070 (neon_vmovn<mode>): Likewise.
3071 (neon_vqmovn<mode>): Likewise.
3072 (neon_vqmovun<mode>): Likewise.
3073 (neon_vmovl<mode>): Likewise.
3074 (neon_vmul_lane<mode>): Likewise.
3075 (neon_vmull_lane<mode>): Likewise.
3076 (neon_vqdmull_lane<mode>): Likewise.
3077 (neon_vqdmulh_lane<mode>): Likewise.
3078 (neon_vmla_lane<mode>): Likewise.
3079 (neon_vmlal_lane<mode>): Likewise.
3080 (neon_vqdmlal_lane<mode>): Likewise.
3081 (neon_vmls_lane<mode>): Likewise.
3082 (neon_vmlsl_lane<mode>): Likewise.
3083 (neon_vqdmlsl_lane<mode>): Likewise.
3084 (neon_vext<mode>): Likewise.
3085 (neon_vrev64<mode>): Likewise.
3086 (neon_vrev32<mode>): Likewise.
3087 (neon_vrev16<mode>): Likewise.
3088 (neon_vbsl<mode>_internal): Likewise.
3089 (neon_vshl<mode>): Likewise.
3090 (neon_vqshl<mode>): Likewise.
3091 (neon_vshr_n<mode>): Likewise.
3092 (neon_vshrn_n<mode>): Likewise.
3093 (neon_vqshrn_n<mode>): Likewise.
3094 (neon_vqshrun_n<mode>): Likewise.
3095 (neon_vshl_n<mode>): Likewise.
3096 (neon_vqshl_n<mode>): Likewise.
3097 (neon_vqshlu_n<mode>): Likewise.
3098 (neon_vshll_n<mode>): Likewise.
3099 (neon_vsra_n<mode>): Likewise.
3100 (neon_vsri_n<mode>): Likewise.
3101 (neon_vsli_n<mode>): Likewise.
3102 (neon_vtbl1v8qi): Likewise.
3103 (neon_vtbl2v8qi): Likewise.
3104 (neon_vtbl3v8qi): Likewise.
3105 (neon_vtbl4v8qi): Likewise.
3106 (neon_vtbx1v8qi): Likewise.
3107 (neon_vtbx2v8qi): Likewise.
3108 (neon_vtbx3v8qi): Likewise.
3109 (neon_vtbx4v8qi): Likewise.
3110 (neon_vtrn<mode>_internal): Likewise.
3111 (neon_vzip<mode>_internal): Likewise.
3112 (neon_vuzp<mode>_internal): Likewise.
3113 (neon_vld1<mode>): Likewise.
3114 (neon_vld1_lane<mode>): Likewise.
3115 (neon_vld1_dup<mode>): Likewise.
3116 (neon_vld1_dupv2di): Likewise.
3117 (neon_vst1<mode>): Likewise.
3118 (neon_vst1_lane<mode>): Likewise.
3119 (neon_vld2<mode>): Likewise.
3120 (neon_vld2_lane<mode>): Likewise.
3121 (neon_vld2_dup<mode>): Likewise.
3122 (neon_vst2<mode>): Likewise.
3123 (neon_vst2_lane<mode>): Likewise.
3124 (neon_vld3<mode>): Likewise.
3125 (neon_vld3qa<mode>): Likewise.
3126 (neon_vld3qb<mode>): Likewise.
3127 (neon_vld3_lane<mode>): Likewise.
3128 (neon_vld3_dup<mode>): Likewise.
3129 (neon_vst3<mode>): Likewise.
3130 (neon_vst3qa<mode>): Likewise.
3131 (neon_vst3qb<mode>): Likewise.
3132 (neon_vst3_lane<mode>): Likewise.
3133 (neon_vld4<mode>): Likewise.
3134 (neon_vld4qa<mode>): Likewise.
3135 (neon_vld4qb<mode>): Likewise.
3136 (neon_vld4_lane<mode>): Likewise.
3137 (neon_vld4_dup<mode>): Likewise.
3138 (neon_vst4<mode>): Likewise.
3139 (neon_vst4qa<mode>): Likewise.
3140 (neon_vst4qb<mode>): Likewise.
3141 (neon_vst4_lane<mode>): Likewise.
3142 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3143 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3144 (neon_vec_<US>mult_lo_<mode>): Likewise.
3145 (neon_vec_<US>mult_hi_<mode>): Likewise.
3146 (neon_vec_<US>shiftl_<mode>): Likewise.
3147 (neon_unpack<US>_<mode>): Likewise.
3148 (neon_vec_<US>mult_<mode>): Likewise.
3149 (vec_pack_trunc_<mode>): Likewise.
3150 (neon_vec_pack_trunk_<mode>): Likewise.
3151 (neon_vabd<mode>_2): Likewise.
3152 (neon_vabd<mode>_3): Likewise.
3153 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3154 (thumb2_movsi_vfp): Likewise.
3155 (movdi_vfp): Likewise.
3156 (movdi_vfp_cortexa8): Likewise.
3157 (movhf_vfp_neon): Likewise.
3158 (movhf_vfp): Likewiwse.
3159 (movsf_vfp): Likewiwse.
3160 (thumb2_movsf_vfp): Likewiwse.
3161 (movdf_vfp): Likewise.
3162 (thumb2_movdf_vfp): Likewise.
3163 (movsfcc_vfp): Likewise.
3164 (thumb2_movsfcc_vfp): Likewise.
3165 (movdfcc_vfp): Likewise.
3166 (thumb2_movdfcc_vfp): Likewise.
3167 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3168 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3169 (v10_v2c): Likewise.
3170 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3172 (cortex_a15_neon_int_2): Likewise.
3173 (cortex_a15_neon_int_3): Likewise.
3174 (cortex_a15_neon_int_4): Likewise.
3175 (cortex_a15_neon_int_5): Likewise.
3176 (cortex_a15_neon_vqneg_vqabs): Likewise.
3177 (cortex_a15_neon_vmov): Likewise.
3178 (cortex_a15_neon_vaba): Likewise.
3179 (cortex_a15_neon_vaba_qqq): Likewise.
3180 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3181 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3182 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3183 scalar_64_32_long_scalar): Likewise.
3184 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3185 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3186 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3187 lotype_qdd_64_32_long): Likewise.
3188 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3189 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3190 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3191 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3192 (cortex_a15_neon_shift_1): Likewise.
3193 (cortex_a15_neon_shift_2): Likewise.
3194 (cortex_a15_neon_shift_3): Likewise.
3195 (cortex_a15_neon_vshl_ddd): Likewise.
3196 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3197 (cortex_a15_neon_vsra_vrsra): Likewise.
3198 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3199 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3200 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3201 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3202 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3203 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3204 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3205 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3206 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3207 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3208 (cortex_a15_neon_bp_simple): Likewise.
3209 (cortex_a15_neon_bp_2cycle): Likewise.
3210 (cortex_a15_neon_bp_3cycle): Likewise.
3211 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3212 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3213 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3214 (cortex_a15_neon_vld2_4_regs): Likewise.
3215 (cortex_a15_neon_vld3_vld4): Likewise.
3216 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3217 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3218 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3219 (cortex_a15_neon_vst3_vst4): Likewise.
3220 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3221 (cortex_a15_neon_vld3_vld4_lane" 10
3222 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3223 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3224 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3225 (cortex_a15_neon_ldm_2): Likewise.0
3226 (cortex_a15_neon_stm_2): Likewise.
3227 (cortex_a15_neon_mcr): Likewise.
3228 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3229 (cortex_a15_neon_mrc): Likewise.
3230 (cortex_a15_neon_mrrc): Likewise.
3231 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3233 (cortex_a15_alu_shift): Likewise.
3234 (cortex_a15_alu_shift_reg): Likewise.
3235 (cortex_a15_mult32): Likewise.
3236 (cortex_a15_mult64): Likewise.
3237 (cortex_a15_block): Likewise.
3238 (cortex_a15_branch): Likewise.
3239 (cortex_a15_load1): Likewise.
3240 (cortex_a15_load3): Likewise.
3241 (cortex_a15_store1): Likewise.
3242 (cortex_a15_store3): Likewise.
3243 (cortex_a15_call): Likewise.
3244 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3245 (cortex_a5_f2r): Likewise.
3246 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3248 (cortex_a53_f2r): Likewise.
3249 * config/arm/cortex-a7.md
3250 (cortex_a7_branch): Update for attribute change.
3251 (cortex_a7_call): Likewise.
3252 (cortex_a7_alu_imm): Likewise.
3253 (cortex_a7_alu_reg): Likewise.
3254 (cortex_a7_alu_shift): Likewise.
3255 (cortex_a7_mul): Likewise.
3256 (cortex_a7_load1): Likewise.
3257 (cortex_a7_store1): Likewise.
3258 (cortex_a7_load2): Likewise.
3259 (cortex_a7_store2): Likewise.
3260 (cortex_a7_load3): Likewise.
3261 (cortex_a7_store3): Likewise.
3262 (cortex_a7_load4): Likewise.
3263 (cortex_a7_store4): Likewise.
3264 (cortex_a7_fpalu): Likewise.
3265 (cortex_a7_fconst): Likewise.
3266 (cortex_a7_fpmuls): Likewise.
3267 (cortex_a7_neon_mul): Likewise.
3268 (cortex_a7_fpmacs): Likewise.
3269 (cortex_a7_neon_mla: Likewise.
3270 (cortex_a7_fpmuld: Likewise.
3271 (cortex_a7_fpmacd: Likewise.
3272 (cortex_a7_fpfmad: Likewise.
3273 (cortex_a7_fdivs: Likewise.
3274 (cortex_a7_fdivd: Likewise.
3275 (cortex_a7_r2f: Likewise.
3276 (cortex_a7_f2r: Likewise.
3277 (cortex_a7_f_flags: Likewise.
3278 (cortex_a7_f_loads: Likewise.
3279 (cortex_a7_f_loadd: Likewise.
3280 (cortex_a7_f_stores: Likewise.
3281 (cortex_a7_f_stored: Likewise.
3282 (cortex_a7_neon): Likewise.
3283 * config/arm/cortex-a8-neon.md
3284 (cortex_a8_neon_mrc): Update for attribute change.
3285 (cortex_a8_neon_mrrc): Likewise.
3286 (cortex_a8_neon_int_1): Likewise.
3287 (cortex_a8_neon_int_2): Likewise.
3288 (cortex_a8_neon_int_3): Likewise.
3289 (cortex_a8_neon_int_4): Likewise.
3290 (cortex_a8_neon_int_5): Likewise.
3291 (cortex_a8_neon_vqneg_vqabs): Likewise.
3292 (cortex_a8_neon_vmov): Likewise.
3293 (cortex_a8_neon_vaba): Likewise.
3294 (cortex_a8_neon_vaba_qqq): Likewise.
3295 (cortex_a8_neon_vsma): Likewise.
3296 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3297 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3298 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3299 long_scalar): Likewise.
3300 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3301 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3302 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3303 long_scalar_qdd_64_32_long): Likewise.
3304 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3305 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3306 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3307 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3308 (cortex_a8_neon_shift_1): Likewise.
3309 (cortex_a8_neon_shift_2): Likewise.
3310 (cortex_a8_neon_shift_3): Likewise.
3311 (cortex_a8_neon_vshl_ddd): Likewise.
3312 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3313 (cortex_a8_neon_vsra_vrsra): Likewise.
3314 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3315 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3316 (cortex_a8_neon_fp_vsum): Likewise.
3317 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3318 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3319 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3320 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3321 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3322 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3323 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3324 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3325 (cortex_a8_neon_bp_simple): Likewise.
3326 (cortex_a8_neon_bp_2cycle): Likewise.
3327 (cortex_a8_neon_bp_3cycle): Likewise.
3328 (cortex_a8_neon_ldr): Likewise.
3329 (cortex_a8_neon_str): Likewise.
3330 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3331 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3332 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3333 (cortex_a8_neon_vld2_4_regs): Likewise.
3334 (cortex_a8_neon_vld3_vld4): Likewise.
3335 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3336 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3337 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3338 (cortex_a8_neon_vst3_vst4): Likewise.
3339 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3340 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3341 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3342 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3343 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3344 (cortex_a8_neon_mcr): Likewise.
3345 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3346 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3347 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3349 (ca9_neon_mrrc): Likewise.
3350 (cortex_a9_neon_int_1): Likewise.
3351 (cortex_a9_neon_int_2): Likewise.
3352 (cortex_a9_neon_int_3): Likewise.
3353 (cortex_a9_neon_int_4): Likewise.
3354 (cortex_a9_neon_int_5): Likewise.
3355 (cortex_a9_neon_vqneg_vqabs): Likewise.
3356 (cortex_a9_neon_vmov): Likewise.
3357 (cortex_a9_neon_vaba): Likewise.
3358 (cortex_a9_neon_vaba_qqq): Likewise.
3359 (cortex_a9_neon_vsma): Likewise.
3360 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3361 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3362 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3363 long_scalar): Likewise.
3364 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3365 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3366 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3367 long_scalar_qdd_64_32_long): Likewise.
3368 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3369 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3370 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3371 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3372 (cortex_a9_neon_shift_1): Likewise.
3373 (cortex_a9_neon_shift_2): Likewise.
3374 (cortex_a9_neon_shift_3): Likewise.
3375 (cortex_a9_neon_vshl_ddd): Likewise.
3376 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3377 (cortex_a9_neon_vsra_vrsra): Likewise.
3378 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3379 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3380 (cortex_a9_neon_fp_vsum): Likewise.
3381 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3382 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3383 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3384 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3385 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3386 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3387 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3388 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3389 (cortex_a9_neon_bp_simple): Likewise.
3390 (cortex_a9_neon_bp_2cycle): Likewise.
3391 (cortex_a9_neon_bp_3cycle): Likewise.
3392 (cortex_a9_neon_ldr): Likewise.
3393 (cortex_a9_neon_str): Likewise.
3394 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3395 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3396 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3397 (cortex_a9_neon_vld2_4_regs): Likewise.
3398 (cortex_a9_neon_vld3_vld4): Likewise.
3399 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3400 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3401 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3402 (cortex_a9_neon_vst3_vst4): Likewise.
3403 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3404 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3405 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3406 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3407 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3408 (cortex_a9_neon_mcr): Likewise.
3409 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3410 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3411 (cortex_a9_fps): Likewise.
3412 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3414 (cortex_m4_fmuls): Likewise.
3415 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3417 (cortex_r4_mrc): Likewise.
3418 * config/arm/iterators.md: Update comment referring to neon_type.
3419 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3420 (iwmmxt_movsi_insn): Likewise.
3421 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3423 (pj4_core_to_vfp): Likewise.
3424 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3426 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3427 (vfp_fstore): Likewise.
3428 * doc/md.texi: Change references to neon_type to refer to type.
3430 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3432 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3434 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3436 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3439 2013-09-04 Easwaran Raman <eraman@google.com>
3442 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3443 (build_and_add_sum): Use it.
3444 (appears_later_in_bb): Simplify code.
3446 2013-09-04 Teresa Johnson <tejohnson@google.com>
3448 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3450 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3452 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3454 2013-09-04 Jan Hubicka <jh@suse.cz>
3456 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3457 * ipa-devirt.c: Include diganostic.h
3458 (odr_type_d): Add types and types_set.
3459 (hash_type_name): Work for types with vtables during LTO.
3460 (odr_hasher::remove): Fix comment; destroy types_set.
3461 (add_type_duplicate): New function,
3462 (get_odr_type): Use it.
3463 (dump_type_inheritance_graph): Dump type duplicates.
3464 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3466 * tree.c (types_same_for_odr): Give exact answers on types with
3469 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3471 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3472 explaining their differences.
3474 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3476 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3478 2013-09-03 Jeff Law <law@redhat.com>
3480 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3481 when not threading through a joiner block. Pass joiner/no joiner
3482 state to register_jump_thread.
3483 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3484 state from argument rather than implying on path length.
3485 Dump the entire jump thread path into debugging dump.
3486 * tree-flow.h (register_jump_thread): Update prototype.
3488 2013-08-29 Xinliang David Li <davidxl@google.com>
3490 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3491 Remove a trivial gcc_assert.
3493 2013-08-29 Xinliang David Li <davidxl@google.com>
3495 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3496 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3497 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3498 Delay base decl alignment adjustment.
3499 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3500 * tree-vectorizer.h: New data structure.
3501 (set_dr_misalignment): New function.
3502 (dr_misalignment): Ditto.
3503 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3504 (vectorizable_load): Ditto.
3505 (ensure_base_align): New function.
3506 (vectorize_loops): Add dbg_cnt support.
3507 (execute_vect_slp): Ditto.
3508 * dbgcnt.def: New debug counter.
3509 * Makefile: New dependency.
3511 2013-09-03 Meador Inge <meadori@codesourcery.com>
3515 2013-08-30 Meador Inge <meadori@codesourcery.com>
3517 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3519 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3521 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3522 function descriptor.
3524 2013-09-03 Richard Biener <rguenther@suse.de>
3526 * tree-affine.c (add_elt_to_tree): Fix association issue,
3527 avoid useless converts and make sure to always return a
3528 properly typed result.
3530 2013-09-03 Richard Biener <rguenther@suse.de>
3533 * fold-const.c (negate_expr_p): Fix division case.
3534 (negate_expr): Likewise.
3536 2013-09-03 Richard Biener <rguenther@suse.de>
3539 * tree-streamer-out.c: Include tm.h.
3540 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3542 2013-09-03 Jan Hubicka <jh@suse.cz>
3544 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3546 2013-09-03 Alan Modra <amodra@gmail.com>
3548 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3549 multilibs to big-endian target and vice versa.
3550 * config/rs6000/t-linux64: Use := assignment on all vars.
3551 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3552 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3553 * config/rs6000/t-linux64le: New file.
3554 * config/rs6000/t-linux64bele: New file.
3555 * config/rs6000/t-linux64lebe: New file.
3557 2013-09-02 Jan Hubicka <jh@suse.cz>
3559 * ipa-inline-transform.c (inline_transform): Do not
3560 optimize_inline_calls when not optimizing.
3562 2013-09-02 Jan Hubicka <jh@suse.cz>
3564 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3565 duplicated nodes for assembler names.
3566 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3567 hard registers from assembler name hash.
3569 2013-09-02 Jan Hubicka <jh@suse.cz>
3571 * ipa-split.c (execute_split_functions): Split externally visible
3572 functions called once.
3574 2013-09-02 Martin Jambor <mjambor@suse.cz>
3577 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3578 linked list. When finding the correct duplicate, also consider also
3579 the caller in additon to its inlined_to node.
3581 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3583 * config/aarch64/aarch64-simd-builtins.def
3584 (dup_lane_scalar): Remove.
3585 * config/aarch64/aarch64-simd.md
3586 (aarch64_simd_dup): Add 'w->w' alternative.
3587 (aarch64_dup_lane<mode>): Allow for VALL.
3588 (aarch64_dup_lane_scalar<mode>): Remove.
3589 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3590 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3591 (aarch64_get_lane_unsigned<mode>): Likewise.
3592 (aarch64_get_lane<mode>): Likewise.
3593 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3594 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3595 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3596 (VCON): Change container of V2SF.
3597 (vswap_width_name): Likewise.
3598 * config/aarch64/arm_neon.h
3599 (__aarch64_vdup_lane_any): New.
3600 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3601 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3602 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3604 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3607 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3608 the source or the destination is a single hard register.
3610 2013-09-02 Richard Biener <rguenther@suse.de>
3613 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3616 2013-09-02 Richard Biener <rguenther@suse.de>
3618 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3619 arithmetic to sizetype.
3621 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3623 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3624 Find auto-increment use both before and after candidate.
3626 2013-09-02 Marek Polacek <polacek@redhat.com>
3628 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3630 2013-09-01 Jan Hubicka <jh@suse.cz>
3632 * Makefile.in: Add ipa-profile.o
3633 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3634 * cgraph.c (struct cgraph_propagate_frequency_data,
3635 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3636 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3637 * cgraph.h (cgraph_propagate_frequency): Remove.
3638 * ipa-inline-analysis.c: Include ipa-utils.h;
3639 drop duplicated cfgloop.h.
3640 (inline_update_callee_summaries): Update.
3641 * ipa-profile.c: New file.
3642 * ipa-utils.h (ipa_propagate_frequency): Declare.
3643 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3644 data-streamer.h, value-prof.h.
3645 (symtab_remove_unreachable_nodes): Update profile.
3646 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3647 account_time_size, cmp_counts, dump_histogram,
3648 ipa_profile_generate_summary, ipa_profile_write_summary,
3649 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3650 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3651 Move to ipa-profile.c.
3653 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3655 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3657 2013-09-01 Jan Hubicka <jh@suse.cz>
3659 * common.opt (fdevirtualize-speculatively): New function.
3660 * invoke.texi (fdevirtualize-speculatively): Document.
3661 * ipa-devirt.c: Include ipa-inline.h
3662 (likely_target_p): New function.
3663 (ipa_devirt): New function.
3664 (gate_ipa_devirt): New function.
3665 (pass_data_ipa_devirt): New static var.
3666 (pass_ipa_devirt): Likewise.
3667 (make_pass_ipa_devirt): New function.
3668 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3669 (common_handle_option): Disable devirtualization when
3670 value range profiling is available.
3671 * passes.def (pass_ipa_devirt): Add.
3672 * timever.def (TV_IPA_DEVIRT): New timevar.
3673 * tree-pass.h (make_pass_ipa_devirt):
3675 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3677 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3678 include sanitize(undefined).
3680 2013-08-31 Diego Novillo <dnovillo@google.com>
3682 * Makefile.in (TREE_CORE_H): Define.
3684 (GTFILES): Add tree-core.h.
3685 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3687 * tree-core.h: New file.
3688 Move all data structures, enum, typedefs, global
3689 declarations and constants from ...
3692 2013-08-31 Jan Hubicka <jh@suse.cz>
3694 * bulitins.c (expand_builtin): Do not early exit for gcov
3695 instrumented functions.
3697 2013-08-31 Marek Polacek <polacek@redhat.com>
3699 * ubsan.c: Include tm_p.h.
3701 2013-08-31 Jan Hubicka <jh@suse.cz>
3703 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3706 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3707 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3708 * tree-cfg.c (verify_gimple_label): ... here.
3709 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3710 (ipa_merge_profiles): New function.
3711 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3712 (lto_input_function_body): Likewise.
3713 * ipa-utils.h (ipa_merge_profiles): Declare.
3714 * lto-streamer.h (lto_input_function_body): Update prototype.
3715 (emit_label_in_global_context_p): Remove.
3716 * lto-symtab.c: Include ipa-utils.h
3717 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3719 2013-08-31 Jan Hubicka <jh@suse.cz>
3721 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3723 2013-08-31 Jan Hubicka <jh@suse.cz>
3725 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3727 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3729 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3730 "cmp" RTX before signed_comparison_operator check to account
3733 2013-08-30 Jan Hubicka <jh@suse.cz>
3735 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3736 (detect_type_change_1): Rename to ...
3737 (detect_type_change): ... this one; early return on non-polymorphic
3739 (detect_type_change_ssa): Add comp_type parameter; update
3740 use of detect_type_change.
3741 (compute_complex_assign_jump_func): Add param_type parameter;
3742 update use of detect_type_change_ssa.
3743 (compute_complex_ancestor_jump_func): Likewise.
3744 (ipa_get_callee_param_type): New function.
3745 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3746 update calls to the jump function computation functions.
3748 2013-08-30 Teresa Johnson <tejohnson@google.com>
3749 Steven Bosscher <steven@gcc.gnu.org>
3751 * cfgrtl.c (fixup_new_cold_bb): New routine.
3752 (commit_edge_insertions): Invoke fixup_partitions.
3753 (find_partition_fixes): New routine.
3754 (fixup_partitions): Ditto.
3755 (verify_hot_cold_block_grouping): Update comments.
3756 (rtl_verify_edges): Invoke find_partition_fixes.
3757 (rtl_verify_bb_pointers): Update comments.
3758 (rtl_verify_bb_layout): Ditto.
3759 * basic-block.h (probably_never_executed_edge_p): Declare.
3760 (fixup_partitions): Ditto.
3761 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3762 * bb-reorder.c (sanitize_hot_paths): New function.
3763 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3765 * predict.c (probably_never_executed_edge_p): New routine.
3766 * cfg.c (check_bb_profile): Add partition insanity warnings.
3768 2013-08-30 Meador Inge <meadori@codesourcery.com>
3770 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3772 2013-08-30 Marek Polacek <polacek@redhat.com>
3774 * Makefile.in (ubsan.o): Add.
3775 (c-family/c-ubsan.o): Add.
3776 (builtins.o): Add ubsan.h dependency.
3777 * ubsan.h: New file.
3778 * ubsan.c: New file.
3779 * common.opt: Add -fsanitize=undefined option.
3780 (flag_sanitize): Add variable.
3781 (fsanitize=): Add option. Add Driver.
3782 (fsanitize=thread): Remove option.
3783 (fsanitize=address): Likewise.
3784 (static-libubsan): New option.
3785 * doc/invoke.texi: Document the new flag and -static-libubsan.
3786 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3787 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3788 * builtin-attrs.def (ATTR_COLD): Define.
3789 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3790 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3791 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3792 * flag-types.h (sanitize_code): New enum.
3793 * opts.c (common_handle_option): Parse command line arguments
3794 of -fsanitize=. Add -fsanitize=unreachable option.
3795 * varasm.c (get_variable_section): Adjust.
3796 (assemble_noswitch_variable): Likewise.
3797 (assemble_variable): Likewise.
3798 (output_constant_def_contents): Likewise.
3799 (categorize_decl_for_section): Likewise.
3800 (place_block_symbol): Likewise.
3801 (output_object_block): Likewise.
3802 * builtins.def: Likewise.
3803 * toplev.c (compile_file): Likewise.
3804 (process_options): Likewise.
3805 * cppbuiltin.c: Likewise.
3806 * tsan.c (tsan_pass): Likewise.
3807 (tsan_gate): Likewise.
3808 (tsan_gate_O0): Likewise.
3809 * cfgexpand.c (partition_stack_vars): Likewise.
3810 (expand_stack_vars): Likewise.
3811 (defer_stack_allocation): Likewise.
3812 (expand_used_vars): Likewise.
3813 * cfgcleanup.c (old_insns_match_p): Likewise.
3814 * asan.c (asan_finish_file): Likewise.
3815 (asan_instrument): Likewise.
3816 (gate_asan): Likewise.
3817 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3818 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3819 (asan_global_struct): Use pointer_sized_int_node instead
3820 calling build_nonstandard_integer_type.
3821 (initialize_sanitizer_builtins): Likewise.
3822 (asan_finish_file): Likewise.
3823 * gcc.c: Document %{%:function(args):X}.
3824 (static_spec_functions): Add sanitize.
3825 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3826 store funcval != NULL there.
3827 (do_spec_1): Adjust handle_spec_function caller.
3828 (handle_braces): Allow %:function(args) as condition.
3829 (sanitize_spec_function): New function.
3830 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3831 (LIBUBSAN_SPEC): Likewise.
3832 (LIBUBSAN_EARLY_SPEC): Likewise.
3833 (SANITIZER_SPEC): Handle libubsan.
3834 (SANITIZER_EARLY_SPEC): Likewise.
3835 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3836 instead of fsanitize=address.
3837 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3838 instead of fsanitize=address*.
3839 * builtins.c: Include ubsan.h.
3840 (fold_builtin_0): Instrument __builtin_unreachable.
3841 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3842 instead of flag_asan.
3843 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3844 (pointer_sized_int_node): Define.
3845 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3847 2013-08-30 Mike Stump <mikestump@comcast.net>
3849 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3852 2013-08-29 Jan Hubicka <jh@suse.cz>
3854 * cgraph.c (cgraph_function_body_availability): Handle weakref
3856 * passes.def: Remove pass_fixup_cfg.
3857 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3858 track when we need to remove functions.
3859 (gate_ipa_inline): Execute inlining always; add comment why.
3860 (pass_data_ipa_inline): Remove TODO_remove_functions.
3861 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3862 do not produce summaries.
3863 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3864 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3865 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3866 constant pool and vtable.
3868 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3870 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3871 New arm_neon.h's internal macros to specify 64-bit constants.
3872 Avoid using stdint.h's macros.
3874 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3876 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3877 before applying REG_P.
3879 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3881 PR tree-optimization/58277
3882 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3883 after seeing too many stmts with vdef in between dombb and current
3884 bb, invalidate everything.
3886 2013-08-30 Richard Biener <rguenther@suse.de>
3888 * fold-const.c (fold_single_bit_test): Fix overflow test.
3890 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3892 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3893 and which can live in a register, always retrieve the value on entry.
3894 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3895 passed by invisible reference specially.
3896 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3897 (vt_add_function_parameter): Correctly deal with a parameter passed by
3898 invisible reference.
3900 2013-08-30 Jan Hubicka <jh@suse.cz>
3902 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3904 2013-08-30 Richard Biener <rguenther@suse.de>
3906 PR tree-optimization/58228
3907 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3908 allow invariant loads in nested loop vectorization.
3910 2013-08-30 Richard Biener <rguenther@suse.de>
3912 PR tree-optimization/58223
3913 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3914 (has_anti_or_output_dependence): ... this and adjust to also
3915 look for output dependences.
3916 (mark_nodes_having_upstream_mem_writes): Adjust.
3917 (rdg_flag_uses): Likewise.
3919 2013-08-30 Richard Biener <rguenther@suse.de>
3921 PR tree-optimization/58010
3922 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3923 assert that we have a loop-closed PHI.
3925 2013-08-29 Jan Hubicka <jh@suse.cz>
3927 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3928 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3929 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3930 (lto_free_function_in_decl_state_for_node): New function.
3932 2013-08-29 Xinliang David Li <davidxl@google.com>
3934 * loop-unroll.c (report_unroll_peel): Minor message change.
3935 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3936 Emit alignment peeling message with default -fopt-info.
3937 (vect_loop_versioning): Emit loop version info message.
3938 * tree-vectorizer.c (vectorize_loops): Minor message change.
3939 (execute_vect_slp): Ditto.
3941 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3943 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3944 of the clone from the DECL_NAME of the original function.
3946 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3948 * passes.c (register_pass): Add overload.
3949 * tree-pass.h (register_pass): Forward declare it. Add comment.
3951 2013-08-29 Jan Hubicka <jh@suse.cz>
3953 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
3954 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
3955 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
3956 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3957 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
3958 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
3959 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3960 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
3962 2013-08-29 Teresa Johnson <tejohnson@google.com>
3964 * dumpfile.c (dump_loc): Output column number.
3965 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
3966 * doc/invoke.texi: Document optall -fopt-info flag.
3967 * profile.c (read_profile_edge_counts): Use new dump framework.
3968 (compute_branch_probabilities): Ditto.
3969 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
3970 when pass not in any opt group.
3971 * pass_manager.h (pass_manager::get_pass_profile): New method.
3972 * value-prof.c (check_counter): Use new dump framework.
3973 (check_ic_target): Ditto.
3974 * coverage.c (get_coverage_counts): Ditto.
3975 (coverage_init): Setup new dump framework.
3977 2013-08-29 Richard Biener <rguenther@suse.de>
3979 PR tree-optimization/58246
3980 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
3981 handle the dominance check inside a basic-block.
3983 2013-08-29 Richard Biener <rguenther@suse.de>
3986 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
3987 of default defs that appear in abnormal PHI nodes.
3989 2013-08-29 Richard Biener <rguenther@suse.de>
3991 PR tree-optimization/57685
3992 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
3993 single-use operands to avoid exponential complexity.
3995 2013-08-28 Dehao Chen <dehao@google.com>
3997 * ipa-inline.c (edge_badness): Fix integer underflow.
3999 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4001 * gtm-builtins.def (_ITM_free): Declare leaf.
4003 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4006 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4007 (*tls_local_dynamic_base_64_largepic): Likewise.
4008 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4009 Remove predicate from call operand.
4010 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4011 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4013 2013-08-28 Jeff Law <law@redhat.com>
4015 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4016 checks for the number of predecessors and successors allowed.
4017 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4018 which require copying a joiner block if there is a request which
4019 is a subpath that requires no joiner block copying.
4021 2013-08-28 Jan Hubicka <jh@suse.cz>
4023 * lto-streamer-out.c (DFS_write_tree_body): Drop
4024 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4025 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4026 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4027 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4028 Do not read DECL_ERROR_ISSUED.
4029 (unpack_ts_decl_with_vis_value_fields): Do not read
4031 (lto_input_ts_binfo_tree_pointers): Do not read
4032 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4033 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4034 write DECL_ERROR_ISSUED..
4035 (pack_ts_decl_with_vis_value_fields): Do not write
4037 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4038 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4039 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4040 * tree.h (tree_decl_common): Update comment.
4041 (DECL_ERROR_ISSUED): Remove.
4043 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4046 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4048 2013-08-28 Jan Hubicka <jh@suse.cz>
4050 * builtins.def (free): Declare leaf.
4052 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4055 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4056 * configure: Regenerate.
4058 2013-08-27 Martin Jambor <mjambor@suse.cz>
4060 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4061 (ipa_ancestor_jf_data): Likewise.
4062 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4063 (ipa_get_jf_pass_through_type_preserved): New function.
4064 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4065 (ipa_get_jf_ancestor_type_preserved): New function.
4066 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4067 (ipa_get_jf_ancestor_result): Likewise.
4068 (propagate_vals_accross_pass_through): Use
4069 ipa_get_jf_pass_through_result to do all the value mappings.
4070 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4071 type_preserved flag.
4072 (ipa_set_jf_cst_copy): New function.
4073 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4074 (ipa_set_jf_arith_pass_through): Likewise.
4075 (ipa_set_ancestor_jf): Likewise.
4076 (compute_complex_assign_jump_func): Set type_preserved instead of
4078 (ipa_compute_jump_functions_for_edge): Likewise.
4079 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4080 (update_jump_functions_after_inlining): Update type_preserved.
4081 Explicitely create jump functions when combining one with pass_through.
4082 (ipa_write_jump_function): Stream the type_preserved flags.
4083 (ipa_read_jump_function): Likewise.
4085 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4086 Aldy Hernandez <aldyh@redhat.com>
4088 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4089 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4090 * function.h (struct function): Add has_force_vect_loops and
4092 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4093 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4095 * gimple.def: Update CLAUSES comments.
4096 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4097 (gimple_build_omp_for): Add argument to prototype.
4098 (gimple_omp_for_kind): New.
4099 (gimple_omp_for_set_kind): New.
4100 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4101 GOVD_DATA_SHARE_CLASS.
4102 (enum omp_region_type): Add ORT_SIMD.
4103 (gimple_add_tmp_var): Handle ORT_SIMD.
4104 (gimplify_var_or_parm_decl): Same.
4105 (is_gimple_stmt): Same.
4106 (omp_firstprivatize_variable): Same.
4107 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4108 (omp_notice_variable): Handle ORT_SIMD.
4109 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4110 (omp_check_private): Handle ORT_SIMD.
4111 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4113 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4114 Handle OMP_CLAUSE_LASTPRIVATE.
4115 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4117 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4118 (gimplify_expr): Handle OMP_SIMD.
4119 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4120 (expand_GOMP_SIMD_VF): New.
4121 (expand_GOMP_SIMD_LAST_LANE): New.
4122 * internal-fn.def (GOMP_SIMD_LANE): New.
4123 (GOMP_SIMD_VF): New.
4124 (GOMP_SIMD_LAST_LANE): New.
4125 * omp-low.c: Include target.h.
4126 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4128 (check_omp_nesting_restrictions): Same.
4130 (lower_rec_simd_input_clauses): New.
4131 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4133 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4134 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4135 (expand_omp_build_assign): New.
4136 (expand_omp_for_init_counts): New.
4137 (expand_omp_for_init_vars): New.
4138 (extract_omp_for_update_vars): New.
4139 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4140 and rewrite accordingly.
4141 (expand_omp_simd): New.
4142 (expand_omp_for): Use expand_omp_simd.
4143 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4144 (lower_omp_for): Do not lower the body.
4145 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4147 * tree-flow.h (find_omp_clause): Remove prototype.
4148 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4149 forcing vectorization of the loop, or if flag_tree_vectorize.
4150 (gate_tree_if_conversion): Similarly.
4151 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4152 gimple_build_omp_for.
4153 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4154 * tree-parloops (create_parallel_loop): Pass kind argument to
4155 gimple_build_omp_for.
4156 * tree-pretty-print.c (dump_omp_clause): Add cases for
4157 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4158 OMP_CLAUSE__SIMDUID_.
4159 (dump_generic_node): Handle OMP_SIMD.
4160 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4161 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4162 unroll OMP_SIMD loops here.
4163 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4164 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4166 (vect_analyze_data_refs): Handle simd loops.
4167 * tree-vect-loop.c (vectorizable_live_operation): Handle
4169 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4170 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4171 (vectorizable_load): Same.
4172 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4173 (struct simduid_to_vf): New.
4174 (simduid_to_vf::hash): New.
4175 (simduid_to-vf::equal): New.
4176 (struct simd_array_to_simduid): New.
4177 (simd_array_to_simduid::hash): New.
4178 (simd_array_to_simduid::equal): New.
4179 (adjust_simduid_builtins): New.
4180 (struct note_simd_array_uses_struct): New.
4181 (note_simd_array_uses_cb): New.
4182 (note_simd_array_uses): New.
4183 (vectorize_loops): Handle simd hints and adjust simd builtins
4185 * tree-vectorizer.h (struct _stmt_vec_info): Add
4186 simd_lane_access_p field.
4187 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4188 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4189 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4190 (omp_clause_code_name): Same.
4191 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4192 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4193 * tree.def (OMP_SIMD): New entry.
4194 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4195 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4196 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4197 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4198 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4199 (OMP_CLAUSE_LINEAR_STEP): New.
4200 (OMP_CLAUSE_SAFELEN_EXPR): New.
4201 (OMP_CLAUSE__SIMDUID__DECL): New.
4202 (find_omp_clause): New prototype.
4204 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4206 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4207 Haswell processor detection.
4209 2013-08-27 Christian Widmer <shadow@umbrox.de>
4212 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4213 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4214 AVX2 capable processors.
4216 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4218 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4219 of vget_low_* with implementations in terms of other intrinsics.
4221 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4224 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4225 values to -1, 0 and 1.
4227 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4229 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4230 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4231 (<optab><mode>3_insn): Remove.
4232 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4233 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4234 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4235 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4236 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4237 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4238 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4239 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4240 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4241 (ror<mode>3_insn): Likewise.
4242 * config/aarch64/predicates.md (aarch64_simd_register): New.
4244 2013-08-27 Richard Biener <rguenther@suse.de>
4246 PR tree-optimization/57521
4247 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4248 one edge is non-critical.
4249 (find_phi_replacement_condition): Make sure to use a non-critical
4250 edge. Cleanup and remove old bug workarounds.
4251 (bb_postdominates_preds): Remove.
4252 (if_convertible_loop_p_1): Do not compute post-dominators.
4253 (combine_blocks): Do not free post-dominators.
4254 (main_tree_if_conversion): Likewise.
4255 (pass_data_if_conversion): Add TODO_verify_ssa.
4257 2013-08-27 DJ Delorie <dj@redhat.com>
4259 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4261 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4263 * function.c (assign_parm_find_data_types): Set passed_mode and
4264 nominal_mode to the TYPE_MODE of nominal_type for the built
4265 pointer type in case of the struct-pass-by-reference.
4267 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4269 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4270 (UINT16_TYPE): Change default to "unsigned int".
4272 * config/avr/avr.opt (mfract-convert-truncate): New option.
4273 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4274 is set, round negative fractional integers according to n1169
4275 when converting to integer types.
4277 2013-08-26 Jan Hubicka <jh@suse.cz>
4279 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4280 methods can not be called indirectly when their address is not taken.
4282 2013-08-26 Jan Hubicka <jh@suse.cz>
4284 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4287 2013-08-26 Jan Hubicka <jh@suse.cz>
4289 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4292 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4294 * reload.c (find_valid_class): Allow classes that do not include
4295 FIRST_PSEUDO_REGISTER - 1.
4297 2013-08-26 Jan Hubicka <jh@suse.cz>
4299 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4300 fix edge count/frequency when speculation failed; fix type check
4301 for the direct call.
4303 2013-08-26 Jan Hubicka <jh@suse.cz>
4305 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4307 2013-08-26 Jan Hubicka <jh@suse.cz>
4309 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4310 to be changed by edge redirection.
4312 2013-08-26 Jan Hubicka <jh@suse.cz>
4314 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4315 formating; add sanity check.
4316 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4317 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4318 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4319 (resolve_noninline_speculation): Update callee keys, too.
4321 2013-08-26 Jan Hubicka <jh@suse.cz>
4323 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4324 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4326 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4328 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4331 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4333 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4334 (x86_64_elf_select_section): Ditto.
4335 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4337 (ix86_pass_by_reference): Ditto.
4338 (output_set_got): Ditto.
4339 (ix86_unary_operator_ok): Ditto.
4340 (ix86_expand_builtin): Ditto.
4342 2013-08-23 Jan Hubicka <jh@suse.cz>
4344 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4346 2013-08-23 Jan Hubicka <jh@suse.cz>
4348 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4349 (tree_decl_with_vis): Add FINAL field.
4351 2013-08-23 Jeff Law <law@redhat.com>
4353 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4354 the debugging dump when the expression is fully redundant.
4356 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4358 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4359 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4360 * pretty-print.c (pp_formatted_text_data): Likewise.
4361 (pp_write_text_to_stream): Likewise.
4362 (pp_write_text_as_dot_label_to_stream): Likewise.
4363 (pp_append_r): Likewise.
4364 (pp_format): Likewise.
4365 (pp_flush): Likewise.
4366 (pp_clear_output_area): Likewise.
4367 (pp_append_text): Likewise.
4368 (pp_formatted_text): Likewise.
4369 (pp_remaining_character_count_for_line): Likewise.
4370 (pp_newline): Likewise.
4371 (pp_character): Likewise.
4372 (output_buffer::~output_buffer): Define.
4373 (pretty_printer::~pretty_printer): Destruct output buffer.
4374 * pretty-print.h (output_buffer::~output_buffer): Declare.
4375 (pretty_printer::~pretty_printer): Declare virtual.
4377 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4380 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4381 HOST_WIDE_INT_M1U): New macros.
4382 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4383 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4384 unsigned -1 for lshift.
4385 * cse.c (cse_insn): Likewise.
4386 * double-int.c (rshift_double, lshift_double): Likewise.
4387 * builtins.c (fold_builtin_bitop): Likewise.
4388 * combine.c (force_to_mode): Likewise.
4389 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4390 * simplify-rtx.c (simplify_const_unary_operation,
4391 simplify_const_binary_operation): Likewise.
4392 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4393 check_va_list_escapes): Likewise.
4394 * rtlanal.c (nonzero_bits1): Likewise.
4395 * expmed.c (expand_smod_pow2): Likewise.
4396 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4398 2013-08-23 Jan Hubicka <jh@suse.cz>
4400 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4401 as having address taken.
4403 2013-08-23 Jan Hubicka <jh@suse.cz>
4405 * ipa-utils.h (method_class_type): Declare.
4406 * ipa-devirt.c (method_class_type): Export.
4408 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4409 do not walk base classes of anonymous types.
4411 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4413 PR rtl-optimization/58220
4415 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4416 handle SEQUENCE insns properly.
4418 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4420 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4422 (pp_newline_and_indent): Likewise.
4423 (pp_separate_with): Likewise.
4424 * pretty-print.c (pp_newline_and_flush): Define.
4425 (pp_newline_and_indent): Likewise.
4426 (pp_separate_with): Likewise.
4428 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4431 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4432 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4434 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4436 * config/i386/predicates.md (ext_sse_reg_operand): New.
4437 * config/i386/i386.md (*movti_internal): Use
4438 predicate to determine if EVEX is needed.
4439 (*movsi_internal): Ditto.
4440 (*movdf_internal): Ditto.
4441 (*movsf_internal): Ditto.
4442 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4444 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4446 PR tree-optimization/58209
4447 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4448 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4449 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4450 emit POINTER_PLUS_EXPR.
4451 (create_tailcall_accumulator): For pointer result type accumulate in
4454 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4456 * configure.ac: Add backslashes missing from the last change.
4457 * configure: Regenerate.
4459 2013-08-22 Jan Hubicka <jh@susue.cz>
4461 * ipa.c (function_and_variable_visibility): First remember function
4462 was global and then make it local.
4464 2013-08-22 Julian Brown <julian@codesourcery.com>
4466 * configure.ac: Add aarch64 to list of arches which use "nop" in
4468 * configure: Regenerate.
4470 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4472 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4473 gnu_libc_has_function.
4474 * config/s390/tpf.h: Likewise.
4476 2013-08-22 Jan Hubicka <jh@susue.cz>
4478 * timevar.c (validate_phases): Add cast.
4480 2013-08-22 Jan Hubicka <jh@susue.cz>
4482 * timevar.c (validate_phases): Use size_t for memory.
4483 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4485 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4487 * pretty-print.h (output_buffer::output_buffer): Declare.
4488 (pretty_printer::pretty_printer): Likewise.
4489 (pp_construct): Remove.
4490 * pretty-print.c (output_buffer::output_buffer): Define.
4491 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4492 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4493 (print_gimple_expr): Likewise.
4494 (print_gimple_seq): Likewise.
4495 (gimple_dump_bb): Likewise.
4496 * sched-vis.c (dump_value_slim): Likewise.
4497 (dump_insn_slim): Likewise.
4498 (dump_rtl_slim): Likewise.
4499 (str_pattern_slim): Likewise.
4500 * tree-mudflap.c (mf_varname_tree): Likewise.
4501 * graph.c (print_graph_cfg): Likewise.
4502 (start_graph_dump): Likewise.
4503 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4505 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4506 pretty printer initialization.
4507 * coretypes.h (diagnostic_context): Remove superflous type alias
4509 (pretty_printer): Likewise. Declare directly as a class.
4510 (pretty_print_info): Remove declaration as class.
4511 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4512 and pp_clear_output_area.
4513 (asan_add_global): Likewise.
4515 2013-08-22 Jan Hubicka <jh@suse.cz>
4517 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4518 * ipa-utils.h (update_type_inheritance_graph): Declare.
4519 (possible_polymorphic_call_target_p): Declare.
4520 (possible_polymorphic_call_target_p): New.
4521 * ipa-devirt.c: Update toplevel comments.
4522 (cached_polymorphic_call_targets): Move up.
4523 (odr_type_d): Move ID down.
4524 (polymorphic_type_binfo_p): Update comment.
4525 (odr_hasher::remove): Likewise;
4526 (get_odr_type): Set anonymous_namespace.
4527 (dump_odr_type): Dump it.
4528 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4529 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4530 (record_binfo): Add comment.
4531 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4533 (devirt_node_removal_hook): Do not iCE when cache is freed.
4534 (possible_polymorphic_call_target_p): New predicate.
4535 (update_type_inheritance_graph): New function.
4537 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4538 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4539 Sergey Lega <sergey.s.lega@intel.com>
4540 Anna Tikhonova <anna.tikhonova@intel.com>
4541 Ilya Tocar <ilya.tocar@intel.com>
4542 Andrey Turetskiy <andrey.turetskiy@intel.com>
4543 Ilya Verbin <ilya.verbin@intel.com>
4544 Kirill Yukhin <kirill.yukhin@intel.com>
4545 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4547 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4548 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4549 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4550 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4551 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4552 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4553 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4554 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4555 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4556 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4557 OPT_mavx512pf, OPT_mavx512er cases.
4558 * config/i386/constraints.md (v): New constraint.
4559 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4560 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4561 (bit_AVX512CD): New.
4562 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4563 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4564 * config/i386/i386-c.c (ix86_target_macros_internal):
4565 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4567 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4568 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4569 * config/i386/i386.c (regclass_map, dbx_register_map)
4570 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4571 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4572 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4574 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4575 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4576 -mavx512pf options. Fix formatting.
4577 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4578 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4579 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4580 -mavx512cd, -mavx512pf options.
4581 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4582 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4583 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4585 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4586 (ix86_set_reg_reg_cost): Ditto.
4587 (x86_order_regs_for_local_alloc): Ditto.
4588 (MAX_VECT_LEN): Extend to 64-byte.
4589 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4590 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4591 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4592 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4593 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4594 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4595 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4596 (SSE_REG_MODE_P): Support new modes.
4597 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4598 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4599 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4600 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4601 (REG_CLASS_CONTENTS): Add new registers.
4602 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4603 (EXT_REX_SSE_REGNO_P): New.
4604 (HI_REGISTER_NAMES): Add new registers.
4605 * config/i386/i386.md: Define constants for new registers.
4606 (mode): Add new 512-bit modes.
4607 (prefix): Support evex prefix.
4608 (isa): Support avx512f, noavx512f, fma_avx512f.
4609 (ssemodesuffix): Add new 512-bit modes.
4611 (*movxi_internal_avx512f): Ditto.
4612 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4614 (*movsi_internal): Likewise.
4615 (*movdf_internal): Likewise.
4616 (*movsf_internal): Likewise.
4617 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4618 (<code><mode>3): Likewise.
4619 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4621 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4622 with the new constraint "v".
4623 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4625 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4626 with the new constraint "v".
4627 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4628 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4629 (*<plusminus_insn><mode>3): Likewise.
4630 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4631 (*mul<mode>3): Likewise.
4632 (<sse>_vmmul<mode>3): Likewise.
4633 (<sse>_div<mode>3): Likewise.
4634 (<sse>_vmdiv<mode>3): Likewise.
4635 (<sse>_sqrt<mode>2): Likewise.
4636 (<sse>_vmsqrt<mode>2): Likewise.
4637 (*<code><mode>3_finite): Likewise.
4638 (*<code><mode>3) <smaxmin>: Likewise.
4639 (<sse>_vm<code><mode>3): Likewise.
4640 (*<code><mode>3) <any_logic>: Likewise.
4641 (*fma_fmadd_<mode>): Likewise.
4642 (*fma_fmsub_<mode>): Likewise.
4643 (*fma_fnmadd_<mode>): Likewise.
4644 (*fma_fnmsub_<mode>): Likewise.
4645 (*fma_fmaddsub_<mode>): Likewise.
4646 (*fma_fmsubadd_<mode>): Likewise.
4647 (*fmai_fmadd_<mode>): Likewise.
4648 (*fmai_fmsub_<mode>): Likewise.
4649 (*fmai_fnmadd_<mode>): Likewise.
4650 (*fmai_fnmsub_<mode>): Likewise.
4651 (sse_cvtsi2ss): Likewise.
4652 (sse_cvtsi2ssq): Likewise.
4653 (sse_cvtss2si): Likewise.
4654 (sse_cvtss2si_2): Likewise.
4655 (sse_cvtss2siq): Likewise.
4656 (sse_cvtss2siq_2): Likewise.
4657 (sse_cvttss2si): Likewise.
4658 (sse_cvtss2siq_2): Likewise.
4659 (float<sseintvecmodelower><mode>2): Likewise.
4660 (sse2_cvtsd2si_2): Likewise.
4661 (sse2_cvtsd2siq_2): Likewise.
4662 (*<plusminus_insn><mode>3): Likewise.
4663 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4664 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4665 (ashr<mode>3): Likewise.
4666 (<shift_insn><mode>3): Likewise.
4667 (avx2_<code><mode>3): Likewise.
4668 (*avx2_<code><mode>3): Likewise.
4669 (*andnot<mode>3): Likewise.
4670 (*<code><mode>3) <any_logic>: Likewise.
4671 (abs<mode>2): Likewise.
4672 (avx2_permvar<mode>): Likewise.
4673 (avx2_perm<mode>_1): Likewise.
4674 (*avx_vpermilp<mode>): Likewise.
4675 (avx_vpermilvar<mode>3): Likewise.
4676 (avx2_ashrv<mode>): Likewise.
4677 (avx2_<shift_insn>v<mode>): Likewise.
4678 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4680 * doc/rtl.texi: Document XImode.
4682 2013-08-21 Jeff Law <law@redhat.com>
4684 * tree-flow.h (register_jump_thread): Pass vector of edges
4685 instead of each important edge.
4686 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4687 thread path into a vector and pass that to register_jump_thread.
4688 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4689 passed in edge vector to the current 3-edge form.
4692 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4694 * dce.c (fini_dce): Call df_analyze again just in case
4695 delete_unmarked_insns removed anything.
4697 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4699 * reload.h (struct reg_equivs): Rename to ..
4700 (struct reg_equivs_s): .. this.
4702 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4704 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4706 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4708 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4710 2013-08-21 Jeff Law <law@redhat.com>
4712 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4713 simplify assignments too. If the RHS collapses to a singleton
4714 range, then return the value for the range.
4716 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4718 * config/i386/sse.md (V16): Rename to...
4720 (mov<mode>): Update iterator name.
4721 (*mov<mode>_internal): Ditto.
4722 (push<mode>1): Ditto.
4723 (movmisalign<mode>): Ditto.
4725 2013-08-20 Jan Hubicka <jh@suse.cz>
4728 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4729 entry for direct edges.
4730 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4732 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4734 Revert my last two changes, r201865 and r201864:
4737 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4739 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4740 instances can own GC refs.
4742 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4743 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4744 (gcc::context::gt_pch_nx): Likewise.
4745 (gcc::context::gt_pch_nx): Likewise.
4746 * ggc.h (gt_ggc_mx <T>): New.
4747 (gt_pch_nx_with_op <T>): New.
4748 (gt_pch_nx <T>): New.
4749 * passes.c (opt_pass::gt_ggc_mx): New.
4750 (opt_pass::gt_pch_nx): New.
4751 (opt_pass::gt_pch_nx_with_op): New.
4752 (pass_manager::gt_ggc_mx): New.
4753 (pass_manager::gt_pch_nx): New.
4754 (pass_manager::gt_pch_nx_with_op): New.
4755 (pass_manager::operator new): Use
4756 ggc_internal_cleared_alloc_stat rather than xcalloc.
4757 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4758 (pass_manager::gt_ggc_mx): New.
4759 (pass_manager::gt_pch_nx): New.
4760 (pass_manager::gt_pch_nx_with_op): New.
4761 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4762 (opt_pass::operator new): New.
4763 (opt_pass::gt_ggc_mx): New.
4764 (opt_pass::gt_pch_nx): New.
4765 (opt_pass::gt_pch_nx_with_op): New.
4768 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4770 * Makefile.in (GTFILES): Add context.h.
4771 * context.c (gcc::context::operator new): New.
4772 (gcc::context::gt_ggc_mx): New.
4773 (gcc::context::gt_pch_nx): New.
4774 (gcc::context::gt_pch_nx): New.
4775 * context.h (gcc::context): Add GTY((user)) marking.
4776 (gcc::context::operator new): New.
4777 (gcc::context::gt_ggc_mx): New.
4778 (gcc::context::gt_pch_nx): New.
4779 (gcc::context::gt_pch_nx): New.
4780 (g): Add GTY marking.
4781 (gt_ggc_mx (gcc::context *)): New.
4782 (gt_pch_nx (gcc::context *)): New.
4783 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4784 void *cookie)): New.
4785 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4787 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4789 * dce.c (fini_dce): Call df_analyze again just in case
4790 delete_unmarked_insns removed anything.
4792 2013-08-20 Teresa Johnson <tejohnson@google.com>
4794 PR rtl-optimizations/57451
4795 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4796 from crossing split section boundaries.
4798 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4800 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4801 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4802 with MULTLIB_DEFAULTS.
4804 2013-08-20 Nick Clifton <nickc@redhat.com>
4806 * target.def (narrow_volatile_bitfield): Note that the default
4807 value is false, not !TARGET_STRICT_ALIGN.
4808 * doc/tm.texi: Regenerate.
4810 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4812 Fix LIB_SPEC for systems without libpthread.
4814 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4815 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4817 * config/i386/linux-common.h: Likewise.
4818 * config/mips/linux-common.h: Likewise.
4820 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4822 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4825 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4827 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4828 instances can own GC refs.
4830 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4831 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4832 (gcc::context::gt_pch_nx): Likewise.
4833 (gcc::context::gt_pch_nx): Likewise.
4834 * ggc.h (gt_ggc_mx <T>): New.
4835 (gt_pch_nx_with_op <T>): New.
4836 (gt_pch_nx <T>): New.
4837 * passes.c (opt_pass::gt_ggc_mx): New.
4838 (opt_pass::gt_pch_nx): New.
4839 (opt_pass::gt_pch_nx_with_op): New.
4840 (pass_manager::gt_ggc_mx): New.
4841 (pass_manager::gt_pch_nx): New.
4842 (pass_manager::gt_pch_nx_with_op): New.
4843 (pass_manager::operator new): Use
4844 ggc_internal_cleared_alloc_stat rather than xcalloc.
4845 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4846 (pass_manager::gt_ggc_mx): New.
4847 (pass_manager::gt_pch_nx): New.
4848 (pass_manager::gt_pch_nx_with_op): New.
4849 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4850 (opt_pass::operator new): New.
4851 (opt_pass::gt_ggc_mx): New.
4852 (opt_pass::gt_pch_nx): New.
4853 (opt_pass::gt_pch_nx_with_op): New.
4855 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4857 * Makefile.in (GTFILES): Add context.h.
4858 * context.c (gcc::context::operator new): New.
4859 (gcc::context::gt_ggc_mx): New.
4860 (gcc::context::gt_pch_nx): New.
4861 (gcc::context::gt_pch_nx): New.
4862 * context.h (gcc::context): Add GTY((user)) marking.
4863 (gcc::context::operator new): New.
4864 (gcc::context::gt_ggc_mx): New.
4865 (gcc::context::gt_pch_nx): New.
4866 (gcc::context::gt_pch_nx): New.
4867 (g): Add GTY marking.
4868 (gt_ggc_mx (gcc::context *)): New.
4869 (gt_pch_nx (gcc::context *)): New.
4870 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4871 void *cookie)): New.
4872 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4874 2013-08-20 Alan Modra <amodra@gmail.com>
4877 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4878 (rs6000_emit_epilogue): Likewise.
4880 2013-08-19 Dehao Chen <dehao@google.com>
4882 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4884 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4885 Jakub Jelinek <jakub@redhat.com>
4887 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4888 (BUILT_IN_FABSD64): Likewise.
4889 (BUILT_IN_FABSD128): Likewise.
4890 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4891 (fold_builtin_1): Likewise.
4892 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4893 destination and source operands.
4894 (*abstd2_fpr): Likewise.
4895 (*nabstd2_fpr): Likewise.
4897 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4899 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4902 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4904 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4907 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4909 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4910 * builtins.c (default_libc_has_function): New.
4911 (gnu_libc_has_function): Ditto.
4912 (no_c99_libc_has_function): Ditto.
4913 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4914 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4915 (fold_builtin_sincos): Likewise.
4916 (fold_builtin_cexp): Likewise.
4917 * builtins.def (DEF_C94_BUILTIN): Likewise.
4918 (DEF_C99_BUILTIN): Likewise.
4919 (DEF_C99_C90RES_BUILTIN): Likewise.
4920 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4921 definitions to using this define.
4922 * config/darwin-protos.h (darwin_libc_has_function): New.
4923 * config/darwin.c (darwin_libc_has_function): Ditto.
4924 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4925 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4926 * config/darwin.h: Ditto.
4927 * config/elfos.h: Ditto.
4928 * config/freebsd.h: Ditto.
4929 * config/i386/cygming.h: Ditto.
4930 * config/i386/djgpp.h: Ditto.
4931 * config/i386/i386-interix.h: Ditto.
4932 * config/microblaze/microblaze.h: Ditto.
4933 * config/mmix/mmix.h: Ditto.
4934 * config/gnu-user.h: Ditto.
4935 * config/ia64/hpux.h: Ditto.
4936 * config/pa/pa-hpux.h: Ditto.
4937 * config/pdp11/pdp11.h: Ditto.
4938 * config/picochip/picochip.h: Ditto.
4939 * config/linux.h: Ditto.
4940 * config/netbsd.h: Ditto.
4941 * config/openbsd.h: Ditto.
4942 * config/rs6000/aix43.h: Ditto.
4943 * config/rs6000/aix51.h: Ditto.
4944 * config/rs6000/aix52.h: Ditto.
4945 * config/rs6000/aix53.h: Ditto.
4946 * config/rs6000/aix61.h: Ditto.
4947 * config/rs6000/darwin.h: Ditto.
4948 * config/rs6000/linux.h: Ditto.
4949 * config/rs6000/linux64.h: Ditto.
4950 * config/s390/tpf.h: Ditto.
4951 * config/sol2-10.h: Ditto.
4952 * config/sol2.h: Ditto.
4953 * config/vms/vms.h: Ditto.
4954 * config/vxworks.h: Ditto.
4955 * config/linux-android.c (linux_android_libc_has_function):
4956 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
4957 * config/linux-protos.h (linux_android_libc_has_function):
4959 * config/i386/i386.c (ix86_libc_has_function): New.
4960 * config/i386/i386-protos.h
4961 (ix86_libc_has_function): New declaration.
4962 * config/i386/i386.md
4963 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
4964 ("isinf<mode>2): Likewise.
4965 * convert.c (convert_to_integer): Using new target hook
4966 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
4967 TARGET_C99_FUNCTIONS.
4968 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
4969 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
4970 * coretypes.h (function_class): New enum for different
4971 classes of functions.
4972 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
4973 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
4974 (TARGET_HAS_SINCOS): Likewise.
4975 (TARGET_LIBC_HAS_FUNCTION): New.
4976 * doc/tm.texi: Regenerated.
4977 * targhooks.h (default_libc_has_function): New declaration.
4978 (no_c99_libc_has_function): Ditto.
4979 (gnu_libc_has_function): Ditto.
4980 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
4981 and TARGET_HAS_SINCOS.
4983 2013-08-18 Jan Hubicka <jh@suse.cz>
4985 * Makeifle-in (ipa-devirt.o): New.
4986 (GTFILES): Add ipa-utils.h and ipa-devirt.c
4987 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
4988 (analyze_functions): Look into possible targets of polymorphic call.
4989 * dumpfile.c (dump_files): Add type-inheritance dump.
4990 * dumpfile.h (TDI_inheritance): New.
4991 * ipa-devirt.c: New file.
4992 * ipa-utils.h (odr_type_d): Forward declare.
4993 (odr_type): New type.
4994 (build_type_inheritance_graph): Declare.
4995 (possible_polymorphic_call_targets): Declare and introduce inline
4996 variant when only edge is pased.
4997 (dump_possible_polymorphic_call_targets): Likewise.
4998 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
4999 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5000 (types_same_for_odr): ... here.
5001 * tree.h (type_in_anonymous_namespace_p): Declare.
5003 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5005 PR tree-optimization/58006
5006 * tree-parloops.c (take_address_of): Don't ICE if get_name
5008 (eliminate_local_variables_stmt): Remove clobber stmts.
5010 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5012 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5013 error message is issued for an alias to undefined symbol.
5015 2013-08-18 Jan Hubicka <jh@suse.cz>
5017 * cgraph.c (cgraph_create_indirect_edge): Discover
5018 polymorphic calls and record basic info into indirect_info.
5019 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5020 devirtualization, ignore objc function calls.
5021 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5022 call with no parm index info.
5023 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5024 * tree.c (virtual_method_call_p): New function.
5025 * tree.h (virtual_method_call_p): Declare.
5027 2013-08-16 Jan Hubicka <jh@suse.cz>
5030 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5032 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5034 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5037 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5039 * gengtype.c (type_for_name): Add special-case support for
5040 locating types within the "gcc::" namespace.
5041 (open_base_files): Emit a "using namespace gcc" directive.
5043 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5046 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5047 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5049 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5050 array instead of each individual operand as a separate argument.
5051 (emit_fusion_gpr_load): Likewise.
5052 (expand_fusion_gpr_load): Add new function declaration.
5054 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5055 signature to have the operands passed as an array, instead of as
5056 separate arguments. Allow ZERO_EXTEND to be in the memory
5057 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5058 depend on the register live/dead flags when peepholes are run.
5059 (expand_fusion_gpr_load): New function to be called from the
5060 peephole2 pass, to change the register that addis sets to be the
5062 (emit_fusion_gpr_load): Change the calling signature to have the
5063 operands passed as an array, instead of as separate arguments.
5064 Allow ZERO_EXTEND to be in the memory address, and also
5065 SIGN_EXTEND if -mpower8-fusion-sign.
5067 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5069 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5070 adjust the register addis loads up in the peephole2 pass. Do not
5071 depend on the register live/dead state when the peephole pass is done.
5073 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5075 * gengtype.c (create_user_defined_type): Ensure that the kind
5076 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5077 declaration is seen before the GTY((user)) marking.
5079 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5082 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5084 2013-08-16 Jan Hubicka <jh@suse.cz>
5086 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5087 arugment expected_type.
5088 (gimple_fold_call): Use it.
5089 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5090 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5091 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5092 (try_make_edge_direct_virtual_call): Likewise.
5093 * tree.c (obj_type_ref_class): New.
5094 * tree.h (obj_type_ref_class): Use it.
5096 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5098 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5099 (rtl_slim_pp): Likewise.
5100 (init_rtl_slim_pretty_print): Likewise.
5101 (dump_value_slim): Don't call it. Use local pretty printer.
5102 (dump_insn_slim): Likewise.
5103 (dump_rtl_slim): Likewise.
5104 (str_pattern_slim): Likewise.
5105 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5108 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5110 PR tree-optimization/58164
5111 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5112 walk gimple_goto_dest of GIMPLE_GOTO.
5114 PR tree-optimization/58165
5115 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5116 bi_call must be the last stmt in a bb, don't split_block, instead
5117 use fallthru edge from it and give up if there is none.
5118 Release conds vector when returning early.
5120 2013-08-14 Xinliang David Li <davidxl@google.com>
5122 * config/i386/i386.c (ix86_option_override_internal):
5123 Remove unused variable and field.
5125 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5128 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5129 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5130 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5131 and Linux, correct BLKmode alignment when 128-bit alignment is
5132 required and compatibility flag is not set.
5133 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5134 for zero-size arguments when compatibility flag is not set.
5136 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5138 PR tree-optimization/58145
5139 * tree-sra.c (build_ref_for_offset): If prev_base has
5140 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5142 2013-08-14 Xinliang David Li <davidxl@google.com>
5144 * config/i386/i386.c (ix86_option_override_internal):
5145 Fix uninitialized variable error.
5147 2013-08-14 Xinliang David Li <davidxl@google.com>
5149 * config/i386/i386.opt: Define two new options.
5150 * config/i386/x86-tune.def: Add arch selector field in macros.
5151 * config/i386/i386.h: Adjust macro definition.
5152 * config/i386/i386.c (ix86_option_override_internal):
5154 (parse_mtune_ctrl_str): New function.
5155 (set_ix86_tune_features): New function.
5156 (ix86_function_specific_restore): Call the new helper function.
5158 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5160 PR rtl-optimization/57662
5161 * sel-sched.c (code_motion_process_successors): When the current insn
5162 is removed after the recursive traversal, break from the loop.
5163 Add comments and debug printouts.
5165 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5166 Alexandre Oliva <aoliva@redhat.com>
5169 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5170 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5171 there also UNSPEC_PLTOFF.
5173 2013-08-14 Marek Polacek <polacek@redhat.com>
5175 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5176 NUM_CONDITIONS bit positions.
5178 2013-08-13 Cary Coutant <ccoutant@google.com>
5180 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5181 (attr_checksum): Hash vector contents instead of pointer.
5182 (attr_checksum_ordered): Likewise.
5184 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5186 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5187 when Pmode != word_mode. Add length_address attribute.
5188 (sse3_monitor_<mode>): Merge from sse3_monitor and
5189 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5190 Pmode != word_mode. Update insn length attribute.
5191 * config/i386/i386.c (ix86_option_override_internal): Update
5192 ix86_gen_monitor selection for merged sse3_monitor insn.
5194 2013-08-13 Julian Brown <julian@codesourcery.com>
5196 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5197 perform invalid legitimization on greater-than-word-size modes for
5200 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5202 * ira.c (setup_class_translate_array): Use aclass instead of cl
5203 for classes not fully covered by allocno classes.
5205 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5207 PR tree-optimization/57661
5208 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5209 * tree-inline.c (tree_function_versioning): Initialize it.
5210 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5211 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5212 that is not being copied.
5215 * asan.c (instrument_strlen_call): Fix typo in comment.
5216 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5218 2013-08-13 Steve Ellcey <sellcey@mips.com>
5220 * config/mips/mips.md (prefetch): Use lw instead of ld on
5221 loongson in 32bit mode.
5223 2013-08-13 Nick Clifton <nickc@redhat.com>
5225 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5227 2013-08-13 Jan Hubicka <jh@suse.cz>
5229 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5230 introduced edge; fix typo in sanity check.
5231 (cgraph_resolve_speculation): Export; improve diagnostic.
5232 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5233 speculation at type mismatch.
5234 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5235 (cgraph_resolve_speculation): Declare.
5236 (symtab_can_be_discarded): New function.
5237 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5238 * ipa-inline-transform.c (speculation_removed): New global var.
5239 (clone_inlined_nodes): See if speculation can be removed.
5240 (inline_call): If speculations was removed, we growths may not match.
5241 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5242 (speculation_useful_p): New function.
5243 (resolve_noninline_speculation): New function.
5244 (inline_small_functions): Resolve useless speculations.
5245 * ipa-inline.h (speculation_useful_p): Declare
5246 * ipa.c (can_replace_by_local_alias): Simplify.
5247 (ipa_profile): Produce speculative calls in non-lto, too;
5248 add simple cost model; produce local aliases.
5250 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5252 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5255 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5257 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5259 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5261 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5262 iterator instead of X87MODEF.
5264 2013-08-12 Perez Read <netfirewall@gmail.com>
5267 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5268 operand 0 for intel asm alternative.
5269 (*movabs<mode>_2): Ditto for operand 1.
5271 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5273 * config/aarch64/arm_none.h
5274 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5276 2013-08-12 Nick Clifton <nickc@redhat.com>
5278 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5280 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5282 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5283 expand for QI/HImode operand to produce more effictive code for
5284 unsigned char(short) --> float(double) conversion.
5286 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5288 * doc/invoke.texi: Mention that -ftls-model does not force the final
5291 2013-08-12 Marek Polacek <polacek@redhat.com>
5292 Marc Glisse <marc.glisse@inria.fr>
5294 PR tree-optimization/57980
5295 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5296 when creating -1 constant.
5298 2013-08-10 Jan Hubicka <jh@suse.cz>
5300 Workaround binutils PR14342.
5301 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5302 (gimple_init_edge_profiler): Likewise.
5303 (gimple_gen_ic_func_profiler): Likewise.
5305 2013-08-09 Jan Hubicka <jh@suse.cz>
5307 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5309 2013-08-09 Xinliang David Li <davidxl@google.com>
5311 * config/i386/stringop.def: New file.
5312 * config/i386/stringop.opt: New file.
5313 * config/i386/i386-opts.h: Include stringopt.def.
5314 * config/i386/i386.opt: Include stringopt.opt.
5315 * config/i386/i386.c (ix86_option_override_internal):
5316 Override default size based stringop inline strategies with options.
5317 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5320 2013-08-09 Jan Hubicka <jh@suse.cz>
5322 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5324 2013-08-09 Jan Hubicka <jh@suse.cz>
5326 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5328 (dump_cgraph_node): Dump profile-id.
5329 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5330 and common_target_probability.
5331 * lto-cgraph.c (lto_output_edge): Stream common targets.
5332 (lto_output_node): Stream profile ids.
5333 (input_node): Stream profile ids.
5334 (input_edge): Stream common targets.
5335 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5336 * ipa.c: Include value-prof.h
5337 (ipa_profile_generate_summary): Turn indirect call statement histograms
5338 into common targets.
5339 (ipa_profile): Turn common targets into speculative edges.
5341 2013-08-09 Jan Hubicka <jh@suse.cz>
5343 * cgraph.h (cgraph_node): Add profile_id.
5344 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5345 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5346 (del_node_map): Update.
5347 (find_func_by_funcdef_no): Replace by ...
5348 (find_func_by_profile_id): ... this one.
5349 (gimple_ic_transform): Do not remove useful histograms when
5350 speculation is not done; dump info when indirect call removal
5352 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5353 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5354 (__gcov_indirect_call_profiler_v2): .. this one.
5355 * profile.h (init_node_map): Update.
5356 * coverage.c (coverage_compute_profile_id): New function.
5357 * coverage.h (coverage_compute_profile_id): Declare.
5358 * tree-profile.c (init_ic_make_global_vars): Make
5359 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5360 (gimple_init_edge_profiler): Update prototype of
5361 __gcov_indirect_call_profiler.
5362 (gimple_gen_ic_func_profiler): Simplify.
5363 (tree_profiling): Use init_node_map
5365 2013-08-09 Jan Hubicka <jh@suse.cz>
5367 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5368 non-speculative refs.
5369 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5370 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5371 (cgraph_set_call_stmt): Likewise.
5372 (cgraph_create_edge_1): Fix release checking compilatoin;
5374 (cgraph_free_edge): Free indirect info.
5375 (cgraph_turn_edge_to_speculative): New function.
5376 (cgraph_speculative_call_info): New function.
5377 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5378 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5379 (dump_cgraph_node): Dump speculation.
5380 (verify_edge_count_and_frequency): Accept speculative edges.
5381 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5382 (verify_cgraph_node): Handle speculation.
5383 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5384 (cgraph_set_call_stmt): Update prototype.
5385 (cgraph_make_edge_direct): Update prototype.
5386 (cgraph_speculative_call_info): Declare.
5387 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5388 to change; update call of ipa_find_references.
5389 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5390 lto_stmt_uid and speculative flags.
5391 (ipa_dump_references): Dump speculation.
5392 (ipa_clone_references): Clone speculative flag.
5393 (ipa_clone_referring): Likewise.
5394 (ipa_clone_ref): New function.
5395 (ipa_find_reference): Look into lto_stmt_uids
5396 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5397 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5398 (ipa_find_reference): Update declaration.
5399 (ipa_clone_ref): Declare.
5400 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5401 stream speculative flag.
5402 (lto_output_ref): Stream statements uids and speculation.
5403 (input_ref): Likewise.
5404 (input_edge): Stream speuclation.
5405 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5406 (cgraph_set_call_stmt_including_clones): Handle speculation.
5407 * ipa-inline.c (heap_edge_removal_hook): New function.
5408 (inline_small_functions): Register it.
5409 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5410 also initialize refs.
5411 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5413 (try_make_edge_direct_simple_call): Likewise.
5414 (try_make_edge_direct_simple_call): Likewise.
5415 (update_indirect_edges_after_inlining): Likewise.
5416 (remove_described_reference): Look proper lto_stmt_uid.
5417 (propagate_controlled_uses): Likewise.
5418 (propagate_controlled_uses): Liekwise.
5419 * tree-inline.c (copy_bb): Copy speculative edges.
5420 (redirect_all_calls): New function.
5421 (copy_cfg_body): Do redirection after loop info is updated.
5422 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5424 2013-08-09 Jan Hubicka <jh@suse.cz>
5426 * lto-streamer-out.c (output_function): Renumber PHIs.
5427 * lto-streamer-in.c (input_function): Likewise.
5429 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5431 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5432 (get_lane_unsigned): Likewise.
5433 (dup_lane_scalar): Likewise.
5434 (get_lane): enable for VALL.
5435 * config/aarch64/aarch64-simd.md
5436 (aarch64_dup_lane_scalar<mode>): Remove.
5437 (aarch64_get_lane_signed<mode>): Likewise.
5438 (aarch64_get_lane_unsigned<mode>): Likewise.
5439 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5440 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5441 (aarch64_get_lane<mode>): Enable for all vector modes.
5442 (aarch64_get_lanedi): Remove misleading constraints.
5443 * config/aarch64/arm_neon.h
5444 (__aarch64_vget_lane_any): Define.
5445 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5446 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5447 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5448 * config/aarch64/iterators.md (VDQQH): New.
5452 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5454 * configure.ac: Add GAS check for LEON instructions on SPARC.
5455 * configure: Regenerate.
5456 * config.in: Likewise.
5457 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5459 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5460 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5462 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5463 (AS_LEON_FLAG): New macro.
5464 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5465 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5466 Deal with LEON and LEON3 for the memory model.
5467 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5468 (atomic_compare_and_swap<mode>_1): Likewise.
5469 (*atomic_compare_and_swap<mode>_1): Likewise.
5471 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5473 * config/arm/neon.md (vcond): Fix floating-point vector
5474 comparisons against 0.
5476 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5478 * lra-constraints.c (emit_spill_move): Remove assert.
5479 (process_alt_operands): Add more debugging
5480 output. Increase reject for spilling into memory. Decrease
5481 reject for reloading scratch.
5482 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5484 2013-08-08 Steve Ellcey <sellcey@mips.com>
5486 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5487 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5488 micromips incompatible. Add nan2008.
5489 (MULTILIB_DIRNAMES): Add nan2008.
5490 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5491 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5492 and micromips incompatible. Add nan2008.
5493 (MULTILIB_DIRNAMES): Add nan2008.
5494 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5496 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5498 PR rtl-optimization/58079
5499 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5500 simplify_comparison has widened a comparison with an integer.
5502 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5504 * config/arm/neon.md (movmisalign<mode>): Disable when we
5505 don't allow unaligned accesses.
5506 (*movmisalign<mode>_neon_store): Likewise.
5507 (*movmisalign<mode>_neon_load): Likewise.
5508 (*movmisalign<mode>_neon_store): Likewise.
5509 (*movmisalign<mode>_neon_load): Likewise.
5511 2013-08-08 Jan Hubicka <jh@suse.cz>
5513 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5514 (make_pass_rebuild_cgraph_edges): Also clear references.
5515 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5516 * ipa-inline-transform.c (inline_transform): Remove all references
5518 * cgraphunit.c (expand_function): Remove all references after
5520 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5521 (ipa_find_reference): Rewrite to iterator.
5522 (remove_stmt_references): Likewise.
5523 (ipa_clear_stmts_in_references): New function.
5524 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5525 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5527 * ipa-split.c (split_function): Remove references in split function.
5529 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5532 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5533 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5535 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5538 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5539 suggested mode for the assignment isn't compatible with the
5542 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5545 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5547 2013-08-07 Xinliang David Li <davidxl@google.com>
5549 * config/i386/i386.opt: New option -mtune-ctrl=.
5550 * config/i386/x86-tune.def: New file.
5551 * config/i386/i386.h: include x86-tune.def.
5552 * config/i386/i386.c (ix86_option_override_internal):
5553 Parsing -mtune-ctrl= option and set tune features.
5555 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5558 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5559 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5562 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5564 * rtl.h (update_alignments): Declare.
5565 * final.c (grow_label_align): New function extracted from...
5566 (shorten_branches): ...here. Call it.
5567 (update_alignments): New function.
5568 * reorg.c (sibling_labels): New variable.
5569 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5570 the new label along with it onto the sibling_labels vector.
5571 (fill_simple_delay_slots): Adjust call to get_label_before.
5572 (fill_slots_from_thread): Likewise.
5573 (relax_delay_slots): Likewise.
5574 (make_return_insns): Likewise.
5575 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5577 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5579 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5580 document its semantics.
5581 (diagnostic_report_diagnostic): Adjust accordingly.
5583 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5585 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5586 (sparc_option_override): ...and port to new C++ pass API.
5587 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5589 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5591 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5593 2013-08-06 Caroline Tice <cmtice@google.com>
5595 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5596 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5597 * tree-pass.h: Add pass_vtable_verify.
5598 * varasm.c (assemble_variable): Add code to properly set the comdat
5599 section and name for the .vtable_map_vars section.
5600 (assemble_vtyv_preinit_initializer): New function.
5601 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5603 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5604 * vtable-verify.c: New file.
5605 * vtable-verify.h: New file.
5606 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5607 initialiation levels.
5608 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5609 * passes.def: Insert pass_vtable_verify.
5610 * aclocal.m4: Reorder includes.
5611 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5612 -fvtv-counts options.
5613 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5614 as appropriate, if -fvtable-verify=... is used.
5615 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5616 -fvtable-verify=... is used.
5617 * Makefile.in (OBJS): Add vtable-verify.o to list.
5618 (vtable-verify.o): Add new build rule.
5619 (GTFILES): Add vtable-verify.c to list.
5620 * common.opt (fvtable-verify=): New flag.
5621 (vtv_priority): Values for fvtable-verify= flag.
5622 (fvtv-counts): New flag.
5623 (fvtv-debug): New flag.
5624 * tree.h (save_vtable_map_decl): New extern function decl.
5626 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5628 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5629 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5630 (pass_data_rl78_devirt): ...new pass_data instance and...
5631 (make_pass_rl78_devirt): ...new function.
5632 (rl78_asm_file_start): Port pass registration to new C++ API.
5634 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5636 * coretypes.h (rtl_opt_pass): Add.
5637 (gcc::context): Add.
5638 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5639 (epiphany_init): Port to new C++ pass API.
5640 (epiphany_optimize_mode_switching): Likewise.
5641 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5642 (pass_manager::get_pass_mode_switching): New.
5643 (pass_manager::get_pass_peephole2): New.
5644 * mode-switching.c (pass_mode_switching): Add clone method.
5645 * recog.c (pass_peephole2): Add clone method.
5646 (pass_split_all_insns): Add clone method.
5648 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5650 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5651 (mips_option_override): ...here, porting to new C++ API for passes.
5653 2013-08-06 Jan Hubicka <jh@suse.cz>
5655 * cgraph.c (cgraph_get_body): New function based on lto.c
5657 * cgraph.h (cgraph_get_body): Declare.
5658 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5660 * cgraphunit.c (expand_function): Get body prior expanding.
5661 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5662 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5664 * passes.c (do_per_function_toporder): Get body.
5665 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5666 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5668 2013-08-06 Martin Jambor <mjambor@suse.cz>
5671 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5672 is not re-finalized. Rename second parameter to no_collect.
5674 2013-08-06 Martin Jambor <mjambor@suse.cz>
5677 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5678 MEM_REF has proper alignment information.
5680 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5683 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5685 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5686 rtx (*) (rtx, ...) with insn_gen_fn.
5687 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5688 insn_gen_fn::stored_funcptr. Add initializer braces.
5690 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5692 Rewrite how instances of passes are cloned to remove assumptions
5693 about their sizes (thus allowing pass subclasses to have
5694 additional data fields, albeit non-GC-managed ones at this point).
5696 * passes.c (make_pass_instance): Now that passes have clone
5697 methods, rewrite this function to eliminate XNEW and memcpy
5698 calls that used hardcoded sizes. Since this function no longer
5699 creates pass instances, rename it to...
5700 (add_pass_instance): ...this. Document the old way that passes were
5701 numbered and flagged, and rework this function to continue using it.
5702 (next_pass_1): Add an initial_pass argument for use by
5704 (position_pass): When adding multiple instances of a pass, use
5705 the pass's clone method, rather than relying on the XNEW/memcpy
5706 within the former make_pass_instance (now add_pass_instance).
5707 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5708 the initial instance of the current pass within the pass manager.
5710 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5712 This is the automated part of the conversion of passes from C
5713 structs to C++ classes.
5715 Patch autogenerated by refactor_passes.py from
5716 https://github.com/davidmalcolm/gcc-refactoring-scripts
5717 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5719 * asan.c (pass_asan): Convert from a global struct to a subclass of
5720 gimple_opt_pass along with...
5721 (pass_data_asan): ...new pass_data instance and...
5722 (make_pass_asan): ...new function.
5723 (pass_asan_O0): Convert from a global struct to a subclass of
5724 gimple_opt_pass along with...
5725 (pass_data_asan_O0): ...new pass_data instance and...
5726 (make_pass_asan_O0): ...new function.
5727 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5728 subclass of rtl_opt_pass along with...
5729 (pass_data_inc_dec): ...new pass_data instance and...
5730 (make_pass_inc_dec): ...new function.
5731 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5732 a subclass of rtl_opt_pass along with...
5733 (pass_data_reorder_blocks): ...new pass_data instance and...
5734 (make_pass_reorder_blocks): ...new function.
5735 (pass_duplicate_computed_gotos): Convert from a global struct to a
5736 subclass of rtl_opt_pass along with...
5737 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5738 (make_pass_duplicate_computed_gotos): ...new function.
5739 (pass_partition_blocks): Convert from a global struct to a subclass of
5740 rtl_opt_pass along with...
5741 (pass_data_partition_blocks): ...new pass_data instance and...
5742 (make_pass_partition_blocks): ...new function.
5743 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5744 struct to a subclass of rtl_opt_pass along with...
5745 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5747 (make_pass_branch_target_load_optimize1): ...new function.
5748 (pass_branch_target_load_optimize2): Convert from a global struct to a
5749 subclass of rtl_opt_pass along with...
5750 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5752 (make_pass_branch_target_load_optimize2): ...new function.
5753 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5754 of rtl_opt_pass along with...
5755 (pass_data_jump): ...new pass_data instance and...
5756 (make_pass_jump): ...new function.
5757 (pass_jump2): Convert from a global struct to a subclass of
5758 rtl_opt_pass along with...
5759 (pass_data_jump2): ...new pass_data instance and...
5760 (make_pass_jump2): ...new function.
5761 * cfgexpand.c (pass_expand): Convert from a global struct to a
5762 subclass of rtl_opt_pass along with...
5763 (pass_data_expand): ...new pass_data instance and...
5764 (make_pass_expand): ...new function.
5765 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5766 of rtl_opt_pass along with...
5767 (pass_data_free_cfg): ...new pass_data instance and...
5768 (make_pass_free_cfg): ...new function.
5769 (pass_into_cfg_layout_mode): Convert from a global struct to a
5770 subclass of rtl_opt_pass along with...
5771 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5772 (make_pass_into_cfg_layout_mode): ...new function.
5773 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5774 subclass of rtl_opt_pass along with...
5775 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5776 (make_pass_outof_cfg_layout_mode): ...new function.
5777 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5778 struct to a subclass of gimple_opt_pass along with...
5779 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5780 (make_pass_build_cgraph_edges): ...new function.
5781 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5782 subclass of gimple_opt_pass along with...
5783 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5784 (make_pass_rebuild_cgraph_edges): ...new function.
5785 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5786 subclass of gimple_opt_pass along with...
5787 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5789 (make_pass_remove_cgraph_callee_edges): ...new function.
5790 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5791 struct to a subclass of rtl_opt_pass along with...
5792 (pass_data_stack_adjustments): ...new pass_data instance and...
5793 (make_pass_stack_adjustments): ...new function.
5794 * combine.c (pass_combine): Convert from a global struct to a subclass
5795 of rtl_opt_pass along with...
5796 (pass_data_combine): ...new pass_data instance and...
5797 (make_pass_combine): ...new function.
5798 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5799 global struct to a subclass of rtl_opt_pass along with...
5800 (pass_data_compare_elim_after_reload): ...new pass_data instance
5802 (make_pass_compare_elim_after_reload): ...new function.
5803 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5804 of rtl_opt_pass along with...
5805 (pass_data_rtl_cprop): ...new pass_data instance and...
5806 (make_pass_rtl_cprop): ...new function.
5807 * cse.c (pass_cse): Convert from a global struct to a subclass of
5808 rtl_opt_pass along with...
5809 (pass_data_cse): ...new pass_data instance and...
5810 (make_pass_cse): ...new function.
5811 (pass_cse2): Convert from a global struct to a subclass of
5812 rtl_opt_pass along with...
5813 (pass_data_cse2): ...new pass_data instance and...
5814 (make_pass_cse2): ...new function.
5815 (pass_cse_after_global_opts): Convert from a global struct to a
5816 subclass of rtl_opt_pass along with...
5817 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5818 (make_pass_cse_after_global_opts): ...new function.
5819 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5820 of rtl_opt_pass along with...
5821 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5822 (make_pass_ud_rtl_dce): ...new function.
5823 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5824 rtl_opt_pass along with...
5825 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5826 (make_pass_fast_rtl_dce): ...new function.
5827 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5828 a subclass of rtl_opt_pass along with...
5829 (pass_data_df_initialize_opt): ...new pass_data instance and...
5830 (make_pass_df_initialize_opt): ...new function.
5831 (pass_df_initialize_no_opt): Convert from a global struct to a
5832 subclass of rtl_opt_pass along with...
5833 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5834 (make_pass_df_initialize_no_opt): ...new function.
5835 (pass_df_finish): Convert from a global struct to a subclass of
5836 rtl_opt_pass along with...
5837 (pass_data_df_finish): ...new pass_data instance and...
5838 (make_pass_df_finish): ...new function.
5839 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5840 rtl_opt_pass along with...
5841 (pass_data_rtl_dse1): ...new pass_data instance and...
5842 (make_pass_rtl_dse1): ...new function.
5843 (pass_rtl_dse2): Convert from a global struct to a subclass of
5844 rtl_opt_pass along with...
5845 (pass_data_rtl_dse2): ...new pass_data instance and...
5846 (make_pass_rtl_dse2): ...new function.
5847 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5848 subclass of rtl_opt_pass along with...
5849 (pass_data_dwarf2_frame): ...new pass_data instance and...
5850 (make_pass_dwarf2_frame): ...new function.
5851 * except.c (pass_set_nothrow_function_flags): Convert from a global
5852 struct to a subclass of rtl_opt_pass along with...
5853 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5855 (make_pass_set_nothrow_function_flags): ...new function.
5856 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5857 subclass of rtl_opt_pass along with...
5858 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5860 (make_pass_convert_to_eh_region_ranges): ...new function.
5861 * final.c (pass_compute_alignments): Convert from a global struct to a
5862 subclass of rtl_opt_pass along with...
5863 (pass_data_compute_alignments): ...new pass_data instance and...
5864 (make_pass_compute_alignments): ...new function.
5865 (pass_final): Convert from a global struct to a subclass of
5866 rtl_opt_pass along with...
5867 (pass_data_final): ...new pass_data instance and...
5868 (make_pass_final): ...new function.
5869 (pass_shorten_branches): Convert from a global struct to a subclass of
5870 rtl_opt_pass along with...
5871 (pass_data_shorten_branches): ...new pass_data instance and...
5872 (make_pass_shorten_branches): ...new function.
5873 (pass_clean_state): Convert from a global struct to a subclass of
5874 rtl_opt_pass along with...
5875 (pass_data_clean_state): ...new pass_data instance and...
5876 (make_pass_clean_state): ...new function.
5877 * function.c (pass_instantiate_virtual_regs): Convert from a global
5878 struct to a subclass of rtl_opt_pass along with...
5879 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5880 (make_pass_instantiate_virtual_regs): ...new function.
5881 (pass_leaf_regs): Convert from a global struct to a subclass of
5882 rtl_opt_pass along with...
5883 (pass_data_leaf_regs): ...new pass_data instance and...
5884 (make_pass_leaf_regs): ...new function.
5885 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5886 subclass of rtl_opt_pass along with...
5887 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5889 (make_pass_thread_prologue_and_epilogue): ...new function.
5890 (pass_match_asm_constraints): Convert from a global struct to a
5891 subclass of rtl_opt_pass along with...
5892 (pass_data_match_asm_constraints): ...new pass_data instance and...
5893 (make_pass_match_asm_constraints): ...new function.
5894 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5895 subclass of rtl_opt_pass along with...
5896 (pass_data_rtl_fwprop): ...new pass_data instance and...
5897 (make_pass_rtl_fwprop): ...new function.
5898 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5899 rtl_opt_pass along with...
5900 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5901 (make_pass_rtl_fwprop_addr): ...new function.
5902 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5903 rtl_opt_pass along with...
5904 (pass_data_rtl_pre): ...new pass_data instance and...
5905 (make_pass_rtl_pre): ...new function.
5906 (pass_rtl_hoist): Convert from a global struct to a subclass of
5907 rtl_opt_pass along with...
5908 (pass_data_rtl_hoist): ...new pass_data instance and...
5909 (make_pass_rtl_hoist): ...new function.
5910 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5911 subclass of gimple_opt_pass along with...
5912 (pass_data_lower_cf): ...new pass_data instance and...
5913 (make_pass_lower_cf): ...new function.
5914 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5915 from a global struct to a subclass of gimple_opt_pass along with...
5916 (pass_data_strength_reduction): ...new pass_data instance and...
5917 (make_pass_strength_reduction): ...new function.
5918 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5919 of rtl_opt_pass along with...
5920 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5921 (make_pass_rtl_ifcvt): ...new function.
5922 (pass_if_after_combine): Convert from a global struct to a subclass of
5923 rtl_opt_pass along with...
5924 (pass_data_if_after_combine): ...new pass_data instance and...
5925 (make_pass_if_after_combine): ...new function.
5926 (pass_if_after_reload): Convert from a global struct to a subclass of
5927 rtl_opt_pass along with...
5928 (pass_data_if_after_reload): ...new pass_data instance and...
5929 (make_pass_if_after_reload): ...new function.
5930 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5931 a subclass of rtl_opt_pass along with...
5932 (pass_data_initialize_regs): ...new pass_data instance and...
5933 (make_pass_initialize_regs): ...new function.
5934 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5935 of ipa_opt_pass_d along with...
5936 (pass_data_ipa_cp): ...new pass_data instance and...
5937 (make_pass_ipa_cp): ...new function.
5938 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5939 global struct to a subclass of gimple_opt_pass along with...
5940 (pass_data_inline_parameters): ...new pass_data instance and...
5941 (make_pass_inline_parameters): ...new function.
5942 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5943 subclass of gimple_opt_pass along with...
5944 (pass_data_early_inline): ...new pass_data instance and...
5945 (make_pass_early_inline): ...new function.
5946 (pass_ipa_inline): Convert from a global struct to a subclass of
5947 ipa_opt_pass_d along with...
5948 (pass_data_ipa_inline): ...new pass_data instance and...
5949 (make_pass_ipa_inline): ...new function.
5950 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5951 struct to a subclass of gimple_opt_pass along with...
5952 (pass_data_local_pure_const): ...new pass_data instance and...
5953 (make_pass_local_pure_const): ...new function.
5954 (pass_ipa_pure_const): Convert from a global struct to a subclass of
5955 ipa_opt_pass_d along with...
5956 (pass_data_ipa_pure_const): ...new pass_data instance and...
5957 (make_pass_ipa_pure_const): ...new function.
5958 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
5959 to a subclass of ipa_opt_pass_d along with...
5960 (pass_data_ipa_reference): ...new pass_data instance and...
5961 (make_pass_ipa_reference): ...new function.
5962 * ipa-split.c (pass_split_functions): Convert from a global struct to
5963 a subclass of gimple_opt_pass along with...
5964 (pass_data_split_functions): ...new pass_data instance and...
5965 (make_pass_split_functions): ...new function.
5966 (pass_feedback_split_functions): Convert from a global struct to a
5967 subclass of gimple_opt_pass along with...
5968 (pass_data_feedback_split_functions): ...new pass_data instance and...
5969 (make_pass_feedback_split_functions): ...new function.
5970 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
5971 global struct to a subclass of simple_ipa_opt_pass along with...
5972 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
5974 (make_pass_ipa_function_and_variable_visibility): ...new function.
5975 (pass_ipa_free_inline_summary): Convert from a global struct to a
5976 subclass of simple_ipa_opt_pass along with...
5977 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
5978 (make_pass_ipa_free_inline_summary): ...new function.
5979 (pass_ipa_whole_program_visibility): Convert from a global struct to a
5980 subclass of ipa_opt_pass_d along with...
5981 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
5983 (make_pass_ipa_whole_program_visibility): ...new function.
5984 (pass_ipa_profile): Convert from a global struct to a subclass of
5985 ipa_opt_pass_d along with...
5986 (pass_data_ipa_profile): ...new pass_data instance and...
5987 (make_pass_ipa_profile): ...new function.
5988 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
5989 ipa_opt_pass_d along with...
5990 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
5991 (make_pass_ipa_cdtor_merge): ...new function.
5992 * ira.c (pass_ira): Convert from a global struct to a subclass of
5993 rtl_opt_pass along with...
5994 (pass_data_ira): ...new pass_data instance and...
5995 (make_pass_ira): ...new function.
5996 (pass_reload): Convert from a global struct to a subclass of
5997 rtl_opt_pass along with...
5998 (pass_data_reload): ...new pass_data instance and...
5999 (make_pass_reload): ...new function.
6000 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6001 subclass of rtl_opt_pass along with...
6002 (pass_data_cleanup_barriers): ...new pass_data instance and...
6003 (make_pass_cleanup_barriers): ...new function.
6004 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6005 of rtl_opt_pass along with...
6006 (pass_data_loop2): ...new pass_data instance and...
6007 (make_pass_loop2): ...new function.
6008 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6009 rtl_opt_pass along with...
6010 (pass_data_rtl_loop_init): ...new pass_data instance and...
6011 (make_pass_rtl_loop_init): ...new function.
6012 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6013 rtl_opt_pass along with...
6014 (pass_data_rtl_loop_done): ...new pass_data instance and...
6015 (make_pass_rtl_loop_done): ...new function.
6016 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6017 subclass of rtl_opt_pass along with...
6018 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6019 (make_pass_rtl_move_loop_invariants): ...new function.
6020 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6021 rtl_opt_pass along with...
6022 (pass_data_rtl_unswitch): ...new pass_data instance and...
6023 (make_pass_rtl_unswitch): ...new function.
6024 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6025 subclass of rtl_opt_pass along with...
6026 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6028 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6029 (pass_rtl_doloop): Convert from a global struct to a subclass of
6030 rtl_opt_pass along with...
6031 (pass_data_rtl_doloop): ...new pass_data instance and...
6032 (make_pass_rtl_doloop): ...new function.
6033 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6034 a subclass of rtl_opt_pass along with...
6035 (pass_data_lower_subreg): ...new pass_data instance and...
6036 (make_pass_lower_subreg): ...new function.
6037 (pass_lower_subreg2): Convert from a global struct to a subclass of
6038 rtl_opt_pass along with...
6039 (pass_data_lower_subreg2): ...new pass_data instance and...
6040 (make_pass_lower_subreg2): ...new function.
6041 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6042 struct to a subclass of ipa_opt_pass_d along with...
6043 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6044 (make_pass_ipa_lto_gimple_out): ...new function.
6045 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6046 of ipa_opt_pass_d along with...
6047 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6048 (make_pass_ipa_lto_finish_out): ...new function.
6049 * mode-switching.c (pass_mode_switching): Convert from a global struct
6050 to a subclass of rtl_opt_pass along with...
6051 (pass_data_mode_switching): ...new pass_data instance and...
6052 (make_pass_mode_switching): ...new function.
6053 * modulo-sched.c (pass_sms): Convert from a global struct to a
6054 subclass of rtl_opt_pass along with...
6055 (pass_data_sms): ...new pass_data instance and...
6056 (make_pass_sms): ...new function.
6057 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6058 subclass of gimple_opt_pass along with...
6059 (pass_data_expand_omp): ...new pass_data instance and...
6060 (make_pass_expand_omp): ...new function.
6061 (pass_lower_omp): Convert from a global struct to a subclass of
6062 gimple_opt_pass along with...
6063 (pass_data_lower_omp): ...new pass_data instance and...
6064 (make_pass_lower_omp): ...new function.
6065 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6066 of gimple_opt_pass along with...
6067 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6068 (make_pass_diagnose_omp_blocks): ...new function.
6069 * passes.c (pass_early_local_passes): Convert from a global struct to
6070 a subclass of simple_ipa_opt_pass along with...
6071 (pass_data_early_local_passes): ...new pass_data instance and...
6072 (make_pass_early_local_passes): ...new function.
6073 (pass_all_early_optimizations): Convert from a global struct to a
6074 subclass of gimple_opt_pass along with...
6075 (pass_data_all_early_optimizations): ...new pass_data instance and...
6076 (make_pass_all_early_optimizations): ...new function.
6077 (pass_all_optimizations): Convert from a global struct to a subclass
6078 of gimple_opt_pass along with...
6079 (pass_data_all_optimizations): ...new pass_data instance and...
6080 (make_pass_all_optimizations): ...new function.
6081 (pass_all_optimizations_g): Convert from a global struct to a subclass
6082 of gimple_opt_pass along with...
6083 (pass_data_all_optimizations_g): ...new pass_data instance and...
6084 (make_pass_all_optimizations_g): ...new function.
6085 (pass_rest_of_compilation): Convert from a global struct to a subclass
6086 of rtl_opt_pass along with...
6087 (pass_data_rest_of_compilation): ...new pass_data instance and...
6088 (make_pass_rest_of_compilation): ...new function.
6089 (pass_postreload): Convert from a global struct to a subclass of
6090 rtl_opt_pass along with...
6091 (pass_data_postreload): ...new pass_data instance and...
6092 (make_pass_postreload): ...new function.
6093 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6094 subclass of rtl_opt_pass along with...
6095 (pass_data_gcse2): ...new pass_data instance and...
6096 (make_pass_gcse2): ...new function.
6097 * postreload.c (pass_postreload_cse): Convert from a global struct to
6098 a subclass of rtl_opt_pass along with...
6099 (pass_data_postreload_cse): ...new pass_data instance and...
6100 (make_pass_postreload_cse): ...new function.
6101 * predict.c (pass_profile): Convert from a global struct to a subclass
6102 of gimple_opt_pass along with...
6103 (pass_data_profile): ...new pass_data instance and...
6104 (make_pass_profile): ...new function.
6105 (pass_strip_predict_hints): Convert from a global struct to a subclass
6106 of gimple_opt_pass along with...
6107 (pass_data_strip_predict_hints): ...new pass_data instance and...
6108 (make_pass_strip_predict_hints): ...new function.
6109 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6110 of rtl_opt_pass along with...
6111 (pass_data_peephole2): ...new pass_data instance and...
6112 (make_pass_peephole2): ...new function.
6113 (pass_split_all_insns): Convert from a global struct to a subclass of
6114 rtl_opt_pass along with...
6115 (pass_data_split_all_insns): ...new pass_data instance and...
6116 (make_pass_split_all_insns): ...new function.
6117 (pass_split_after_reload): Convert from a global struct to a subclass
6118 of rtl_opt_pass along with...
6119 (pass_data_split_after_reload): ...new pass_data instance and...
6120 (make_pass_split_after_reload): ...new function.
6121 (pass_split_before_regstack): Convert from a global struct to a
6122 subclass of rtl_opt_pass along with...
6123 (pass_data_split_before_regstack): ...new pass_data instance and...
6124 (make_pass_split_before_regstack): ...new function.
6125 (pass_split_before_sched2): Convert from a global struct to a subclass
6126 of rtl_opt_pass along with...
6127 (pass_data_split_before_sched2): ...new pass_data instance and...
6128 (make_pass_split_before_sched2): ...new function.
6129 (pass_split_for_shorten_branches): Convert from a global struct to a
6130 subclass of rtl_opt_pass along with...
6131 (pass_data_split_for_shorten_branches): ...new pass_data instance
6133 (make_pass_split_for_shorten_branches): ...new function.
6134 * ree.c (pass_ree): Convert from a global struct to a subclass of
6135 rtl_opt_pass along with...
6136 (pass_data_ree): ...new pass_data instance and...
6137 (make_pass_ree): ...new function.
6138 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6139 subclass of rtl_opt_pass along with...
6140 (pass_data_stack_regs): ...new pass_data instance and...
6141 (make_pass_stack_regs): ...new function.
6142 (pass_stack_regs_run): Convert from a global struct to a subclass of
6143 rtl_opt_pass along with...
6144 (pass_data_stack_regs_run): ...new pass_data instance and...
6145 (make_pass_stack_regs_run): ...new function.
6146 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6147 subclass of rtl_opt_pass along with...
6148 (pass_data_cprop_hardreg): ...new pass_data instance and...
6149 (make_pass_cprop_hardreg): ...new function.
6150 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6151 subclass of rtl_opt_pass along with...
6152 (pass_data_reginfo_init): ...new pass_data instance and...
6153 (make_pass_reginfo_init): ...new function.
6154 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6155 of rtl_opt_pass along with...
6156 (pass_data_regmove): ...new pass_data instance and...
6157 (make_pass_regmove): ...new function.
6158 * regrename.c (pass_regrename): Convert from a global struct to a
6159 subclass of rtl_opt_pass along with...
6160 (pass_data_regrename): ...new pass_data instance and...
6161 (make_pass_regrename): ...new function.
6162 * reorg.c (pass_delay_slots): Convert from a global struct to a
6163 subclass of rtl_opt_pass along with...
6164 (pass_data_delay_slots): ...new pass_data instance and...
6165 (make_pass_delay_slots): ...new function.
6166 (pass_machine_reorg): Convert from a global struct to a subclass of
6167 rtl_opt_pass along with...
6168 (pass_data_machine_reorg): ...new pass_data instance and...
6169 (make_pass_machine_reorg): ...new function.
6170 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6171 of rtl_opt_pass along with...
6172 (pass_data_sched): ...new pass_data instance and...
6173 (make_pass_sched): ...new function.
6174 (pass_sched2): Convert from a global struct to a subclass of
6175 rtl_opt_pass along with...
6176 (pass_data_sched2): ...new pass_data instance and...
6177 (make_pass_sched2): ...new function.
6178 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6179 to a subclass of rtl_opt_pass along with...
6180 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6181 (make_pass_stack_ptr_mod): ...new function.
6182 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6183 to a subclass of rtl_opt_pass along with...
6184 (pass_data_rtl_store_motion): ...new pass_data instance and...
6185 (make_pass_rtl_store_motion): ...new function.
6186 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6187 of gimple_opt_pass along with...
6188 (pass_data_tracer): ...new pass_data instance and...
6189 (make_pass_tracer): ...new function.
6190 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6191 to a subclass of gimple_opt_pass along with...
6192 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6193 (make_pass_diagnose_tm_blocks): ...new function.
6194 (pass_lower_tm): Convert from a global struct to a subclass of
6195 gimple_opt_pass along with...
6196 (pass_data_lower_tm): ...new pass_data instance and...
6197 (make_pass_lower_tm): ...new function.
6198 (pass_tm_init): Convert from a global struct to a subclass of
6199 gimple_opt_pass along with...
6200 (pass_data_tm_init): ...new pass_data instance and...
6201 (make_pass_tm_init): ...new function.
6202 (pass_tm_mark): Convert from a global struct to a subclass of
6203 gimple_opt_pass along with...
6204 (pass_data_tm_mark): ...new pass_data instance and...
6205 (make_pass_tm_mark): ...new function.
6206 (pass_tm_edges): Convert from a global struct to a subclass of
6207 gimple_opt_pass along with...
6208 (pass_data_tm_edges): ...new pass_data instance and...
6209 (make_pass_tm_edges): ...new function.
6210 (pass_tm_memopt): Convert from a global struct to a subclass of
6211 gimple_opt_pass along with...
6212 (pass_data_tm_memopt): ...new pass_data instance and...
6213 (make_pass_tm_memopt): ...new function.
6214 (pass_ipa_tm): Convert from a global struct to a subclass of
6215 simple_ipa_opt_pass along with...
6216 (pass_data_ipa_tm): ...new pass_data instance and...
6217 (make_pass_ipa_tm): ...new function.
6218 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6219 subclass of gimple_opt_pass along with...
6220 (pass_data_call_cdce): ...new pass_data instance and...
6221 (make_pass_call_cdce): ...new function.
6222 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6223 subclass of gimple_opt_pass along with...
6224 (pass_data_build_cfg): ...new pass_data instance and...
6225 (make_pass_build_cfg): ...new function.
6226 (pass_split_crit_edges): Convert from a global struct to a subclass of
6227 gimple_opt_pass along with...
6228 (pass_data_split_crit_edges): ...new pass_data instance and...
6229 (make_pass_split_crit_edges): ...new function.
6230 (pass_warn_function_return): Convert from a global struct to a
6231 subclass of gimple_opt_pass along with...
6232 (pass_data_warn_function_return): ...new pass_data instance and...
6233 (make_pass_warn_function_return): ...new function.
6234 (pass_warn_function_noreturn): Convert from a global struct to a
6235 subclass of gimple_opt_pass along with...
6236 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6237 (make_pass_warn_function_noreturn): ...new function.
6238 (pass_warn_unused_result): Convert from a global struct to a subclass
6239 of gimple_opt_pass along with...
6240 (pass_data_warn_unused_result): ...new pass_data instance and...
6241 (make_pass_warn_unused_result): ...new function.
6242 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6243 a subclass of gimple_opt_pass along with...
6244 (pass_data_merge_phi): ...new pass_data instance and...
6245 (make_pass_merge_phi): ...new function.
6246 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6247 a subclass of gimple_opt_pass along with...
6248 (pass_data_lower_complex): ...new pass_data instance and...
6249 (make_pass_lower_complex): ...new function.
6250 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6251 gimple_opt_pass along with...
6252 (pass_data_lower_complex_O0): ...new pass_data instance and...
6253 (make_pass_lower_complex_O0): ...new function.
6254 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6255 subclass of gimple_opt_pass along with...
6256 (pass_data_lower_eh): ...new pass_data instance and...
6257 (make_pass_lower_eh): ...new function.
6258 (pass_refactor_eh): Convert from a global struct to a subclass of
6259 gimple_opt_pass along with...
6260 (pass_data_refactor_eh): ...new pass_data instance and...
6261 (make_pass_refactor_eh): ...new function.
6262 (pass_lower_resx): Convert from a global struct to a subclass of
6263 gimple_opt_pass along with...
6264 (pass_data_lower_resx): ...new pass_data instance and...
6265 (make_pass_lower_resx): ...new function.
6266 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6267 of gimple_opt_pass along with...
6268 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6269 (make_pass_lower_eh_dispatch): ...new function.
6270 (pass_cleanup_eh): Convert from a global struct to a subclass of
6271 gimple_opt_pass along with...
6272 (pass_data_cleanup_eh): ...new pass_data instance and...
6273 (make_pass_cleanup_eh): ...new function.
6274 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6275 to a subclass of simple_ipa_opt_pass along with...
6276 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6277 (make_pass_ipa_lower_emutls): ...new function.
6278 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6279 a subclass of gimple_opt_pass along with...
6280 (pass_data_if_conversion): ...new pass_data instance and...
6281 (make_pass_if_conversion): ...new function.
6282 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6283 subclass of gimple_opt_pass along with...
6284 (pass_data_build_ssa): ...new pass_data instance and...
6285 (make_pass_build_ssa): ...new function.
6286 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6287 global struct to a subclass of gimple_opt_pass along with...
6288 (pass_data_loop_distribution): ...new pass_data instance and...
6289 (make_pass_loop_distribution): ...new function.
6290 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6291 subclass of gimple_opt_pass along with...
6292 (pass_data_mudflap_1): ...new pass_data instance and...
6293 (make_pass_mudflap_1): ...new function.
6294 (pass_mudflap_2): Convert from a global struct to a subclass of
6295 gimple_opt_pass along with...
6296 (pass_data_mudflap_2): ...new pass_data instance and...
6297 (make_pass_mudflap_2): ...new function.
6298 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6299 subclass of gimple_opt_pass along with...
6300 (pass_data_mudflap_1): ...new pass_data instance and...
6301 (make_pass_mudflap_1): ...new function.
6302 (pass_mudflap_2): Convert from a global struct to a subclass of
6303 gimple_opt_pass along with...
6304 (pass_data_mudflap_2): ...new pass_data instance and...
6305 (make_pass_mudflap_2): ...new function.
6306 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6307 gimple_opt_pass along with...
6308 (pass_data_nrv): ...new pass_data instance and...
6309 (make_pass_nrv): ...new function.
6310 (pass_return_slot): Convert from a global struct to a subclass of
6311 gimple_opt_pass along with...
6312 (pass_data_return_slot): ...new pass_data instance and...
6313 (make_pass_return_slot): ...new function.
6314 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6315 to a subclass of gimple_opt_pass along with...
6316 (pass_data_object_sizes): ...new pass_data instance and...
6317 (make_pass_object_sizes): ...new function.
6318 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6319 global struct to a subclass of gimple_opt_pass along with...
6320 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6322 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6323 (pass_fixup_cfg): Convert from a global struct to a subclass of
6324 gimple_opt_pass along with...
6325 (pass_data_fixup_cfg): ...new pass_data instance and...
6326 (make_pass_fixup_cfg): ...new function.
6327 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6328 (make_pass_mudflap_1): ...new function.
6329 (pass_mudflap_2): Replace declaration with that of...
6330 (make_pass_mudflap_2): ...new function.
6331 (pass_asan): Replace declaration with that of...
6332 (make_pass_asan): ...new function.
6333 (pass_asan_O0): Replace declaration with that of...
6334 (make_pass_asan_O0): ...new function.
6335 (pass_tsan): Replace declaration with that of...
6336 (make_pass_tsan): ...new function.
6337 (pass_tsan_O0): Replace declaration with that of...
6338 (make_pass_tsan_O0): ...new function.
6339 (pass_lower_cf): Replace declaration with that of...
6340 (make_pass_lower_cf): ...new function.
6341 (pass_refactor_eh): Replace declaration with that of...
6342 (make_pass_refactor_eh): ...new function.
6343 (pass_lower_eh): Replace declaration with that of...
6344 (make_pass_lower_eh): ...new function.
6345 (pass_lower_eh_dispatch): Replace declaration with that of...
6346 (make_pass_lower_eh_dispatch): ...new function.
6347 (pass_lower_resx): Replace declaration with that of...
6348 (make_pass_lower_resx): ...new function.
6349 (pass_build_cfg): Replace declaration with that of...
6350 (make_pass_build_cfg): ...new function.
6351 (pass_early_tree_profile): Replace declaration with that of...
6352 (make_pass_early_tree_profile): ...new function.
6353 (pass_cleanup_eh): Replace declaration with that of...
6354 (make_pass_cleanup_eh): ...new function.
6355 (pass_sra): Replace declaration with that of...
6356 (make_pass_sra): ...new function.
6357 (pass_sra_early): Replace declaration with that of...
6358 (make_pass_sra_early): ...new function.
6359 (pass_early_ipa_sra): Replace declaration with that of...
6360 (make_pass_early_ipa_sra): ...new function.
6361 (pass_tail_recursion): Replace declaration with that of...
6362 (make_pass_tail_recursion): ...new function.
6363 (pass_tail_calls): Replace declaration with that of...
6364 (make_pass_tail_calls): ...new function.
6365 (pass_tree_loop): Replace declaration with that of...
6366 (make_pass_tree_loop): ...new function.
6367 (pass_tree_loop_init): Replace declaration with that of...
6368 (make_pass_tree_loop_init): ...new function.
6369 (pass_lim): Replace declaration with that of...
6370 (make_pass_lim): ...new function.
6371 (pass_tree_unswitch): Replace declaration with that of...
6372 (make_pass_tree_unswitch): ...new function.
6373 (pass_predcom): Replace declaration with that of...
6374 (make_pass_predcom): ...new function.
6375 (pass_iv_canon): Replace declaration with that of...
6376 (make_pass_iv_canon): ...new function.
6377 (pass_scev_cprop): Replace declaration with that of...
6378 (make_pass_scev_cprop): ...new function.
6379 (pass_empty_loop): Replace declaration with that of...
6380 (make_pass_empty_loop): ...new function.
6381 (pass_record_bounds): Replace declaration with that of...
6382 (make_pass_record_bounds): ...new function.
6383 (pass_graphite): Replace declaration with that of...
6384 (make_pass_graphite): ...new function.
6385 (pass_graphite_transforms): Replace declaration with that of...
6386 (make_pass_graphite_transforms): ...new function.
6387 (pass_if_conversion): Replace declaration with that of...
6388 (make_pass_if_conversion): ...new function.
6389 (pass_loop_distribution): Replace declaration with that of...
6390 (make_pass_loop_distribution): ...new function.
6391 (pass_vectorize): Replace declaration with that of...
6392 (make_pass_vectorize): ...new function.
6393 (pass_slp_vectorize): Replace declaration with that of...
6394 (make_pass_slp_vectorize): ...new function.
6395 (pass_complete_unroll): Replace declaration with that of...
6396 (make_pass_complete_unroll): ...new function.
6397 (pass_complete_unrolli): Replace declaration with that of...
6398 (make_pass_complete_unrolli): ...new function.
6399 (pass_parallelize_loops): Replace declaration with that of...
6400 (make_pass_parallelize_loops): ...new function.
6401 (pass_loop_prefetch): Replace declaration with that of...
6402 (make_pass_loop_prefetch): ...new function.
6403 (pass_iv_optimize): Replace declaration with that of...
6404 (make_pass_iv_optimize): ...new function.
6405 (pass_tree_loop_done): Replace declaration with that of...
6406 (make_pass_tree_loop_done): ...new function.
6407 (pass_ch): Replace declaration with that of...
6408 (make_pass_ch): ...new function.
6409 (pass_ccp): Replace declaration with that of...
6410 (make_pass_ccp): ...new function.
6411 (pass_phi_only_cprop): Replace declaration with that of...
6412 (make_pass_phi_only_cprop): ...new function.
6413 (pass_build_ssa): Replace declaration with that of...
6414 (make_pass_build_ssa): ...new function.
6415 (pass_build_alias): Replace declaration with that of...
6416 (make_pass_build_alias): ...new function.
6417 (pass_build_ealias): Replace declaration with that of...
6418 (make_pass_build_ealias): ...new function.
6419 (pass_dominator): Replace declaration with that of...
6420 (make_pass_dominator): ...new function.
6421 (pass_dce): Replace declaration with that of...
6422 (make_pass_dce): ...new function.
6423 (pass_dce_loop): Replace declaration with that of...
6424 (make_pass_dce_loop): ...new function.
6425 (pass_cd_dce): Replace declaration with that of...
6426 (make_pass_cd_dce): ...new function.
6427 (pass_call_cdce): Replace declaration with that of...
6428 (make_pass_call_cdce): ...new function.
6429 (pass_merge_phi): Replace declaration with that of...
6430 (make_pass_merge_phi): ...new function.
6431 (pass_split_crit_edges): Replace declaration with that of...
6432 (make_pass_split_crit_edges): ...new function.
6433 (pass_pre): Replace declaration with that of...
6434 (make_pass_pre): ...new function.
6435 (pass_profile): Replace declaration with that of...
6436 (make_pass_profile): ...new function.
6437 (pass_strip_predict_hints): Replace declaration with that of...
6438 (make_pass_strip_predict_hints): ...new function.
6439 (pass_lower_complex_O0): Replace declaration with that of...
6440 (make_pass_lower_complex_O0): ...new function.
6441 (pass_lower_complex): Replace declaration with that of...
6442 (make_pass_lower_complex): ...new function.
6443 (pass_lower_vector): Replace declaration with that of...
6444 (make_pass_lower_vector): ...new function.
6445 (pass_lower_vector_ssa): Replace declaration with that of...
6446 (make_pass_lower_vector_ssa): ...new function.
6447 (pass_lower_omp): Replace declaration with that of...
6448 (make_pass_lower_omp): ...new function.
6449 (pass_diagnose_omp_blocks): Replace declaration with that of...
6450 (make_pass_diagnose_omp_blocks): ...new function.
6451 (pass_expand_omp): Replace declaration with that of...
6452 (make_pass_expand_omp): ...new function.
6453 (pass_expand_omp_ssa): Replace declaration with that of...
6454 (make_pass_expand_omp_ssa): ...new function.
6455 (pass_object_sizes): Replace declaration with that of...
6456 (make_pass_object_sizes): ...new function.
6457 (pass_strlen): Replace declaration with that of...
6458 (make_pass_strlen): ...new function.
6459 (pass_fold_builtins): Replace declaration with that of...
6460 (make_pass_fold_builtins): ...new function.
6461 (pass_stdarg): Replace declaration with that of...
6462 (make_pass_stdarg): ...new function.
6463 (pass_early_warn_uninitialized): Replace declaration with that of...
6464 (make_pass_early_warn_uninitialized): ...new function.
6465 (pass_late_warn_uninitialized): Replace declaration with that of...
6466 (make_pass_late_warn_uninitialized): ...new function.
6467 (pass_cse_reciprocals): Replace declaration with that of...
6468 (make_pass_cse_reciprocals): ...new function.
6469 (pass_cse_sincos): Replace declaration with that of...
6470 (make_pass_cse_sincos): ...new function.
6471 (pass_optimize_bswap): Replace declaration with that of...
6472 (make_pass_optimize_bswap): ...new function.
6473 (pass_optimize_widening_mul): Replace declaration with that of...
6474 (make_pass_optimize_widening_mul): ...new function.
6475 (pass_warn_function_return): Replace declaration with that of...
6476 (make_pass_warn_function_return): ...new function.
6477 (pass_warn_function_noreturn): Replace declaration with that of...
6478 (make_pass_warn_function_noreturn): ...new function.
6479 (pass_cselim): Replace declaration with that of...
6480 (make_pass_cselim): ...new function.
6481 (pass_phiopt): Replace declaration with that of...
6482 (make_pass_phiopt): ...new function.
6483 (pass_forwprop): Replace declaration with that of...
6484 (make_pass_forwprop): ...new function.
6485 (pass_phiprop): Replace declaration with that of...
6486 (make_pass_phiprop): ...new function.
6487 (pass_tree_ifcombine): Replace declaration with that of...
6488 (make_pass_tree_ifcombine): ...new function.
6489 (pass_dse): Replace declaration with that of...
6490 (make_pass_dse): ...new function.
6491 (pass_nrv): Replace declaration with that of...
6492 (make_pass_nrv): ...new function.
6493 (pass_rename_ssa_copies): Replace declaration with that of...
6494 (make_pass_rename_ssa_copies): ...new function.
6495 (pass_sink_code): Replace declaration with that of...
6496 (make_pass_sink_code): ...new function.
6497 (pass_fre): Replace declaration with that of...
6498 (make_pass_fre): ...new function.
6499 (pass_check_data_deps): Replace declaration with that of...
6500 (make_pass_check_data_deps): ...new function.
6501 (pass_copy_prop): Replace declaration with that of...
6502 (make_pass_copy_prop): ...new function.
6503 (pass_vrp): Replace declaration with that of...
6504 (make_pass_vrp): ...new function.
6505 (pass_uncprop): Replace declaration with that of...
6506 (make_pass_uncprop): ...new function.
6507 (pass_return_slot): Replace declaration with that of...
6508 (make_pass_return_slot): ...new function.
6509 (pass_reassoc): Replace declaration with that of...
6510 (make_pass_reassoc): ...new function.
6511 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6512 (make_pass_rebuild_cgraph_edges): ...new function.
6513 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6514 (make_pass_remove_cgraph_callee_edges): ...new function.
6515 (pass_build_cgraph_edges): Replace declaration with that of...
6516 (make_pass_build_cgraph_edges): ...new function.
6517 (pass_local_pure_const): Replace declaration with that of...
6518 (make_pass_local_pure_const): ...new function.
6519 (pass_tracer): Replace declaration with that of...
6520 (make_pass_tracer): ...new function.
6521 (pass_warn_unused_result): Replace declaration with that of...
6522 (make_pass_warn_unused_result): ...new function.
6523 (pass_diagnose_tm_blocks): Replace declaration with that of...
6524 (make_pass_diagnose_tm_blocks): ...new function.
6525 (pass_lower_tm): Replace declaration with that of...
6526 (make_pass_lower_tm): ...new function.
6527 (pass_tm_init): Replace declaration with that of...
6528 (make_pass_tm_init): ...new function.
6529 (pass_tm_mark): Replace declaration with that of...
6530 (make_pass_tm_mark): ...new function.
6531 (pass_tm_memopt): Replace declaration with that of...
6532 (make_pass_tm_memopt): ...new function.
6533 (pass_tm_edges): Replace declaration with that of...
6534 (make_pass_tm_edges): ...new function.
6535 (pass_split_functions): Replace declaration with that of...
6536 (make_pass_split_functions): ...new function.
6537 (pass_feedback_split_functions): Replace declaration with that of...
6538 (make_pass_feedback_split_functions): ...new function.
6539 (pass_strength_reduction): Replace declaration with that of...
6540 (make_pass_strength_reduction): ...new function.
6541 (pass_ipa_lower_emutls): Replace declaration with that of...
6542 (make_pass_ipa_lower_emutls): ...new function.
6543 (pass_ipa_function_and_variable_visibility): Replace declaration with
6545 (make_pass_ipa_function_and_variable_visibility): ...new function.
6546 (pass_ipa_tree_profile): Replace declaration with that of...
6547 (make_pass_ipa_tree_profile): ...new function.
6548 (pass_early_local_passes): Replace declaration with that of...
6549 (make_pass_early_local_passes): ...new function.
6550 (pass_ipa_whole_program_visibility): Replace declaration with that
6552 (make_pass_ipa_whole_program_visibility): ...new function.
6553 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6554 (make_pass_ipa_lto_gimple_out): ...new function.
6555 (pass_ipa_increase_alignment): Replace declaration with that of...
6556 (make_pass_ipa_increase_alignment): ...new function.
6557 (pass_ipa_inline): Replace declaration with that of...
6558 (make_pass_ipa_inline): ...new function.
6559 (pass_ipa_free_lang_data): Replace declaration with that of...
6560 (make_pass_ipa_free_lang_data): ...new function.
6561 (pass_ipa_free_inline_summary): Replace declaration with that of...
6562 (make_pass_ipa_free_inline_summary): ...new function.
6563 (pass_ipa_cp): Replace declaration with that of...
6564 (make_pass_ipa_cp): ...new function.
6565 (pass_ipa_reference): Replace declaration with that of...
6566 (make_pass_ipa_reference): ...new function.
6567 (pass_ipa_pure_const): Replace declaration with that of...
6568 (make_pass_ipa_pure_const): ...new function.
6569 (pass_ipa_pta): Replace declaration with that of...
6570 (make_pass_ipa_pta): ...new function.
6571 (pass_ipa_lto_finish_out): Replace declaration with that of...
6572 (make_pass_ipa_lto_finish_out): ...new function.
6573 (pass_ipa_tm): Replace declaration with that of...
6574 (make_pass_ipa_tm): ...new function.
6575 (pass_ipa_profile): Replace declaration with that of...
6576 (make_pass_ipa_profile): ...new function.
6577 (pass_ipa_cdtor_merge): Replace declaration with that of...
6578 (make_pass_ipa_cdtor_merge): ...new function.
6579 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6581 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6582 (pass_init_datastructures): Replace declaration with that of...
6583 (make_pass_init_datastructures): ...new function.
6584 (pass_fixup_cfg): Replace declaration with that of...
6585 (make_pass_fixup_cfg): ...new function.
6586 (pass_expand): Replace declaration with that of...
6587 (make_pass_expand): ...new function.
6588 (pass_instantiate_virtual_regs): Replace declaration with that of...
6589 (make_pass_instantiate_virtual_regs): ...new function.
6590 (pass_rtl_fwprop): Replace declaration with that of...
6591 (make_pass_rtl_fwprop): ...new function.
6592 (pass_rtl_fwprop_addr): Replace declaration with that of...
6593 (make_pass_rtl_fwprop_addr): ...new function.
6594 (pass_jump): Replace declaration with that of...
6595 (make_pass_jump): ...new function.
6596 (pass_jump2): Replace declaration with that of...
6597 (make_pass_jump2): ...new function.
6598 (pass_lower_subreg): Replace declaration with that of...
6599 (make_pass_lower_subreg): ...new function.
6600 (pass_cse): Replace declaration with that of...
6601 (make_pass_cse): ...new function.
6602 (pass_fast_rtl_dce): Replace declaration with that of...
6603 (make_pass_fast_rtl_dce): ...new function.
6604 (pass_ud_rtl_dce): Replace declaration with that of...
6605 (make_pass_ud_rtl_dce): ...new function.
6606 (pass_rtl_dce): Replace declaration with that of...
6607 (make_pass_rtl_dce): ...new function.
6608 (pass_rtl_dse1): Replace declaration with that of...
6609 (make_pass_rtl_dse1): ...new function.
6610 (pass_rtl_dse2): Replace declaration with that of...
6611 (make_pass_rtl_dse2): ...new function.
6612 (pass_rtl_dse3): Replace declaration with that of...
6613 (make_pass_rtl_dse3): ...new function.
6614 (pass_rtl_cprop): Replace declaration with that of...
6615 (make_pass_rtl_cprop): ...new function.
6616 (pass_rtl_pre): Replace declaration with that of...
6617 (make_pass_rtl_pre): ...new function.
6618 (pass_rtl_hoist): Replace declaration with that of...
6619 (make_pass_rtl_hoist): ...new function.
6620 (pass_rtl_store_motion): Replace declaration with that of...
6621 (make_pass_rtl_store_motion): ...new function.
6622 (pass_cse_after_global_opts): Replace declaration with that of...
6623 (make_pass_cse_after_global_opts): ...new function.
6624 (pass_rtl_ifcvt): Replace declaration with that of...
6625 (make_pass_rtl_ifcvt): ...new function.
6626 (pass_into_cfg_layout_mode): Replace declaration with that of...
6627 (make_pass_into_cfg_layout_mode): ...new function.
6628 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6629 (make_pass_outof_cfg_layout_mode): ...new function.
6630 (pass_loop2): Replace declaration with that of...
6631 (make_pass_loop2): ...new function.
6632 (pass_rtl_loop_init): Replace declaration with that of...
6633 (make_pass_rtl_loop_init): ...new function.
6634 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6635 (make_pass_rtl_move_loop_invariants): ...new function.
6636 (pass_rtl_unswitch): Replace declaration with that of...
6637 (make_pass_rtl_unswitch): ...new function.
6638 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6639 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6640 (pass_rtl_doloop): Replace declaration with that of...
6641 (make_pass_rtl_doloop): ...new function.
6642 (pass_rtl_loop_done): Replace declaration with that of...
6643 (make_pass_rtl_loop_done): ...new function.
6644 (pass_web): Replace declaration with that of...
6645 (make_pass_web): ...new function.
6646 (pass_cse2): Replace declaration with that of...
6647 (make_pass_cse2): ...new function.
6648 (pass_df_initialize_opt): Replace declaration with that of...
6649 (make_pass_df_initialize_opt): ...new function.
6650 (pass_df_initialize_no_opt): Replace declaration with that of...
6651 (make_pass_df_initialize_no_opt): ...new function.
6652 (pass_reginfo_init): Replace declaration with that of...
6653 (make_pass_reginfo_init): ...new function.
6654 (pass_inc_dec): Replace declaration with that of...
6655 (make_pass_inc_dec): ...new function.
6656 (pass_stack_ptr_mod): Replace declaration with that of...
6657 (make_pass_stack_ptr_mod): ...new function.
6658 (pass_initialize_regs): Replace declaration with that of...
6659 (make_pass_initialize_regs): ...new function.
6660 (pass_combine): Replace declaration with that of...
6661 (make_pass_combine): ...new function.
6662 (pass_if_after_combine): Replace declaration with that of...
6663 (make_pass_if_after_combine): ...new function.
6664 (pass_ree): Replace declaration with that of...
6665 (make_pass_ree): ...new function.
6666 (pass_partition_blocks): Replace declaration with that of...
6667 (make_pass_partition_blocks): ...new function.
6668 (pass_match_asm_constraints): Replace declaration with that of...
6669 (make_pass_match_asm_constraints): ...new function.
6670 (pass_regmove): Replace declaration with that of...
6671 (make_pass_regmove): ...new function.
6672 (pass_split_all_insns): Replace declaration with that of...
6673 (make_pass_split_all_insns): ...new function.
6674 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6675 (make_pass_fast_rtl_byte_dce): ...new function.
6676 (pass_lower_subreg2): Replace declaration with that of...
6677 (make_pass_lower_subreg2): ...new function.
6678 (pass_mode_switching): Replace declaration with that of...
6679 (make_pass_mode_switching): ...new function.
6680 (pass_sms): Replace declaration with that of...
6681 (make_pass_sms): ...new function.
6682 (pass_sched): Replace declaration with that of...
6683 (make_pass_sched): ...new function.
6684 (pass_ira): Replace declaration with that of...
6685 (make_pass_ira): ...new function.
6686 (pass_reload): Replace declaration with that of...
6687 (make_pass_reload): ...new function.
6688 (pass_clean_state): Replace declaration with that of...
6689 (make_pass_clean_state): ...new function.
6690 (pass_branch_prob): Replace declaration with that of...
6691 (make_pass_branch_prob): ...new function.
6692 (pass_value_profile_transformations): Replace declaration with that
6694 (make_pass_value_profile_transformations): ...new function.
6695 (pass_postreload_cse): Replace declaration with that of...
6696 (make_pass_postreload_cse): ...new function.
6697 (pass_gcse2): Replace declaration with that of...
6698 (make_pass_gcse2): ...new function.
6699 (pass_split_after_reload): Replace declaration with that of...
6700 (make_pass_split_after_reload): ...new function.
6701 (pass_branch_target_load_optimize1): Replace declaration with that
6703 (make_pass_branch_target_load_optimize1): ...new function.
6704 (pass_thread_prologue_and_epilogue): Replace declaration with that
6706 (make_pass_thread_prologue_and_epilogue): ...new function.
6707 (pass_stack_adjustments): Replace declaration with that of...
6708 (make_pass_stack_adjustments): ...new function.
6709 (pass_peephole2): Replace declaration with that of...
6710 (make_pass_peephole2): ...new function.
6711 (pass_if_after_reload): Replace declaration with that of...
6712 (make_pass_if_after_reload): ...new function.
6713 (pass_regrename): Replace declaration with that of...
6714 (make_pass_regrename): ...new function.
6715 (pass_cprop_hardreg): Replace declaration with that of...
6716 (make_pass_cprop_hardreg): ...new function.
6717 (pass_reorder_blocks): Replace declaration with that of...
6718 (make_pass_reorder_blocks): ...new function.
6719 (pass_branch_target_load_optimize2): Replace declaration with that
6721 (make_pass_branch_target_load_optimize2): ...new function.
6722 (pass_leaf_regs): Replace declaration with that of...
6723 (make_pass_leaf_regs): ...new function.
6724 (pass_split_before_sched2): Replace declaration with that of...
6725 (make_pass_split_before_sched2): ...new function.
6726 (pass_compare_elim_after_reload): Replace declaration with that of...
6727 (make_pass_compare_elim_after_reload): ...new function.
6728 (pass_sched2): Replace declaration with that of...
6729 (make_pass_sched2): ...new function.
6730 (pass_stack_regs): Replace declaration with that of...
6731 (make_pass_stack_regs): ...new function.
6732 (pass_stack_regs_run): Replace declaration with that of...
6733 (make_pass_stack_regs_run): ...new function.
6734 (pass_df_finish): Replace declaration with that of...
6735 (make_pass_df_finish): ...new function.
6736 (pass_compute_alignments): Replace declaration with that of...
6737 (make_pass_compute_alignments): ...new function.
6738 (pass_duplicate_computed_gotos): Replace declaration with that of...
6739 (make_pass_duplicate_computed_gotos): ...new function.
6740 (pass_variable_tracking): Replace declaration with that of...
6741 (make_pass_variable_tracking): ...new function.
6742 (pass_free_cfg): Replace declaration with that of...
6743 (make_pass_free_cfg): ...new function.
6744 (pass_machine_reorg): Replace declaration with that of...
6745 (make_pass_machine_reorg): ...new function.
6746 (pass_cleanup_barriers): Replace declaration with that of...
6747 (make_pass_cleanup_barriers): ...new function.
6748 (pass_delay_slots): Replace declaration with that of...
6749 (make_pass_delay_slots): ...new function.
6750 (pass_split_for_shorten_branches): Replace declaration with that of...
6751 (make_pass_split_for_shorten_branches): ...new function.
6752 (pass_split_before_regstack): Replace declaration with that of...
6753 (make_pass_split_before_regstack): ...new function.
6754 (pass_convert_to_eh_region_ranges): Replace declaration with that
6756 (make_pass_convert_to_eh_region_ranges): ...new function.
6757 (pass_shorten_branches): Replace declaration with that of...
6758 (make_pass_shorten_branches): ...new function.
6759 (pass_set_nothrow_function_flags): Replace declaration with that of...
6760 (make_pass_set_nothrow_function_flags): ...new function.
6761 (pass_dwarf2_frame): Replace declaration with that of...
6762 (make_pass_dwarf2_frame): ...new function.
6763 (pass_final): Replace declaration with that of...
6764 (make_pass_final): ...new function.
6765 (pass_rtl_seqabstr): Replace declaration with that of...
6766 (make_pass_rtl_seqabstr): ...new function.
6767 (pass_release_ssa_names): Replace declaration with that of...
6768 (make_pass_release_ssa_names): ...new function.
6769 (pass_early_inline): Replace declaration with that of...
6770 (make_pass_early_inline): ...new function.
6771 (pass_inline_parameters): Replace declaration with that of...
6772 (make_pass_inline_parameters): ...new function.
6773 (pass_update_address_taken): Replace declaration with that of...
6774 (make_pass_update_address_taken): ...new function.
6775 (pass_convert_switch): Replace declaration with that of...
6776 (make_pass_convert_switch): ...new function.
6777 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6778 to a subclass of simple_ipa_opt_pass along with...
6779 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6780 (make_pass_ipa_tree_profile): ...new function.
6781 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6782 subclass of gimple_opt_pass along with...
6783 (pass_data_sra_early): ...new pass_data instance and...
6784 (make_pass_sra_early): ...new function.
6785 (pass_sra): Convert from a global struct to a subclass of
6786 gimple_opt_pass along with...
6787 (pass_data_sra): ...new pass_data instance and...
6788 (make_pass_sra): ...new function.
6789 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6790 gimple_opt_pass along with...
6791 (pass_data_early_ipa_sra): ...new pass_data instance and...
6792 (make_pass_early_ipa_sra): ...new function.
6793 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6794 subclass of gimple_opt_pass along with...
6795 (pass_data_ccp): ...new pass_data instance and...
6796 (make_pass_ccp): ...new function.
6797 (pass_fold_builtins): Convert from a global struct to a subclass of
6798 gimple_opt_pass along with...
6799 (pass_data_fold_builtins): ...new pass_data instance and...
6800 (make_pass_fold_builtins): ...new function.
6801 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6802 subclass of gimple_opt_pass along with...
6803 (pass_data_copy_prop): ...new pass_data instance and...
6804 (make_pass_copy_prop): ...new function.
6805 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6806 global struct to a subclass of gimple_opt_pass along with...
6807 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6808 (make_pass_rename_ssa_copies): ...new function.
6809 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6810 subclass of gimple_opt_pass along with...
6811 (pass_data_dce): ...new pass_data instance and...
6812 (make_pass_dce): ...new function.
6813 (pass_dce_loop): Convert from a global struct to a subclass of
6814 gimple_opt_pass along with...
6815 (pass_data_dce_loop): ...new pass_data instance and...
6816 (make_pass_dce_loop): ...new function.
6817 (pass_cd_dce): Convert from a global struct to a subclass of
6818 gimple_opt_pass along with...
6819 (pass_data_cd_dce): ...new pass_data instance and...
6820 (make_pass_cd_dce): ...new function.
6821 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6822 subclass of gimple_opt_pass along with...
6823 (pass_data_dominator): ...new pass_data instance and...
6824 (make_pass_dominator): ...new function.
6825 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6826 gimple_opt_pass along with...
6827 (pass_data_phi_only_cprop): ...new pass_data instance and...
6828 (make_pass_phi_only_cprop): ...new function.
6829 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6830 subclass of gimple_opt_pass along with...
6831 (pass_data_dse): ...new pass_data instance and...
6832 (make_pass_dse): ...new function.
6833 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6834 a subclass of gimple_opt_pass along with...
6835 (pass_data_forwprop): ...new pass_data instance and...
6836 (make_pass_forwprop): ...new function.
6837 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6838 struct to a subclass of gimple_opt_pass along with...
6839 (pass_data_tree_ifcombine): ...new pass_data instance and...
6840 (make_pass_tree_ifcombine): ...new function.
6841 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6842 subclass of gimple_opt_pass along with...
6843 (pass_data_ch): ...new pass_data instance and...
6844 (make_pass_ch): ...new function.
6845 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6846 subclass of gimple_opt_pass along with...
6847 (pass_data_tree_loop): ...new pass_data instance and...
6848 (make_pass_tree_loop): ...new function.
6849 (pass_tree_loop_init): Convert from a global struct to a subclass of
6850 gimple_opt_pass along with...
6851 (pass_data_tree_loop_init): ...new pass_data instance and...
6852 (make_pass_tree_loop_init): ...new function.
6853 (pass_lim): Convert from a global struct to a subclass of
6854 gimple_opt_pass along with...
6855 (pass_data_lim): ...new pass_data instance and...
6856 (make_pass_lim): ...new function.
6857 (pass_tree_unswitch): Convert from a global struct to a subclass of
6858 gimple_opt_pass along with...
6859 (pass_data_tree_unswitch): ...new pass_data instance and...
6860 (make_pass_tree_unswitch): ...new function.
6861 (pass_predcom): Convert from a global struct to a subclass of
6862 gimple_opt_pass along with...
6863 (pass_data_predcom): ...new pass_data instance and...
6864 (make_pass_predcom): ...new function.
6865 (pass_vectorize): Convert from a global struct to a subclass of
6866 gimple_opt_pass along with...
6867 (pass_data_vectorize): ...new pass_data instance and...
6868 (make_pass_vectorize): ...new function.
6869 (pass_graphite): Convert from a global struct to a subclass of
6870 gimple_opt_pass along with...
6871 (pass_data_graphite): ...new pass_data instance and...
6872 (make_pass_graphite): ...new function.
6873 (pass_graphite_transforms): Convert from a global struct to a subclass
6874 of gimple_opt_pass along with...
6875 (pass_data_graphite_transforms): ...new pass_data instance and...
6876 (make_pass_graphite_transforms): ...new function.
6877 (pass_check_data_deps): Convert from a global struct to a subclass of
6878 gimple_opt_pass along with...
6879 (pass_data_check_data_deps): ...new pass_data instance and...
6880 (make_pass_check_data_deps): ...new function.
6881 (pass_iv_canon): Convert from a global struct to a subclass of
6882 gimple_opt_pass along with...
6883 (pass_data_iv_canon): ...new pass_data instance and...
6884 (make_pass_iv_canon): ...new function.
6885 (pass_scev_cprop): Convert from a global struct to a subclass of
6886 gimple_opt_pass along with...
6887 (pass_data_scev_cprop): ...new pass_data instance and...
6888 (make_pass_scev_cprop): ...new function.
6889 (pass_record_bounds): Convert from a global struct to a subclass of
6890 gimple_opt_pass along with...
6891 (pass_data_record_bounds): ...new pass_data instance and...
6892 (make_pass_record_bounds): ...new function.
6893 (pass_complete_unroll): Convert from a global struct to a subclass of
6894 gimple_opt_pass along with...
6895 (pass_data_complete_unroll): ...new pass_data instance and...
6896 (make_pass_complete_unroll): ...new function.
6897 (pass_complete_unrolli): Convert from a global struct to a subclass of
6898 gimple_opt_pass along with...
6899 (pass_data_complete_unrolli): ...new pass_data instance and...
6900 (make_pass_complete_unrolli): ...new function.
6901 (pass_parallelize_loops): Convert from a global struct to a subclass
6902 of gimple_opt_pass along with...
6903 (pass_data_parallelize_loops): ...new pass_data instance and...
6904 (make_pass_parallelize_loops): ...new function.
6905 (pass_loop_prefetch): Convert from a global struct to a subclass of
6906 gimple_opt_pass along with...
6907 (pass_data_loop_prefetch): ...new pass_data instance and...
6908 (make_pass_loop_prefetch): ...new function.
6909 (pass_iv_optimize): Convert from a global struct to a subclass of
6910 gimple_opt_pass along with...
6911 (pass_data_iv_optimize): ...new pass_data instance and...
6912 (make_pass_iv_optimize): ...new function.
6913 (pass_tree_loop_done): Convert from a global struct to a subclass of
6914 gimple_opt_pass along with...
6915 (pass_data_tree_loop_done): ...new pass_data instance and...
6916 (make_pass_tree_loop_done): ...new function.
6917 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6918 struct to a subclass of gimple_opt_pass along with...
6919 (pass_data_cse_reciprocals): ...new pass_data instance and...
6920 (make_pass_cse_reciprocals): ...new function.
6921 (pass_cse_sincos): Convert from a global struct to a subclass of
6922 gimple_opt_pass along with...
6923 (pass_data_cse_sincos): ...new pass_data instance and...
6924 (make_pass_cse_sincos): ...new function.
6925 (pass_optimize_bswap): Convert from a global struct to a subclass of
6926 gimple_opt_pass along with...
6927 (pass_data_optimize_bswap): ...new pass_data instance and...
6928 (make_pass_optimize_bswap): ...new function.
6929 (pass_optimize_widening_mul): Convert from a global struct to a
6930 subclass of gimple_opt_pass along with...
6931 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6932 (make_pass_optimize_widening_mul): ...new function.
6933 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6934 subclass of gimple_opt_pass along with...
6935 (pass_data_phiopt): ...new pass_data instance and...
6936 (make_pass_phiopt): ...new function.
6937 (pass_cselim): Convert from a global struct to a subclass of
6938 gimple_opt_pass along with...
6939 (pass_data_cselim): ...new pass_data instance and...
6940 (make_pass_cselim): ...new function.
6941 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6942 subclass of gimple_opt_pass along with...
6943 (pass_data_phiprop): ...new pass_data instance and...
6944 (make_pass_phiprop): ...new function.
6945 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6946 subclass of gimple_opt_pass along with...
6947 (pass_data_pre): ...new pass_data instance and...
6948 (make_pass_pre): ...new function.
6949 (pass_fre): Convert from a global struct to a subclass of
6950 gimple_opt_pass along with...
6951 (pass_data_fre): ...new pass_data instance and...
6952 (make_pass_fre): ...new function.
6953 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
6954 subclass of gimple_opt_pass along with...
6955 (pass_data_reassoc): ...new pass_data instance and...
6956 (make_pass_reassoc): ...new function.
6957 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
6958 subclass of gimple_opt_pass along with...
6959 (pass_data_sink_code): ...new pass_data instance and...
6960 (make_pass_sink_code): ...new function.
6961 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
6962 subclass of gimple_opt_pass along with...
6963 (pass_data_strlen): ...new pass_data instance and...
6964 (make_pass_strlen): ...new function.
6965 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
6966 struct to a subclass of gimple_opt_pass along with...
6967 (pass_data_build_alias): ...new pass_data instance and...
6968 (make_pass_build_alias): ...new function.
6969 (pass_build_ealias): Convert from a global struct to a subclass of
6970 gimple_opt_pass along with...
6971 (pass_data_build_ealias): ...new pass_data instance and...
6972 (make_pass_build_ealias): ...new function.
6973 (pass_ipa_pta): Convert from a global struct to a subclass of
6974 simple_ipa_opt_pass along with...
6975 (pass_data_ipa_pta): ...new pass_data instance and...
6976 (make_pass_ipa_pta): ...new function.
6977 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
6978 subclass of gimple_opt_pass along with...
6979 (pass_data_uncprop): ...new pass_data instance and...
6980 (make_pass_uncprop): ...new function.
6981 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
6982 global struct to a subclass of gimple_opt_pass along with...
6983 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
6984 (make_pass_late_warn_uninitialized): ...new function.
6985 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
6986 to a subclass of gimple_opt_pass along with...
6987 (pass_data_init_datastructures): ...new pass_data instance and...
6988 (make_pass_init_datastructures): ...new function.
6989 (pass_early_warn_uninitialized): Convert from a global struct to a
6990 subclass of gimple_opt_pass along with...
6991 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
6992 (make_pass_early_warn_uninitialized): ...new function.
6993 (pass_update_address_taken): Convert from a global struct to a
6994 subclass of gimple_opt_pass along with...
6995 (pass_data_update_address_taken): ...new pass_data instance and...
6996 (make_pass_update_address_taken): ...new function.
6997 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
6998 struct to a subclass of gimple_opt_pass along with...
6999 (pass_data_release_ssa_names): ...new pass_data instance and...
7000 (make_pass_release_ssa_names): ...new function.
7001 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7002 subclass of gimple_opt_pass along with...
7003 (pass_data_stdarg): ...new pass_data instance and...
7004 (make_pass_stdarg): ...new function.
7005 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7006 global struct to a subclass of gimple_opt_pass along with...
7007 (pass_data_convert_switch): ...new pass_data instance and...
7008 (make_pass_convert_switch): ...new function.
7009 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7010 to a subclass of gimple_opt_pass along with...
7011 (pass_data_tail_recursion): ...new pass_data instance and...
7012 (make_pass_tail_recursion): ...new function.
7013 (pass_tail_calls): Convert from a global struct to a subclass of
7014 gimple_opt_pass along with...
7015 (pass_data_tail_calls): ...new pass_data instance and...
7016 (make_pass_tail_calls): ...new function.
7017 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7018 struct to a subclass of gimple_opt_pass along with...
7019 (pass_data_lower_vector): ...new pass_data instance and...
7020 (make_pass_lower_vector): ...new function.
7021 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7022 gimple_opt_pass along with...
7023 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7024 (make_pass_lower_vector_ssa): ...new function.
7025 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7026 to a subclass of gimple_opt_pass along with...
7027 (pass_data_slp_vectorize): ...new pass_data instance and...
7028 (make_pass_slp_vectorize): ...new function.
7029 (pass_ipa_increase_alignment): Convert from a global struct to a
7030 subclass of simple_ipa_opt_pass along with...
7031 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7032 (make_pass_ipa_increase_alignment): ...new function.
7033 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7034 gimple_opt_pass along with...
7035 (pass_data_vrp): ...new pass_data instance and...
7036 (make_pass_vrp): ...new function.
7037 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7038 subclass of simple_ipa_opt_pass along with...
7039 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7040 (make_pass_ipa_free_lang_data): ...new function.
7041 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7042 gimple_opt_pass along with...
7043 (pass_data_tsan): ...new pass_data instance and...
7044 (make_pass_tsan): ...new function.
7045 (pass_tsan_O0): Convert from a global struct to a subclass of
7046 gimple_opt_pass along with...
7047 (pass_data_tsan_O0): ...new pass_data instance and...
7048 (make_pass_tsan_O0): ...new function.
7049 * var-tracking.c (pass_variable_tracking): Convert from a global
7050 struct to a subclass of rtl_opt_pass along with...
7051 (pass_data_variable_tracking): ...new pass_data instance and...
7052 (make_pass_variable_tracking): ...new function.
7053 * web.c (pass_web): Convert from a global struct to a subclass of
7054 rtl_opt_pass along with...
7055 (pass_data_web): ...new pass_data instance and...
7056 (make_pass_web): ...new function.
7057 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7058 declaration with that of...
7059 (make_pass_mode_switch_use): ...new function.
7060 (pass_resolve_sw_modes): Replace declaration with that of...
7061 (make_pass_resolve_sw_modes): ...new function.
7062 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7063 from a global struct to a subclass of rtl_opt_pass along with...
7064 (pass_data_mode_switch_use): ...new pass_data instance and...
7065 (make_pass_mode_switch_use): ...new function.
7066 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7067 from a global struct to a subclass of rtl_opt_pass along with...
7068 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7069 (make_pass_resolve_sw_modes): ...new function.
7070 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7071 struct to a subclass of rtl_opt_pass along with...
7072 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7073 (make_pass_insert_vzeroupper): ...new function.
7074 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7075 global struct to a subclass of rtl_opt_pass along with...
7076 (pass_data_work_around_errata): ...new pass_data instance and...
7077 (make_pass_work_around_errata): ...new function.
7078 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7079 struct to a subclass of rtl_opt_pass along with...
7080 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7081 (make_pass_mips_machine_reorg2): ...new function.
7083 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7085 * passes.c (pass_manager::operator new): New.
7087 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7089 Handwritten part of conversion of passes to C++ classes.
7091 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7092 (toplev.o): Add dep on PASS_MANAGER_H.
7093 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7094 of early local pases to reflect this moving from a global to a
7095 member of gcc::pass_manager.
7096 (cgraph_add_new_function): Likewise.
7097 * lto-cgraph.c (lto_output_node): Update for conversion of
7098 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7099 * passes.c (opt_pass::clone): New.
7100 (opt_pass::gate): New.
7101 (opt_pass::execute): New.
7102 (opt_pass::opt_pass): New.
7103 (pass_manager::execute_early_local_passes): New.
7104 (pass_manager::execute_pass_mode_switching): new.
7105 (finish_optimization_passes): Convert to...
7106 (pass_manager::finish_optimization_passes): ...this.
7107 (finish_optimization_passes): Update for conversion of passes to
7109 (register_dump_files_1): Use has_gate since we cannot portably
7110 check a vtable entry against NULL.
7111 (dump_one_pass): Likewise.
7112 (ipa_write_summaries_2): Likewise.
7113 (ipa_write_optimization_summaries_1): Likewise.
7114 (ipa_read_summaries_1): Likewise.
7115 (ipa_read_optimization_summaries_1): Likewise.
7116 (execute_ipa_stmt_fixups): Likewise.
7117 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7118 pass-creation functions rather than wiring up globals, and
7119 storing the results in fields of pass_manager generated using
7121 (pass_manager::dump_profile_report): Update for conversion of
7122 passes to C++ classes.
7123 (pass_manager::execute_ipa_summary_passes): Likewise.
7124 (execute_one_ipa_transform_pass): Likewise.
7125 (execute_one_pass): Use has_gate and has_execute since we cannot
7126 portably check a vtable entry against NULL.
7127 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7128 (pass_manager): Use pass-instances.def to add fields for the
7129 various pass instances.
7130 * toplev.c (finalize): Update for move of
7131 finish_optimization_passes to a method of gcc::pass_manager.
7132 * toplev.h (finish_optimization_passes): Move to method of class
7134 * tree-pass.h (struct pass_data): New.
7135 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7136 (opt_pass::gate): Convert to virtual function.
7137 (opt_pass::~opt_pass): New.
7138 (opt_pass::clone): New.
7139 (opt_pass::execute): Convert to virtual function.
7140 (opt_pass::opt_pass): New.
7141 (opt_pass::ctxt_): new.
7142 (gimple_opt_pass): Convert to subclass of opt_pass.
7143 (gimple_opt_pass::gimple_opt_pass): New.
7144 (rtl_opt_pass): Convert to subclass of opt_pass.
7145 (rtl_opt_pass::rtl_opt_pass): New.
7146 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7147 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7148 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7149 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7150 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7151 invocation of pass_mode_switching to reflect this moving from a
7152 global to a member of gcc::pass_manager.
7153 (ix86_option_override): Rework how pass_insert_vzeroupper is
7154 added to the pass_manager to reflect autogenerated changes.
7155 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7157 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7159 PR rtl-optimization/57708
7160 * recog.c (peep2_find_free_register): Validate all regs in a
7163 2013-08-05 Jan Hubicka <jh@suse.cz>
7166 * cgraph.c (verify_cgraph_node): Accept local flags from other
7168 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7169 (function_and_variable_visibility): Likewise.
7170 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7172 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7174 * graph.c (init_graph_slim_pretty_print): Remove.
7175 (print_graph_cfg): Do not call it. Use local pretty printer.
7176 (start_graph_dump): Likewise.
7178 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7180 * gimple-pretty-print.c (buffer): Remove.
7181 (initialized): Likewise.
7182 (maybe_init_pretty_print): Likewise.
7183 (print_gimple_stmt): Do not call it. Use non-static local
7184 pretty_printer variable.
7185 (print_gimple_expr): Likewise.
7186 (print_gimple_seq): Likewise.
7187 (gimple_dump_bb): Likewise.
7189 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7191 * asan.c (asan_pp): Remove.
7192 (asan_pp_initialized): Likewise.
7193 (asan_pp_initialize): Likewise.
7194 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7195 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7196 (asan_add_global): Likewise.
7198 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7200 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7201 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7202 * pretty-print.c (pp_formatted_text_data): Likewise.
7203 (pp_indent): Rename from pp_base_indent.
7204 (pp_format): Rename from pp_base_format.
7205 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7206 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7207 (pp_flush): Rename from pp_base_flush.
7208 (pp_set_line_maximum_length): Rename from
7209 pp_base_set_line_maximum_length.
7210 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7211 (pp_set_prefix): Rename from pp_base_set_prefix.
7212 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7213 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7214 (pp_append_text): Rename from pp_base_append_text.
7215 (pp_formatted_text): Rename from pp_base_formatted_text.
7216 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7217 (pp_remaining_character_count_for_line): Rename from
7218 pp_base_remaining_character_count_for_line.
7219 (pp_newline): Rename from pp_base_newline.
7220 (pp_character): Rename from pp_base_character.
7221 (pp_string): Rename from pp_base_string.
7222 (pp_maybe_space): Rename from pp_base_maybe_space.
7223 * asan.c (asan_pp_string): Adjust.
7224 (asan_emit_stack_protection): Likewise.
7225 (asan_add_global): Likewise.
7226 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7227 * tree-mudflap.c (mf_varname_tree): Likewise.
7228 * tree-pretty-print.c (pp_tree_identifier): Rename from
7229 pp_base_tree_identifier.
7230 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7231 Declare as function.
7233 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7235 * pretty-print.h (pp_bar_bar): New.
7236 (pp_ampersand_ampersand): Likewise.
7237 (pp_less_equal): Likewise.
7238 (pp_greater_equal): Likewise.
7239 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7240 printer functions instead of pp_string or operators and punctuators.
7241 (dump_gimple_call): Likewise.
7242 (dump_gimple_omp_for): Likewise.
7243 (dump_gimple_transaction): Likewise.
7244 (dump_gimple_phi): Likewise.
7245 (pp_gimple_stmt_1): Likewise.
7246 * sched-vis.c (print_insn): Likewise.
7247 * tree-mudflap.c (mf_varname_tree): Likewise.
7248 * tree-pretty-print.c (dump_block_node): Likewise.
7249 (dump_generic_node): Likewise.
7251 2013-08-02 Jan Hubicka <jh@suse.cz>
7253 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7255 * lto-streamer-out.c (tree_is_indexable): Results decls and
7256 parm decls are not indexable.
7257 (DFS_write_tree_body): Do not follow args and results.
7258 (hash_tree): Likewise.
7259 (output_functions): Rearrange so struct function is needed
7260 only when real body is output; be able to also ouptut abstract
7261 functions; output DECL_ARGUMENTS and DECL_RESULT.
7262 (lto_output): When not in WPA, ale store abstract functions.
7263 (write_symbol): Do not care about RESULT_DECL.
7264 (output_symbol_p): Handle correctly sbtract decls.
7265 * lto-streamer-in.c (input_function): Rearrange so struct
7266 function can be NULL at entry; allow streaming of
7267 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7268 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7269 sanity check during LTO.
7270 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7271 RESULT_DECl and DECL_ARGUMENTS.
7272 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7275 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7277 * pretty-print.h (pp_underscore): New.
7279 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7280 printer functions instead of pp_character.
7281 (dump_binary_rhs): Likewise.
7282 (dump_ternary_rhs): Likewise.
7283 (dump_gimple_call_args): Likewise.
7284 (pp_points_to_solution): Likewise.
7285 (dump_gimple_call): Likewise.
7286 (dump_gimple_switch): Likewise.
7287 (dump_gimple_cond): Likewise.
7288 (dump_gimple_bind): Likewise.
7289 (dump_gimple_try): Likewise.
7290 (dump_gimple_omp_for): Likewise.
7291 (dump_gimple_omp_continue): Likewise.
7292 (dump_gimple_omp_single): Likewise.
7293 (dump_gimple_omp_sections): Likewise.
7294 (dump_gimple_omp_block): Likewise.
7295 (dump_gimple_omp_critical): Likewise.
7296 (dump_gimple_transaction): Likewise.
7297 (dump_gimple_asm): Likewise.
7298 (dump_gimple_phi): Likewise.
7299 (dump_gimple_omp_parallel): Likewise.
7300 (dump_gimple_omp_task): Likewise.
7301 (dump_gimple_omp_atomic_load): Likewise.
7302 (dump_gimple_omp_atomic_store): Likewise.
7303 (dump_gimple_mem_ops): Likewise.
7304 (pp_gimple_stmt_1): Likewise.
7305 (pp_cfg_jump): Likewise.
7306 (dump_implicit_edges): Likewise.
7307 (gimple_dump_bb_for_graph): Likewise.
7308 * graph.c (draw_cfg_node): Likewise.
7309 * langhooks.c (lhd_print_error_function): Likewise.
7310 * sched-vis.c (print_exp): Likewise.
7311 (print_value): Likewise.
7312 (print_pattern): Likewise.
7313 (print_insn): Likewise.
7314 (rtl_dump_bb_for_graph): Likewise.
7315 * tree-pretty-print.c (dump_function_declaration): Likewise.
7316 (dump_array_domain): Likewise.
7317 (dump_omp_clause): Likewise.
7318 (dump_location): Likewise.
7319 (dump_generic_node): Likewise.
7320 (print_struct_decl): Likewise.
7321 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7323 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7325 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7326 candidate table when replacing a candidate statement.
7327 (replace_rhs_if_not_dup): Likewise.
7328 (replace_one_candidate): Likewise.
7330 2013-08-02 Jan Hubicka <jh@suse.cz>
7331 Martin Liska <marxin.liska@gmail.com>
7333 * cgraphunit.c (add_new_function): Fix logic when adding from
7335 (assemble_thunk): Rename to ...
7336 (expand_thunk); .. this one; export; get it working with
7337 general functions; make produced gimple valid.
7338 * cgraph.h (expand_thunk): Declare.
7340 2013-08-02 Jan Hubicka <jh@suse.cz>
7342 * ipa-cp.c (gather_context_independent_values): Use
7343 ipa_get_param_move_cost.
7344 (get_replacement_map): Remove PARAM; move parameter folding
7346 (create_specialized_node): Update.
7347 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7348 assert that we have gimple body; update move_cost.
7349 (count_formal_params): Assert that we have gimple body.
7350 (ipa_dump_param): New function.
7351 (ipa_alloc_node_params): Break out from ...
7352 (ipa_initialize_node_params): ... here.
7353 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7354 (ipa_write_node_info): Stream move costs.
7355 (ipa_read_node_info): Read move costs.
7356 (ipa_update_after_lto_read): Do not recompute node params.
7357 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7358 (ipa_get_param): Check we are not in WPA.
7359 (ipa_get_param_move_cost): New.
7360 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7361 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7362 parm numbers to be present.
7364 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7366 PR rtl-optimization/58048
7367 * lra-constraints.c (process_alt_operands): Don't check asm
7368 operand on register.
7370 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7372 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7373 the implied StoreLoad barrier for atomic operations if before.
7375 2013-08-02 Jan Hubicka <jh@suse.cz>
7376 Martin Liska <marxin.liska@gmail.com>
7378 * cgraph.c (cgraph_function_body_availability): Do not check
7380 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7381 symtab_node_availability): Declare.
7382 * ipa.c (can_replace_by_local_alias): New.
7383 (function_and_variable_visibility): Use it.
7384 * symtab.c (symtab_for_node_and_aliases,
7385 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7387 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7389 PR rtl-optimization/57963
7390 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7391 (lra_constraints): Use them.
7393 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7395 * config/arm/types.md (define_attr "type"): Add "load_acq"
7397 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7399 (cortex_a53_store1): Likewise.
7401 2013-08-01 Jan Hubicka <jh@suse.cz>
7403 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7404 partitions are not needed.
7406 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7408 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7409 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7410 MAYBE_NON_Q_CLASS_P where appropriate.
7412 2013-08-01 Jan Hubicka <jh@suse.cz>
7414 * cgraph.h (release_function_body): Declare.
7415 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7416 of unused delcarations.
7418 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7420 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7421 RTL form when subtracting a constant.
7423 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7425 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7426 Generate canonical plus rtx with negated immediate instead of minus
7428 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7430 2013-08-01 Jan Hubicka <jh@suse.cz>
7432 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7433 (cgraph_release_function_body): Likewise.
7434 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7435 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7436 to used_as_abstract_origin.
7437 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7438 symbols used as abstract origins.
7439 * cgraphunit.c (analyze_functions): Update.
7440 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7441 used_as_abstract_origin.
7442 * tree-inline.c (tree_function_versioning): Update
7443 used_as_abstract_origin; be ready for DECL_RESULT and
7444 DECL_ARGUMENTS to be NULL.
7446 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7447 for abstract functions.
7448 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7451 2013-08-01 Jan Hubicka <jh@suse.cz>
7453 * profile.c (compute_value_histograms): Fix thinko.
7455 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7457 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7458 aarch-common-protos.h to extra_headers.
7459 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7460 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7461 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7463 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7465 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7466 Include "../arm/types.md". Define "type" attribute for all patterns.
7467 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7470 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7472 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7473 to support power8 load fusion.
7474 (fusion_gpr_mem_load): Likewise.
7476 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7478 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7479 declarations for power8 load fusion.
7480 (emit_fusion_gpr_load): Likewise.
7482 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7483 tuning for power8, turn on fusion mode by default. Turn on sign
7484 extending fusion mode if normal fusion mode is on, and we are at
7486 (fusion_gpr_load_p): New function, return true if we can fuse an
7487 addis instruction with a dependent load to a GPR.
7488 (emit_fusion_gpr_load): Emit the instructions for power8 load
7491 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7492 (VSX load fusion peepholes): New peepholes to fuse together an
7493 addi instruction with a VSX load instruction.
7495 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7496 peepholes to fuse an addis instruction with a load to a GPR base
7497 register. If we are supporting sign extending fusions, convert
7498 sign extending loads to zero extending loads and add an explicit
7501 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7503 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7504 aarch-common-protos.h to extra_headers.
7505 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7506 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7507 (arm_early_store_addr_dep): Likewise.
7508 (arm_no_early_alu_shift_dep): Likewise.
7509 (arm_no_early_alu_shift_value_dep): Likewise.
7510 (arm_no_early_mul_dep): Likewise.
7511 (arm_no_early_store_addr_dep): Likewise.
7512 (arm_mac_accumulator_is_mul_result): Likewise.
7513 (arm_mac_accumulator_is_result): Likewise.
7514 * config/arm/aarch-common.c: ... here. New file.
7515 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7517 (arm_early_store_addr_dep): Likewise.
7518 (arm_no_early_alu_shift_dep): Likewise.
7519 (arm_no_early_alu_shift_value_dep): Likewise.
7520 (arm_no_early_mul_dep): Likewise.
7521 (arm_no_early_store_addr_dep): Likewise.
7522 (arm_mac_accumulator_is_mul_result): Likewise.
7523 (arm_mac_accumulator_is_result): Likewise.
7524 * config/arm/aarch-common-protos.h: ... here. New file.
7525 * config/arm/t-arm (aarch-common.o): Define.
7527 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7529 * config/arm/arm.md: Include new file "types.md".
7530 (define_attr "type"): Move from here to ...
7531 (define_attr "mul32"): Likewise.
7532 (define_attr "mul64"): Likewise.
7533 * config/arm/types.md: ... here. New file.
7535 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7537 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7538 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7540 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7542 * gen-pass-instances.awk: Fix offset of substr().
7544 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7546 * Makefile.in (pass-instances.def): New.
7547 (passes.o): Replace dependency on passes.def with one on
7550 * gen-pass-instances.awk: New.
7552 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7553 rather than passes.def, updating local definition of NEXT_PASS
7554 macro to add an extra NUM parameter (currently unused).
7556 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7558 * Makefile.in (PASS_MANAGER_H): New.
7559 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7560 (passes.o): Likewise.
7561 (statistics.o): Likewise.
7562 (cgraphunit.o): Likewise.
7563 (context.o): Depend on PASS_MANAGER_H.
7565 * pass_manager.h: New.
7567 * cgraphunit.c (cgraph_add_new_function): Update for moves
7568 of globals to fields of pass_manager.
7569 (analyze_function): Likewise.
7570 (expand_function): Likewise.
7571 (ipa_passes): Likewise.
7572 (compile): Likewise.
7574 * context.c (context::context): New.
7575 * context.h (context::context): New.
7576 (context::get_passes): New.
7577 (context::passes_): New.
7579 * lto-cgraph.c (input_node): Update for moves of globals to
7580 fields of pass_manager.
7582 * passes.c (all_passes): Remove, in favor of a field of the
7583 same name within the new class pass_manager.
7584 (all_small_ipa_passes): Likewise.
7585 (all_lowering_passes): Likewise.
7586 (all_regular_ipa_passes): Likewise.
7587 (all_late_ipa_passes): Likewise.
7588 (all_lto_gen_passes): Likewise.
7589 (passes_by_id): Likewise.
7590 (passes_by_id_size): Likewise.
7591 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7592 the new class pass_manager.
7593 (set_pass_for_id): Convert to...
7594 (pass_manager::set_pass_for_id): ...method.
7595 (get_pass_for_id): Convert to...
7596 (pass_manager::get_pass_for_id): ...method.
7597 (register_one_dump_file): Move body of implementation into...
7598 (pass_manager::register_one_dump_file): ...here.
7599 (register_dump_files_1): Convert to...
7600 (pass_manager::register_dump_files_1): ...method.
7601 (register_dump_files): Convert to...
7602 (pass_manager::register_dump_files): ...method.
7603 (create_pass_tab): Update for moves of globals to fields of
7605 (dump_passes): Move body of implementation into...
7606 (pass_manager::dump_passes): ...here.
7607 (register_pass): Move body of implementation into...
7608 (pass_manager::register_pass): ...here.
7609 (init_optimization_passes): Convert into...
7610 (pass_manager::pass_manager): ...constructor for new
7611 pass_manager class, and initialize the pass_lists array.
7612 (check_profile_consistency): Update for moves of globals to
7613 fields of pass_manager.
7614 (dump_profile_report): Move body of implementation into...
7615 (pass_manager::dump_profile_report): ...here.
7616 (ipa_write_summaries_1): Update for moves of pass lists from
7617 being globals to fields of pass_manager.
7618 (ipa_write_optimization_summaries): Likewise.
7619 (ipa_read_summaries): Likewise.
7620 (ipa_read_optimization_summaries): Likewise.
7621 (execute_all_ipa_stmt_fixups): Likewise.
7623 * statistics.c (statistics_fini): Update for moves of globals to
7624 fields of pass_manager.
7626 * toplev.c (general_init): Replace call to
7627 init_optimization_passes with construction of the pass_manager
7630 * tree-pass.h (all_passes): Remove, in favor of a field of the
7631 same name within the new class pass_manager.
7632 (all_small_ipa_passes): Likewise.
7633 (all_lowering_passes): Likewise.
7634 (all_regular_ipa_passes): Likewise.
7635 (all_lto_gen_passes): Likewise.
7636 (all_late_ipa_passes): Likewise.
7637 (passes_by_id): Likewise.
7638 (passes_by_id_size): Likewise.
7639 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7640 the new class pass_manager.
7641 (get_pass_for_id): Remove.
7643 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7645 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7648 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7650 * arm.md (mulhi3): New expand pattern.
7652 2013-07-30 Jan Hubicka <jh@suse.cz>
7653 Martin Liska <marxin.liska@gmail.com>
7655 * profile.c (compute_value_histograms): Do not ICE when
7656 there is mismatch only on some counters.
7658 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7660 PR rtl-optimization/57637
7661 * function.c (move_insn_for_shrink_wrap): Also check the
7662 GEN set of the LIVE problem for the liveness analysis
7663 if it exists, otherwise give up.
7665 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7667 PR tree-optimization/57993
7668 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7669 replaced statement in the candidate table.
7670 (phi_add_costs): Return infinite cost when the hidden basis does
7671 not dominate all phis on which the candidate is dependent.
7672 (replace_one_candidate): Record replaced statement in the
7675 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7677 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7678 (ashlv2si3): New expander.
7679 (*ashlv2si3_i): New define_insn_and_split.
7680 * predicates.md (float_operation): Allow patterns with three
7683 PR rtl-optimization/58021
7684 * mode-switching.c (create_pre_exit): Always split off preceding
7685 insns if we are not at the basic block head.
7687 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7689 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7690 (UCLIBC_DYNAMIC_LINKER): New macro.
7691 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7693 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7694 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7696 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7697 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7698 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7699 for SF and DF modes. Use ieee_quad_format for TF mode.
7700 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7701 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7702 (mips_option_override): Handle `-mnan=legacy'.
7703 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7704 `-mabs=2008' and `-mnan=2008'.
7705 (OPTION_DEFAULT_SPECS): Add "nan" default.
7706 (ASM_SPEC): Handle `-mnan='.
7707 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7708 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7709 comment accordingly.
7710 (neg<mode>2): Likewise.
7711 * config/mips/mips.opt (mabs, mnan): New options.
7712 * doc/install.texi (Configuration): Document `--with-nan=' option.
7713 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7715 (MIPS Options): Document them.
7716 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7717 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7718 * configure: Regenerate.
7719 * config.in: Regenerate.
7721 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7723 * config/i386/i386.md (float post-reload splitters): Do not check
7724 for subregs of SSE registers.
7726 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7727 H.J. Lu <hongjiu.lu@intel.com>
7731 * config/i386/i386.md (post-reload splitter
7732 to avoid partial SSE reg dependency stalls): New pattern.
7734 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7736 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7738 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7740 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7741 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7743 2013-07-26 Cary Coutant <ccoutant@google.com>
7745 * dwarf2out.c (die_checksum_ordered): Don't include template
7746 instantiations in signature.
7747 (is_template_parameter): New function.
7748 (is_template_instantiation): New function.
7749 (generate_skeleton_bottom_up): Don't include template instantiations
7751 (generate_skeleton): Likewise.
7752 (break_out_comdat_types): Move recursive call to break out nested
7754 (prune_unused_types_mark_generic_parms_dies): Call
7755 is_template_parameter.
7757 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7759 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7760 uses vector registers.
7761 * config/aarch64/iterators.md: Add attributes rtn and vas.
7763 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7764 Richard Earnshaw <richard.earnshaw@arm.com>
7766 * combine.c (simplify_comparison): Re-canonicalize operands
7768 * config/arm/arm.md (movcond_addsi): New splitter.
7770 2013-07-25 Sterling Augustine <saugustine@google.com>
7772 * dwarf2out.c (size_of_pubnames): Move code to...
7773 (include_pubname_in_output): ...here. New.
7774 (want_pubnames): Rearrange.
7775 (output_pubnames): Call include_pubname_in_output. Move assertion.
7777 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7779 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7781 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7784 * doc/extend.texi: Remove obsolete builtins. Fix
7785 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7787 2013-07-25 Jan Hubicka <jh@suse.cz>
7789 * cgraph.c (release_function_body): Break out from ...
7790 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7792 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7793 old_tree in the map.
7794 (create_specialized_node): Update.
7795 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7797 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7798 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7799 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7800 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7803 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7805 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7806 addsi3_carryin_alt2_<optab>): Correct output template.
7808 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7810 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7811 Adjust for arm_restrict_it.
7812 Remove trailing whitespace.
7814 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7816 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7817 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7819 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7821 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7823 PR rtl-optimization/57960
7824 * lra-constraints.c (process_alt_operands): Use the right mode
7825 when checking strict_low.
7827 2013-07-25 Jan Hubicka <jh@suse.cz>
7829 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7830 * cgraph.c (cgraph_remove_node): Do not release function body
7831 when in cgraph streaming.
7832 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7833 in other partitions are not considered reachable; fix handling of
7836 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7838 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7840 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7845 * config/arm/arm.md ("*sibcall_insn): Replace use of
7846 Ss with US. Adjust output for v5 and v4t.
7847 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7849 * config/arm/constraints.md ("Ss"): Rename to US.
7851 2013-07-25 Terry Guo <terry.guo@arm.com>
7853 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7854 shift_add/shift_sub0/shift_sub1 RTXs.
7856 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7857 Anton Blanchard <anton@au1.ibm.com>
7859 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7860 (altivec_vpks<VI_char>ss): Likewise.
7861 (altivec_vpks<VI_char>us): Likewise.
7862 (altivec_vpku<VI_char>us): Likewise.
7863 (altivec_vpku<VI_char>um): Likewise.
7865 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7867 Introduce context class.
7869 * Makefile.in (CONTEXT_H): New.
7870 (OBJS): Add context.o.
7871 (toplev.o): Add CONTEXT_H to dependencies.
7874 * toplev.c (general_init): Create the singleton gcc::context instance.
7880 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7882 PR rtl-optimization/57968
7883 * mode-switching.c (create_pre_exit): Allow instructions that
7884 don't set a return register to need a non-exit mode.
7886 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7887 Anton Blanchard <anton@au1.ibm.com>
7889 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7890 operands to vperm for little endian.
7891 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7892 of lvsl to create the control mask for a vperm for little endian.
7894 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7895 Anton Blanchard <anton@au1.ibm.com>
7897 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7898 two operands for little-endian.
7900 2013-07-23 Steve Ellcey <sellcey@mips.com>
7902 * config/mips/mips.c (mips_case_values_threshold): New.
7903 (TARGET_CASE_VALUES_THRESHOLD): Define.
7905 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7906 Anton Blanchard <anton@au1.ibm.com>
7908 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7909 selection of field for vector splat in little endian mode.
7911 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7913 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7914 expanders to rs6000.md.
7915 (ior<mode>3): Likewise.
7916 (and<mode>3): Likewise.
7917 (one_cmpl<mode>2): Likewise.
7918 (nor<mode>3): Likewise.
7919 (andc<mode>3): Likewise.
7920 (eqv<mode>3): Likewise.
7921 (nand<mode>3): Likewise.
7922 (orc<mode>3): Likewise.
7924 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7927 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7928 to split multi-word logical operations.
7929 (rs6000_split_logical_di): Likewise.
7930 (rs6000_split_logical): Likewise.
7932 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7933 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7934 and allow TImode operations in 32-bit.
7935 (vsx_and<mode>3_64bit): Likewise.
7936 (vsx_ior<mode>3_32bit): Likewise.
7937 (vsx_ior<mode>3_64bit): Likewise.
7938 (vsx_xor<mode>3_32bit): Likewise.
7939 (vsx_xor<mode>3_64bit): Likewise.
7940 (vsx_one_cmpl<mode>2_32bit): Likewise.
7941 (vsx_one_cmpl<mode>2_64bit): Likewise.
7942 (vsx_nor<mode>3_32bit): Likewise.
7943 (vsx_nor<mode>3_64bit): Likewise.
7944 (vsx_andc<mode>3_32bit): Likewise.
7945 (vsx_andc<mode>3_64bit): Likewise.
7946 (vsx_eqv<mode>3_32bit): Likewise.
7947 (vsx_eqv<mode>3_64bit): Likewise.
7948 (vsx_nand<mode>3_32bit): Likewise.
7949 (vsx_nand<mode>3_64bit): Likewise.
7950 (vsx_orc<mode>3_32bit): Likewise.
7951 (vsx_orc<mode>3_64bit): Likewise.
7953 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
7954 logical types in GPRs.
7956 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
7957 logical insns to rs6000.md, and allow TImode operations in
7959 (altivec_ior<mode>3): Likewise.
7960 (altivec_xor<mode>3): Likewise.
7961 (altivec_one_cmpl<mode>2): Likewise.
7962 (altivec_nor<mode>3): Likewise.
7963 (altivec_andc<mode>3): Likewise.
7965 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
7966 attributes for moving the 128-bit logical operations into
7968 (BOOL_REGS_OUTPUT): Likewise.
7969 (BOOL_REGS_OP1): Likewise.
7970 (BOOL_REGS_OP2): Likewise.
7971 (BOOL_REGS_UNARY): Likewise.
7972 (BOOL_REGS_AND_CR0): Likewise.
7973 (one_cmpl<mode>2): Add support for DI logical operations on
7974 32-bit, splitting the operations to 32-bit.
7978 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
7979 changes to combine the 32/64-bit code, allow logical operations on
7980 TI mode in 32-bit, and to use similar match_operator patterns like
7981 scalar mode uses. Combine the Altivec and VSX code for logical
7982 operations, and move it here.
7983 (ior<mode>3, 128-bit types): Likewise.
7984 (xor<mode>3, 128-bit types): Likewise.
7985 (one_cmpl<mode>3, 128-bit types): Likewise.
7986 (nor<mode>3, 128-bit types): Likewise.
7987 (andc<mode>3, 128-bit types): Likewise.
7988 (eqv<mode>3, 128-bit types): Likewise.
7989 (nand<mode>3, 128-bit types): Likewise.
7990 (orc<mode>3, 128-bit types): Likewise.
7991 (and<mode>3_internal): Likewise.
7992 (bool<mode>3_internal): Likewise.
7993 (boolc<mode>3_internal1): Likewise.
7994 (boolc<mode>3_internal2): Likewise.
7995 (boolcc<mode>3_internal1): Likewise.
7996 (boolcc<mode>3_internal2): Likewise.
7997 (eqv<mode>3_internal1): Likewise.
7998 (eqv<mode>3_internal2): Likewise.
7999 (one_cmpl1<mode>3_internal): Likewise.
8001 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8003 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8004 Rename flag_stack_usage to flag_stack_usage_info.
8006 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8008 * config/microblaze/sync.md: New file.
8009 * config/microblaze/microblaze.md: Include sync.md
8010 * config/microblaze/microblaze.c: Add print_operand 'y'.
8011 * config/microblaze/constraints.md: Add memory_contraint
8012 'Q' which is a single register.
8014 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8016 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8018 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8020 * reload.c (find_reloads): Exit loop once we find this operand
8021 cannot be reloaded somehow for this alternative.
8023 * reload.c (find_reloads): Exit loop once we find a hard register.
8025 * rtlanal.c (computed_jump_p): Exit loop once we find label
8028 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8030 * cfgloopmanip.c (remove_path): Exit loop after setting
8033 * gensupport.c (subst_dup): Avoid loop if code is not
8034 MATCH_DUP nor MATCH_OP_DUP.
8036 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8038 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8040 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8042 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8043 true for SP_REGNUM if mode == ptr_mode.
8044 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8045 with value R0_REGNUM + 31.
8047 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8049 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8050 pad pointer-typed argument downward.
8052 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8054 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8055 and __ILP32__ when the ILP32 model is in use.
8057 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8059 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8060 (aarch64_load_symref_appropriately): In the case of
8061 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8062 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8063 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8064 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8065 if 'mode' doesn't equal to 'ptr_mode'.
8066 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8067 'vcall_offset'; change to call aarch64_emit_move differently depending
8068 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8069 to calculate the upper bound of 'vcall_offset'.
8070 (aarch64_cannot_force_const_mem): Change to also return true if
8072 (aarch64_legitimize_reload_address): In the case of large
8073 displacements, add new local variable 'xmode' and an assertion
8074 based on it; change to use 'xmode' to generate the new rtx and
8076 (aarch64_asm_trampoline_template): Change to generate the template
8077 differently depending on TARGET_ILP32 or not; change to use
8078 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8079 (aarch64_trampoline_size): Removed.
8080 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8081 and replace immediate literals with it. Change to use 'ptr_mode'
8082 instead of 'DImode' and call convert_memory_address if the mode
8083 of 'fnaddr' doesn't equal to 'ptr_mode'.
8084 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8086 (aarch64_elf_asm_destructor): Likewise.
8087 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8088 on TARGET_ILP32 instead of aarch64_trampoline_size.
8089 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8090 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8091 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8092 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8093 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8094 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8095 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8096 depending on the value of 'mode'.
8097 (add_losym_<mode>): New.
8098 (ldr_got_small_<mode>): New, based on ldr_got_small.
8099 (ldr_got_small): Remove.
8100 (ldr_got_small_sidi): New.
8101 * config/aarch64/iterators.md (P): New.
8102 (PTR): Change to 'ptr_mode' in the condition.
8104 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8106 * config.gcc (aarch64*-*-*): Support --with-abi.
8107 (aarch64*-*-elf): Support --with-multilib-list.
8108 (aarch64*-*-linux*): Likewise.
8109 (supported_defaults): Add abi to aarch64*-*-*.
8110 * configure.ac: Mention AArch64 for --with-multilib-list.
8111 * configure: Re-generated.
8112 * config/aarch64/biarchilp32.h: New file.
8113 * config/aarch64/biarchlp64.h: New file.
8114 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8116 (MULTILIB_DEFAULTS): Ditto.
8117 (DRIVER_SELF_SPECS): Ditto.
8118 (ASM_SPEC): Update to also substitute -mabi.
8119 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8120 file whose name depends on -mabi= and -mbig-endian.
8121 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8123 (POINTER_SIZE): New define.
8124 (POINTERS_EXTEND_UNSIGNED): Ditto.
8125 (enum aarch64_abi_type): New enumeration tag.
8126 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8127 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8128 (TARGET_ILP32): New define.
8129 * config/aarch64/aarch64.opt (mabi): New.
8131 (ilp32, lp64): New values for -mabi.
8132 * config/aarch64/t-aarch64 (comma): New define.
8133 (MULTILIB_OPTIONS): Ditto.
8134 (MULTILIB_DIRNAMES): Ditto.
8135 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8136 * doc/invoke.texi: Document -mabi for AArch64.
8138 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8140 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8142 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8143 Anton Blanchard <anton@au1.ibm.com>
8145 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8146 endianness when selecting field to splat.
8148 2013-07-22 Eric Christopher <echristo@gmail.com>
8150 * dwarf2out.c (die_odr_checksum): New function to use
8151 CHECKSUM_ macros and ULEB128 for DIE tag.
8152 (generate_type_signature): Use.
8154 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8156 * config.gcc (sparc*-*-*): Accept leon3 processor.
8157 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8158 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8159 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8160 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8161 (mfix-ut699): Adjust comment.
8162 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8163 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8164 (CPP_CPU_SPEC): Likewise.
8165 (ASM_CPU_SPEC): Likewise.
8166 * config/sparc/sparc.c (leon3_cost): New constant.
8167 (sparc_option_override): Add leon3 support.
8168 (mem_ref): New function.
8169 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8170 (sparc_do_work_around_errata): Look into the instruction in the delay
8171 slot and adjust accordingly. Add fix for the data cache nullify issues
8172 of the UT699. Change insertion position for the NOP.
8173 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8174 (leon3_load): New reservation.
8175 (leon_store): Bump latency to 2.
8176 (grfpu): New automaton.
8177 (grfpu_alu): New unit.
8178 (grfpu_ds): Likewise.
8179 (leon_fp_alu): Adjust.
8180 (leon_fp_mult): Delete.
8181 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8182 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8183 * config/sparc/sparc.md (cpu): Add leon3.
8184 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8186 (atomic_test_and_set): Likewise.
8189 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8191 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8192 default for R5900 targets.
8193 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8194 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8195 * config/mips/mips.c (mips_option_override): Report an error for
8196 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8197 for -march=r5900 -mhard-float.
8199 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8201 * df-problems.c (can_move_insns_across): Exit loop once we
8202 find a non-fixed, non-global register.
8204 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8207 * omega.c (omega_eliminate_red): Break after setting red_found.
8208 (omega_problem_has_red_equations): Similarly after setting found.
8209 (omega_query_variable): Similarly after setting coupled.
8211 2013-07-22 Marek Polacek <polacek@redhat.com>
8213 * gimplify.c: Don't include gimple.h twice.
8215 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8217 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8218 instead of TARGET_THUMB1.
8219 (Pz): New constraint.
8220 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8222 (compare_negsi_si): Likewise.
8223 (compare_addsi2_op0): Likewise.
8224 (compare_addsi2_op1): Likewise.
8225 (addsi3_carryin_<optab>): Likewise.
8226 (addsi3_carryin_alt2_<optab>): Likewise.
8227 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8228 for arm_restrict_it.
8229 (subsi3_carryin): Likewise.
8230 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8231 (minmax_arithsi): Disable for arm_restrict_it.
8232 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8233 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8234 (satsi_<SAT:code>_shift): Likewise.
8235 (arm_shiftsi3): Add alternative for 16-bit encoding.
8236 (arm32_movhf): Disable for arm_restrict_it.
8237 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8238 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8240 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8242 * config/arm/arm.md (attribute "insn"): Delete.
8243 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8244 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8245 (not_shiftsi): Update for attribute change.
8246 (not_shiftsi_compare0): Likewise.
8247 (not_shiftsi_compare0_scratch): Likewise.
8248 (arm_one_cmplsi2): Likewise.
8249 (thumb1_one_cmplsi2): Likewise.
8250 (notsi_compare0): Likewise.
8251 (notsi_compare0_scratch): Likewise.
8252 (thumb1_movdi_insn): Likewise.
8253 (arm_movsi_insn): Likewise.
8254 (movhi_insn_arch4): Likewise.
8255 (movhi_bytes): Likewise.
8256 (arm_movqi_insn): Likewise.
8257 (thumb1_movqi_insn): Likewise.
8258 (arm32_movhf): Likewise.
8259 (thumb1_movhf): Likewise.
8260 (arm_movsf_soft_insn): Likewise.
8261 (thumb1_movsf_insn): Likewise.
8262 (thumb_movdf_insn): Likewise.
8263 (movsicc_insn): Likewise.
8264 (movsfcc_soft_insn): Likewise.
8265 (and_scc): Likewise.
8266 (cond_move): Likewise.
8267 (if_move_not): Likewise.
8268 (if_not_move): Likewise.
8269 (if_shift_move): Likewise.
8270 (if_move_shift): Likewise.
8271 (if_shift_shift): Likewise.
8272 (if_not_arith): Likewise.
8273 (if_arith_not): Likewise.
8274 (cond_move_not): Likewise.
8275 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8276 (neon_mov<mode>): Likewise.
8277 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8278 (thumb2_movsi_vfp): Likewise.
8279 (movsf_vfp): Likewise.
8280 (thumb2_movsf_vfp): Likewise.
8281 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8283 (cortexa7_older_only): Likewise.
8284 (cortexa7_younger): Likewise.
8285 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8286 (1020alu_shift_op): Likewise.
8287 (1020alu_shift_reg_op): Likewise.
8288 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8289 (alu_shift_op): Likewise.
8290 (alu_shift_reg_op): Likewise.
8291 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8292 (11_alu_shift_op): Likewise.
8293 (11_alu_shift_reg_op): Likewise.
8294 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8295 (9_alu_shift_reg_op): Likewise.
8296 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8298 (cortex_a15_alu_shift): Likewise.
8299 (cortex_a15_alu_shift_reg): Likewise.
8300 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8301 (cortex_a5_alu_shift): Likewise.
8302 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8304 (cortex_a53_alu_shift): Likewise.
8305 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8307 (cortex_a7_alu_reg): Likewise.
8308 (cortex_a7_alu_shift): Likewise.
8309 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8310 (cortex_a8_alu_shift): Likewise.
8311 (cortex_a8_alu_shift_reg): Likewise.
8312 (cortex_a8_mov): Likewise.
8313 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8314 (cortex_a9_dp_shift): Likewise.
8315 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8316 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8317 (cortex_r4_mov): Likewise.
8318 (cortex_r4_alu_shift): Likewise.
8319 (cortex_r4_alu_shift_reg): Likewise.
8320 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8321 (526_alu_shift_op): Likewise.
8322 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8323 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8324 (626te_alu_shift_op): Likewise.
8325 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8326 (726te_alu_op): Likewise.
8327 (726te_alu_shift_op): Likewise.
8328 (726te_alu_shift_reg_op): Likewise.
8329 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8330 (mp626_alu_shift_op): Likewise.
8331 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8332 (pj4_alu_e1_conds): Likewise.
8333 (pj4_alu): Likewise.
8334 (pj4_alu_conds): Likewise.
8335 (pj4_shift): Likewise.
8336 (pj4_shift_conds): Likewise.
8337 (pj4_alu_shift): Likewise.
8338 (pj4_alu_shift_conds): Likewise.
8340 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8342 * config/arm/predicates.md (shiftable_operator_strict_it):
8344 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8345 Disable cond_exec version for arm_restrict_it.
8346 (thumb2_smaxsi3): Convert to generate cond_exec.
8347 (thumb2_sminsi3): Likewise.
8348 (thumb32_umaxsi3): Likewise.
8349 (thumb2_uminsi3): Likewise.
8350 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8351 (thumb2_neg_abssi2): Likewise.
8352 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8353 (thumb2_movsicc_insn): Adjust alternatives.
8354 (thumb2_mov_negscc): Disable for arm_restrict_it.
8355 (thumb2_mov_negscc_strict_it): New pattern.
8356 (thumb2_mov_notscc_strict_it): New pattern.
8357 (thumb2_mov_notscc): Disable for arm_restrict_it.
8358 (thumb2_ior_scc): Likewise.
8359 (thumb2_ior_scc_strict_it): New pattern.
8360 (thumb2_cond_move): Adjust for arm_restrict_it.
8361 (thumb2_cond_arith): Disable for arm_restrict_it.
8362 (thumb2_cond_arith_strict_it): New pattern.
8363 (thumb2_cond_sub): Adjust for arm_restrict_it.
8364 (thumb2_movcond): Likewise.
8365 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8366 (thumb2_zero_extendhisi2_v6): Likewise.
8367 (thumb2_zero_extendqisi2_v6): Likewise.
8368 (orsi_notsi_si): Likewise.
8369 (orsi_not_shiftsi_si): Likewise.
8371 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8373 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8374 instruction sequence is 1 byte shorter.
8376 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8378 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8379 it is not needed after split.
8381 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8384 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8385 second label for nonlocal goto receivers. Don't output pic base labels
8386 unless we're producing PIC; mark that action unreachable().
8387 (ix86_save_reg): If the function contains a nonlocal label, save the
8389 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8390 * config/darwin.c (emitted_pic_label_num): New GTY.
8391 (update_pic_label_number_if_needed): New.
8392 (machopic_output_function_base_name): Adjust for nonlocal receiver
8394 (machopic_should_output_picbase_label): New.
8395 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8396 (nonlocal_goto_receiver): New insn and split.
8398 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8400 * config/aarch64/aarch64-builtins.c
8401 (aarch64_fold_builtin): Fold abs in all modes.
8402 * config/aarch64/aarch64-simd-builtins.def
8403 (abs): Enable for all modes.
8404 * config/aarch64/arm_neon.h
8405 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8406 (vabs_f64): Add missing intrinsic.
8408 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8410 * config/aarch64/arm_neon.h (vabs_s64): New function
8412 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8415 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8416 * config/avr/avr.md (adjust_len): Add `round'.
8417 * config/avr/avr-protos.h (avr_out_round): New prototype.
8418 (avr_out_plus): Add `out_label' argument.
8419 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8420 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8421 Handle the case where `insn' is just a pattern.
8422 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8423 (avr_out_round): New function.
8424 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8426 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8428 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8429 Add check for flag_stack_usage to handle -fstack-usage support
8431 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8433 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8434 interaction for new Power8 flags and VSX.
8436 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8439 * tree-inline.c (expand_call_inline): Emit errors during
8440 early_inlining only if optimization is not turned on.
8442 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8446 * passes.c (init_optimization_passes): Move the construction of
8447 the pass hierarchy into a new passes.def file.
8449 * Makefile.in (passes.o): Add dependency on passes.def.
8451 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8453 * passes.c (init_optimization_passes): Introduce macros for
8454 constructing the tree of passes (INSERT_PASSES_AFTER,
8455 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8456 TERMINATE_PASS_LIST).
8458 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8459 Wei Mi <wmi@google.com>
8461 PR rtl-optimization/57878
8462 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8464 (reload_pseudo_compare_func): Check nregs first for reload
8467 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8469 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8471 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8473 * read-rtl.c (validate_const_int): Once an invalid character is
8474 seen, quit the loop.
8476 * gengtype.c (write_roots): Similarly once we find the "deletable"
8477 or "if_marked" option.
8479 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8481 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8482 "xtab" and "sat". Move value "clz" from here to ...
8483 (attriubte "type"): ... here.
8484 (satsi_<SAT:code>): Delete "insn" attribute.
8485 (satsi_<SAT:code>_shift): Likewise.
8486 (arm_zero_extendqisi2addsi): Likewise.
8487 (arm_extendqisi2addsi): Likewise.
8488 (clzsi2): Update for attribute changes.
8489 (rbitsi2): Likewise.
8490 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8492 (arm_usatsihi): Likewise.
8493 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8495 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8497 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8498 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8499 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8500 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8501 in alphabetical order.
8502 (attribute "core_cycles"): Update for attribute changes.
8503 (arm_addsi3): Likewise.
8504 (addsi3_compare0): Likewise.
8505 (addsi3_compare0_scratch): Likewise.
8506 (addsi3_compare_op1): Likewise.
8507 (addsi3_compare_op2): Likewise.
8508 (compare_addsi2_op0): Likewise.
8509 (compare_addsi2_op1): Likewise.
8510 (addsi3_carryin_shift_<optab>): Likewise.
8511 (subsi3_carryin_shift): Likewise.
8512 (rsbsi3_carryin_shift): Likewise.
8513 (arm_subsi3_insn): Likewise.
8514 (subsi3_compare0): Likewise.
8515 (subsi3_compare): Likewise.
8516 (arm_andsi3_insn): Likewise.
8517 (thumb1_andsi3_insn): Likewise.
8518 (andsi3_compare0): Likewise.
8519 (andsi3_compare0_scratch): Likewise.
8520 (zeroextractsi_compare0_scratch
8521 (andsi_not_shiftsi_si): Likewise.
8522 (iorsi3_insn): Likewise.
8523 (iorsi3_compare0): Likewise.
8524 (iorsi3_compare0_scratch): Likewise.
8525 (arm_xorsi3): Likewise.
8526 (thumb1_xorsi3_insn): Likewise.
8527 (xorsi3_compare0): Likewise.
8528 (xorsi3_compare0_scratch): Likewise.
8529 (satsi_<SAT:code>_shift): Likewise.
8531 (arm_shiftsi3): Likewise.
8532 (shiftsi3_compare0): Likewise.
8533 (not_shiftsi): Likewise.
8534 (not_shiftsi_compare0): Likewise.
8535 (not_shiftsi_compare0_scratch): Likewise.
8536 (arm_one_cmplsi2): Likewise.
8537 (thumb_one_complsi2): Likewise.
8538 (notsi_compare0): Likewise.
8539 (notsi_compare0_scratch): Likewise.
8540 (thumb1_zero_extendhisi2): Likewise.
8541 (arm_zero_extendhisi2): Likewise.
8542 (arm_zero_extendhisi2_v6): Likewise.
8543 (arm_zero_extendhisi2addsi): Likewise.
8544 (thumb1_zero_extendqisi2): Likewise.
8545 (thumb1_zero_extendqisi2_v6): Likewise.
8546 (arm_zero_extendqisi2): Likewise.
8547 (arm_zero_extendqisi2_v6): Likewise.
8548 (arm_zero_extendqisi2addsi): Likewise.
8549 (thumb1_extendhisi2): Likewise.
8550 (arm_extendhisi2): Likewise.
8551 (arm_extendhisi2_v6): Likewise.
8552 (arm_extendqisi): Likewise.
8553 (arm_extendqisi_v6): Likewise.
8554 (arm_extendqisi2addsi): Likewise.
8555 (thumb1_extendqisi2): Likewise.
8556 (thumb1_movdi_insn): Likewise.
8557 (arm_movsi_insn): Likewise.
8558 (movsi_compare0): Likewise.
8559 (movhi_insn_arch4): Likewise.
8560 (movhi_bytes): Likewise.
8561 (arm_movqi_insn): Likewise.
8562 (thumb1_movqi_insn): Likewise.
8563 (arm32_movhf): Likewise.
8564 (thumb1_movhf): Likewise.
8565 (arm_movsf_soft_insn): Likewise.
8566 (thumb1_movsf_insn): Likewise.
8567 (movdf_soft_insn): Likewise.
8568 (thumb_movdf_insn): Likewise.
8569 (arm_cmpsi_insn): Likewise.
8570 (cmpsi_shiftsi): Likewise.
8571 (cmpsi_shiftsi_swp): Likewise.
8572 (arm_cmpsi_negshiftsi_si): Likewise.
8573 (movsicc_insn): Likewise.
8574 (movsfcc_soft_insn): Likewise.
8575 (arith_shiftsi): Likewise.
8576 (arith_shiftsi_compare0
8577 (arith_shiftsi_compare0_scratch
8578 (sub_shiftsi): Likewise.
8579 (sub_shiftsi_compare0
8580 (sub_shiftsi_compare0_scratch
8581 (and_scc): Likewise.
8582 (cond_move): Likewise.
8583 (if_plus_move): Likewise.
8584 (if_move_plus): Likewise.
8585 (if_move_not): Likewise.
8586 (if_not_move): Likewise.
8587 (if_shift_move): Likewise.
8588 (if_move_shift): Likewise.
8589 (if_shift_shift): Likewise.
8590 (if_not_arith): Likewise.
8591 (if_arith_not): Likewise.
8592 (cond_move_not): Likewise.
8593 (thumb1_ashlsi3): Set type attribute.
8594 (thumb1_ashrsi3): Likewise.
8595 (thumb1_lshrsi3): Likewise.
8596 (thumb1_rotrsi3): Likewise.
8597 (shiftsi3_compare0_scratch): Likewise.
8598 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8599 (neon_mov<mode>): Likewise.
8600 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8602 (thumb2_movsi_insn): Likewise.
8603 (thumb2_cmpsi_neg_shiftsi): Likewise.
8604 (thumb2_extendqisi_v6): Likewise.
8605 (thumb2_zero_extendhisi2_v6): Likewise.
8606 (thumb2_zero_extendqisi2_v6): Likewise.
8607 (thumb2_shiftsi3_short): Likewise.
8608 (thumb2_addsi3_compare0_scratch): Likewise.
8609 (orsi_not_shiftsi_si): Likewise.
8610 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8611 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8613 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8614 (1020alu_shift_op): Likewise.
8615 (1020alu_shift_reg_op): Likewise.
8616 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8617 (alu_shift_op): Likewise.
8618 (alu_shift_reg_op): Likewise.
8619 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8620 (11_alu_shift_op): Likewise.
8621 (11_alu_shift_reg_op): Likewise.
8622 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8623 (9_alu_shift_reg_op): Likewise.
8624 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8626 (cortex_a15_alu_shift): Likewise.
8627 (cortex_a15_alu_shift_reg): Likewise.
8628 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8630 (cortex_a5_alu_shift): Likewise.
8631 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8633 (cortex_a53_alu_shift): Likewise.
8634 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8636 (cortex_a7_alu_reg): Likewise.
8637 (cortex_a7_alu_shift): Likewise.
8638 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8640 (cortex_a8_alu_shift): Likewise.
8641 (cortex_a8_alu_shift_reg): Likewise.
8642 (cortex_a8_mov): Likewise.
8643 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8644 (cortex_a9_dp_shift): Likewise.
8645 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8647 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8649 (cortex_r4_mov): Likewise.
8650 (cortex_r4_alu_shift): Likewise.
8651 (cortex_r4_alu_shift_reg): Likewise.
8652 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8653 (526_alu_shift_op): Likewise.
8654 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8655 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8656 (626te_alu_shift_op): Likewise.
8657 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8658 (726te_alu_op): Likewise.
8659 (726te_alu_shift_op): Likewise.
8660 (726te_alu_shift_reg_op): Likewise.
8661 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8662 (mp626_alu_shift_op): Likewise.
8663 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8664 (pj4_alu_e1_conds): Likewise.
8665 (pj4_alu): Likewise.
8666 (pj4_alu_conds): Likewise.
8667 (pj4_shift): Likewise.
8668 (pj4_shift_conds): Likewise.
8669 (pj4_alu_shift): Likewise.
8670 (pj4_alu_shift_conds): Likewise.
8671 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8673 (cortexa7_older_only): Likewise.
8674 (cortexa7_younger): Likewise.
8676 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8678 * ipa-pure-const.c (generate_summary): Rename to...
8679 (pure_const_generate_summary): ... this.
8681 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8683 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8685 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8688 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8691 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8693 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8694 enabled without -march=zEC12.
8695 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8698 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8700 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8701 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8702 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8703 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8704 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8705 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8706 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8708 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8709 and ISA_HAS_NMADD3_NMSUB3.
8710 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8711 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8712 (nmsub4<mode>, nmsub3<mode>): Likewise.
8713 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8715 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8717 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8718 TARGET_MIPS5400 checking.
8720 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8721 Peter Bergner <bergner@vnet.ibm.com>
8723 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8724 registers in the comment.
8725 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8726 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8727 rather than FIRST_PSEUDO_REGISTERS.
8729 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8731 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8732 enable extra ISA flags with TARGET_HTM.
8734 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8736 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8739 2013-07-15 Cong Hou <congh@google.com>
8741 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8742 in compare function for sorting.
8744 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8746 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8747 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8748 * config/rs6000/rs6000.opt: Add -mhtm option.
8749 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8750 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8751 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8752 __HTM__ if the HTM instructions are available.
8753 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8754 htm_spr_reg_operand): New define_predicates.
8755 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8756 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8758 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8759 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8760 HTM builtin functions.
8761 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8762 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8763 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8764 (rs6000_builtin_mask_calculate): Likewise.
8765 (rs6000_option_override_internal): Likewise.
8766 (bdesc_htm): Add new HTM builtin support.
8767 (htm_spr_num): New function.
8768 (htm_spr_regno): Likewise.
8769 (rs6000_htm_spr_icode): Likewise.
8770 (htm_expand_builtin): Likewise.
8771 (htm_init_builtins): Likewise.
8772 (rs6000_expand_builtin): Add support for HTM builtin functions.
8773 (rs6000_init_builtins): Likewise.
8774 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8776 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8777 (TARGET_HTM, MASK_HTM): Define macros.
8778 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8779 (FIXED_REGISTERS): Likewise.
8780 (CALL_USED_REGISTERS): Likewise.
8781 (CALL_REALLY_USED_REGISTERS): Likewise.
8782 (REG_ALLOC_ORDER): Likewise.
8783 (enum reg_class): Likewise.
8784 (REG_CLASS_NAMES): Likewise.
8785 (REG_CLASS_CONTENTS): Likewise.
8786 (REGISTER_NAMES): Likewise.
8787 (ADDITIONAL_REGISTER_NAMES): Likewise.
8788 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8789 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8790 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8791 * config/rs6000/htm.md: New file.
8792 * config/rs6000/htmintrin.h: New file.
8793 * config/rs6000/htmxlintrin.h: New file.
8795 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8797 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8798 Define SYMBOL_TINY_GOT, update comment.
8799 * config/aarch64/aarch64.c
8800 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8801 (aarch64_expand_mov_immediate): Likewise.
8802 (aarch64_print_operand): Likewise.
8803 (aarch64_classify_symbol): Likewise.
8804 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8805 (ldr_got_tiny): Define.
8807 2013-07-13 Tobias Grosser <tobias@grosser.es>
8809 PR tree-optimization/54094
8810 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8811 scheduling dimension for the parallelism check from the polyhedral
8812 information in the AST.
8813 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8816 2013-07-13 Jason Merrill <jason@redhat.com>
8818 * print-tree.c (debug_vec_tree): Use debug_raw.
8819 (debug_raw (vec<tree, va_gc> &)): New.
8820 (debug_raw (vec<tree, va_gc> *)): New.
8821 * tree.h: Declare them.
8823 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8825 * ifcvt.c (ifcvt_after_combine): New static variable.
8826 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8828 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8829 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8830 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8832 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8834 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8836 2013-07-12 Michael Matz <matz@suse.de>
8839 * convert.c (convert_to_real): Reject non-float inner types.
8841 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8843 * config/aarch64/aarch64-protos.h
8844 (aarch64_simd_immediate_valid_for_move): Remove.
8845 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8846 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8848 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8850 2013-07-11 Steve Ellcey <sellcey@mips.com>
8852 * config/mips/mips.c (mips_conditional_register_usage): Do not
8853 use t[0-7] registers in MIPS16 mode when optimizing for size.
8855 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8857 * config/i386/i386.c (dispatch_function_versions): Fix array
8858 indexing of function_version_info to match actual_versions.
8860 2013-07-11 Teresa Johnson <tejohnson@google.com>
8862 * vec.h (struct va_gc): Move release out-of-line.
8863 (va_gc::release): Call ggc_free on released vec.
8865 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8867 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8868 Require GOT register as additional operand in UNSPEC.
8869 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8870 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8871 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8872 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8873 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8874 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8875 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8877 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8880 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8881 name seen by assembler/linker rather if available.
8883 2013-07-11 Andreas Schwab <schwab@suse.de>
8885 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8887 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8889 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8891 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8893 * doc/tm.texi.in: Move hook documentation to ....
8894 * target.def: ... here.
8896 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8898 * doc/tm.texi: Regenerate.
8900 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8903 * doc/invoke.texi: Document -Wconditionally-supported.
8905 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8908 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8911 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8914 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8915 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8916 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8917 Remove duplicate devices.
8918 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8919 * config/avr/t-multilib: Regenerate.
8920 * config/avr/avr-tables.opt: Regenerate.
8921 * doc/avr-mmcu.texi: Regenerate.
8923 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8926 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8928 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8930 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8931 the cost of MULT when optimizing for size.
8933 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8935 * config/cr16/cr16-protos.h: Don't include target.h.
8937 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8939 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8940 adjust register size for TDmode and TFmode for VSX registers.
8942 2013-07-08 Kai Tietz <ktietz@redhat.com>
8945 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8946 hook_bool_const_tree_true.
8948 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8950 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8951 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
8952 * config/s390/s390.md: Define FPR*_REGNUM constants.
8953 Fix FPR2_REGNUM constant (18 -> 17).
8954 ("*trunc<BFP:mode><DFP_ALL:mode>2")
8955 ("*trunc<DFP_ALL:mode><BFP:mode>2")
8956 ("trunc<BFP:mode><DFP_ALL:mode>2")
8957 ("trunc<DFP_ALL:mode><BFP:mode>2")
8958 ("*extend<BFP:mode><DFP_ALL:mode>2")
8959 ("*extend<DFP_ALL:mode><BFP:mode>2")
8960 ("extend<BFP:mode><DFP_ALL:mode>2")
8961 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
8964 2013-07-08 Graham Stott <graham.stott@btinternet.com>
8966 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
8968 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8970 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
8971 and cfun_fpr_bit_p to cfun_fpr_save_p.
8972 (s390_frame_area, s390_register_info, s390_frame_info)
8973 (s390_emit_prologue, s390_emit_epilogue)
8974 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
8976 * config/s390/s390.h: Define *_REGNUM macros for floating point
8979 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
8981 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
8983 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
8985 PR rtl-optimization/57786
8986 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
8987 and break out of the loop when it is set to false.
8989 2013-07-08 Jakub Jelinek <jakub@redhat.com>
8992 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
8993 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
8994 (const_int 63)) 0)).
8995 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
8996 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
8997 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
8999 PR rtl-optimization/57829
9000 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9001 mask bits outside of mode are just sign-extension from mode to HWI.
9003 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9005 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9006 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9007 adjust_address instead of change_address to keep info about alignment.
9008 (emit_strmov): Remove.
9009 (emit_memmov): New function.
9010 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9011 (expand_movmem_epilogue): Likewise and return updated rtx for
9013 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9014 destination and source.
9015 (decide_alignment): Refactor, handle vector_loop.
9016 (ix86_expand_movmem): Likewise.
9017 (ix86_expand_setmem): Likewise.
9018 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9020 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9022 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9023 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9025 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9027 * config/i386/sse.md (sse_movlhps): Change alternative 3
9028 of operand 2 to "m".
9030 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9033 * config/i386/sse.md (iptr): New mode attribute.
9034 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9035 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9036 (<sse>_vmmul<mode>3): Ditto.
9037 (<sse>_vmdiv<mode>3): Ditto.
9038 (sse_vmrcpv4sf2): Ditto.
9039 (<sse>_vmsqrt<mode>2): Ditto.
9040 (sse_vmrsqrtv4sf2): Ditto.
9041 (<sse>_vm<code><mode>3): Ditto.
9042 (avx_vmcmp<mode>3): Ditto.
9043 (<sse>_vmmaskcmp<mode>3): Ditto.
9044 (<sse>_comi): Ditto.
9045 (<sse>_ucomi): Ditto.
9046 (*xop_vmfrcz_<mode>): Ditto.
9047 (*fmai_fmadd_<mode>): Ditto.
9048 (*fmai_fmsub_<mode>): Ditto.
9049 (*fmai_fnmadd_<mode>): Ditto.
9050 (*fmai_fnmsub_<mode>): Ditto.
9051 (*fma4i_vmfmadd_<mode>): Ditto.
9052 (*fma4i_vmfmsub_<mode>): Ditto.
9053 (*fma4i_vmfnmadd_<mode>): Ditto.
9054 (*fma4i_vmfnmsub_<mode>): Ditto.
9055 (*xop_vmfrcz_<mode>): Ditto.
9056 (sse_cvtps2pi): Ditto.
9057 (sse_cvttps2pi): Ditto.
9058 (sse_cvtss2si): Ditto.
9059 (sse_cvtss2si_2): Ditto.
9060 (sse_cvtss2siq_2): Ditto.
9061 (sse_cvttss2si): Ditto.
9062 (sse_cvttss2siq): Ditto.
9063 (sse_cvtsd2si): Ditto.
9064 (sse_cvtsd2si_2): Ditto.
9065 (sse_cvtsd2siq_2): Ditto.
9066 (sse_cvttsd2si): Ditto.
9067 (sse_cvttsd2siq): Ditto.
9068 (sse_cvtsd2ss): Ditto.
9069 (sse_cvtss2sd): Ditto.
9070 (avx2_pbroadcast<mode>): Ditto.
9071 (avx2_pbroadcast<mode>_1): Ditto.
9072 (*avx_vperm_broadcast_v4sf): Ditto.
9074 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9075 (sse_movlhps): Ditto.
9076 (sse_storehps): Ditto.
9077 (sse_loadhps): Ditto.
9078 (sse_storelps): Ditto.
9079 (sse_loadlps): Ditto.
9080 (*vec_concatv4sf): Ditto.
9081 (*vec_interleave_highv2df): Ditto.
9082 (*vec_interleave_lowv2df): Ditto.
9083 (*vec_extractv2df_1_sse): Ditto.
9084 (*vec_extractv2df_0_sse): Ditto.
9085 (sse2_storelpd): Ditto.
9086 (sse2_loadlpd): Ditto.
9087 (sse2_movsd): Ditto.
9088 (*vec_concatv4si): Ditto.
9089 (vec_concatv2di): Ditto.
9091 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9092 for Intel asm dialect.
9093 (mmx_punpcklwd): Ditto.
9094 (mmx_punpckldq): Ditto.
9096 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9097 for intel assembler dialect.
9099 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9102 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9103 for BUILT_IN_C{LZ,LRSB}*.
9104 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9105 * tree-vrp.c (extract_range_basic): Handle
9106 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9107 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9108 fall thru to code calling set_value*.
9109 * builtins.c (expand_builtin): Remove *IMAX cases.
9110 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9111 if width is bigger than 2*HWI.
9113 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9115 PR rtl-optimization/55342
9116 * lra-int.h (lra_subreg_reload_pseudos): New.
9117 * lra.c: Add undoing optional reloads to the block diagram.
9118 (lra_subreg_reload_pseudos): New.
9119 (lra_optional_reload_pseudos): Change comments.
9120 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9121 lra_optional_reload_pseudos after undo transformations.
9122 * lra-assigns.c (pseudo_prefix_title): New.
9123 (lra_setup_reg_renumber): Use it.
9124 (spill_for): Ditto. Check subreg reload pseudos too.
9125 (assign_by_spills): Consider subreg reload pseudos too.
9126 * lra-constraints.c (simplify_operand_subreg): Use
9127 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9128 (curr_insn_transform): Recognize and do optional reloads.
9129 (undo_optional_reloads): New.
9130 (lra_undo_inheritance): Call undo_optional_reloads.
9132 2013-07-05 Thomas Quinot <quinot@adacore.com>
9134 * tree-complex.c (expand_complex_operations_1): Fix typo.
9136 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9138 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9139 (tune_params): New member 'const vec_costs'.
9140 * config/aarch64/aarch64.c (generic_vector_cost): New.
9141 (generic_tunings): New member 'generic_vector_cost'.
9142 (aarch64_builtin_vectorization_cost): New.
9143 (aarch64_add_stmt_cost): New.
9144 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9145 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9147 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9150 * config/i386/predicates.md (vsib_address_operand): Disallow
9151 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9153 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9156 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9157 expand_builtin_setjmp_receiver.
9158 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9159 with NULL for the label parameter.
9160 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9161 the frame-pointer. Adjust comments.
9162 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9163 only if LABEL is non-NULL.
9165 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9167 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9168 (ARM_ABI_AAPCS64): Ditto.
9170 (ARM_DEFAULT_ABI): Ditto.
9172 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9174 * config/aarch64/aarch64-builtins.c
9175 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9176 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9178 * config/aarch64/aarch64-simd.md
9179 (aarch64_ld1<VALL:mode>): New.
9180 (aarch64_st1<VALL:mode>): Likewise.
9181 * config/aarch64/arm_neon.h
9182 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9184 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9186 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9188 (ix86_option_override_internal):Turn on all -mavx target flags by
9189 default as they are dependent on AVX anyway.
9191 2013-07-02 Cary Coutant <ccoutant@google.com>
9193 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9195 (loc_checksum_ordered): Likewise.
9196 (hash_loc_operands): Remove inline keyword.
9198 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9200 PR tree-optimization/57741
9201 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9202 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9203 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9204 Allow REAL_CST step_exprs if flag_associative_math.
9205 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9207 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9209 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9211 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9213 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9215 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9217 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9219 (iorsi3_insn): Likewise.
9220 (arm_xorsi3): Likewise.
9222 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9224 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9226 (attribute "type"): ... here, and prefix with "wmmx_".
9227 (attribute "core_cycles"): Update for attribute changes.
9228 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9229 (tbcstv4hi): Likewise.
9230 (tbcstv2si): Likewise.
9231 (iwmmxt_iordi3): Likewise.
9232 (iwmmxt_xordi3): Likewise.
9233 (iwmmxt_anddi3): Likewise.
9234 (iwmmxt_nanddi3): Likewise.
9235 (iwmmxt_arm_movdi): Likewise.
9236 (iwmmxt_movsi_insn): Likewise.
9237 (mov<mode>_internal): Likewise.
9238 (and<mode>3_iwmmxt): Likewise.
9239 (ior<mode>3_iwmmxt): Likewise.
9240 (xor<mode>3_iwmmxt): Likewise.
9241 (add<mode>3_iwmmxt): Likewise.
9242 (ssaddv8qi3): Likewise.
9243 (ssaddv4hi3): Likewise.
9244 (ssaddv2si3): Likewise.
9245 (usaddv8qi3): Likewise.
9246 (usaddv4hi3): Likewise.
9247 (usaddv2si3): Likewise.
9248 (sub<mode>3_iwmmxt): Likewise.
9249 (sssubv8qi3): Likewise.
9250 (sssubv4hi3): Likewise.
9251 (sssubv2si3): Likewise.
9252 (ussubv8qi3): Likewise.
9253 (ussubv4hi3): Likewise.
9254 (ussubv2si3): Likewise.
9255 (mulv4hi3_iwmmxt): Likewise.
9256 (smulv4hi3_highpart): Likewise.
9257 (umulv4hi3_highpart): Likewise.
9258 (iwmmxt_wmacs): Likewise.
9259 (iwmmxt_wmacsz): Likewise.
9260 (iwmmxt_wmacu): Likewise.
9261 (iwmmxt_wmacuz): Likewise.
9262 (iwmmxt_clrdi): Likewise.
9263 (iwmmxt_clrv8qi): Likewise.
9264 (iwmmxt_clr4hi): Likewise.
9265 (iwmmxt_clr2si): Likewise.
9266 (iwmmxt_uavgrndv8qi3): Likewise.
9267 (iwmmxt_uavgrndv4hi3): Likewise.
9268 (iwmmxt_uavgv8qi3): Likewise.
9269 (iwmmxt_uavgv4hi3): Likewise.
9270 (iwmmxt_tinsrb): Likewise.
9271 (iwmmxt_tinsrh): Likewise.
9272 (iwmmxt_tinsrw): Likewise.
9273 (iwmmxt_textrmub): Likewise.
9274 (iwmmxt_textrmsb): Likewise.
9275 (iwmmxt_textrmuh): Likewise.
9276 (iwmmxt_textrmsh): Likewise.
9277 (iwmmxt_textrmw): Likewise.
9278 (iwmxxt_wshufh): Likewise.
9279 (eqv8qi3): Likewise.
9280 (eqv4hi3): Likewise.
9281 (eqv2si3): Likewise.
9282 (gtuv8qi3): Likewise.
9283 (gtuv4hi3): Likewise.
9284 (gtuv2si3): Likewise.
9285 (gtv8qi3): Likewise.
9286 (gtv4hi3): Likewise.
9287 (gtv2si3): Likewise.
9288 (smax<mode>3_iwmmxt): Likewise.
9289 (umax<mode>3_iwmmxt): Likewise.
9290 (smin<mode>3_iwmmxt): Likewise.
9291 (umin<mode>3_iwmmxt): Likewise.
9292 (iwmmxt_wpackhss): Likewise.
9293 (iwmmxt_wpackwss): Likewise.
9294 (iwmmxt_wpackdss): Likewise.
9295 (iwmmxt_wpackhus): Likewise.
9296 (iwmmxt_wpackwus): Likewise.
9297 (iwmmxt_wpackdus): Likewise.
9298 (iwmmxt_wunpckihb): Likewise.
9299 (iwmmxt_wunpckihh): Likewise.
9300 (iwmmxt_wunpckihw): Likewise.
9301 (iwmmxt_wunpckilb): Likewise.
9302 (iwmmxt_wunpckilh): Likewise.
9303 (iwmmxt_wunpckilw): Likewise.
9304 (iwmmxt_wunpckehub): Likewise.
9305 (iwmmxt_wunpckehuh): Likewise.
9306 (iwmmxt_wunpckehuw): Likewise.
9307 (iwmmxt_wunpckehsb): Likewise.
9308 (iwmmxt_wunpckehsh): Likewise.
9309 (iwmmxt_wunpckehsw): Likewise.
9310 (iwmmxt_wunpckelub): Likewise.
9311 (iwmmxt_wunpckeluh): Likewise.
9312 (iwmmxt_wunpckeluw): Likewise.
9313 (iwmmxt_wunpckelsb): Likewise.
9314 (iwmmxt_wunpckelsh): Likewise.
9315 (iwmmxt_wunpckelsw): Likewise.
9316 (ror<mode>3): Likewise.
9317 (ashr<mode>3_iwmmxt): Likewise.
9318 (lshr<mode>3_iwmmxt): Likewise.
9319 (ashl<mode>3_iwmmxt): Likewise.
9320 (ror<mode>3_di): Likewise.
9321 (ashr<mode>3_di): Likewise.
9322 (lshr<mode>3_di): Likewise.
9323 (ashl<mode>3_di): Likewise.
9324 (iwmmxt_wmadds): Likewise.
9325 (iwmmxt_wmaddu): Likewise.
9326 (iwmmxt_tmia): Likewise.
9327 (iwmmxt_tmiaph): Likewise.
9328 (iwmmxt_tmiabb): Likewise.
9329 (iwmmxt_tmiatb): Likewise.
9330 (iwmmxt_tmiabt): Likewise.
9331 (iwmmxt_tmiatt): Likewise.
9332 (iwmmxt_tmovmskb): Likewise.
9333 (iwmmxt_tmovmskh): Likewise.
9334 (iwmmxt_tmovmskw): Likewise.
9335 (iwmmxt_waccb): Likewise.
9336 (iwmmxt_wacch): Likewise.
9337 (iwmmxt_waccw): Likewise.
9338 (iwmmxt_waligni): Likewise.
9339 (iwmmxt_walignr): Likewise.
9340 (iwmmxt_walignr0): Likewise.
9341 (iwmmxt_walignr1): Likewise.
9342 (iwmmxt_walignr2): Likewise.
9343 (iwmmxt_walignr3): Likewise.
9344 (iwmmxt_wsadb): Likewise.
9345 (iwmmxt_wsadh): Likewise.
9346 (iwmmxt_wsadbz): Likewise.
9347 (iwmmxt_wsadhz): Likewise.
9348 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9349 (iwmmxt_wabsdiffb): Likewise.
9350 (iwmmxt_wabsdiffh): Likewise.
9351 (iwmmxt_wabsdiffw): Likewise.
9352 (iwmmxt_waddsubhx): Likewise
9353 (iwmmxt_wsubaddhx): Likewise.
9354 (addc<mode>3): Likewise.
9355 (iwmmxt_avg4): Likewise.
9356 (iwmmxt_avg4r): Likewise.
9357 (iwmmxt_wmaddsx): Likewise.
9358 (iwmmxt_wmaddux): Likewise.
9359 (iwmmxt_wmaddsn): Likewise.
9360 (iwmmxt_wmaddun): Likewise.
9361 (iwmmxt_wmulwsm): Likewise.
9362 (iwmmxt_wmulwum): Likewise.
9363 (iwmmxt_wmulsmr): Likewise.
9364 (iwmmxt_wmulumr): Likewise.
9365 (iwmmxt_wmulwsmr): Likewise.
9366 (iwmmxt_wmulwumr): Likewise.
9367 (iwmmxt_wmulwl): Likewise.
9368 (iwmmxt_wqmulm): Likewise.
9369 (iwmmxt_wqmulwm): Likewise.
9370 (iwmmxt_wqmulmr): Likewise.
9371 (iwmmxt_wqmulwmr): Likewise.
9372 (iwmmxt_waddbhusm): Likewise.
9373 (iwmmxt_waddbhusl): Likewise.
9374 (iwmmxt_wqmiabb): Likewise.
9375 (iwmmxt_wqmiabt): Likewise.
9376 (iwmmxt_wqmiatb): Likewise.
9377 (iwmmxt_wqmiatt): Likewise.
9378 (iwmmxt_wqmiabbn): Likewise.
9379 (iwmmxt_wqmiabtn): Likewise.
9380 (iwmmxt_wqmiatbn): Likewise.
9381 (iwmmxt_wqmiattn): Likewise.
9382 (iwmmxt_wmiabb): Likewise.
9383 (iwmmxt_wmiabt): Likewise.
9384 (iwmmxt_wmiatb): Likewise.
9385 (iwmmxt_wmiatt): Likewise.
9386 (iwmmxt_wmiabbn): Likewise.
9387 (iwmmxt_wmiabtn): Likewise.
9388 (iwmmxt_wmiatbn): Likewise.
9389 (iwmmxt_wmiattn): Likewise.
9390 (iwmmxt_wmiawbb): Likewise.
9391 (iwmmxt_wmiawbt): Likewise.
9392 (iwmmxt_wmiawtb): Likewise.
9393 (iwmmxt_wmiawtt): Likewise.
9394 (iwmmxt_wmiawbbn): Likewise.
9395 (iwmmxt_wmiawbtn): Likewise.
9396 (iwmmxt_wmiawtbn): Likewise.
9397 (iwmmxt_wmiawttn): Likewise.
9398 (iwmmxt_wmerge): Likewise.
9399 (iwmmxt_tandc<mode>3): Likewise.
9400 (iwmmxt_torc<mode>3): Likewise.
9401 (iwmmxt_torvsc<mode>3): Likewise.
9402 (iwmmxt_textrc<mode>3): Likewise.
9403 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9404 (wmmxt_pack): Likewise.
9405 (wmmxt_mult_c1): Likewise.
9406 (wmmxt_mult_c2): Likewise.
9407 (wmmxt_alu_c1): Likewise.
9408 (wmmxt_alu_c2): Likewise.
9409 (wmmxt_alu_c3): Likewise.
9410 (wmmxt_transfer_c1): Likewise.
9411 (wmmxt_transfer_c2): Likewise.
9412 (wmmxt_transfer_c3): Likewise.
9413 (marvell_f_iwmmxt_wstr): Likewise.
9414 (marvell_f_iwmmxt_wldr): Likewise.
9416 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9418 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9420 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9423 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9424 * lra-constraints.c (need_for_split_p): Check call used hard regs
9425 living through calls.
9427 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9428 call used regs for call insn.
9430 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9433 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9434 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9436 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9438 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9440 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9442 * lra-constraints.c (need_for_split_p): Check call used hard regs
9443 living through calls.
9445 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9448 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9449 to tie with any other modes. Eliminate Altivec vector mode tests,
9450 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9451 code, to return 0 if testing MODE2 for a condition, if we've
9452 already tested MODE1 for the same condition.
9454 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9456 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9459 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9461 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9462 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9464 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9466 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9468 (aarch64_symbolic_constant_p): Remove.
9469 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9470 static. Fix line length and white space.
9471 (aarch64_symbolic_constant_p): Remove.
9472 * config/aarch64/predicates.md (aarch64_valid_symref):
9473 Use aarch64_classify_symbol_expression.
9475 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9477 * config/arm/constraints.md (Ts): New constraint.
9478 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9480 (compare_scc): Use "Ts" constraint for operand 0.
9481 (ior_scc_scc): Likewise.
9482 (and_scc_scc): Likewise.
9483 (and_scc_scc_nodom): Likewise.
9484 (ior_scc_scc_cmp): Likewise for operand 7.
9485 (and_scc_scc_cmp): Likewise.
9486 * config/arm/thumb2.md (thumb2_movsi_insn):
9487 Add alternatives for 16-bit encodings.
9488 (thumb2_movhi_insn): Likewise.
9489 (thumb2_movsicc_insn): Likewise.
9490 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9491 (thumb2_negscc): Use "Ts" constraint.
9492 Move mvn instruction outside cond_exec block.
9493 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9494 for 16-bit encodings.
9496 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9498 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9500 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9501 (mulsi3subsi): Likewise.
9502 (mulsidi3adddi): Likewise.
9503 (mulsidi3_v6): Likewise.
9504 (umulsidi3_v6): Likewise.
9505 (umulsidi3adddi_v6): Likewise.
9506 (smulsi3_highpart_v6): Likewise.
9507 (umulsi3_highpart_v6): Likewise.
9508 (mulhisi3tb): Likewise.
9509 (mulhisi3bt): Likewise.
9510 (mulhisi3tt): Likewise.
9511 (maddhisi4): Likewise.
9512 (maddhisi4tb): Likewise.
9513 (maddhisi4tt): Likewise.
9514 (maddhidi4): Likewise.
9515 (maddhidi4tb): Likewise.
9516 (maddhidi4tt): Likewise.
9517 (zeroextractsi_compare0_scratch): Likewise.
9518 (insv_zero): Likewise.
9519 (insv_t2): Likewise.
9520 (anddi_notzesidi_di): Likewise.
9521 (anddi_notsesidi_di): Likewise.
9522 (andsi_notsi_si): Likewise.
9523 (iordi_zesidi_di): Likewise.
9524 (xordi_zesidi_di): Likewise.
9525 (andsi_iorsi3_notsi): Likewise.
9527 (smax_m1): Likewise.
9529 (not_shiftsi): Likewise.
9530 (unaligned_loadsi): Likewise.
9531 (unaligned_loadhis): Likewise.
9532 (unaligned_loadhiu): Likewise.
9533 (unaligned_storesi): Likewise.
9534 (unaligned_storehi): Likewise.
9535 (extv_reg): Likewise.
9536 (extzv_t2): Likewise.
9538 (udivsi3): Likewise.
9539 (arm_zero_extendhisi2addsi): Likewise.
9540 (arm_zero_extendqisi2addsi): Likewise.
9541 (compareqi_eq0): Likewise.
9542 (arm_extendhisi2_v6): Likewise.
9543 (arm_extendqisi2addsi): Likewise.
9544 (arm_movt): Likewise.
9545 (thumb2_ldrd): Likewise.
9546 (thumb2_ldrd_base): Likewise.
9547 (thumb2_ldrd_base_neg): Likewise.
9548 (thumb2_strd): Likewise.
9549 (thumb2_strd_base): Likewise.
9550 (thumb2_strd_base_neg): Likewise.
9551 (arm_negsi2): Add alternative for 16-bit encoding.
9552 (arm_one_cmplsi2): Likewise.
9554 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9556 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9557 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9558 (movdfcc): Likewise.
9559 * config/arm/vfp.md (*thumb2_movsf_vfp):
9560 Disable predication for arm_restrict_it.
9561 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9562 (*thumb2_movdfcc_vfp): Likewise.
9563 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9564 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9565 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9566 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9567 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9568 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9569 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9570 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9571 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9572 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9573 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9574 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9575 Disable predication for arm_restrict_it.
9577 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9579 * config/i386/bmiintrin.h (_bextr_u32): New.
9580 (_bextr_u64): Ditto.
9582 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9584 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9585 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9586 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9587 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9588 * config/mips/n32-elf.h: ...this new file.
9590 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9593 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9594 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9596 2013-06-27 Catherine Moore <clm@codesourcery.com>
9598 * config/mips/mips-tables.opt: Regenerate.
9599 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9600 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9601 * doc/invoke.texi: Add -m14kc.
9603 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9606 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9607 constraints of operand 1 and 2.
9610 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9611 to match RTL canonicalization. Swap predicates and
9612 constraints of operand 1 and 2.
9614 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9616 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9617 Process OP_INOUT regs for splitting too.
9619 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9621 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9622 decl before the loop, initialize to NULL.
9623 (vectorizable_load): Initialize ptr_incr to NULL.
9625 2013-06-27 Martin Jambor <mjambor@suse.cz>
9628 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9629 * ipa-ref.c (ipa_maybe_record_reference): New function.
9630 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9631 * ipa-cp.c (create_specialized_node): Record potential references from
9633 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9635 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9637 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9638 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9639 'mode' to force_reg.
9640 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9641 (aarch64_expand_mov_immediate): Likewise.
9643 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9645 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9646 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9648 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9650 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9651 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9652 (struct machine_function): Add tbegin_p.
9653 (s390_canonicalize_comparison): Fold CC mode compares to
9654 conditional jump if possible.
9655 (s390_emit_jump): Return the emitted jump.
9656 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9657 Handle CCRAWmode compares.
9658 (s390_option_override): Default to -mhtm if available.
9659 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9660 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9661 FPRs instead of df_regs_ever_live_p.
9662 (s390_optimize_nonescaping_tx): New function.
9663 (s390_init_frame_layout): Extend clobbered_regs array to cover
9665 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9666 (s390_expand_tbegin): New function.
9667 (enum s390_builtin): New enum definition.
9668 (code_for_builtin): New array definition.
9669 (s390_init_builtins): New function.
9670 (s390_expand_builtin): New function.
9671 (TARGET_INIT_BUILTINS): Define.
9672 (TARGET_EXPAND_BUILTIN): Define.
9673 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9674 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9675 (s390_alc_comparison): Likewise.
9676 * config/s390/s390-modes.def: Add CCRAWmode.
9677 * config/s390/s390.h (processor_flags): Add PF_TX.
9678 (TARGET_CPU_HTM): Define macro.
9679 (TARGET_HTM): Define macro.
9680 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9681 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9682 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9683 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9684 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9685 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9686 ("*cc_to_int"): Move up.
9687 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9688 constants other than 0.
9689 ("*ccraw_to_int"): New insn and splitter definition.
9690 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9691 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9692 ("tx_assist"): New expander.
9693 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9694 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9695 * config/s390/s390.opt: Add -mhtm option.
9696 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9697 * config/s390/htmxlintrin.h: New file.
9698 * config/s390/htmintrin.h: New file.
9699 * config/s390/s390intrin.h: New file.
9700 * doc/extend.texi: Document htm builtins.
9701 * config.gcc: Add the new header files to extra_headers.
9703 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9705 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9706 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9708 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9709 Pat Haugen <pthaugen@us.ibm.com>
9710 Peter Bergner <bergner@vnet.ibm.com>
9712 * config/rs6000/power8.md: New.
9713 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9714 setting for power8 entry.
9715 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9716 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9717 test for Power4/Power5 only.
9718 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9720 (force_new_group): Adjust comment.
9721 * config/rs6000/rs6000.md: Include power8.md.
9723 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9725 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9726 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9728 (tune_params): Update comment.
9729 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9730 (arm_max_conditional_execute): New function.
9731 (thumb2_final_prescan_insn): Use max_insn_skipped and
9732 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9734 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9736 PR tree-optimization/57705
9737 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9738 SSA_NAME step, provided that it is not defined inside the loop.
9739 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9740 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9742 2013-06-25 Martin Jambor <mjambor@suse.cz>
9745 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9746 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9748 (ipa_note_param_call): Initialize member_ptr flag.
9749 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9750 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9751 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9752 (ipa_read_indirect_edge_info): Likewise.
9754 2013-06-25 Richard Biener <rguenther@suse.de>
9757 * passes.c (init_optimization_passes): Move pass_fold_builtins
9758 and pass_dce earlier with -Og.
9760 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9762 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9763 <BIT_FIELD_REF>: Remove trailing TAB.
9764 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9767 2013-06-24 Martin Jambor <mjambor@suse.cz>
9769 PR tree-optimization/57358
9770 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9771 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9772 (ipa_analyze_params_uses): Generate pessimistic info when true.
9774 2013-06-24 Martin Jambor <mjambor@suse.cz>
9776 PR tree-optimization/57539
9777 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9778 global.inlined_to of the new node to it. All callers changed.
9779 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9780 inlining_into, pass it to cgraph_clone_node.
9781 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9782 ipa_free_edge_args_substructures.
9783 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9784 rdesc linked list. Do not assert rdesc edges have inlined caller.
9785 Assert we have found an rdesc in the rdesc list.
9787 2013-06-24 Richard Biener <rguenther@suse.de>
9789 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9790 (pointer_set_lookup): Declare.
9791 (class pointer_map): New template class implementing a
9792 generic pointer to T map.
9793 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9794 pointer_map<T>::contains, pointer_map<T>::insert,
9795 pointer_map<T>::traverse): New functions.
9796 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9797 (pointer_set_lookup): New function.
9798 (pointer_set_contains): Use pointer_set_lookup.
9799 (pointer_set_insert): Likewise.
9800 (insert_aux): Remove.
9801 (struct pointer_map_t): Embed a pointer_set_t.
9802 (pointer_map_create): Adjust.
9803 (pointer_map_destroy): Likewise.
9804 (pointer_map_contains): Likewise.
9805 (pointer_map_insert): Likewise.
9806 (pointer_map_traverse): Likewise.
9807 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9808 pointer_map<unsigned> instead of a pointer_map_t.
9809 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9810 (streamer_tree_cache_lookup): Likewise.
9811 (streamer_tree_cache_create): Likewise.
9812 (streamer_tree_cache_delete): Likewise.
9813 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9814 pointer_map<unsigned> instead of a pointer_map_t.
9815 (lto_init_tree_ref_encoder): Adjust.
9816 (lto_destroy_tree_ref_encoder): Likewise.
9817 * lto-section-out.c (lto_output_decl_index): Likewise.
9818 (lto_record_function_out_decl_state): Likewise.
9819 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9821 2013-06-24 Richard Biener <rguenther@suse.de>
9823 PR tree-optimization/57488
9824 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9826 2013-06-24 Alan Modra <amodra@gmail.com>
9828 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9829 (gen_easy_altivec_constant): Likewise.
9830 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9831 easy_vector_constant_msb): Likewise.
9833 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9836 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9837 add missing return true.
9839 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9842 * config/sh/predicates.md (general_extend_operand): Invoke
9843 general_movsrc_operand for memory operands.
9844 (general_movsrc_operand): Allow reg+reg addressing, do not use
9845 general_operand for memory operands.
9847 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9849 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9850 when current target options does not apply.
9851 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9852 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9853 * config/i386/bmiintrin.h: Pass appropriate target
9854 attributes to header.
9855 * config/i386/mmintrin.h: Ditto.
9856 * config/i386/nmmintrin.h: Ditto.
9857 * config/i386/avx2intrin.h: Ditto.
9858 * config/i386/fxsrintrin.h: Ditto.
9859 * config/i386/tbmintrin.h: Ditto.
9860 * config/i386/xsaveintrin.h: Ditto.
9861 * config/i386/f16cintrin.h: Ditto.
9862 * config/i386/xtestintrin.h: Ditto.
9863 * config/i386/xsaveoptintrin.h: Ditto.
9864 * config/i386/bmi2intrin.h: Ditto.
9865 * config/i386/lzcntintrin.h: Ditto.
9866 * config/i386/smmintrin.h: Ditto.
9867 * config/i386/wmmintrin.h: Ditto.
9868 * config/i386/x86intrin.h: Remove all header include guards.
9869 * config/i386/prfchwintrin.h: Ditto.
9870 * config/i386/pmmintrin.h: Ditto.
9871 * config/i386/tmmintrin.h: Ditto.
9872 * config/i386/xmmintrin.h: Ditto.
9873 * config/i386/popcntintrin.h: Ditto.
9874 * config/i386/rdseedintrin.h: Ditto.
9875 * config/i386/ammintrin.h: Ditto.
9876 * config/i386/emmintrin.h: Ditto.
9877 * config/i386/immintrin.h: Remove all header include guards.
9878 * config/i386/fma4intrin.h: Ditto.
9879 * config/i386/lwpintrin.h: Ditto.
9880 * config/i386/xopintrin.h: Ditto.
9881 * config/i386/ia32intrin.h: Ditto.
9882 * config/i386/avxintrin.h: Ditto.
9883 * config/i386/rtmintrin.h: Ditto.
9884 * config/i386/fmaintrin.h: Ditto.
9885 * config/i386/mm3dnow.h: Ditto.
9887 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9889 * common/config/i386/i386-common.c: Handle LZCNT.
9891 2013-06-22 Andi Kleen <ak@linux.intel.com>
9893 * doc/extend.texi: Use __atomic_store_n instead of
9894 __atomic_store in HLE example.
9896 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9898 * config/sh/sh.c: Remove <cstdlib> workaround.
9900 2013-06-21 Andi Kleen <ak@linux.intel.com>
9902 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9904 2013-06-21 Andi Kleen <ak@linux.intel.com>
9906 * doc/extend.texi: Document that __atomic_clear and
9907 __atomic_test_and_set should only be used with bool.
9909 2013-06-20 Jan Hubicka <jh@suse.cz>
9911 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9913 * tree.c (decls_same_for_odr): New function.
9914 (same_for_edr): New function.
9915 (types_same_for_odr): New function.
9916 (get_binfo_at_offset): Use it.
9917 * tree.h (types_same_for_odr): Declare.
9919 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9920 Jason Merrill <jason@redhat.com>
9922 * system.h: Include <cstdlib> as well as <stdlib.h>.
9924 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9927 * config/i386/i386.c (construct_container): Report error if
9928 long double is used with disabled x87 float returns.
9930 2013-06-20 Jan Hubicka <jh@suse.cz>
9932 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9934 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9936 PR rtl-optimization/57425
9937 PR rtl-optimization/57569
9938 * alias.c (write_dependence_p): Remove parameters mem_mode and
9939 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9940 Changed all callers.
9941 (canon_anti_dependence): Get comments and semantics in sync.
9942 Add parameter mem_canonicalized. Changed all callers.
9943 * rtl.h (canon_anti_dependence): Update prototype.
9945 2013-06-20 Richard Biener <rguenther@suse.de>
9947 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9948 byte case, inline streamer_read_uchar and defer section
9951 2013-06-20 Richard Biener <rguenther@suse.de>
9953 PR tree-optimization/57584
9954 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
9955 SSA names into the expanded expression that take part in
9956 abnormal coalescing.
9958 2013-06-19 Sharad Singhai <singhai@google.com>
9960 * gcov.c (print_usage): Handle new option.
9961 (process_args): Ditto.
9962 (get_gcov_intermediate_filename): New function.
9963 (output_intermediate_file): New function.
9964 (output_gcov_file): New function
9965 (generate_results): Handle new option.
9966 (release_function): Relase demangled name.
9967 (read_graph_file): Handle demangled name.
9968 (output_lines): Ditto.
9969 * doc/gcov.texi: Document gcov intermediate format.
9971 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
9974 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
9975 (lra_emit_add): Use the functions. Add comment about Y as an
9978 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
9981 * collect2.c (collect_atexit): New.
9982 (collect_exit): Delete.
9983 (main): Register collect_atexit with atexit.
9984 (collect_wait): Change collect_exit to exit.
9986 * collect2.h (collect_exit): Delete.
9987 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
9989 2013-06-19 Wei Mi <wmi@google.com>
9991 PR rtl-optimization/57518
9992 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
9993 if regno is used in paradoxical subreg.
9994 (update_equiv_regs): Check pdx_subregs[regno] before
9995 set a reg to be equivalent with a mem.
9997 2013-06-19 Matthias Klose <doko@ubuntu.com>
10000 * file-find.h (find_a_file): Add a mode parameter.
10001 * file-find.c (find_a_file): Likewise.
10002 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10003 with X_OK for the executables.
10004 * collect2.c (main): Call find_a_file with X_OK.
10006 2013-06-19 Steve Ellcey <sellcey@mips.com>
10009 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10010 Use NEXT_INSN instead of next_real_insn.
10012 2013-06-19 Jan Hubicka <jh@suse.cz>
10014 * cgraph.h (const_value_known_p): Replace by ...
10015 (ctor_for_folding): .. this one.
10016 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10017 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10018 * expr.c (expand_expr_real_1): Likewise.
10019 (string_constant): Likewise.
10020 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10021 * ipa.c (process_references): Likewise.
10022 (symtab_remove_unreachable_nodes): Likewise.
10023 * ipa-inline-analysis.c (param_change_prob): Likewise.
10024 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10025 (get_base_constructor): Likwise.
10026 * varpool.c (varpool_remove_node): Likewise.
10027 (varpool_remove_initializer): LIkewise.
10028 (dump_varpool_node): LIkwise.
10029 (const_value_known_p): Rewrite to ...
10030 (ctor_for_folding): ... this one.
10032 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10035 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10036 PERSONALITY in $PATH derived prefixes.
10038 2013-06-19 Jeff Law <law@redhat.com>
10040 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10043 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10044 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10047 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10049 * config/arm/vfp.md: Move VFP instruction classification documentation
10051 * config/arm/arm.md: ... here. Update instruction classification
10054 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10056 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10057 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10058 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10060 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10063 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10064 with NEXT_INSN. Use tablejump_p to check for jump table data
10067 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10070 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10071 that now in C++ the value is correct per the C++ standards.
10073 2013-06-19 Richard Biener <rguenther@suse.de>
10075 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10076 for global context.
10078 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10081 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10084 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10085 with next_active_insn.
10087 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10089 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10090 functions are inlined during failures to flag an error.
10091 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10092 in early inline pass.
10094 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10096 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10099 2013-06-18 Julian Brown <julian@codesourcery.com>
10101 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10102 Permit virtual register pre-reload if !strict.
10103 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10105 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10107 * config/arm/neon.md (movmisalign<mode>): Use
10108 neon_perm_struct_or_reg_operand instead of
10109 neon_struct_or_register_operand.
10110 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10111 neon_permissive_struct_operand instead of neon_struct_operand.
10112 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10113 neon_vector_mem_operand.
10114 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10115 neon_vector_mem_operand.
10116 (neon_permissive_struct_operand): New.
10117 (neon_struct_or_register_operand): Rename to...
10118 (neon_perm_struct_or_reg_operand): This. Adjust call to
10119 neon_vector_mem_operand.
10121 2013-06-18 Richard Biener <rguenther@suse.de>
10123 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10124 * lto-streamer.h: Include pointer-set.h.
10125 (struct lto_decl_slot): Remove.
10126 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10127 Remove next_index entry.
10128 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10129 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10130 (lto_init_tree_ref_encoder): Adjust.
10131 (lto_destroy_tree_ref_encoder): Likewise.
10132 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10133 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10134 (lto_output_decl_index): Adjust.
10135 (lto_new_out_decl_state): Likewise.
10136 (lto_record_function_out_decl_state): Likewise.
10137 * lto-streamer-out.c (copy_function): Likewise.
10139 2013-06-18 Richard Biener <rguenther@suse.de>
10141 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10142 * cgraphunit.c: Include cfgloop.h.
10143 (init_lowered_empty_function): Initialize the loop tree.
10144 (assemble_thunk): Insert new BBs into loops.
10146 2013-06-18 Richard Biener <rguenther@suse.de>
10148 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10149 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10150 the map from cache entry to cache index optional.
10151 (streamer_tree_cache_replace_tree): Adjust accordingly.
10152 (streamer_tree_cache_append): Likewise.
10153 (streamer_tree_cache_delete): Likewise.
10154 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10155 streamer cache map from cache entry to cache index.
10156 * lto-streamer-out.c (create_output_block): Adjust.
10158 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10160 * config/arm/arm.md (attribute "insn"): Move multiplication and
10161 division attributes to...
10162 (attribute "type"): ... here. Remove mult.
10163 (attribute "mul32"): New attribute.
10164 (attribute "mul64"): Add umaal.
10165 (*arm_mulsi3): Update attributes.
10166 (*arm_mulsi3_v6): Likewise.
10167 (*thumb_mulsi3): Likewise.
10168 (*thumb_mulsi3_v6): Likewise.
10169 (*mulsi3_compare0): Likewise.
10170 (*mulsi3_compare0_v6): Likewise.
10171 (*mulsi_compare0_scratch): Likewise.
10172 (*mulsi_compare0_scratch_v6): Likewise.
10173 (*mulsi3addsi): Likewise.
10174 (*mulsi3addsi_v6): Likewise.
10175 (*mulsi3addsi_compare0): Likewise.
10176 (*mulsi3addsi_compare0_v6): Likewise.
10177 (*mulsi3addsi_compare0_scratch): Likewise.
10178 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10179 (*mulsi3subsi): Likewise.
10180 (*mulsidi3adddi): Likewise.
10181 (*mulsi3addsi_v6): Likewise.
10182 (*mulsidi3adddi_v6): Likewise.
10183 (*mulsidi3_nov6): Likewise.
10184 (*mulsidi3_v6): Likewise.
10185 (*umulsidi3_nov6): Likewise.
10186 (*umulsidi3_v6): Likewise.
10187 (*umulsidi3adddi): Likewise.
10188 (*umulsidi3adddi_v6): Likewise.
10189 (*smulsi3_highpart_nov6): Likewise.
10190 (*smulsi3_highpart_v6): Likewise.
10191 (*umulsi3_highpart_nov6): Likewise.
10192 (*umulsi3_highpart_v6): Likewise.
10193 (mulhisi3): Likewise.
10194 (*mulhisi3tb): Likewise.
10195 (*mulhisi3bt): Likewise.
10196 (*mulhisi3tt): Likewise.
10197 (maddhisi4): Likewise.
10198 (*maddhisi4tb): Likewise.
10199 (*maddhisi4tt): Likewise.
10200 (maddhidi4): Likewise.
10201 (*maddhidi4tb): Likewise.
10202 (*maddhidi4tt): Likewise.
10203 (divsi3): Likewise.
10204 (udivsi3): Likewise.
10205 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10206 (thumb2_mulsi_short_compare0): Likewise.
10207 (thumb2_mulsi_short_compare0_scratch): Likewise.
10208 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10209 (1020mult2): Likewise.
10210 (1020mult3): Likewise.
10211 (1020mult4): Likewise.
10212 (1020mult5): Likewise.
10213 (1020mult6): Likewise.
10214 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10216 (cortex_a15_mult64): Likewise.
10217 (cortex_a15_sdiv): Likewise.
10218 (cortex_a15_udiv): Likewise.
10219 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10225 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10226 (pj4_ir_div): Likewise.
10227 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10228 (11_mult2): Likewise.
10229 (11_mult3): Likewise.
10230 (11_mult4): Likewise.
10231 (11_mult5): Likewise.
10232 (11_mult6): Likewise.
10233 (11_mult7): Likewise.
10234 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10235 (cortex_a8_mla): Likewise.
10236 (cortex_a8_mull): Likewise.
10237 (cortex_a8_smulwy): Likewise.
10238 (cortex_a8_smlald): Likewise.
10239 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10240 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10241 (cortex_r4_mul_3): Likewise.
10242 (cortex_r4_mla_4): Likewise.
10243 (cortex_r4_mla_3): Likewise.
10244 (cortex_r4_smlald): Likewise.
10245 (cortex_r4_mull): Likewise.
10246 (cortex_r4_sdiv): Likewise.
10247 (cortex_r4_udiv): Likewise.
10248 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10249 (cortex_a7_idiv): Likewise.
10250 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10251 (9_mult2): Likewise.
10252 (9_mult3): Likewise.
10253 (9_mult4): Likewise.
10254 (9_mult5): Likewise.
10255 (9_mult6): Likewise.
10256 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10257 (cortex_a53_sdiv): Likewise.
10258 (cortex_a53_udiv): Likewise.
10259 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10260 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10261 (mp626_mult2): Likewise.
10262 (mp626_mult3): Likewise.
10263 (mp626_mult4): Likewise.
10264 * config/arm/fa526.md (526_mult1): Update attribute change.
10265 (526_mult2): Likewise.
10266 * config/arm/arm-generic.md (mult): Update attribute change.
10267 (mult_ldsched_strongarm): Likewise.
10268 (mult_ldsched): Likewise.
10269 (multi_cycle): Likewise.
10270 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10271 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10272 (606te_mult2): Likewise.
10273 (606te_mult3): Likewise.
10274 (606te_mult4): Likewise.
10275 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10276 (cortex_a9_mac16): Likewise.
10277 (cortex_a9_multiply): Likewise.
10278 (cortex_a9_mac): Likewise.
10279 (cortex_a9_multiply_long): Likewise.
10280 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10281 (626te_mult2): Likewise.
10282 (626te_mult3): Likewise.
10283 (626te_mult4): Likewise.
10285 2013-06-18 Richard Biener <rguenther@suse.de>
10288 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10290 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10293 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10294 with next_active_insn.
10296 2013-06-18 Alan Modra <amodra@gmail.com>
10298 * config/rs6000/rs6000.h (enum data_align): New.
10299 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10300 (DATA_ABI_ALIGNMENT): Define.
10301 (CONSTANT_ALIGNMENT): Correct comment.
10302 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10303 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10305 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10307 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10308 ATTRIBUTE_UNUSED marking.
10310 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10312 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10313 alternative and update.
10314 (aarch64_dup_lanedi): Delete.
10315 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10316 * config/aarch64/aarch64-simd-builtins.def: Update.
10318 2013-06-17 Richard Biener <rguenther@suse.de>
10320 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10321 (lto_input_scc): Declare.
10322 (lto_input_tree_1): Likewise.
10323 (struct lto_stats_d): Add num_tree_bodies_output and
10324 num_pickle_refs_output.
10325 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10326 (lto_read_tree_1): Split out from ...
10327 (lto_read_tree): ... this.
10328 (lto_input_scc): New function.
10329 (lto_input_tree_1): Split out from ...
10330 (lto_input_tree): ... this. Handle LTO_tree_scc.
10331 (lto_data_in_create): Create the streamer cache without hashes.
10332 * lto-streamer-out.c (create_output_block): Create the streamer
10333 cache with hashes when not doing WPA.
10334 (lto_write_tree_1): Split out from ...
10335 (lto_write_tree): ... this.
10336 (get_symbol_initial_value): New function.
10337 (lto_output_tree_1): Split out from ...
10338 (lto_output_tree): ... this. Write trees as series of SCCs
10339 using a DFS walk via DFS_write_tree.
10340 (struct sccs, struct scc_entry): New types.
10341 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10342 (DFS_write_tree_body): New function.
10343 (DFS_write_tree): Likewise.
10344 (hash_tree): Likewise.
10345 (scc_entry_compare): Likewise.
10346 (hash_scc): Likewise.
10347 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10348 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10349 TREE_CHAIN as regular reference.
10350 (streamer_read_integer_cst): Remove.
10351 (streamer_get_pickled_tree): Adjust.
10352 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10353 of DECL_EXTERNALs in BLOCK_VARS for now.
10354 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10356 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10357 Add hash value argument and record that if hashes are recorded
10359 (streamer_tree_cache_insert_1): Adjust.
10360 (streamer_tree_cache_insert): Likewise.
10361 (streamer_tree_cache_insert_at): Rename to ...
10362 (streamer_tree_cache_replace_tree): ... this and adjust.
10363 (streamer_tree_cache_append): Adjust.
10364 (record_common_node): Likewise.
10365 (streamer_tree_cache_create): Add argument whether to
10366 record hash values together with trees.
10367 (streamer_tree_cache_delete): Adjust.
10368 * tree-streamer.h (struct streamer_tree_cache_d): Add
10370 (streamer_read_integer_cst): Remove.
10371 (streamer_tree_cache_insert): Adjust.
10372 (streamer_tree_cache_append): Likewise.
10373 (streamer_tree_cache_insert_at): Rename to ...
10374 (streamer_tree_cache_replace_tree): ... this and adjust.
10375 (streamer_tree_cache_create): Add argument whether to record hashes.
10376 (streamer_tree_cache_get): Rename to ...
10377 (streamer_tree_cache_get_tree): ... this.
10378 (streamer_tree_cache_get_hash): New function.
10379 * tree.c (cache_integer_cst): New function.
10380 * tree.h (cache_integer_cst): Declare.
10381 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10382 * lto-symtab.c (lto_varpool_replace_node): Only release
10383 DECL_INITIAL of non-prevailing decls.
10384 * varpool.c (varpool_remove_initializer): Do not release
10385 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10387 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10389 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10390 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10391 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10392 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10393 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10394 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10395 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10396 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10397 instead of TARGET_64BIT.
10398 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10399 Require ISA_HAS_<D>DIV.
10401 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10403 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10404 (mips*-*-linux*): Move default with_llsc setting to where other
10406 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10408 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10409 Likewise. Remove default with_tune setting. Move default float
10410 setting to its own block. Handle with_llsc in the same block as above.
10412 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10414 PR rtl-optimization/57425
10415 PR rtl-optimization/57569
10416 * alias.c (write_dependence_p): Add new parameters mem_mode,
10417 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10418 Changed all callers.
10419 (canon_anti_dependence): New function.
10420 * cse.c (check_dependence): Use canon_anti_dependence.
10421 * cselib.c (cselib_invalidate_mem): Likewise.
10422 * rtl.h (canon_anti_dependence): Declare.
10424 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10426 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10427 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10428 ".set mips3" for 64-bit targets.
10430 2013-06-15 Dehao Chen <dehao@google.com>
10432 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10433 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10434 (gimple_check_call_args): Likewise.
10435 * value-prof.c (check_ic_target): Likewise.
10436 * ipa-inline.c (early_inliner): Likewise.
10437 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10438 * cgraph.c (cgraph_create_edge_1): Likewise.
10439 (cgraph_make_edge_direct): Likewise.
10441 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10444 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10445 rs6000_output_move_128bit to handle emitting quad memory
10446 operations. Set attribute length to 8 bytes.
10448 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10450 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10452 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10453 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10454 (aarch64_<su>mlsl<mode>): Likewise.
10456 2013-06-14 Mike Stump <mikestump@comcast.net>
10458 * Makefile.in (TARGET_H): Add insn-codes.h.
10460 2013-06-14 Alan Modra <amodra@gmail.com>
10462 PR middle-end/57134
10463 PR middle-end/57586
10464 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10465 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10466 bitfield expansion when EXPAND_MEMORY.
10467 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10469 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10471 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10472 test for clearing quad memory on 32-bit later.
10474 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10476 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10477 (fold_negate_expr): Likewise.
10478 (fold_real_zero_addition_p): Handle vectors.
10479 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10481 2013-06-14 Alan Modra <amodra@gmail.com>
10483 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10485 2013-06-13 Jan Hubicka <jh@suse.cz>
10487 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10488 Local comdats are not externally visible.
10489 * symtab.c (dump_symtab_base): Dump externally visible.
10490 (verify_symtab_base): Verify back links in the symtab hash.
10492 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10494 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10495 CONVERT_EXPR as equal nodes.
10497 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10499 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10501 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10503 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10504 Generalize to complex and vector.
10505 * tree.c (build_all_ones_cst): New function.
10506 * tree.h (build_all_ones_cst): Declare it.
10508 2013-06-13 Alan Modra <amodra@gmail.com>
10510 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10511 * config/rs6000/rs6000.md (signbittf2): New insn.
10512 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10513 (abstf2_internal, cmptf_internal2): Likewise.
10514 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10516 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10517 Pat Haugen <pthaugen@us.ibm.com>
10518 Peter Bergner <bergner@vnet.ibm.com>
10520 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10521 power8 byte, half-word, and quad-word atomic instructions.
10522 (emit_store_conditional): Likewise.
10523 (rs6000_expand_atomic_compare_and_swap): Likewise.
10524 (rs6000_expand_atomic_op): Likewise.
10526 * config/rs6000/sync.md (larx): Add new modes for power8.
10528 (AINT): New mode iterator to include TImode as well as normal
10529 integer modes on power8.
10530 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10531 that VSX registers are not considered. Use AINT mode iterator
10532 instead of INT1 to allow inclusion of quad word atomic operations
10534 (load_locked<mode>): Likewise.
10535 (store_conditional<mode>): Likewise.
10536 (atomic_compare_and_swap<mode>): Likewise.
10537 (atomic_exchange<mode>): Likewise.
10538 (atomic_nand<mode>): Likewise.
10539 (atomic_fetch_<fetchop_name><mode>): Likewise.
10540 (atomic_nand_fetch<mode>): Likewise.
10541 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10543 (ATOMIC): On power8, add QImode, HImode modes.
10544 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10545 modes that promote to SImode.
10546 (load_lockedti): Convert TImode arguments to PTImode, so that we
10547 get a guaranteed even/odd register pair.
10548 (load_lockedpti): Likewise.
10549 (store_conditionalti): Likewise.
10550 (store_conditionalpti): Likewise.
10552 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10553 atomic load/store instructions.
10556 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10558 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10560 (insn_count): New attribute, with most cases extracted from...
10561 (length): ...here. Redefine most cases in terms of insn_count.
10562 (single_insn): Delete.
10563 (can_delay): Use insn_count to check for single instructions.
10564 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10565 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10566 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10567 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10568 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10569 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10570 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10571 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10572 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10573 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10574 rather than "length".
10575 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10576 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10577 Use "insn_count" rather than "length".
10578 * config/mips/mips-dsp.md
10579 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10580 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10583 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10585 PR tree-optimization/57361
10586 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10588 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10590 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10592 (aarch64_simd_combine<mode>): New instruction expansion.
10593 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10594 function prototype.
10595 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10596 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10598 2013-06-12 Jan Hubicka <jh@suse.cz>
10600 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10601 decl has when in streaming stage.
10602 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10603 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10605 2013-06-12 Roland Stigge <stigge@antcom.de>
10608 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10610 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10612 PR tree-optimization/57537
10613 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10614 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10616 2013-06-12 Richard Biener <rguenther@suse.de>
10618 * data-streamer.h (streamer_write_char_stream): CSE
10619 obs->current_pointer.
10620 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10621 streamer_write_char_stream manually and optimize the resulting loop.
10622 (streamer_write_hwi_stream): Likewise.
10624 2013-06-12 Jan Hubicka <jh@suse.cz>
10626 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10627 * cgraph.h (varpool_create_empty_node): Declare.
10628 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10630 * symtab.c (symtab_unregister_node): Be lax about missin entries
10632 (symtab_get_node): Update comment.
10633 * varpool.c (varpool_create_empty_node): Break out from ...
10634 (varpool_node_for_decl): ... here.
10635 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10637 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10639 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10640 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10641 part. Use straight-line flow at the end.
10642 <COMPONENT_REF>: Remove superfluous else.
10643 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10645 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10648 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10649 target hook even for !TREE_PUBLIC decls. If no resolution info
10650 is available, return false for common and external decls.
10652 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10654 * config/rl78/constraints.md (U): New constraint.
10655 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10658 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10661 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10662 to allow returning address to AT_PLATFORM name.
10664 2013-06-11 Jan Hubicka <jh@suse.cz>
10666 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10667 * cgraph.h (symtab_node_base): Add weakref flag.
10668 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10669 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10670 (output_weakrefs): Use weakref flag.
10671 * fold-const.c (simple_operand_p): Handle WEAK.
10672 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10673 * ipa.c (varpool_externally_visible_p): Drop weakref.
10674 (function_and_variable_visibility): Update comment; fix weakref
10675 sanity checks; do not clear DECL_WEAK on them.
10676 * lto-cgraph.c (lto_output_node): update.
10677 (lto_output_varpool_node): Update.
10678 (input_overwrite_node): Update.
10679 (input_node): Update.
10680 (input_varpool_node): Update.
10681 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10682 (lto_symtab_merge_symbols): Add sanity check.
10683 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10684 * passes.c (rest_of_decl_compilation): Set static flag, too.
10685 * symtab.c (dump_symtab_base): Dump weakref.
10686 (verify_symtab_base): Sanity check weakrefs.
10687 (symtab_make_decl_local): Remove duplicated code.
10688 (symtab_alias_ultimate_target): Simplify.
10689 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10691 2013-06-11 Tom de Vries <tom@codesourcery.com>
10693 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10694 sequence_vect == NULL.
10696 2013-06-11 DJ Delorie <dj@redhat.com>
10698 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10699 (rl78_unwind_word_mode): New.
10701 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10703 * final.c (debug_prefix_maps): Make static.
10705 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10707 * function.c (initial_trampoline): Remove stray copy.
10709 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10711 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10713 2013-06-11 Martin Jambor <mjambor@suse.cz>
10715 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10716 within bounds at the beginning of the function.
10718 2013-06-11 Alan Modra <amodra@gmail.com>
10720 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10722 (get_named_section): Don't NULL !DECL_P decl.
10724 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10726 * doc/invoke.texi (core-avx2): Document.
10728 (atom): Updated with MOVBE.
10730 2013-06-11 Richard Biener <rguenther@suse.de>
10732 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10734 2013-06-11 Anton Blanchard <anton@samba.org>
10736 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10737 correct shift value in little-endian mode.
10739 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10742 * varasm.c (get_variable_align): Move #endif to the right place.
10744 2013-06-10 Cary Coutant <ccoutant@google.com>
10746 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10747 for hash so that hash table traversal order is deterministic.
10749 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10750 Pat Haugen <pthaugen@us.ibm.com>
10751 Peter Bergner <bergner@vnet.ibm.com>
10753 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10754 of vectors in GPRS if they are direct moves or quad word load or
10757 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10759 (direct_move_p): Likewise.
10760 (quad_load_store_p): Likewise.
10762 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10763 classes into bins based on the physical register type.
10764 (reg_class_to_reg_type): Likewise.
10765 (IS_STD_REG_TYPE): Likewise.
10766 (IS_FP_VECT_REG_TYPE): Likewise.
10767 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10768 use direct move instructions.
10769 (reload_gpr_vsx): Likewise.
10770 (reload_vsx_gpr): Likewise.
10771 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10772 information that is a simplification of register classes. Also
10773 precalculate direct move reload helpers.
10774 (direct_move_p): New function to return true if the operation can
10775 be done as a direct move instruciton.
10776 (quad_load_store_p): New function to return true if the operation
10777 is a quad memory operation.
10778 (rs6000_legitimize_address): If quad memory, only allow register
10779 indirect for TImode addresses.
10780 (rs6000_legitimate_address_p): Likewise.
10781 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10782 (rs6000_reload_register_type): Likewise.
10783 (register_to_reg_type): Return register type.
10784 (rs6000_secondary_reload_simple_move): New helper function for
10785 secondary reload and secondary memory needed to identify anything
10786 that is a simple move, and does not need reloading.
10787 (rs6000_secondary_reload_direct_move): New helper function for
10788 secondary reload to identify cases that can be done with several
10789 instructions via the direct move instructions.
10790 (rs6000_secondary_reload_move): New helper function for secondary
10791 reload to identify moves between register types that can be done.
10792 (rs6000_secondary_reload): Add support for quad memory operations
10793 and for direct move.
10794 (rs6000_secondary_memory_needed): Likewise.
10795 (rs6000_debug_secondary_memory_needed): Change argument names.
10796 (rs6000_output_move_128bit): New function to return the move to
10797 use for 128-bit moves, including knowing about the various
10798 limitations of quad memory operations.
10800 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10801 memory operations. call rs6000_output_move_128bit for the actual
10802 instruciton(s) to generate.
10803 (vsx_movti_64bit): Likewise.
10805 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10806 (UNSPEC_P8V_MTVSRWZ): Likewise.
10807 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10808 (UNSPEC_P8V_MTVSRD): Likewise.
10809 (UNSPEC_P8V_XXPERMDI): Likewise.
10810 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10811 (UNSPEC_FUSION_GPR): Likewise.
10812 (FMOVE128_GPR): New iterator for direct move.
10813 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10814 (f32_sv): Likewise.
10815 (f32_dm): Likewise.
10816 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10817 loads and direct move instructions.
10818 (zero_extendsidi2_lfiwzx): Likewise.
10819 (extendsidi2_lfiwax): Likewise.
10820 (extendsidi2_nocell): Likewise.
10821 (floatsi<mode>2_lfiwax): Likewise.
10822 (lfiwax): Likewise.
10823 (floatunssi<mode>2_lfiwzx): Likewise.
10824 (lfiwzx): Likewise.
10825 (fix_trunc<mode>_stfiwx): Likewise.
10826 (fixuns_trunc<mode>_stfiwx): Likewise.
10827 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10828 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10829 (parity<mode>2_cmpb): Set length/type attr.
10830 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10831 for 'mr.' to fast_compare.
10832 (bpermd_<mode>): Change type attr to popcnt.
10833 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10834 (p8_mtvsrwz_1): Likewise.
10835 (p8_mtvsrwz_2): Likewise.
10836 (reload_fpr_from_gpr<mode>): Likewise.
10837 (p8_mtvsrd_1): Likewise.
10838 (p8_mtvsrd_2): Likewise.
10839 (p8_xxpermdi_<mode>): Likewise.
10840 (reload_vsx_from_gpr<mode>): Likewise.
10841 (reload_vsx_from_gprsf): Likewise.
10842 (p8_mfvsrd_3_<mode>): LIkewise.
10843 (reload_gpr_from_vsx<mode>): Likewise.
10844 (reload_gpr_from_vsxsf): Likewise.
10845 (p8_mfvsrd_4_disf): Likewise.
10846 (multi-word GPR splits): Do not split direct moves or quad memory
10849 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10851 * tree-into-ssa.c (interesting_blocks): Make static.
10853 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10856 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10857 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10858 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10859 (get_variable_align): New function.
10860 (get_variable_section, emit_bss, emit_common,
10861 assemble_variable_contents, place_block_symbol): Use
10862 get_variable_align instead of DECL_ALIGN.
10863 (assemble_noswitch_variable): Add align argument, use it
10864 instead of DECL_ALIGN.
10865 (assemble_variable): Adjust caller. Use get_variable_align
10866 instead of DECL_ALIGN.
10867 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10869 (DATA_ABI_ALIGNMENT): Define.
10870 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10871 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10872 opt is false, only return the psABI mandated alignment increase.
10873 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10874 (DATA_ABI_ALIGNMENT): ... this.
10875 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10876 (DATA_ABI_ALIGNMENT): ... this.
10877 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10878 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10879 (DATA_ABI_ALIGNMENT): ... this.
10880 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10881 * doc/tm.texi: Regenerated.
10883 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10885 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10886 cmp_code to construct REG_EQUAL note.
10888 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10891 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10892 that operands[2] doesn't overlap with operands[0].
10894 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10895 Jan Hubicka <jh@suse.cz>
10897 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10898 hack to mark symbols as used.
10900 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10902 PR rtl-optimization/57559
10903 * lra-constraints.c (process_alt_operands): Don't discourage
10904 memory with known offset for offsetable memory constraint.
10905 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10907 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10909 * varasm.c (struct oc_local_state): Reorder fields.
10910 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10911 and adjust accordingly.
10912 (output_constructor): Reorder initialization code and adjust call to
10913 output_constructor_bitfield.
10915 2013-06-07 Jan Hubicka <jh@suse.cz>
10917 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10919 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10921 * tree-object-size.c (unknown): Make const.
10923 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10925 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10926 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10927 for last alternative in the cpu_facility attribute.
10929 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10932 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10933 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10934 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10935 * config/arm/constraints.md (Dg): New constraint.
10936 * config/arm/neon.md (xordi3_neon): Remove.
10937 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10938 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10940 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10942 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10943 Clean up alternatives.
10945 2013-06-07 Alan Modra <amodra@gmail.com>
10947 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10950 2013-06-07 Alan Modra <amodra@gmail.com>
10952 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
10954 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10956 * config/arm/constraints.md (Df): New constraint.
10957 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
10958 Correct length attribute for last two alternatives.
10960 2013-06-07 Alan Modra <amodra@gmail.com>
10962 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10963 override user -mfp-in-toc.
10964 (offsettable_ok_by_alignment): Consider just the current access
10965 rather than the whole object, unless BLKmode. Handle
10966 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
10967 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
10968 for -mcmodel=medium.
10969 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
10970 override user -mfp-in-toc or -msum-in-toc. Default to
10971 -mno-fp-in-toc for -mcmodel=medium.
10973 2013-06-06 DJ Delorie <dj@redhat.com>
10975 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
10976 TARGET_VALID_POINTER_MODE.
10978 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
10979 Pat Haugen <pthaugen@us.ibm.com>
10980 Peter Bergner <bergner@vnet.ibm.com>
10982 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
10983 Document new power8 builtins.
10985 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
10986 condition code register, to allow 128-bit logical operations to be
10987 done in the VSX or GPR registers.
10988 (nor<mode>3): Use the canonical form for nor.
10989 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
10990 vclz*, and vpopcnt* vector instructions.
10991 (nand<mode>3): Likewise.
10992 (orc<mode>3): Likewise.
10993 (clz<mode>2): LIkewise.
10994 (popcount<mode>2): Likewise.
10996 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
10997 that only the GPRs are recognized.
10999 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11000 support for new power8 builtins.
11002 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11004 (xscvdpspn): Likewise.
11010 (vpopcnt): Likewise.
11011 (vpopcntb): Likewise.
11012 (vpopcnth): Likewise.
11013 (vpopcntw): Likewise.
11014 (vpopcntd): Likewise.
11016 (vmrgew): Likewise.
11017 (vmrgow): Likewise.
11019 (eqv_v16qi3): Likewise.
11020 (eqv_v8hi3): Likewise.
11021 (eqv_v4si3): Likewise.
11022 (eqv_v2di3): Likewise.
11023 (eqv_v4sf3): Likewise.
11024 (eqv_v2df3): Likewise.
11026 (nand_v16qi3): Likewise.
11027 (nand_v8hi3): Likewise.
11028 (nand_v4si3): Likewise.
11029 (nand_v2di3): Likewise.
11030 (nand_v4sf3): Likewise.
11031 (nand_v2df3): Likewise.
11033 (orc_v16qi3): Likewise.
11034 (orc_v8hi3): Likewise.
11035 (orc_v4si3): Likewise.
11036 (orc_v2di3): Likewise.
11037 (orc_v4sf3): Likewise.
11038 (orc_v2df3): Likewise.
11040 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11041 allow power8 quad mode in 64-bit.
11042 (rs6000_builtin_vectorized_function): Add support to vectorize
11043 ISA 2.07 count leading zeros, population count builtins.
11044 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11045 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11046 (builtin_function_type): Add vgbbd builtin function which takes an
11048 (altivec_expand_vec_perm_const): Add support for new power8 merge
11051 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11052 that does not include TImdoe for use with 32-bit.
11053 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11055 (UNSPEC_VSX_CVDPSPN): Likewise.
11056 (vsx_xscvdpspn): Likewise.
11057 (vsx_xscvspdpn): Likewise.
11058 (vsx_xscvdpspn_scalar): Likewise.
11059 (vsx_xscvspdpn_directmove): Likewise.
11060 (vsx_and<mode>3): Split logical operations into 32-bit and
11061 64-bit. Add support to do logical operations on TImode as well as
11062 VSX vector types. Allow logical operations to be done in either
11063 VSX registers or in general purpose registers in 64-bit mode. Add
11064 splitters if GPRs were used. For AND, add clobber of CCmode to
11065 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11067 (vsx_and<mode>3_32bit): Likewise.
11068 (vsx_and<mode>3_64bit): Likewise.
11069 (vsx_ior<mode>3): Likewise.
11070 (vsx_ior<mode>3_32bit): Likewise.
11071 (vsx_ior<mode>3_64bit): Likewise.
11072 (vsx_xor<mode>3): Likewise.
11073 (vsx_xor<mode>3_32bit): Likewise.
11074 (vsx_xor<mode>3_64bit): Likewise.
11075 (vsx_one_cmpl<mode>2): Likewise.
11076 (vsx_one_cmpl<mode>2_32bit): Likewise.
11077 (vsx_one_cmpl<mode>2_64bit): Likewise.
11078 (vsx_nor<mode>3): Likewise.
11079 (vsx_nor<mode>3_32bit): Likewise.
11080 (vsx_nor<mode>3_64bit): Likewise.
11081 (vsx_andc<mode>3): Likewise.
11082 (vsx_andc<mode>3_32bit): Likewise.
11083 (vsx_andc<mode>3_64bit): Likewise.
11084 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11085 and xxlorc instructions.
11086 (vsx_eqv<mode>3_64bit): Likewise.
11087 (vsx_nand<mode>3_32bit): Likewise.
11088 (vsx_nand<mode>3_64bit): Likewise.
11089 (vsx_orc<mode>3_32bit): Likewise.
11090 (vsx_orc<mode>3_64bit): Likewise.
11092 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11094 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11096 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11097 (p8_vmrgow): Likewise.
11098 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11099 GPRs to be split under VSX.
11100 (p8v_clz<mode>2): Add power8 count leading zero support.
11101 (p8v_popcount<mode>2): Add power8 population count support.
11102 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11105 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11108 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11110 (vec_nand): Likewise.
11111 (vec_vclz): Likewise.
11112 (vec_vclzb): Likewise.
11113 (vec_vclzd): Likewise.
11114 (vec_vclzh): Likewise.
11115 (vec_vclzw): Likewise.
11116 (vec_vgbbd): Likewise.
11117 (vec_vmrgew): Likewise.
11118 (vec_vmrgow): Likewise.
11119 (vec_vpopcnt): Likewise.
11120 (vec_vpopcntb): Likewise.
11121 (vec_vpopcntd): Likewise.
11122 (vec_vpopcnth): Likewise.
11123 (vec_vpopcntw): Likewise.
11125 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11127 PR rtl-optimization/57468
11128 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11131 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11133 PR rtl-optimization/57459
11134 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11135 type when setting live regs.
11137 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11139 * config/s390/s390.opt (mlra): New option.
11140 * config/s390/s390.c (s390_decompose_address): Check displacement
11141 for all registers for LRA.
11142 (s390_secondary_reload): Don't used secondary reloads for LRA.
11143 (s390_lra_p): New function.
11144 (TARGET_LRA_P): Define.
11145 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11146 of attribute cpu_facility to zarch for the last alternative.
11147 (*cmpmem_short): Ditto.
11149 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11151 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11152 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11153 (arm_expand_prologue): Likewise.
11155 2013-06-06 Teresa Johnson <tejohnson@google.com>
11158 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11159 as this is now done by redirect_edge_and_branch_force.
11160 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11161 barriers, and fix interaction with splitting.
11162 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11163 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11164 reflect changes made in the routine.
11165 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11166 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11167 since this is called in cfglayout mode, and replace partition fixup
11168 with assert as that is now done by force_nonfallthru_and_redirect.
11169 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11170 already be marked with region crossing note.
11171 (insert_section_boundary_note): Make non-static, gate on flag
11172 has_bb_partition, rewrite to also check for multiple partitions.
11173 (rest_of_handle_reorder_blocks): Remove call to
11174 insert_section_boundary_note, now done later during free_cfg.
11175 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11176 * bb-reorder.h (insert_section_boundary_note): Declare.
11177 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11178 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11179 invoke insert_section_boundary_note.
11180 (try_redirect_by_replacing_jump): Remove unnecessary
11181 check for region crossing note.
11182 (fixup_partition_crossing): New function.
11183 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11184 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11185 in non-cfglayout mode.
11186 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11187 remove old code that tried to do this. Emit barrier correctly
11188 when we are in cfglayout mode.
11189 (last_bb_in_partition): New function.
11190 (rtl_split_edge): Correctly fixup partition boundaries.
11191 (commit_one_edge_insertion): Remove old code that tried to
11192 fixup region crossing edge since this is now handled in
11193 split_block, and set up insertion point correctly since
11194 block may now end in a jump.
11195 (verify_hot_cold_block_grouping): Guard against checking when not in
11196 linearized RTL mode.
11197 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11199 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11200 rtl_verify_flow_info, so not called in cfglayout mode.
11201 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11202 (fixup_reorder_chain): Remove old code that attempted to fixup region
11203 crossing note as this is now handled in force_nonfallthru_and_redirect.
11204 (duplicate_insn_chain): Don't duplicate switch section notes.
11205 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11207 * basic-block.h (emit_barrier_after_bb): Declare.
11209 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11211 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11212 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11213 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11215 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11217 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11219 * config/arm/ldmstm.md: Regenerate.
11221 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11223 * config/arm/sync.md (atomic_loaddi_1):
11224 Disable predication for arm_restrict_it.
11225 (arm_load_exclusive<mode>): Likewise.
11226 (arm_load_exclusivesi): Likewise.
11227 (arm_load_exclusivedi): Likewise.
11228 (arm_load_acquire_exclusive<mode>): Likewise.
11229 (arm_load_acquire_exclusivesi): Likewise.
11230 (arm_load_acquire_exclusivedi): Likewise.
11231 (arm_store_exclusive<mode>): Likewise.
11232 (arm_store_exclusive<mode>): Likewise.
11233 (arm_store_release_exclusivedi): Likewise.
11234 (arm_store_release_exclusive<mode>): Likewise.
11236 2013-06-06 Richard Biener <rguenther@suse.de>
11238 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11240 (lto_tag_is_tree_code_p): Adjust.
11241 (lto_tag_is_gimple_code_p): Likewise.
11242 (lto_gimple_code_to_tag): Likewise.
11243 (lto_tag_to_gimple_code): Likewise.
11244 (lto_tree_code_to_tag): Likewise.
11245 (lto_tag_to_tree_code): Likewise.
11246 * data-streamer.h (streamer_write_hwi_in_range): Use
11247 uhwi streaming to stream the normalized range.
11248 (streamer_read_hwi_in_range): Likewise.
11250 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11252 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11253 (predicable_short_it): Likewise.
11254 (predicated): Likewise.
11255 (enabled): Handle above.
11256 (define_cond_exec): Set predicated attribute to yes.
11258 2013-06-05 Mike Stump <mikestump@comcast.net>
11260 * gdbinit.in (__FUNCTION__): Add.
11262 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11264 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11265 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11267 2013-06-05 Jan Hubicka <jh@suse.cz>
11269 * varasm.c (mark_decl_referenced): Revert the removal until targets
11272 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11274 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11275 instead of mark_decl_referenced.
11277 2013-06-05 Jan Hubicka <jh@suse.cz>
11279 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11280 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11281 and symtab_used_from_object_file_p.
11282 (cgraph_make_node_local_1): Clear forced_by_abi.
11283 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11284 * cgraph.h (symtab_node_base): Add forced_by_abi.
11285 (decide_is_variable_needed): Remove.
11286 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11287 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11288 (decide_is_symbol_needed): ... this one; handle symbols in general;
11289 always analyze virtuals; honnor forced_by_abi.
11290 (cgraph_finalize_function): Update.
11291 (varpool_finalize_decl): Update.
11292 (symbol_defined_and_needed): Remove.
11293 (analyze_functions): Update.
11294 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11295 output_refs, input_overwrite_node): Handle forced_by_abi.
11296 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11297 (address_taken_from_non_vtable_p): ... this one.
11298 (comdat_can_be_unshared_p_1): New function.
11299 (cgraph_comdat_can_be_unshared_p): Rename to ...
11300 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11301 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11302 (function_and_variable_visibility): Clear forced_by_abi as needed.
11303 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11304 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11305 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11306 * varpool.c (decide_is_variable_needed): Remove.
11308 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11310 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11311 (arm_option_override): Override arm_restrict_it where appropriate.
11312 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11313 * config/arm/arm.opt (mrestrict-it): New command-line option.
11314 * doc/invoke.texi: Document -mrestrict-it.
11316 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11318 * tsan.c (tsan_atomic_table): Make const.
11320 2013-06-05 Richard Biener <rguenther@suse.de>
11322 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11323 index associated with the tree we are supposed to replace.
11324 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11325 TREE_ASM_WRITTEN as zero for everything but SSA names.
11327 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11329 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11331 2013-06-04 Jan Hubicka <jh@suse.cz>
11333 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11334 (input_node, input_varpool_node): Handle correctly external same
11336 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11337 nodes at ltrans stage.
11339 2013-06-04 Jan Hubicka <jh@suse.cz>
11341 * ipa-inline.c (update_caller_keys): Fix availability test.
11342 (update_callee_keys): Likewise.
11343 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11344 to follow ELF standard.
11346 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11348 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11349 (mips64r5900el-*-elf*): New configurations.
11350 * config/mips/mips-cpus.def (r5900): New processor.
11351 * config/mips/mips-tables.opt: Regenerate.
11352 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11353 (mips_issue_rate): Handle PROCESSOR_R5900.
11354 (mips_reorg_process_insns): Force reorder mode for the R5900.
11355 * config/mips/mips.h (TARGET_MIPS5900): Define.
11356 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11358 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11360 * config/mips/mips.md (processor): Add r5900.
11361 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11363 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11365 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11366 into function to generate MOVI instruction.
11367 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11368 (aarch64_preferred_simd_mode): Turn into wrapper.
11369 (aarch64_output_scalar_simd_mov_immediate): New function.
11370 * config/aarch64/aarch64-protos.h: Add prototype for above.
11372 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11374 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11375 element_char member.
11376 (sizetochar): Return signed char.
11377 (aarch64_simd_valid_immediate): Remove elchar and other
11378 unnecessary variables.
11379 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11380 Calculate element_char as required.
11381 * config/aarch64/aarch64-protos.h: Update and move prototype
11382 for aarch64_output_simd_mov_immediate.
11383 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11386 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11388 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11389 information completed by aarch64_simd_valid_immediate.
11390 (aarch64_legitimate_constant_p): Update arguments.
11391 (aarch64_simd_valid_immediate): Work with struct rather than many
11393 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11394 (aarch64_simd_make_constant): Update arguments.
11395 (aarch64_output_simd_mov_immediate): Work with struct rather than
11396 many pointers. Output immediate directly rather than as operand.
11397 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11399 * config/aarch64/constraints.md (Dn): Update arguments.
11401 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11403 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11405 (aarch64_simd_immediate_valid_for_move): Remove.
11406 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11407 (aarch64_simd_make_constant): Update call.
11408 (aarch64_output_simd_mov_immediate): Update call.
11409 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11411 * config/aarch64/constraints.md (Dn): Update call.
11413 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11415 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11416 return type to bool for prototype.
11417 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11418 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11419 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11421 2013-06-04 Catherine Moore <clm@codesourcery.com>
11423 * config/mips/mips.opt (meva): New.
11424 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11425 (ASM_SPEC): Handle -meva.
11426 * doc/invoke.texi (meva): Document.
11428 2013-06-04 Alan Modra <amodra@gmail.com>
11430 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11433 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11435 * rtl.def: Add extra fourth optional field to define_cond_exec.
11436 * gensupport.c (process_one_cond_exec): Process attributes from
11438 * doc/md.texi: Document fourth field in define_cond_exec.
11440 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11442 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11443 out the processing order as in store_bit_field_1.
11445 2013-06-04 Jan Hubicka <jh@suse.cz>
11447 PR middle-end/57500
11448 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11449 non-VAR_DECL node if it does not exist yet.
11451 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11453 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11454 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11455 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11456 target_cpu_default setting.
11458 2013-06-03 Teresa Johnson <tejohnson@google.com>
11460 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11461 default to -fopt-info=optimized instead of all.
11462 * doc/invoke.texi: Ditto.
11463 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11464 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11465 (execute_vect_slp): Emit BB vectorization success under
11466 MSG_OPTIMIZED_LOCATIONS.
11467 * tree-vect-slp.c (vect_slp_transform_bb): Change
11468 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11469 * tree-vect-loop.c (vect_transform_loop): Ditto.
11471 2013-06-03 Jason Merrill <jason@redhat.com>
11474 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11475 Use TARGET_EXPR for C++.
11477 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11479 PR rtl-optimization/57268
11480 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11481 if DEBUG_INSN_P (insn).
11484 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11486 PR rtl-optimization/57268
11487 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11488 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11490 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11492 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11493 (ix86_avoid_lea_for_addr): Likewise.
11494 (exact_dependency_1): Likewise.
11495 (ix86_adjust_cost): Likewise.
11496 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11498 (do_reorder_for_imul): Fix typo, formatting and
11499 !reload_completed check removed.
11500 (ix86_sched_reorder): Fix typo and formatting.
11501 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11504 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11506 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11508 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11510 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11511 <STRING_CST>: Likewise.
11512 <VECTOR_CST>: Likewise.
11514 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11515 Mikael Morin <mikael@gcc.gnu.org>
11517 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11518 * config.in: Regenerated.
11519 * configure: Regenerated.
11521 2013-06-01 Jan Hubicka <jh@suse.cz>
11523 PR middle-end/57366
11524 * cgraphunit.c (compile): When weakref is not supported,
11525 set up transparent aliases before final output pass.
11526 * varasm.c (assemble_alias): Do not try to do it here.
11528 2013-06-01 Jan Hubicka <jh@suse.cz>
11530 PR middle-end/57467
11531 * passes.c (for_per_function): Skip unanalyzed functions.
11533 2013-06-01 Jan Hubicka <jh@suse.cz>
11535 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11536 (lto_symtab_merge_symbols_1): ... this one.
11537 (lto_symtab_merge_cgraph_nodes): Rename to ...
11538 (lto_symtab_merge_symbols): ... this one; simplify.
11539 * cgraph.c (same_body_aliases_done): Rename to ...
11540 (cpp_implicit_aliases_done): ... this one.
11541 (cgraph_create_function_alias): Update.
11542 (cgraph_same_body_alias): Update.
11543 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11544 (verify_edge_corresponds_to_fndecl): Simplify.
11545 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11546 (cgraph_node): Remove same_body_alias.
11547 (varpool_node): Remove alias_of and extra_name_alias.
11548 (same_body_aliases_done): Rename to ..
11549 (cpp_implicit_aliases_done): ... this one.
11550 (symtab_alias_ultimate_target): Add default parameter.
11551 (symtab_resolve_alias): New function.
11552 (fixup_same_cpp_alias_visibility): Declare.
11553 (cgraph_function_node): Add default parameter.
11554 (cgraph_node_asm_name): Likewise.
11555 (cgraph_function_or_thunk_node): Add default parameter; do
11556 not ICE when it is NULL.
11557 (varpool_variable_node): Likewise.
11558 * tree-emutls.c (create_emultls_var): Update.
11559 (ipa_lower_emutls): Update.
11560 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11561 (cgraph_reset_node): Reset alias info.
11562 (cgraph_finalize_function): Update.
11563 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11564 (analyze_function): Simplify.
11565 (cgraph_process_same_body_aliases): Simplify.
11566 (analyze_functions): Fixup same body aliases.
11567 (handle_alias_pairs): Simplify.
11568 (assemble_thunk): Update.
11569 (assemble_thunks_and_aliases): Update.
11570 (output_weakrefs): Rewrite.
11571 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11572 (lto_output_varpool_node): Likewise.
11573 (compute_ltrans_boundary): Remve assert.
11574 (get_alias_symbol): New functoin.
11575 (input_node): Rewrite alias handling.
11576 (input_varpool_node): Likewise.
11577 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11578 * ipa.c (process_references): Handle weakrefs correctly.
11579 (symtab_remove_unreachable_nodes): Likewise.
11580 * trans-mem.c (get_cg_data): Update.
11581 (ipa_tm_create_version_alias): Update.
11582 (ipa_tm_execute): Update.
11583 * symtab.c (dump_symtab_base): Dump aliases.
11584 (verify_symtab_base): Verify aliases.
11585 (symtab_node_availability): New function.
11586 (symtab_alias_ultimate_target): Simplify.
11587 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11588 handle all the fixup cases.
11589 (symtab_resolve_alias): New function.
11590 * passes.c (ipa_write_summaries): Handle weakrefs.
11591 * varpool.c (varpool_analyze_node): Simplify.
11592 (assemble_aliases): Update.
11593 (varpool_create_variable_alias): Simplify.
11594 (varpool_extra_name_alias): Simplify.
11595 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11596 (lto_symtab_merge_symbols): ... this one.
11598 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11601 PR rtl-optimization/57268
11602 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11603 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11605 2013-06-01 Tobias Burnus <burnus@net-b.de>
11607 Partially reverted:
11608 2013-05-31 Tobias Burnus <burnus@net-b.de>
11610 PR middle-end/57073
11611 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11614 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11616 PR rtl-optimization/57268
11617 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11618 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11620 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11622 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11623 unordered comparison operators when -fno-trapping-math is in effect
11625 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11626 and implement unordered comparison operators properly on the e500.
11628 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11630 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11631 for constant scalar integers.
11632 (simplify_relational_operation_1): Likewise.
11634 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11636 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11637 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11640 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11641 Igor Zamyatin <igor.zamyatin@intel.com>
11643 Silvermont (SLM) architecture performance tuning.
11644 * config/i386/i386.h (enum ix86_tune_indices): Add
11645 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11646 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11648 * config/i386/i386.c (initial_ix86_tune_features)
11649 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11650 (ix86_lea_outperforms): Handle Silvermont tuning.
11651 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11653 (ix86_use_lea_for_mov): Likewise.
11654 (ix86_avoid_lea_for_addr): Likewise.
11655 (ix86_lea_for_add_ok): Likewise.
11656 (exact_dependency_1): New function.
11657 (exact_store_load_dependency): Likewise.
11658 (ix86_adjust_cost): Handle Silvermont tuning.
11659 (do_reoder_for_imul): Likewise.
11660 (swap_top_of_ready_list): New function.
11661 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11663 * config/i386/i386.md (peepholes that split memory operand in fp
11666 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11668 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11669 Remove un-necessary braces.
11671 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11673 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11674 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11676 2013-05-31 Tobias Burnus <burnus@net-b.de>
11678 PR middle-end/57073
11679 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11681 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11684 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11685 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11686 * config/arm/neon.md (iordi3_neon): Remove.
11687 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11688 * config/arm/predicates.md (imm_for_neon_logic_operand):
11689 Move to earlier in the file.
11690 (neon_logic_op2): Likewise.
11691 (arm_iordi_operand_neon): New predicate.
11693 2013-05-31 Richard Biener <rguenther@suse.de>
11695 PR tree-optimization/57478
11696 PR tree-optimization/57453
11697 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11700 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11702 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11703 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11705 2013-05-30 Tobias Burnus <burnus@net-b.de>
11706 Thomas Koenig <tkoenig@gcc.gnu.org>
11708 PR middle-end/57073
11709 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11710 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11712 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11714 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11716 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11718 * target.def (register_usage_leveling_p): New hook.
11719 * targhooks.c (default_register_usage_leveling_p): New.
11720 * targhooks.h (default_register_usage_leveling_p): New prototype.
11721 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11722 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11723 * doc/tm.texi: Update.
11724 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11726 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11728 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11729 (*insv_reg<mode>): New define_insn.
11731 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11733 PR rtl-optimization/57439
11734 * postreload.c (move2add_valid_value_p): Check that we have
11735 a zero subreg_regno_offset when accessing the register in
11736 the requested mode.
11738 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11739 Igor Zamyatin <igor.zamyatin@intel.com>
11741 Silvermont (SLM) architecture pipeline model, tuning and
11743 * config.gcc: Add slm config options and target.
11745 * config/i386/slm.md: New.
11747 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11749 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11751 (ix86_target_macros_internal): Likewise.
11753 * config/i386/i386.c (slm_cost): New cost.
11754 (m_SLM): New macro flag.
11755 (initial_ix86_tune_features): Set m_SLM.
11756 (x86_accumulate_outgoing_args): Likewise.
11757 (x86_arch_always_fancy_math_387): Likewise.
11758 (processor_target_table): Add slm cost.
11759 (cpu_names): Add slm cpu name.
11760 (x86_option_override_internal): Set SLM ISA.
11761 (ix86_issue_rate): New case PROCESSOR_SLM.
11762 (ia32_multipass_dfa_lookahead): Likewise.
11763 (fold_builtin_cpu): Add slm.
11765 * config/i386/i386.h (TARGET_SLM): New target macro.
11766 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11767 (processor_type): Add PROCESSOR_SLM.
11769 * config/i386/i386.md (cpu): Add new value "slm".
11770 (slm.md): Include slm.md.
11772 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11773 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11775 * config/arm/arm-protos.h: Add and update function protos.
11776 * config/arm/arm.c (use_simple_return_p): New added.
11777 (thumb2_expand_return): Check simple_return flag.
11778 * config/arm/arm.md: Add simple_return and conditional simple_return.
11779 * config/arm/iterators.md: Add iterator for return and simple_return.
11781 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11783 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11784 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11785 (arm_emit_vfp_multi_reg_pop): Likewise.
11786 (thumb2_emit_ldrd_pop): Likewise.
11787 (arm_expand_epilogue): Add misc REG_CFA notes.
11788 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11790 2013-05-29 Lawrence Crowl <crowl@google.com>
11792 * config/arm/t-arm: Update for below.
11794 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11795 Change type to hash_table. Update dependent calls and types.
11797 * config/i386/t-cygming: Update for below.
11799 * config/i386/t-interix: Update for below.
11801 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11802 Change type to hash_table. Update dependent calls and types.
11803 (i386_find_on_wrapper_list::wrappers): Likewise.
11805 * config/ia64/t-ia64: Update for below.
11807 * config/ia64/ia64.c (bundle_state_table):
11808 Change type to hash_table. Update dependent calls and types.
11810 * config/mips/mips.c (mips_reorg_process_insns::htab):
11811 Change type to hash_table. Update dependent calls and types.
11813 * config/sol2.c (solaris_comdat_htab):
11814 Change type to hash_table. Update dependent calls and types.
11816 * config/t-sol2: Update for above.
11818 2013-05-29 Teresa Johnson <tejohnson@google.com>
11820 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11821 functions are not yet marked as defined.
11823 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11824 Pat Haugen <pthaugen@us.ibm.com>
11825 Peter Bergner <bergner@vnet.ibm.com>
11827 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11831 (vrotl<mode>3): Likewise.
11832 (vashl<mode>3): Likewise.
11833 (vlshr<mode>3): Likewise.
11834 (vashr<mode>3): Likewise.
11836 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11837 support for power8 V2DI builtins.
11839 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11840 power8 V2DI builtins.
11841 (vupkhsw): Likewise.
11842 (vupklsw): Likewise.
11843 (vaddudm): Likewise.
11844 (vminsd): Likewise.
11845 (vmaxsd): Likewise.
11846 (vminud): Likewise.
11847 (vmaxud): Likewise.
11848 (vpkudum): Likewise.
11849 (vpksdss): Likewise.
11850 (vpkudus): Likewise.
11851 (vpksdus): Likewise.
11856 (vsubudm): Likewise.
11857 (vcmpequd): Likewise.
11858 (vcmpgtsd): Likewise.
11859 (vcmpgtud): Likewise.
11860 (vcmpequd_p): Likewise.
11861 (vcmpgtsd_p): Likewise.
11862 (vcmpgtud_p): Likewise.
11863 (vupkhsw): Likewise.
11864 (vupklsw): Likewise.
11865 (vaddudm): Likewise.
11866 (vmaxsd): Likewise.
11867 (vmaxud): Likewise.
11868 (vminsd): Likewise.
11869 (vminud): Likewise.
11870 (vpksdss): Likewise.
11871 (vpksdus): Likewise.
11872 (vpkudum): Likewise.
11873 (vpkudus): Likewise.
11878 (vsubudm): Likewise.
11880 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11881 support for power8 V2DI instructions.
11883 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11884 power8 V2DI instructions. Combine pack and unpack insns to use an
11885 iterator for each mode. Check whether a particular mode supports
11886 Altivec instructions instead of just checking TARGET_ALTIVEC.
11887 (UNSPEC_VPKUWUM): Likewise.
11888 (UNSPEC_VPKSHSS): Likewise.
11889 (UNSPEC_VPKSWSS): Likewise.
11890 (UNSPEC_VPKUHUS): Likewise.
11891 (UNSPEC_VPKSHUS): Likewise.
11892 (UNSPEC_VPKUWUS): Likewise.
11893 (UNSPEC_VPKSWUS): Likewise.
11894 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11895 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11896 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11897 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11898 (UNSPEC_VUPKHSB): Likewise.
11899 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11900 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11901 (UNSPEC_VUPKHSH): Likewise.
11902 (UNSPEC_VUPKLSB): Likewise.
11903 (UNSPEC_VUPKLSH): Likewise.
11905 (VI_char): Likewise.
11906 (VI_scalar): Likewise.
11907 (VI_unit): Likewise.
11909 (VP_small): Likewise.
11910 (VP_small_lc): Likewise.
11911 (VU_char): Likewise.
11912 (add<mode>3): Likewise.
11913 (altivec_vaddcuw): Likewise.
11914 (altivec_vaddu<VI_char>s): Likewise.
11915 (altivec_vadds<VI_char>s): Likewise.
11916 (sub<mode>3): Likewise.
11917 (altivec_vsubcuw): Likewise.
11918 (altivec_vsubu<VI_char>s): Likewise.
11919 (altivec_vsubs<VI_char>s): Likewise.
11920 (altivec_vavgs<VI_char>): Likewise.
11921 (altivec_vcmpbfp): Likewise.
11922 (altivec_eq<mode>): Likewise.
11923 (altivec_gt<mode>): Likewise.
11924 (altivec_gtu<mode>): Likewise.
11925 (umax<mode>3): Likewise.
11926 (smax<mode>3): Likewise.
11927 (umin<mode>3): Likewise.
11928 (smin<mode>3): Likewise.
11929 (altivec_vpkuhum): Likewise.
11930 (altivec_vpkuwum): Likewise.
11931 (altivec_vpkshss): Likewise.
11932 (altivec_vpkswss): Likewise.
11933 (altivec_vpkuhus): Likewise.
11934 (altivec_vpkshus): Likewise.
11935 (altivec_vpkuwus): Likewise.
11936 (altivec_vpkswus): Likewise.
11937 (altivec_vpks<VI_char>ss): Likewise.
11938 (altivec_vpks<VI_char>us): Likewise.
11939 (altivec_vpku<VI_char>us): Likewise.
11940 (altivec_vpku<VI_char>um): Likewise.
11941 (altivec_vrl<VI_char>): Likewise.
11942 (altivec_vsl<VI_char>): Likewise.
11943 (altivec_vsr<VI_char>): Likewise.
11944 (altivec_vsra<VI_char>): Likewise.
11945 (altivec_vsldoi_<mode>): Likewise.
11946 (altivec_vupkhsb): Likewise.
11947 (altivec_vupkhs<VU_char>): Likewise.
11948 (altivec_vupkls<VU_char>): Likewise.
11949 (altivec_vupkhsh): Likewise.
11950 (altivec_vupklsb): Likewise.
11951 (altivec_vupklsh): Likewise.
11952 (altivec_vcmpequ<VI_char>_p): Likewise.
11953 (altivec_vcmpgts<VI_char>_p): Likewise.
11954 (altivec_vcmpgtu<VI_char>_p): Likewise.
11955 (abs<mode>2): Likewise.
11956 (vec_unpacks_hi_v16qi): Likewise.
11957 (vec_unpacks_hi_v8hi): Likewise.
11958 (vec_unpacks_lo_v16qi): Likewise.
11959 (vec_unpacks_hi_<VP_small_lc>): Likewise.
11960 (vec_unpacks_lo_v8hi): Likewise.
11961 (vec_unpacks_lo_<VP_small_lc>): Likewise.
11962 (vec_pack_trunc_v8h): Likewise.
11963 (vec_pack_trunc_v4si): Likewise.
11964 (vec_pack_trunc_<mode>): Likewise.
11966 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
11968 (vec_vmaxsd): Likewise.
11969 (vec_vmaxud): Likewise.
11970 (vec_vminsd): Likewise.
11971 (vec_vminud): Likewise.
11972 (vec_vpksdss): Likewise.
11973 (vec_vpksdus): Likewise.
11974 (vec_vpkudum): Likewise.
11975 (vec_vpkudus): Likewise.
11976 (vec_vrld): Likewise.
11977 (vec_vsld): Likewise.
11978 (vec_vsrad): Likewise.
11979 (vec_vsrd): Likewise.
11980 (vec_vsubudm): Likewise.
11981 (vec_vupkhsw): Likewise.
11982 (vec_vupklsw): Likewise.
11984 2013-05-29 Jan Hubicka <jh@suse.cz>
11986 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
11987 flags; reorder rest of fields in more consistent way.
11988 (varpool_node): Remove analyzed, finalized and alias.
11989 (cgraph_ndoe): Likewise.
11990 (symtab_alias_ultimate_target): New function.
11991 (cgraph_function_node): Move offline.
11992 (cgraph_reset_node): Declare.
11993 (cgraph_comdat_can_be_unshared_p): Remove.
11994 (varpool_remove_initializer): Declare.
11995 (varpool_first_defined_variable, varpool_next_defined_variable
11996 cgraph_first_defined_function, cgraph_next_defined_function): Update.
11997 (cgraph_function_with_gimple_body_p): Update.
11998 (varpool_all_refs_explicit_p): Update.
11999 (symtab_alias_target): New function.
12000 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12001 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12002 (cgraph_function_or_thunk_node): Simplify using
12003 symtab_alias_ultimate_target.
12004 (varpool_variable_node): Likewise.
12005 * cgraph.c (cgraph_create_function_alias): Update.
12006 (cgraph_add_thunk): Update.
12007 (cgraph_remove_node): Update.
12008 (dump_cgraph_node): Do not dump removed flags.
12009 (cgraph_function_body_availability): Update.
12010 (cgraph_propagate_frequency): Update.
12011 (verify_cgraph_node): Check sanity of local flag.
12012 (cgraph_function_node): Move here from cgraph.h; revamp for
12013 cgraph_function_or_thunk_node.
12014 * lto-symtab.c (lto_varpool_replace_node): Update.
12015 (lto_symtab_resolve_can_prevail_p): Update.
12016 (lto_symtab_merge_cgraph_nodes): Update.
12017 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12018 propagate_constants_accross_call, devirtualization_time_bonus,
12019 ipcp_propagate_stage): Update.
12020 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12021 * ipa-inline-transform.c (clone_inlined_nodes,
12022 preserve_function_body_p): Update.
12023 * ipa-reference.c (propagate): Update.
12024 (write_node_summary_p): Update.
12025 * toplev.c (wrapup_global_declaration_2): Update.
12026 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12027 (analyze_function) ... this one.
12028 (cgraph_process_new_functions): Update.
12029 (cgraph_reset_node): Export.
12030 (cgraph_finalize_function): Update.
12031 (cgraph_add_new_function): Update.
12032 (process_function_and_variable_attributes): Update.
12033 (varpool_finalize_decl): Update.
12034 (symbol_finalized): Remove.
12035 (symbol_finalized_and_needed): Rename to ...
12036 (symbol_defined_and_needed): ... update.
12037 (cgraph_analyze_functions): Update.
12038 (handle_alias_pairs): Update.
12039 (mark_functions_to_output): Update.
12040 (assemble_thunk): Update.
12041 (output_in_order): Update.
12042 (output_weakrefs): Update.
12043 (finalize_compilation_unit): Update.
12044 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12045 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12046 input_node, input_varpool_node): Update.
12047 * dbxout.c (dbxout_expand_expr): Update.
12048 * cgraphclones.c (cgraph_clone_node): Update.
12049 (cgraph_copy_node_for_versioning): Update.
12050 (cgraph_materialize_clone): Update.
12051 (cgraph_materialize_all_clones): Update.
12052 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12053 propagate_pure_const, propagate_nothrow): Update.
12054 * lto-streamer-out.c (lto_output, write_symbol): Update.
12055 * ipa-utils.c (ipa_reverse_postorder): Update.
12056 * ipa-inline.c (can_inline_edge_p): Update.
12057 (update_caller_keys, ipa_inline): Update.
12058 * dwarf2out.c (reference_to_unused,
12059 premark_types_used_by_global_vars_helper): Update.
12060 * tree-eh.c (tree_could_trap_p): Update.
12061 * ipa-split.c (consider_split, execute_split_functions): Update.
12062 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12063 has_addr_references_p): Update; move ahead in file for better
12065 (process_references): Simplify.
12066 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12067 bodies are removed.
12068 (cgraph_comdat_can_be_unshared_p): Make static.
12069 (cgraph_externally_visible_p): Update.
12070 (varpool_externally_visible_p): Update.
12071 (function_and_variable_visibility): Update.
12072 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12073 ipa_tm_mark_force_output_node): Update.
12074 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12075 estimate_edge_devirt_benefit, inline_generate_summary,
12076 inline_write_summary): Update.
12077 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12078 * ipa-prop.c (ipa_compute_jump_functions): Update.
12079 (ipa_print_node_params, ipa_prop_read_section,
12080 ipa_update_after_lto_read, read_replacements_section): Update.
12081 * varasm.c (mark_decl_referenced): Update.
12082 (assemble_alias, dump_tm_clone_pairs): Update.
12083 * tree-inline.c (copy_bb): Update.
12084 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12086 * symtab.c (dump_symtab_base): Print new flags.
12087 (verify_symtab_base): Verify new flags.
12088 (symtab_alias_ultimate_target): New function.
12089 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12090 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12092 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12094 * i386.c (ix86_get_function_versions_dispatcher,
12095 ix86_generate_version_dispatcher_body): Update.
12096 (fold_builtin_cpu): Use varpool_add_new_variable.
12097 * varpool.c (varpool_remove_initializer): Break out from ...
12098 (varpool_remove_node): ... this one.
12099 (dump_varpool_node, varpool_node_for_asm,
12100 cgraph_variable_initializer_availability, varpool_analyze_node,
12101 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12102 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12105 2013-05-29 Jan Hubicka <jh@suse.cz>
12107 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12109 2013-05-29 Easwaran Raman <eraman@google.com>
12111 PR tree-optimization/57442
12112 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12113 when control exits the main loop.
12115 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12117 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12119 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12120 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12121 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12124 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12126 PR tree-optimization/57441
12127 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12128 Don't limit size of incr_vec to number of candidates.
12130 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12132 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12133 and mips16 directories.
12134 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12135 (MULTILIB_DIRNAMES): Ditto.
12136 (MULTILIB_EXCEPTIONS): Add new exceptions.
12137 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12138 (MULTILIB_DIRNAMES): Ditto.
12139 (MULTILIB_EXCEPTIONS): Add new exceptions.
12141 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12142 Marcus Shawcroft <marcus.shawcroft@arm.com>
12144 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12145 SYMBOL_TINY_ABSOLUTE.
12146 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12147 SYMBOL_TINY_ABSOLUTE.
12148 (aarch64_expand_mov_immediate): Likewise.
12149 (aarch64_classify_symbol): Likewise.
12150 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12151 Permit SYMBOL_TINY_ABSOLUTE.
12152 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12154 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12155 Marcus Shawcroft <marcus.shawcroft@arm.com>
12157 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12158 Refactor if/switch. Replace gcc_assert with if.
12160 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12162 * config/i386/i386.c (initial_ix86_tune_features): Enable
12163 FP Reassociation for AMD bdver1 and bdver2.
12165 2013-05-29 Martin Jambor <mjambor@suse.cz>
12167 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12168 and IMAGPART_EXPR do not occur within other handled_components.
12170 2013-05-29 Richard Biener <rguenther@suse.de>
12172 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12173 access on whether the use is in the BB we currently try to
12175 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12176 vectorize to vect_bb_slp_scalar_cost.
12178 2013-05-29 Richard Biener <rguenther@suse.de>
12180 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12181 computing scalar cost offsetted by stmts that are kept live
12183 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12184 for computation of scalar cost.
12186 2013-05-28 Steve Ellcey <sellcey@mips.com>
12188 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12190 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12192 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12193 array notation built-in reduction functions.
12194 * doc/passes.texi (Passes): Added documentation about changes done
12196 * doc/invoke.texi (C Dialect Options): Added documentation about
12197 the -fcilkplus flag.
12198 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12199 (BUILTINS_DEF): Depend on cilkplus.def.
12200 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12201 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12202 * cilkplus.def: New file.
12204 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12206 PR rtl-optimization/57439
12207 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12209 2013-05-28 Easwaran Raman <eraman@google.com>
12211 PR tree-optimization/57337
12212 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12213 (find_insert_point): Correctly identify the insertion point
12214 when two statements with the same UID is compared.
12216 2013-05-28 Richard Biener <rguenther@suse.de>
12218 PR tree-optimization/56787
12219 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12220 from the list of data references.
12221 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12223 (vect_analyze_loop_operations): Likewise.
12224 (vect_transform_loop): Remove clobbers.
12226 2013-05-28 Martin Jambor <mjambor@suse.cz>
12228 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12229 and REALPART_EXPRs have scalar type.
12231 2013-05-28 Richard Biener <rguenther@suse.de>
12233 PR tree-optimization/57411
12234 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12236 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12237 virtual operand propagation.
12239 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12241 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12242 destination register for bmasksi_vis.
12243 (vector_init_bshuffle): Likewise.
12244 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12246 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12248 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12249 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12250 mode if the instruction isn't available in the original mode.
12251 * config/sparc/sparc.opt (mfix-ut699): New option.
12252 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12253 (divdf3): Turn into expander.
12254 (divdf3_nofix): New insn.
12255 (divdf3_fix): Likewise.
12256 (divsf3): Disable if -mfix-ut699.
12257 (sqrtdf2): Turn into expander.
12258 (sqrtdf2_nofix): New insn.
12259 (sqrtdf2_fix): Likewise.
12260 (sqrtsf2): Disable if -mfix-ut699.
12262 2013-05-27 Richard Biener <rguenther@suse.de>
12264 PR middle-end/57412
12265 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12266 block for the new loop.
12268 2013-05-27 Richard Biener <rguenther@suse.de>
12270 PR tree-optimization/57343
12271 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12272 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12273 (number_of_iterations_cond): Do not build the folded tree.
12275 2013-05-27 Richard Biener <rguenther@suse.de>
12278 PR middle-end/57381
12279 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12280 OEP_CONSTANT_ADDRESS_OF retained.
12282 PR tree-optimization/57417
12283 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12284 for unchanged base.
12285 (set_ssa_val_to): Compare addresses using
12286 get_addr_base_and_unit_offset.
12288 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12290 PR rtl-optimization/56833
12291 * postreload.c (move2add_record_mode): New function.
12292 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12293 (move2add_use_add2_insn): Use move2add_record_sym_value.
12294 (move2add_use_add3_insn): Likewise.
12295 (reload_cse_move2add): Use move2add_valid_value_p and
12296 move2add_record_mode. Invalidate call-clobbered and REG_INC
12297 affected regs by setting reg_mode to VOIDmode.
12298 (move2add_note_store): Don't pretend the inside of a SUBREG is
12299 the actual destination. Invalidate single/leading registers by
12300 setting reg_mode to VOIDmode.
12301 Use move2add_record_sym_value, move2add_valid_value_p and
12302 move2add_record_mode.
12304 2013-05-27 Richard Biener <rguenther@suse.de>
12306 PR tree-optimization/57396
12307 * tree-affine.c (double_int_constant_multiple_p): Properly
12308 return false for val == 0 and div != 0.
12310 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12312 * config/mips/mips.h: Use #elif in preprocessor conditions.
12314 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12317 * config/mips/constraints.md (kl): New constraint.
12318 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12319 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12320 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12321 from LO for MIPS16.
12322 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12324 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12327 * config/mips/mips.c (mips_can_inline_p): New function.
12328 (TARGET_CAN_INLINE_P): Define.
12330 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12332 * sched-int.h (ds_t, dw_t): Make unsigned int.
12333 Fix documentation that describes how all the ds_t bits are used.
12334 Reserve the last bit for delayed-branch scheduling.
12335 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12336 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12337 (gen_dep_weak_1): Remove prototype.
12338 * sched-deps.c (get_dep_weak_1): Make static.
12339 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12340 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12341 * doc/tm.texi: Regenerate.
12342 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12344 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12347 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12349 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12350 Sandra Loosemore <sandra@codesourcery.com>
12352 * config.gcc (powerpc-*): Allow native for with-cpu.
12354 2013-05-24 Jeff Law <law@redhat.com>
12356 PR tree-optimization/57124
12357 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12358 conversion feeding a condition if the range has an overflow
12359 if -fstrict-overflow. Add warnings for when we do make the
12362 2013-05-24 Dehao Chen <dehao@google.com>
12364 * tree-cfg.c (locus_discrim_map): Fix the typo.
12365 (locus_discrim_hasher): Likewise.
12366 (locus_discrim_hasher::hash): Likewise.
12367 (locus_discrim_hasher::equal): Likewise.
12369 2013-05-24 Martin Jambor <mjambor@suse.cz>
12371 PR tree-optimization/57294
12372 * cgraph.h (ipa_record_stmt_references): Declare.
12373 * cgraphbuild.c (ipa_record_stmt_references): New function.
12374 (build_cgraph_edges): Use ipa_record_stmt_references.
12375 (rebuild_cgraph_edges): Likewise.
12376 (cgraph_rebuild_references): Likewise.
12377 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12378 associated with the old statement and build references from the
12379 newly built statements.
12380 * ipa-ref.c (ipa_remove_stmt_references): New function.
12381 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12383 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12385 * lra-constraints.c (emit_spill_move): Use smaller mode for
12387 (check_and_process_move): Consider mem-reg moves for secondary
12389 (curr_insn_transform): Don't lose insns emitted before for
12390 secondary memory moves.
12391 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12392 reg set up in the current insn.
12394 2013-05-24 Dehao Chen <dehao@google.com>
12396 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12398 (locus_descrim_hasher::equal): Likewise.
12399 (build_gimple_cfg): New discriminator assignment algorithm.
12400 (make_edges): Likewise.
12401 (next_discriminator_for_locus): Likewise.
12402 (same_line_p): Likewise.
12403 (assign_discriminators): Likewise.
12404 (make_cond_expr_edges): Likewise.
12405 (make_gimple_switch_edges): Likewise.
12406 (make_goto_expr_edges): Likewise.
12407 (make_gimple_asm_edges): Likewise.
12409 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12411 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12412 X format specifier to only display bottom 16 bits.
12413 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12414 immediate to match for operand 2, since it will be masked.
12416 2013-05-24 Richard Biener <rguenther@suse.de>
12418 PR tree-optimization/57287
12419 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12420 all SSA names that occur in abnormal PHIs.
12422 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12425 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12426 that index is not negative.
12428 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12430 PR rtl-optimization/55177
12431 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12432 (simplify_byte_swapping_operation): New.
12433 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12434 (simplify_relational_operation_1): Deal with BSWAP.
12436 2013-05-23 Richard Henderson <rth@redhat.com>
12439 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12440 (ix86_reorg): Call it.
12442 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12445 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12446 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12447 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12449 2013-05-23 Christian Bruel <christian.bruel@st.com>
12452 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12454 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12455 Marcus Shawcroft <marcus.shawcroft@arm.com>
12457 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12458 * config/aarch64/constraints.md (Usa): Remove.
12459 * doc/md.texi (AArch64 Usa): Remove.
12461 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12462 Marcus Shawcroft <marcus.shawcroft@arm.com>
12464 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12465 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12466 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12467 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12469 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12471 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12472 instruction (AdvSIMD).
12473 * config/aarch64/aarch64-builtins.c
12474 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12475 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12477 2013-05-23 Martin Jambor <mjambor@suse.cz>
12479 PR middle-end/57347
12480 * tree.h (contains_bitfld_component_ref_p): Declare.
12481 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12482 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12484 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12485 not access a bit-field. Assert all final offsets are byte-aligned.
12487 2013-05-23 Richard Biener <rguenther@suse.de>
12489 PR tree-optimization/57380
12490 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12491 least one invariant or re-used load.
12492 * passes.c (init_optimization_passes): Move pass_phiprop before
12495 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12497 * config/aarch64/aarch64-simd.md
12498 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12500 2013-05-23 Richard Biener <rguenther@suse.de>
12502 PR middle-end/57381
12503 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12504 OEP_CONSTANT_ADDRESS_OF retained.
12506 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12508 PR middle-end/57344
12509 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12510 don't lower unit. Handle unit not being always BITS_PER_WORD.
12512 2013-05-23 Richard Biener <rguenther@suse.de>
12514 PR rtl-optimization/57341
12515 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12516 instead of true_dependence.
12518 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12520 * bb-reorder.c (branch_threshold): Make const.
12521 (exec_threshold): Ditto.
12523 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12524 Pat Haugen <pthaugen@us.ibm.com>
12525 Peter Bergner <bergner@vnet.ibm.com>
12527 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12528 documentation for the power8 crypto builtins.
12530 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12532 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12533 macros for defining power8 builtin functions.
12534 (BU_P8V_AV_2): Likewise.
12535 (BU_P8V_AV_P): Likewise.
12536 (BU_P8V_VSX_1): Likewise.
12537 (BU_P8V_OVERLOAD_1): Likewise.
12538 (BU_P8V_OVERLOAD_2): Likewise.
12539 (BU_CRYPTO_1): Likewise.
12540 (BU_CRYPTO_2): Likewise.
12541 (BU_CRYPTO_3): Likewise.
12542 (BU_CRYPTO_OVERLOAD_1): Likewise.
12543 (BU_CRYPTO_OVERLOAD_2): Likewise.
12544 (XSCVSPDP): Fix typo, point to the correct instruction.
12545 (VCIPHER): Add power8 crypto builtins.
12546 (VCIPHERLAST): Likewise.
12547 (VNCIPHER): Likewise.
12548 (VNCIPHERLAST): Likewise.
12549 (VPMSUMB): Likewise.
12550 (VPMSUMH): Likewise.
12551 (VPMSUMW): Likewise.
12552 (VPERMXOR_V2DI): Likewise.
12553 (VPERMXOR_V4SI: Likewise.
12554 (VPERMXOR_V8HI: Likewise.
12555 (VPERMXOR_V16QI: Likewise.
12556 (VSHASIGMAW): Likewise.
12557 (VSHASIGMAD): Likewise.
12558 (VPMSUM): Likewise.
12559 (VPERMXOR): Likewise.
12560 (VSHASIGMA): Likewise.
12562 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12563 __CRYPTO__ if the crypto instructions are available.
12564 (altivec_overloaded_builtins): Add support for overloaded power8
12567 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12568 support for power8 crypto builtins.
12569 (builtin_function_type): Likewise.
12570 (altivec_init_builtins): Add support for builtins that take vector
12571 long long (V2DI) arguments.
12573 * config/rs6000/crypto.md: New file, define power8 crypto
12576 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12577 Pat Haugen <pthaugen@us.ibm.com>
12578 Peter Bergner <bergner@vnet.ibm.com>
12580 * doc/invoke.texi (Option Summary): Add power8 options.
12581 (RS/6000 and PowerPC Options): Likewise.
12583 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12584 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12585 wm, wn, wr documentation.
12587 * config/rs6000/constraints.md (wm): New constraint for VSX
12588 registers if direct move instructions are enabled.
12589 (wn): New constraint for no registers.
12590 (wq): New constraint for quad word even GPR registers.
12591 (wr): New constraint if 64-bit instructions are enabled.
12592 (wv): New constraint if power8 vector instructions are enabled.
12593 (wQ): New constraint for quad word memory locations.
12595 * config/rs6000/predicates.md (const_0_to_15_operand): New
12596 constraint for 0..15 for crypto instructions.
12597 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12598 as GPR and floating point registers.
12599 (int_reg_operand): New predicate to match only GPR registers.
12600 (base_reg_operand): New predicate to match base registers.
12601 (quad_int_reg_operand): New predicate to match even GPR registers
12602 for quad memory operations.
12603 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12604 operations in both GPR and VSX registers.
12605 (quad_memory_operand): New predicate for quad memory operations.
12606 (reg_or_indexed_operand): New predicate for direct move support.
12608 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12609 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12610 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12611 (POWERPC_MASKS): Add power8 options.
12612 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12615 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12616 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12618 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12619 (-mpower8-fusion): New power8 options.
12620 (-mpower8-fusion-sign): Likewise.
12621 (-mpower8-vector): Likewise.
12622 (-mcrypto): Likewise.
12623 (-mdirect-move): Likewise.
12624 (-mquad-memory): Likewise.
12626 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12627 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12629 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12630 (rs6000_debug_vector_unit): Add p8_vector.
12631 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12632 definitions. Also print fusion state.
12633 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12634 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12635 (rs6000_option_override_internal): Add support for power8.
12636 (rs6000_common_init_builtins): Add debugging for skipped builtins
12637 if -mdebug=builtin.
12638 (rs6000_adjust_cost): Add power8 support.
12639 (rs6000_issue_rate): Likewise.
12640 (insn_must_be_first_in_group): Likewise.
12641 (insn_must_be_last_in_group): Likewise.
12642 (force_new_group): Likewise.
12643 (rs6000_register_move_cost): Likewise.
12644 (rs6000_opt_masks): Likewise.
12646 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12647 power8 capable assembler, default to power7 options.
12648 (TARGET_DIRECT_MOVE): Likewise.
12649 (TARGET_CRYPTO): Likewise.
12650 (TARGET_P8_VECTOR): Likewise.
12651 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12652 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12653 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12654 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12655 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12656 (TARGET_XSCVDPSPN): Likewise.
12657 (TARGET_XSCVSPDPN): Likewsie.
12658 (TARGET_SYNC_HI_QI): Likewise.
12659 (TARGET_SYNC_TI): Likewise.
12660 (MASK_CRYPTO): Likewise.
12661 (MASK_DIRECT_MOVE): Likewise.
12662 (MASK_P8_FUSION): Likewise.
12663 (MASK_P8_VECTOR): Likewise.
12664 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12665 temporary used by some of the direct move instructions to get two FP
12666 temporary registers does not force creation of a stack frame.
12667 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12668 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12669 that any VSX registers are tieable, even if they are also an
12670 Altivec vector mode.
12671 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12672 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12673 (RS6000_BTM_CRYPTO): Likewise.
12674 (RS6000_BTM_COMMON): Likewise.
12676 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12677 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12678 (enum rs6000_vector): Add power8 vector support.
12680 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12684 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12685 (any_sibcall_could_use_r3): this and handle indirect calls.
12686 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12688 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12690 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12692 2013-05-22 Richard Biener <rguenther@suse.de>
12694 PR middle-end/57349
12695 * profile.c (branch_prob): Do not split blocks that are
12696 abnormally receiving from ECF_RETURNS_TWICE functions.
12698 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12700 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12701 address mode. Move pointer mode initialization to the same place.
12703 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12705 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12706 while it has any effect.
12708 2013-05-21 Easwaran Raman <eraman@google.com>
12710 PR tree-optimization/57322
12711 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12712 UID of the statement added to the BB to be 1.
12714 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12716 PR tree-optimization/57331
12717 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12718 of conversion from pointer type to integral type with integer.
12720 2013-05-21 Martin Jambor <mjambor@suse.cz>
12723 * ipa-prop.c (ipa_read_node_info): Process param_used and
12724 controlled_uses in the same order as when writing.
12726 2013-05-21 Magnus Granberg <baldrick@free.fr>
12729 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12731 2013-05-21 Richard Biener <rguenther@suse.de>
12733 PR tree-optimization/57318
12734 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12735 estimate stmts with side-effects as likely eliminated.
12737 2013-05-21 Richard Biener <rguenther@suse.de>
12739 PR tree-optimization/57330
12740 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12741 preserve the call stmts fntype.
12743 2013-05-21 Richard Biener <rguenther@suse.de>
12745 PR tree-optimization/57303
12746 * tree-ssa-sink.c (statement_sink_location): Improve killing
12747 stmt detection and properly handle self-assignments.
12749 2013-05-21 Christian Bruel <christian.bruel@st.com>
12751 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12752 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12753 registers. Set register size out of the PARALLEL loop.
12755 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12758 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12759 (*fmasf4, *fmasf4_media): New insns.
12761 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12763 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12764 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12765 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12766 (mips_idiv_insns): Update the comments to say that the returned
12767 instruction counts are in units of BASE_INSN_LENGTH.
12768 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12769 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12770 using 2 rather than 4 as the length of indirect MIPS16 and
12771 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12772 length of a NOP. Don't divide MIPS16 lengths by 2.
12773 (mips16_split_long_branches): Assume a branch is long if the
12774 length is greater than 4 rather than 8.
12775 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12776 rather than multiplying them by 2. Multiply instruction counts
12777 by BASE_INSN_LENGTH rather than 4.
12778 (*jump_mips16, tls_get_tp_mips16_<mode>)
12779 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12781 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12783 * config/mips/mips.md (extended_mips16): Remove branch case.
12784 (length): Remove duplicated extended_mips16 test.
12786 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12788 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12790 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12792 * recog.h (Recog_data): Rename to...
12793 (recog_data_d): ...this.
12794 (recog_data): Update accordingly.
12795 * recog.c (recog_data): Likewise.
12796 * reload.c (save_recog_data): Likewise.
12797 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12798 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12800 2013-05-17 Julian Brown <julian@codesourcery.com>
12802 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12803 found in a REG_EQUAL note, invalidate it.
12805 2013-05-17 Easwaran Raman <eraman@google.com>
12807 * tree-ssa-reassoc.c (find_insert_point): New function.
12808 (insert_stmt_after): Likewise.
12809 (get_def_stmt): Likewise.
12810 (ensure_ops_are_available): Likewise.
12811 (not_dominated_by): Likewise.
12812 (rewrite_expr_tree): Do not move statements beyond what is
12813 necessary. Remove call to swap_ops_for_binary_stmt...
12814 (reassociate_bb): ... and move it here.
12815 (build_and_add_sum): Assign UIDs for new statements.
12816 (linearize_expr): Likewise.
12817 (do_reassoc): Renumber gimple statement UIDs.
12819 2013-05-17 Jan Hubicka <jh@suse.cz>
12821 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12823 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12824 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12826 (output_weakrefs): Update.
12828 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12829 Martin Jambor <mjambor@suse.cz>
12831 PR middle-end/57276
12832 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12833 value that corresponds to the given aggval is found in values vector.
12835 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12837 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12838 sse, sse2, sse3, ssse3 and sse4a flags to options.
12840 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12842 * gengtype-state.c: (s_expr_writer): New class, to handle
12843 prettifying of output layout of s-expressions.
12844 (state_writer): New class, to write out gtype.state.
12845 (state_written_type_count): Move this variable into member data of
12847 (s_expr_writer::s_expr_writer): New code: constructor for new class
12848 (state_writer::state_writer(): ditto
12849 (s_expr_writer::write_new_line): New function
12850 (s_expr_writer::write_any_indent): ditto
12851 (s_expr_writer::begin_s_expr): ditto
12852 (s_expr_writer::end_s_expr): ditto
12853 (write_state_fileloc): convert to method of state_writer...
12854 (state_writer:: write_state_fileloc): ...and use methods of
12855 s_expr_writer to write indentation into the gtype.state output file
12856 to visually represent the hierarchical structure of the list
12858 (write_state_fields): ditto, renaming to...
12859 (state_writer::write_state_fields)
12860 (write_state_a_string): ditto, renaming to...
12861 (state_writer::write_state_a_string)
12862 (write_state_string_option): ditto, renaming to...
12863 (state_writer::write_state_string_option)
12864 (write_state_type_option): ditto, renaming to...
12865 (state_writer::write_state_type_option)
12866 (write_state_nested_option): ditto, renaming to...
12867 (state_writer::write_state_nested_option)
12868 (write_state_option): ditto, renaming to...
12869 (state_writer::write_state_option)
12870 (write_state_options): ditto, renaming to...
12871 (state_writer::write_state_options)
12872 (write_state_lang_bitmap): ditto, renaming to...
12873 (state_writer::write_state_lang_bitmap)
12874 (write_state_version): ditto, renaming to...
12875 (state_writer::write_state_version)
12876 (write_state_scalar_type): ditto, renaming to...
12877 (state_writer::write_state_scalar_type)
12878 (write_state_string_type): ditto, renaming to...
12879 (state_writer::write_state_string_type)
12880 (write_state_undefined_type): ditto, renaming to...
12881 (state_writer::write_state_undefined_type)
12882 (write_state_struct_union_type): ditto, renaming to...
12883 (state_writer::write_state_struct_union_type)
12884 (write_state_struct_type): ditto, renaming to...
12885 (state_writer::write_state_struct_type)
12886 (write_state_user_struct_type): ditto, renaming to...
12887 (state_writer::write_state_user_struct_type)
12888 (write_state_lang_struct_type): ditto, renaming to...
12889 (state_writer::write_state_lang_struct_type)
12890 (write_state_param_struct_type): ditto, renaming to...
12891 (state_writer::write_state_param_struct_type)
12892 (write_state_pointer_type): ditto, renaming to...
12893 (state_writer::write_state_pointer_type)
12894 (write_state_array_type): ditto, renaming to...
12895 (state_writer::write_state_array_type)
12896 (write_state_gc_used): ditto, renaming to...
12897 (state_writer::write_state_gc_used)
12898 (write_state_common_type_content): ditto, renaming to...
12899 (state_writer::write_state_common_type_content)
12900 (write_state_type): ditto, renaming to...
12901 (state_writer::write_state_type)
12902 (write_state_pair_list): ditto, renaming to...
12903 (state_writer::write_state_pair_list)
12904 (write_state_pair): ditto, renaming to...
12905 (state_writer::write_state_pair)
12906 (write_state_typedefs): ditto, renaming to...
12907 (state_writer::write_state_typedefs)
12908 (write_state_structures): ditto, renaming to...
12909 (state_writer::write_state_structures)
12910 (write_state_param_structs): ditto, renaming to...
12911 (state_writer::write_state_param_structs)
12912 (write_state_variables): ditto, renaming to...
12913 (state_writer::write_state_variables)
12914 (write_state_srcdir): ditto, renaming to...
12915 (state_writer::write_state_srcdir)
12916 (write_state_files_list): ditto, renaming to...
12917 (state_writer::write_state_files_list)
12918 (write_state_languages): ditto, renaming to...
12919 (state_writer::write_state_languages)
12920 (write_state): create a state_writer instance and use it when
12921 writing out the state file
12923 2013-05-17 Mike Stump <mikestump@comcast.net>
12925 PR rtl-optimization/57304
12926 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12927 accessing DF_REF_REAL_LOC.
12929 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12931 PR rtl-optimization/57281
12932 PR rtl-optimization/57300
12933 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12934 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12935 what the other splitter did if the registers are dead.
12937 2013-05-17 Richard Biener <rguenther@suse.de>
12939 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12942 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12944 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12947 2013-05-17 Marek Polacek <polacek@redhat.com>
12949 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12950 length when doing non-zero store of storing '\0' to '\0'.
12952 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12954 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
12955 vect_external_def oprnd1 with loop_vinfo, try to emit
12956 optional cast, negation and and stmts on the loop preheader
12957 edge instead of into the pattern def seq.
12959 PR tree-optimization/57051
12960 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
12961 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
12963 2013-05-16 Nick Clifton <nickc@redhat.com>
12965 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
12966 (rl78_is_naked_func): New function.
12967 (rl78_expand_prologue): Skip prologue generation for naked functions.
12968 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
12969 * doc/extend.texi (naked): Add RL78 to the list of processors
12970 that supports this attribute.
12972 2013-05-16 Jeff Law <law@redhat.com>
12974 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
12976 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12978 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
12979 cache parameters using detect_caches_amd also for CYRIX,
12980 NSC and TM2 signatures.
12982 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12983 Dzianis Kahanovich <mahatma@eu.by>
12987 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12988 VIA/Centaur processors and determine their cache parameters
12989 using detect_caches_amd.
12991 2013-05-16 Teresa Johnson <tejohnson@google.com>
12993 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
12994 (rtl_verify_edges): New function.
12995 (rtl_verify_bb_insns): Ditto.
12996 (rtl_verify_bb_pointers): Ditto.
12997 (rtl_verify_bb_insn_chain): Ditto.
12998 (rtl_verify_fallthru): Ditto.
12999 (rtl_verify_bb_layout): Ditto.
13000 (rtl_verify_flow_info_1): Outline checks into new functions.
13001 (rtl_verify_flow_info): Ditto.
13003 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13005 * cfghooks.c (copy_bbs): Add update_dominance argument.
13006 * cfghooks.h (copy_bbs): Update prototype.
13007 * tree-cfg.c (gimple_duplicate_sese_region):
13008 Add update_dominance argument.
13009 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13010 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13011 gimple_duplicate_sese_region call.
13012 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13013 Update copy_bbs call.
13014 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13015 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13017 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13019 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13020 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13021 vect_recog_rotate_pattern.
13022 (vect_recog_rotate_pattern): New function.
13024 2013-05-16 Jason Merrill <jason@redhat.com>
13026 * Makefile.in (LLINKER): New variable.
13027 (mostlyclean): Remove link mutex.
13028 * configure.ac: Handle --enable-link-mutex.
13029 * lock-and-run.sh: New script.
13031 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13034 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13037 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13039 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13041 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13043 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13044 * config/arm/arm.c (next_consecutive_mem): New function.
13045 (gen_movmem_ldrd_strd): Likewise.
13046 * config/arm/arm.md (movmemqi): Update condition and code.
13047 (unaligned_loaddi, unaligned_storedi): New patterns.
13049 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13051 * config.gcc: Obsolete *-*-solaris2.9*.
13052 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13054 2013-05-16 Richard Biener <rguenther@suse.de>
13056 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13057 earlier, after GRAPHITE transforms and IV canonicalization.
13059 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13061 * omp-low.c (extract_omp_for_data): For collapsed loops,
13062 if at least one of the loops is known at compile time to
13063 iterate zero times, set count to 0.
13064 (expand_omp_regimplify_p): New function.
13065 (expand_omp_for_generic): For collapsed loops, if at least
13066 one of the loops isn't known to iterate at least once,
13067 add runtime check with setting count to 0.
13068 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13069 For unsigned types if it isn't known at compile time that
13070 the loop will iterate at least once, add runtime check to bypass
13071 the whole loop if initial condition isn't true.
13073 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13075 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13077 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13079 PR middle-end/57286
13080 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13081 transformations to avoid an infinite loop.
13083 2013-05-16 Marek Polacek <polacek@redhat.com>
13085 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13087 2013-05-15 Leif Ekblad <leif@rdos.net>
13089 * config/i386/i386.c (ix86_decompose_address): Use
13090 DEFAULT_TLS_SEG_REG to access TLS segment register.
13091 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13092 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13093 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13095 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13098 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13099 sibling calls to functions that would normally be lazily bound,
13100 unless $gp is call-clobbered.
13102 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13104 * config/i386/i386.c (ix86_option_override_internal): Update
13105 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13106 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13107 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13108 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13110 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13112 2013-05-15 Andreas Schwab <schwab@suse.de>
13114 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13115 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13118 2013-05-15 Teresa Johnson <tejohnson@google.com>
13120 * loop-unroll.c (report_unroll_peel): Check decision before
13121 emitting unroll/peel message.
13123 2013-05-15 Teresa Johnson <tejohnson@google.com>
13125 * function.h (has_bb_partition): New rtl_data flag.
13126 (bb_reorder_complete): Ditto.
13127 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13128 instead of flag_reorder_blocks_and_partition.
13129 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13130 with some enhancements.
13131 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13132 * bb-reorder.c (connect_traces): Check for has_bb_partition
13133 instead of flag_reorder_blocks_and_partition.
13134 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13135 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13136 verify_hot_cold_block_grouping.
13137 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13139 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13142 * config/arm/predicates.md (call_insn_operand): New predicate.
13143 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13144 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13145 if insn is not a tail call.
13146 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13148 * config/arm/arm.h (enum reg_class): New caller save register class.
13149 (REG_CLASS_NAMES): Likewise.
13150 (REG_CLASS_CONTENTS): Likewise.
13151 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13154 2013-05-15 Richard Biener <rguenther@suse.de>
13156 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13157 of MSG_OPTIMIZED_LOCATIONS.
13158 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13159 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13161 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13162 of MSG_OPTIMIZED_LOCATIONS.
13163 (execute_vect_slp): Likewise.
13164 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13165 (vect_create_cond_for_alias_checks): Likewise.
13166 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13167 (vect_recog_widen_mult_pattern): Likewise.
13168 (vect_recog_widen_sum_pattern): Likewise.
13169 (vect_recog_over_widening_pattern): Likewise.
13170 (vect_recog_widen_shift_pattern): Likewise.
13171 (vect_recog_vector_vector_shift_pattern): Likewise.
13172 (vect_recog_divmod_pattern): Likewise.
13173 (vect_recog_mixed_size_cond_pattern): Likewise.
13174 (vect_recog_bool_pattern): Likewise.
13175 (vect_pattern_recog_1): Likewise.
13177 2013-05-15 Martin Jambor <mjambor@suse.cz>
13179 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13180 non-functions to builtin_unreachable.
13181 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13182 correct when new direct edges were discovered.
13184 2013-05-15 Martin Jambor <mjambor@suse.cz>
13186 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13187 header, print symbol order instead of node uid, print more information
13188 about indirect edge targets.
13189 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13191 (ipa_make_edge_direct_to_target): Likewise.
13192 (remove_described_reference): Likewise.
13193 (propagate_controlled_uses): Likewise.
13194 (ipa_print_node_params): Also print symbol order.
13195 (ipcp_transform_function): Print symbol order instead of node uids.
13196 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13197 (cgraph_get_create_real_symbol_node): Likewise.
13198 * ipa-cp.c (print_lattice): Likewise.
13199 (print_all_lattices): Likewise.
13200 (determine_versionability): Likewise.
13201 (initialize_node_lattices): Likewise.
13202 (estimate_local_effects): Likewise.
13203 (update_profiling_info): Likewise.
13204 (create_specialized_node): Likewise.
13205 (perhaps_add_new_callers): Likewise.
13206 (decide_about_value): Likewise.
13207 (decide_whether_version_node): Likewise.
13208 (identify_dead_nodes): Likewise.
13209 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13210 (dump_inline_summary): Likewise.
13211 (estimate_node_size_and_time): Likewise.
13212 (inline_analyze_function): Likewise.
13213 * ipa-inline.c (report_inline_failed_reason): Likewise.
13214 (want_early_inline_function_p): Likewise.
13215 (edge_badness): Likewise.
13216 (update_edge_key): Likewise.
13217 (inline_small_functions): Likewise. Add dumping of order to two other
13219 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13220 instead of node uids.
13221 (propagate_pure_const): Likewise.
13222 (propagate_pure_const): Likewise.
13223 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13224 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13225 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13227 * tree-pretty-print.c (dump_function_header): Likewise.
13228 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13229 Print symbol order instead of node uids.
13231 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13233 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13234 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13236 2013-05-15 Richard Biener <rguenther@suse.de>
13238 PR tree-optimization/57275
13239 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13240 return value for fail to do runtime alias checks for gather loads.
13242 2013-05-15 Jan Hubicka <jh@suse.cz>
13246 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13247 weakrefs are not external.
13248 (lto_symtab_merge_decls): Fix thinko when dealing with
13249 non-lto_symtab decls.
13250 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13251 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13252 * varpool.c (dump_varpool_node): Dump more flags.
13254 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13256 * config/i386/i386.c (processor_alias_table): Add instruction
13257 FSGSBASE for AMD bdver3 architecture.
13259 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13261 * tree.c (warn_deprecated_use): Print file:line using locus color.
13262 * diagnostic.c (diagnostic_report_current_module): Print file:line
13263 and file:line:column using locus color.
13265 2013-05-14 Mike Stump <mikestump@comcast.net>
13267 * gdbinit.in: Add __null.
13269 2013-05-14 Mike Stump <mikestump@comcast.net>
13271 * recog.h: Rename struct recog_data to Recog_data.
13272 * recog.c: Likewise.
13273 * reload.c (can_reload_into): Likewise.
13274 * config/picochip/picochip.c: Likewise.
13276 2013-05-14 Mike Stump <mikestump@comcast.net>
13278 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13280 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13282 * resource.h (struct resources): Remove unch_memory member.
13283 (CLEAR_RESOURCE): Don't clear unch_memory.
13284 * resource.c (mark_referenced_resources): Don't set it.
13285 (mark_set_resources): Likewise.
13286 (mark_target_live_regs): Don't clear it.
13287 (init_resource_info): Likewise.
13288 * reorg.c (resource_conflicts_p): Don't compare it.
13289 (redundant_insn): Don't set it.
13291 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13293 * emit-rtl.c (next_label): Remove unused function.
13294 (skip_consecutive_labels, link_cc0_insns): Move to ...
13295 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13296 only place where these functions are used, and make them static.
13298 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13300 * fold-const.c (fold_negate_expr): Handle vectors.
13301 (fold_truth_not_expr): Make it static.
13302 (fold_invert_truthvalue): New static function.
13303 (invert_truthvalue_loc): Handle vectors. Do not call
13304 fold_truth_not_expr directly.
13305 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13306 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13307 (fold_comparison): Handle vector constants.
13308 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13309 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13310 * tree.h (fold_truth_not_expr): Remove declaration.
13312 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13314 * config/aarch64/aarch64-simd.md
13315 (aarch64_vcond_internal<mode>): Rename to...
13316 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13317 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13318 float modes. Clarify all iterator modes.
13319 (vcond<mode><mode>): Use new name for vcond expanders.
13320 (vcond<v_cmp_result><mode>): Likewise.
13321 (vcondu<mode><mode>: Likewise.
13322 * config/aarch64/iterators.md (VDQF_COND): New.
13324 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13327 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13328 variable for the shift amount. Check that we shift by non-negative
13331 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13334 * config/arm/arm.h (EPILOGUE_USES): Only return true
13335 for LR_REGNUM after epilogue_completed.
13337 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13339 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13340 is error_mark_node.
13342 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13345 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13346 and Solaris 11+/x86 with gld.
13347 * configure: Regenerate.
13349 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13351 * expmed.c (expand_shift_1): Canonicalize rotates by
13352 constant bitsize / 2 to bitsize - 1.
13353 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13354 case ROTATERT>: Likewise.
13357 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13359 * config/i386/i386.md (rotateinv): New code attr.
13360 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13361 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13362 roll $31, %eax, etc.
13364 2013-05-14 Richard Biener <rguenther@suse.de>
13366 PR middle-end/57235
13367 * tree-eh.c (sink_clobbers): Give up for successors with
13368 multiple predecessors and no virtual uses.
13370 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13372 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13373 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13375 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13377 PR middle-end/57251
13378 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13379 the case when both op0 and op1 have VOIDmode.
13381 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13383 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13384 in multiply-accumulate mode.
13386 2013-05-13 Guozhi Wei <carrot@google.com>
13388 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13390 2013-05-13 Kai Tietz <ktietz@redhat.com>
13393 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13394 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13395 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13396 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13397 (ix86_expand_prologue): Likewise.
13398 (ix86_expand_split_stack_prologue): Likewise.
13399 (legitimate_pic_address_disp_p): Likewise.
13400 (legitimize_pic_address): Likewise.
13401 (legitimize_tls_address): Likewise.
13402 (legitimize_pe_coff_symbol): Likewise.
13403 (output_pic_addr_const): Likewise.
13404 (construct_plt_address): Likewise.
13405 (ix86_expand_call): Likewise.
13406 (x86_output_mi_thunk): Likewise.
13407 (x86_function_profiler): Likewise.
13409 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13411 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13412 similar switch cases.
13413 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13414 (aarch64_simd_mov_to_<mode>low): Delete.
13415 (aarch64_simd_mov_to_<mode>high): Delete.
13416 (move_lo_quad_<mode>): Add w<-r alternative.
13417 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13418 (aarch64_simd_mov_from_*): Update type attribute.
13419 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13422 2013-05-13 Jan Hubicka <jh@suse.cz>
13424 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13425 * config/i386/i386.c (ix86_compute_frame_layout,
13426 ix86_expand_epilogue, emit_i387_cw_initialization,
13427 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13428 ix86_local_alignment): Fix use of size/speed predicates.
13430 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13432 PR tree-optimization/45216
13433 PR tree-optimization/57157
13434 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13435 the (-Y) & (B - 1) variant if OP is |.
13436 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13437 return shifted. Use (-op1) & (prec - 1) as other_amount
13438 instead of prec - op1.
13440 2013-05-13 Martin Jambor <mjambor@suse.cz>
13442 PR middle-end/42371
13443 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13444 (ipa_constant_data): New type.
13445 (ipa_jump_func): Use ipa_constant_data to hold information about
13446 constant jump functions.
13447 (ipa_get_jf_constant): Adjust to jump function type changes.
13448 (ipa_get_jf_constant_rdesc): New function.
13449 (ipa_param_descriptor): New field controlled_uses.
13450 (ipa_get_controlled_uses): New function.
13451 (ipa_set_controlled_uses): Likewise.
13452 * ipa-ref.h (ipa_find_reference): Declare.
13453 * ipa-prop.c (ipa_cst_ref_desc): New type.
13454 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13456 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13457 New parameter cs. Adjust all callers.
13458 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13459 (remove_described_reference): New function.
13460 (jfunc_rdesc_usable): Likewise.
13461 (try_make_edge_direct_simple_call): Decrement controlled use count,
13462 attempt to remove reference if it hits zero.
13463 (combine_controlled_uses_counters): New function.
13464 (propagate_controlled_uses): Likewise.
13465 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13466 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13467 (ipa_print_node_params): Print described use counter.
13468 (ipa_write_jump_function): Adjust to jump function type changes.
13469 (ipa_read_jump_function): New parameter CS, pass it to
13470 ipa_set_jf_constant. Adjust caller.
13471 (ipa_write_node_info): Stream controlled use count
13472 (ipa_read_node_info): Likewise.
13473 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13475 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13476 count. Remove cloning-added reference if it reaches zero.
13477 * ipa-ref.c (ipa_find_reference): New function.
13479 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13481 * config/i386/i386.c (processor_target_table): Modified default
13482 alignment values for AMD BD and BT architectures.
13484 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13486 * tree-vect-generic.c (uniform_vector_p): Move ...
13487 * tree.c (uniform_vector_p): ... here.
13488 * tree.h (uniform_vector_p): Declare it.
13489 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13492 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13494 PR tree-optimization/57230
13495 * tree-ssa-strlen.c (handle_char_store): Record length for
13496 array store from STRING_CST.
13498 PR tree-optimization/57230
13499 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13502 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13504 * config/epiphany/epiphany.c (epiphany_init): Check size of
13505 NUM_MODES_FOR_MODE_SWITCHING.
13506 (epiphany_expand_prologue):
13507 Remove CONFIG_REGNUM initial value handling code.
13508 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13509 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13510 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13511 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13512 Don't return 1 for FP_MODE_NONE.
13513 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13514 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13515 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13516 * config/epiphany/epiphany.md (save_config): New pattern.
13518 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13520 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13522 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13524 * config/i386/i386.md (memory): Handle sseishft1.
13525 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13526 (*vec_extractv2di_1): Ditto.
13528 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13530 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13533 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13535 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13536 Add mthumb/march=armv7-a multilib.
13537 Add mthumb/march=armv7-r multilib.
13538 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13540 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13542 * config/v850/t-rtems: Add more multilibs.
13544 2013-05-10 Richard Biener <rguenther@suse.de>
13546 PR tree-optimization/57214
13547 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13548 not propagate from SSA names that occur in abnormal PHI nodes.
13550 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13552 * stor-layout.c (element_precision): New function.
13553 * machmode.h (element_precision): Declare it.
13554 * tree.c (build_minus_one_cst): New function.
13555 (element_precision): Likewise.
13556 * tree.h (build_minus_one_cst): Declare new function.
13557 (element_precision): Likewise.
13558 * fold-const.c (operand_equal_p): Use element_precision.
13559 (fold_binary_loc): Handle vector types.
13560 * convert.c (convert_to_integer): Use element_precision.
13561 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13564 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13566 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13567 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13568 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13569 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13570 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13571 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13572 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13573 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13574 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13575 (Uuw8): New constraints.
13576 (Usb4): Move into alphabetical order.
13577 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13578 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13579 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13580 previously unnamed patterns.
13581 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13582 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13583 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13584 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13585 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13586 of specific lengths.
13588 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13590 * config/i386/i386.md (rotateinv): New code attr.
13591 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13592 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13593 roll $31, %eax, etc.
13595 PR tree-optimization/45216
13596 PR tree-optimization/57157
13597 * tree-ssa-forwprop.c (simplify_rotate): New function.
13598 (ssa_forward_propagate_and_combine): Call it.
13600 2013-05-10 Richard Biener <rguenther@suse.de>
13602 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13603 disable peeling when we version for aliasing.
13604 (vector_alignment_reachable_p): Honor explicit user alignment.
13605 (vect_supportable_dr_alignment): Likewise.
13606 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13607 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13608 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13609 then peeling to arrange for the cost-model check to come first.
13611 2013-05-10 Alan Modra <amodra@gmail.com>
13613 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13614 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13615 * configure: Regenerate.
13617 2013-05-10 Alan Modra <amodra@gmail.com>
13620 * varasm.c (default_elf_select_section): Move !DECL_P check..
13621 (get_named_section): ..to here before calling get_section_name.
13623 (default_section_type_flags): Add DECL_P check.
13624 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13625 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13627 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13629 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13630 When using gen_stack_adjust_str with a register offset, add a
13631 REG_FRAME_RELATED_EXPR note.
13633 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13635 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13636 (*vec_extractv4si_zext_mem): Ditto.
13637 (*vec_extractv2di): Add 0->x and x->x alternatives.
13638 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13639 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13641 2013-05-09 Jason Merrill <jason@redhat.com>
13643 N3639 C++1y VLA support
13644 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13647 * tree.c (build_constructor_va): New.
13648 * tree.h: Declare it.
13650 2013-05-09 Martin Jambor <mjambor@suse.cz>
13653 * gimple-fold.c (canonicalize_constructor_val): Call
13654 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13656 2013-05-09 Jan Hubicka <jh@suse.cz>
13657 Richard Biener <rguenther@suse.de>
13660 * symtab.c (symtab_make_decl_local): Do not add private names.
13662 2013-05-09 Jan Hubicka <jh@suse.cz>
13665 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13666 (unlink_from_assembler_name_hash): Likewise.
13667 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13668 symtab_unregister_node, symtab_initialize_asm_name_hash,
13669 change_decl_assembler_name): Update.
13671 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13673 * config/aarch64/aarch64.md: New movtf split.
13674 (*movtf_aarch64): Update.
13675 (aarch64_movdi_tilow): Handle TF modes and rename to
13676 aarch64_movdi_<mode>low.
13677 (aarch64_movdi_tihigh): Handle TF modes and rename to
13678 aarch64_movdi_<mode>high
13679 (aarch64_movtihigh_di): Handle TF modes and rename to
13680 aarch64_mov<mode>high_di
13681 (aarch64_movtilow_di): Handle TF modes and rename to
13682 aarch64_mov<mode>low_di
13683 (aarch64_movtilow_tilow): Remove spurious whitespace.
13684 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13686 (aarch64_print_operand): Update.
13688 2013-05-09 Alan Modra <amodra@gmail.com>
13690 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13692 * configure: Regenerate.
13694 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13696 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13697 splitter preparation statements.
13698 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13699 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13702 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13704 * gimple-ssa-strength-reduction.c (count_candidates): Change
13705 return value to int.
13706 (analyze_candidates_and_replace): Change type of length to int.
13708 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13710 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13711 (*vec_extract<mode>): Use VI12_128 mode iterator.
13712 (*vec_extract<mode>_mem): Ditto.
13713 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13716 2013-05-08 Diego Novillo <dnovillo@google.com>
13721 2012-08-17 Diego Novillo <dnovillo@google.com>
13724 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13725 * config.in: Regenerate.
13726 * configure: Regenerate.
13727 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13729 2013-05-08 Jan Hubicka <jh@suse.cz>
13732 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13733 * cgraph.h (symtab_node_base): Add unique_name.
13734 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13735 input_overwrite_node, input_varpool_node): Stream unique_name.
13736 * cgraphclones.c (cgraph_create_virtual_clone,
13737 cgraph_function_versioning): Set unique_name.
13738 * ipa.c (function_and_variable_visibility): Set unique_name.
13740 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13742 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13743 (alloc_cand_and_find_basis): Restrict conditional candidate
13744 processing to CAND_MULTs.
13746 2013-05-08 Jan Hubicka <jh@suse.cz>
13749 lto-symtab.c (lto_symtab_symbol_p): New function.
13750 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13751 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13752 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13753 Skip static symbols.
13755 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13757 PR tree-optimization/57200
13758 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13759 Only call inform if the preceding warning_at returns true.
13761 2013-05-07 Han Shen <shenhan@google.com>
13763 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13764 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13765 * common.opt (fstack-protector-strong): New option.
13766 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13767 * doc/invoke.texi (Optimization Options): Document
13768 "-fstack-protector-strong".
13769 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13771 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13773 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13775 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13777 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13778 it is not lhs of the insn.
13779 (setup_reg_equiv): Remove insn having equiv memory which it is not
13781 * lra-constraints.c (process_address): Try to improve generation
13782 code for address base + disp.
13783 (lra_constraints): Make correct the code for checking insn setting
13784 up backward equivalence. Remove insn only if it is in the init
13786 * lra-eliminations.c (update_reg_eliminate): Change return value.
13787 (lra_eliminate): Use the result.
13789 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13791 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13792 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13793 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13794 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13795 register target operands.
13796 (*vec_extractv8hi_sse2): New pattern.
13797 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13798 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13799 (*vec_extract<mode>_mem): New insn and split pattern.
13801 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13803 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13804 (TARGET_ASAN_SHADOW_OFFSET): Define.
13805 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13806 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13808 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13810 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13811 (incr_vec_index): Return -1 if increment not found.
13812 (create_add_on_incoming_edge): Assert if increment not found.
13813 (record_increment): Limit number of increments recorded.
13814 (all_phi_incrs_profitable): Return false if an increment not found.
13815 (replace_profitable_candidates): Don't process increments that were
13817 (analyze_candidates_and_replace): Limit size of incr_vec.
13819 2013-05-07 Richard Biener <rguenther@suse.de>
13821 * calls.c (special_function_p): setjmp-like functions are leaf.
13822 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13823 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13825 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13827 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13828 (aarch64_simd_mov<mode>): New expander.
13829 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13830 (aarch64_simd_mov_to_<mode>high): Likewise.
13831 (aarch64_simd_mov_from_<mode>low): Likewise.
13832 (aarch64_simd_mov_from_<mode>high): Likewise.
13833 (aarch64_dup_lane<mode>): Update.
13834 (aarch64_dup_lanedi): New instruction pattern.
13835 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13836 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13838 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13840 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13841 (replace_mult_candidate): Remove unnecessary argument; remove
13842 unnecessary parameter from call to introduce_cast_before_cand.
13843 (replace_unconditional_candidate): Remove unnecessary parameter
13844 from call to replace_mult_candidate.
13845 (replace_conditional_candidate): Likewise.
13846 (insert_initializers): Use make_temp_ssa_name.
13847 (introduce_cast_before_cand): Remove unnecessary argument; use
13848 make_temp_ssa_name.
13849 (replace_one_candidate): Remove unnecessary argument; remove
13850 unnecessary parameter from calls to introduce_cast_before_cand.
13851 (replace_profitable_candidates): Remove unnecessary parameters
13852 from calls to replace_one_candidate.
13854 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13856 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13857 phi def as possibly hiding a basis for a CAND_ADD whose operands
13858 have been commuted in the analysis.
13859 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13861 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13863 * config/aarch64/aarch64.md
13864 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13865 shift value between 0-4.
13867 2013-05-07 Richard Biener <rguenther@suse.de>
13869 * double-int.h (rshift): New overload.
13870 * double-int.c (rshift): New function.
13871 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13872 (create_reference_ops_from_ref): Remove.
13873 (vn_reference_insert): Use shared ops for constructing the
13874 reference and copy it.
13876 2013-05-07 Richard Biener <rguenther@suse.de>
13878 PR middle-end/57190
13879 * tree-eh.c (sink_clobbers): Properly propagate
13880 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13882 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13884 PR tree-optimization/57149
13885 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13886 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13887 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13888 uninit_undefined_value_p instead of ssa_undefined_value_p.
13891 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13892 for modifier == EXPAND_INITIALIZER.
13894 2013-05-07 Anton Blanchard <anton@samba.org>
13896 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13897 for powerpc64 little endian.
13898 * configure: Regenerate.
13900 2013-05-06 Graham Stott <grahams@btinternet.com>
13902 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13903 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13904 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13905 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13907 2013-05-06 Graham Stott <grahams@btinternet.com>
13909 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13910 codes which allow non-lvalues.
13912 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13914 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13915 components are all 1s.
13916 (integer_minus_onep): New function.
13917 * tree.h (integer_minus_onep): Declare it.
13918 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13919 integer_minus_onep instead of integer_all_onesp.
13921 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13924 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13925 variations of these patterns.
13927 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13929 * config/i386/i386.md (isa): Add x64_sse4 member.
13930 (enabled): Handle x64_sse4.
13931 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13932 instruction for 64bit SSE4_1 targets. Update insn attributes.
13933 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13934 instruction for SSE4_1 targets. Update insn attributes.
13935 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13936 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13937 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13939 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13940 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13941 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13943 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13946 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13948 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13950 * final.c (do_assembler_dialects): Don't handle curly braces and
13951 vertical bar escaped by % as dialect delimiters.
13952 (output_asm_insn): Print curly braces and vertical bar if escaped
13953 by % and ASSEMBLER_DIALECT defined.
13954 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
13955 * doc/tm.texi: Regenerated.
13957 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13959 * config/mips/mips.c: Include tree-pass.h.
13960 (mips_reorg): Split in pre- and post-dbr_schedule parts.
13961 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
13962 (pass_mips_machine_reorg2): New machine specific pass.
13963 (insert_pass_mips_machine_reorg2): New pass plugin definition.
13964 (mips_option_override): Register the new pass.
13965 * rtl.h (cleanup_barriers): Remove prototype.
13966 (dbr_schedule): Likewise.
13967 * jump.c (cleanup_barriers): Make static.
13968 * reorg.c (dbr_schedule): Likewise.
13970 2013-05-06 Richard Biener <rguenther@suse.de>
13972 PR tree-optimization/57185
13973 * tree-parloops.c (add_field_for_reduction): Handle anonymous
13974 SSA names properly.
13976 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13979 * config/i386/i386.c (add_parameter_dependencies): Add dependence
13980 between "first_arg" and "insn", not "last" and "insn".
13982 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13984 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
13985 (find_candidates_in_block): Re-enable slsr_process_phi.
13986 (create_phi_basis): Fix double counting of candidate adjustment.
13988 2013-05-06 Richard Biener <rguenther@suse.de>
13990 PR middle-end/57147
13991 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
13992 the edge is also fallthru, preserve it and just clear the
13994 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
13995 also complex, preserve that and just clear the fallthru flag.
13996 * tree-inline.c (update_ssa_across_abnormal_edges): Also
13997 update virtual operands.
13999 2013-05-06 Alan Modra <amodra@gmail.com>
14001 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14002 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14003 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14004 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14005 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14006 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14008 2013-05-06 Alan Modra <amodra@gmail.com>
14010 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14011 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14012 (DEFAULT_ASM_ENDIAN): Define.
14013 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14014 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14015 Update -K PIC clause from sysv4.h.
14016 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14017 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14019 2013-05-06 Alan Modra <amodra@gmail.com>
14021 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14022 twice for little-endian.
14023 (ashrdi3_no_power, ashrdi3): Support little-endian.
14025 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14028 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14029 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14031 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14032 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14033 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14036 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14037 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14039 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14040 * config/pa/pa.opt: Make mbig-switch a no-op.
14041 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14042 (CASE_VECTOR_MODE): Always return SImode.
14043 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14044 for the !TARGET_BIG_SWITCH case.
14045 * config/pa/pa-linux.h: Likewise.
14046 * config/pa/pa-openbsd.h: Likewise.
14047 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14048 * config/pa/pa.md (short_jump): Remove define_insn.
14049 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14050 (casesi0): Remove define_insn.
14051 (type): Remove btable_branch.
14052 (pa_combine_type): Likewise.
14053 (in_nullified_branch_delay): Likewise.
14054 (in_call_delay): Likewise.
14055 (define_delay): Likewise.
14056 (define_insn_reservation "Z3"): Likewise.
14057 (define_insn_reservation "Z4"): Likewise.
14058 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14059 (pa_adjust_insn_length): Remove adjustment for btable branches.
14060 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14063 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14065 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14066 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14067 Add m->r,x alternatives.
14068 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14069 splitters using SWI48x mode iterator.
14070 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14071 TARGET_64BIT. Add m->x alternative.
14072 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14073 Add o->x alternative. Enable for TARGET_SSE.
14074 (sse_storeq): Remove expander.
14075 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14076 with memory input operand.
14077 (*vec_extractv2di_1 splitter): New.
14078 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14079 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14081 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14083 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14084 (INT_LOWPART): Delete.
14085 (extract_MB): Adjust.
14086 (extract_ME): Adjust.
14087 (print_operand): Adjust.
14089 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14091 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14092 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14093 (reg_or_logical_cint_operand, easy_fp_constant,
14094 logical_const_operand): Delete "CONST_DOUBLE" case.
14095 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14096 "HOST_BITS_PER_WIDE_INT == 64" test.
14097 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14098 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14099 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14100 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14102 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14103 CONST_DOUBLE DImode/VOIDmode case.
14104 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14105 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14106 CONST_DOUBLE VOIDmode case.
14107 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14108 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14109 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14110 Delete CONST_DOUBLE case.
14111 (splitters for mov FMOVE64 const_double): Delete
14112 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14113 "HOST_BITS_PER_WIDE_INT >= 64" test.
14114 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14116 (mov DI const_double): Delete.
14118 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14120 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14121 on op shows all bits zero in mode of a lowpart subreg, return zero.
14123 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14126 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14127 to save TFmode registers and DImode to save TImode registers for
14128 caller save operations.
14129 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14130 mark being partially clobbered since they only use the first
14133 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14134 and TDmode only use the upper 64-bits of each VSX register.
14136 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14138 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14139 (find_candidates_in_block): Disable slsr_process_phi.
14141 2013-05-03 Guozhi Wei <carrot@google.com>
14143 * coverage.c (coverage_obj_init): Move the construction of gcov
14145 (build_init_ctor): ... here.
14147 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14149 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14150 (slsr_cand_d): Redefine def_phi.
14151 (stride_status, phi_adjust_status, count_phis_status): New enums.
14152 (find_phi_def): New.
14153 (find_basis_for_base_expr): New.
14154 (find_basis_for_candidate): Handle hidden bases.
14155 (alloc_cand_and_find_basis): Handle phi candidates.
14156 (slsr_process_phi): New.
14157 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14158 (create_mul_imm_cand): Likewise.
14159 (create_add_ssa_cand): Exclude phi base candidates.
14160 (create_add_imm_cand): Likewise.
14161 (slsr_process_cast): Likewise.
14162 (slsr_process_copy): Likewise.
14163 (find_candidates_in_block): Handle phi candidates.
14164 (dump_candidate): Likewise.
14165 (unconditional_cands): Delete.
14166 (unconditional_cands_with_known_stride_p): Delete.
14167 (phi_dependent_cand_p): New.
14168 (cand_increment): Handle phi-dependent candidates.
14169 (replace_dependent): Delete.
14170 (replace_mult_candidate): New.
14171 (replace_unconditional_candidate): New.
14172 (incr_vec_index): Move to avoid forward reference.
14173 (create_add_on_incoming_edge): New.
14174 (create_phi_basis): New.
14175 (replace_dependents): Delete.
14176 (replace_conditional_candidate): New.
14177 (phi_add_costs): New.
14178 (replace_uncond_cands_and_profitable_phis): New.
14179 (record_increment): Handle phi adjustments.
14180 (record_phi_increments): New.
14181 (record_increments): Handle phi adjustments.
14182 (phi_incr_cost): New.
14183 (lowest_cost_path): Handle phis.
14184 (total_savings): Likewise.
14185 (analyze_increments): Likewise.
14186 (ncd_with_phi): New.
14187 (ncd_of_cand_and_phis): New.
14188 (nearest_common_dominator_for_cands): Handle phi increments.
14189 (all_phi_incrs_profitable): New.
14190 (replace_profitable_candidates): Handle phi-dependent candidates.
14191 (analyze_candidates_and_replace): Likewise.
14193 2013-05-03 Teresa Johnson <tejohnson@google.com>
14196 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14197 do not exceed REG_BR_PROB_BASE.
14199 2013-05-03 Jeff Law <law@redhat.com>
14201 PR tree-optimization/57144
14202 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14203 operand of the condition will bit into the new type when eliminating
14204 a cast feeding a condition.
14206 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14208 PR rtl-optimization/57130
14209 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14210 of COMPARE as in_code to the recursive call if needed.
14212 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14214 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14215 (enabled): Handle new members.
14216 * config/i386/sse.md (*vec_concatv2si): Merge from
14217 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14218 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14220 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14222 PR tree-optimization/57027
14223 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14224 for fnms opportunity, check we got the prerequisite kind
14225 of tree / gimple before using accessor functions.
14227 2013-05-03 Richard Biener <rguenther@suse.de>
14229 * double-int.h (lshift): New overload without precision
14230 and arith argument.
14231 (operator *=, operator +=, operator -=): Move ...
14232 * double-int.c (operator *=, operator +=, operator -=): ... here
14233 and implement more efficiently.
14234 (mul_double_with_sign): Remove.
14235 (lshift_double): Adjust to take unsinged shift argument, push
14236 dispatching code to callers.
14237 (mul_double_wide_with_sign): Add early out for callers that
14238 are not interested in high parts or overflow.
14239 (lshift): New function.
14240 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14241 dispatch code here.
14242 (lrotate, rrotate): Use logical shifts.
14243 * expr.c (get_inner_reference): Use lshift.
14244 * fixed-value.c (do_fixed_divide): Likewise.
14245 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14246 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14247 (indirect_refs_may_alias_p): Likewise.
14248 (stmt_kills_ref_p_1): Likewise.
14250 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14252 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14254 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14256 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14257 scalar form of FABD instruction.
14259 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14261 * lra-constraints.c (process_alt_operands): Add checking alt
14262 number to choose the best alternative.
14264 2013-05-02 Richard Biener <rguenther@suse.de>
14266 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14267 bitmap and its handling.
14268 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14270 2013-05-02 Richard Biener <rguenther@suse.de>
14272 PR middle-end/57140
14273 * tree-inline.c (copy_loops): Properly handle removed loops.
14274 (copy_cfg_body): Mark destination loops for fixup if source
14275 loops needed fixup.
14277 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14280 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14281 generating simple_return for naked functions.
14283 2013-05-02 Martin Jambor <mjambor@suse.cz>
14285 PR middle-end/56988
14286 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14287 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14289 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14290 ipa_agg_replacement_value structures.
14291 (known_aggs_to_agg_replacement_list): Likewise.
14292 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14293 (read_agg_replacement_chain): Likewise.
14294 (ipcp_transform_function): Also check that by_ref flags match.
14296 2013-05-02 Richard Biener <rguenther@suse.de>
14298 * graphds.h (struct graph): Add obstack member.
14299 * graphds.c (new_graph): Initialize obstack and allocate
14301 (add_edge): Allocate edge from the obstack.
14302 (free_graph): Free the obstack instead of all edges and vertices.
14304 2013-05-02 Teresa Johnson <tejohnson@google.com>
14306 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14308 * cfg.c (update_bb_profile_for_threading): Ditto.
14309 * tree-inline.c (copy_bb): Ditto.
14310 (copy_edges_for_bb): Ditto.
14311 (initialize_cfun): Ditto.
14312 (copy_cfg_body): Ditto.
14313 (expand_call_inline): Ditto.
14314 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14315 (estimate_node_size_and_time): Ditto.
14316 (inline_merge_summary): Ditto.
14317 * cgraphclones.c (cgraph_clone_edge): Ditto.
14318 (cgraph_clone_node): Ditto.
14319 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14320 (compute_trg_info): Ditto.
14322 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14324 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14325 S reg when fp attribute set.
14326 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14328 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14330 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14332 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14333 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14334 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14336 2013-05-02 Richard Biener <rguenther@suse.de>
14338 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14339 (struct instantiate_cache_entry): New type.
14340 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14341 (struct instantiate_cache_type): New type.
14342 (set_instantiated_value, get_instantiated_value): Remove.
14343 (get_instantiated_value_entry): New function.
14344 (instantiate_scev_name): Use the new cache and adjust.
14345 (instantiate_scev_poly): Adjust.
14346 (instantiate_scev_binary): Likewise.
14347 (instantiate_array_ref): Likewise.
14348 (instantiate_scev_convert): Likewise.
14349 (instantiate_scev_not): Likewise.
14350 (instantiate_scev_3): Likewise.
14351 (instantiate_scev_2): Likewise.
14352 (instantiate_scev_r): Likewise.
14353 (instantiate_scev): Likewise.
14354 (resolve_mixers): Likewise.
14356 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14359 * lra-constraints.c (best_small_class_operands_num): Remove.
14360 (process_alt_operands): Remove small_class_operands_num. Take
14361 small classes operands into losers and only if the operand is not
14362 matched. Modify debugging output.
14363 (curr_insn_transform): Remove best_small_class_operands_num.
14366 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14368 * config/aarch64/aarch64-builtins.c
14369 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14370 * config/aarch64/aarch64-simd-builtins.def
14371 (reduc_splus_): Add new modes.
14372 (reduc_uplus_): New.
14373 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14374 (reduc_uplus_v4sf): Likewise.
14375 (reduc_splus_v4sf): Likewise.
14376 (aarch64_addv<mode>): Likewise.
14377 (reduc_uplus_<mode>): Likewise.
14378 (reduc_splus_<mode>): Likewise.
14379 (aarch64_addvv2di): Likewise.
14380 (reduc_uplus_v2di): Likewise.
14381 (reduc_splus_v2di): Likewise.
14382 (aarch64_addvv2si): Likewise.
14383 (reduc_uplus_v2si): Likewise.
14384 (reduc_splus_v2si): Likewise.
14385 (reduc_<sur>plus_<mode>): New.
14386 (reduc_<sur>plus_v2di): Likewise.
14387 (reduc_<sur>plus_v2si): Likewise.
14388 (reduc_<sur>plus_v4sf): Likewise.
14389 (aarch64_addpv4sf): Likewise.
14390 * config/aarch64/arm_neon.h
14391 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14392 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14393 add UNSPEC_SADDV, UNSPEC_UADDV.
14395 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14397 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14399 * config/aarch64/arm_neon.h
14400 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14402 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14404 * config/aarch64/aarch64-builtins
14405 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14407 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14409 * config/aarch64/aarch64-simd-builtins.def
14410 (reduc_smax_): New.
14411 (reduc_smin_): Likewise.
14412 (reduc_umax_): Likewise.
14413 (reduc_umin_): Likewise.
14414 (reduc_smax_nan_): Likewise.
14415 (reduc_smin_nan_): Likewise.
14418 (smax): Update for V2SF, V4SF and V2DF modes.
14421 (smin_nan): Likewise.
14422 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14423 (<su><maxmin><mode>3): ...This, refactor.
14424 (s<maxmin><mode>3): New.
14425 (<maxmin_uns><mode>3): Likewise.
14426 (reduc_<maxmin_uns>_<mode>): Refactor.
14427 (reduc_<maxmin_uns>_v4sf): Likewise.
14428 (reduc_<maxmin_uns>_v2si): Likewise.
14429 (aarch64_<fmaxmin><mode>: Remove.
14430 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14432 (vmin<q>_f<32,64>): Likewise.
14433 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14435 (su): Add mappings for smax, smin, umax, umin.
14437 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14438 (FMAXMIN): Rename as...
14439 (FMAXMIN_UNS): ...This.
14441 (fmaxminv): Likewise.
14442 (fmaxmin): Likewise.
14444 (maxmin_uns_op): Likewise.
14446 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14448 * config/aarch64/arm_neon.h
14449 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14450 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14451 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14453 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14455 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14456 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14458 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14460 * config/aarch64/aarch64-simd.md
14461 (vcond<mode>_internal): Handle special cases for constant masks.
14462 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14463 (vcondu<mode><mode>): Likewise.
14464 (vcond<v_cmp_result><mode>): New.
14466 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14468 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14469 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14470 * config/aarch64/aarch64-simd-builtins.def
14471 (cmeq): Update to BUILTIN_VALLDI.
14476 * config/aarch64/arm_neon.h
14477 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14478 to builtins or C as appropriate.
14480 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14482 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14484 (cmhi): Rename to...
14486 * config/aarch64/aarch64-simd.md
14487 (simd_mode): Add SF.
14488 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14489 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14490 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14491 (cstore<mode>_neg): ...This.
14492 * config/aarch64/iterators.md
14494 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14495 (COMPARISONS): New.
14496 (UCOMPARISONS): Likewise.
14497 (optab): Add missing comparisons.
14503 (VCMP_S): Likewise.
14504 (VCMP_U): Likewise.
14505 (V_cmp_result): Add DF, SF modes.
14506 (v_cmp_result): Likewise.
14508 (vmtype): Likewise.
14509 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14511 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14513 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14514 define_insn to define_insn_and_split.
14515 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14516 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14517 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14518 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14519 (thumb2_negscc): Likewise.
14521 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14523 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14525 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14527 * config/arm/thumb2.md: Remove trailing whitespaces.
14529 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14531 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14532 Use gen_int_mode rather than GEN_INT.
14534 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14536 * value-prof.c (stream_in_histogram_value): Remove the strayed
14539 2013-04-30 Richard Biener <rguenther@suse.de>
14541 PR middle-end/57122
14542 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14544 2013-04-30 Richard Biener <rguenther@suse.de>
14546 PR middle-end/57107
14547 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14549 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14551 PR rtl-optimization/56957
14552 PR rtl-optimization/57105
14553 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14554 variable. Use just INSN_UID for determining whether an insn
14555 should be only disconnected from the insn stream.
14556 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14558 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14560 PR tree-optimization/57104
14561 * tsan.c (instrument_expr): Don't instrument accesses to
14562 DECL_HARD_REGISTER VAR_DECLs.
14564 2013-04-30 Richard Biener <rguenther@suse.de>
14566 * function.h (loops_for_fn): New inline function.
14567 (set_loops_for_fn): Likewise.
14568 * cfgloop.h (place_new_loop): Add struct function parameter.
14569 (get_loop): Likewise.
14570 (get_loops): Likewise.
14571 (number_of_loops): Likewise.
14572 (fel_next): Adjust.
14573 (fel_init): Likewise.
14574 * cfg.c (get_loop_copy): Adjust.
14575 * cfgloop.c (flow_loops_dump): Likewise.
14576 (record_loop_exits): Likewise.
14577 (verify_loop_structure): Likewise.
14578 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14579 (estimate_reg_pressure_cost): Likewise.
14580 (mark_loop_exit_edges): Likewise.
14581 * cfgloopmanip.c (place_new_loop): Likewise.
14582 (add_loop): Likewise.
14583 (duplicate_loop): Likewise.
14584 * graph.c (draw_cfg_nodes): Likewise.
14585 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14586 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14587 (extract_affine_chrec): Likewise.
14588 (build_scop_iteration_domain): Likewise.
14589 * graphite.c (graphite_initialize): Likewise.
14590 * ira-build.c (create_loop_tree_nodes): Likewise.
14591 (more_one_region_p): Likewise.
14592 (rebuild_regno_allocno_maps): Likewise.
14593 (mark_loops_for_removal): Likewise.
14594 (mark_all_loops_for_removal): Likewise.
14595 (remove_unnecessary_regions): Likewise.
14596 (ira_build): Likewise.
14597 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14598 * loop-init.c (fix_loop_structure): Likewise.
14599 (gate_rtl_move_loop_invariants): Likewise.
14600 (gate_rtl_unswitch): Likewise.
14601 (gate_rtl_unroll_and_peel_loops): Likewise.
14602 (rtl_doloop): Likewise.
14603 * lto-streamer-in.c (input_cfg): Likewise.
14604 * lto-streamer-out.c (output_cfg): Likewise.
14605 * modulo-sched.c (sms_schedule): Likewise.
14606 * predict.c (tree_estimate_probability): Likewise.
14607 (tree_estimate_probability_driver): Likewise.
14608 (estimate_loops): Likewise.
14609 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14610 (move_sese_region_to_fn): Likewise.
14611 (debug_loop_num): Likewise.
14612 * tree-chrec.c (chrec_evaluate): Likewise.
14613 (hide_evolution_in_other_loops_than_loop): Likewise.
14614 (chrec_component_in_loop_num): Likewise.
14615 (reset_evolution_in_loop): Likewise.
14616 (evolution_function_is_invariant_rec_p): Likewise.
14617 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14618 * tree-inline.c (copy_loops): Likewise.
14619 (copy_cfg_body): Likewise.
14620 (tree_function_versioning): Likewise.
14621 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14622 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14624 (add_to_evolution_1): Likewise.
14625 (scev_const_prop): Likewise.
14626 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14627 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14628 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14629 (tree_ssa_lim_initialize): Likewise.
14630 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14631 (verify_loop_closed_ssa): Likewise.
14632 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14633 (tree_ssa_loop_im): Likewise.
14634 (tree_ssa_loop_unswitch): Likewise.
14635 (tree_vectorize): Likewise.
14636 (check_data_deps): Likewise.
14637 (tree_ssa_loop_ivcanon): Likewise.
14638 (tree_ssa_loop_bounds): Likewise.
14639 (tree_complete_unroll): Likewise.
14640 (tree_complete_unroll_inner): Likewise.
14641 (tree_parallelize_loops): Likewise.
14642 (tree_ssa_loop_prefetch): Likewise.
14643 (tree_ssa_loop_ivopts): Likewise.
14644 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14645 * tree-vectorizer.c (vectorize_loops): Likewise.
14647 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14649 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14650 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14651 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14652 with EABI_LINK_SPEC.
14654 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14657 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14660 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14663 * lra-constraints.c (process_alt_operands): Discourage a bit more
14664 using memory for pseudos. Print cost dump for alternatives.
14665 Modify cost values for conflicts with early clobbers.
14666 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14668 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14671 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14673 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14675 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14676 from/to S register.
14677 (movdi_aarch64): Support LDR/STR from/to D register.
14679 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14681 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14682 or higher by default.
14684 2013-04-29 Richard Biener <rguenther@suse.de>
14686 PR middle-end/57075
14687 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14688 even if not adding abnormal edges for calls that can make
14691 2013-04-29 Richard Biener <rguenther@suse.de>
14693 PR middle-end/57103
14694 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14696 (move_stmt_r): Remove redundant checking.
14698 2013-04-29 Teresa Johnson <tejohnson@google.com>
14701 * basic-block.h (apply_scale): New function.
14702 (apply_probability): Use apply_scale.
14703 * gimple-streamer-in.c (input_bb): Ditto.
14704 * lto-streamer-in.c (input_cfg): Ditto.
14705 * lto-cgraph.c (merge_profile_summaries): Ditto.
14706 * tree-optimize.c (execute_fixup_cfg): Ditto.
14707 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14708 (copy_edges_for_bb): Ditto.
14709 (copy_cfg_body): Ditto.
14711 2013-04-29 Tom de Vries <tom@codesourcery.com>
14713 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14714 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14715 (tail_merge_optimize): Handle current_loops == NULL.
14717 2013-04-26 Jeff Law <law@redhat.com>
14719 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14720 (simplify_cond_using_ranges): Generalize code to simplify
14721 COND_EXPRs where one argument is a constant and the other
14722 is an SSA_NAME created by an integral type conversion.
14724 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14726 * config/arm/arm.md (store_minmaxsi): Use only when
14727 optimize_insn_for_size_p.
14729 2013-04-29 Christian Bruel <christian.bruel@st.com>
14732 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14734 2013-04-29 Richard Biener <rguenther@suse.de>
14736 PR middle-end/57089
14737 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14738 loop tree make sure to schedule a fixup for the child as well.
14739 (expand_omp_for_generic): Properly add loops.
14740 (expand_omp_for_static_nochunk): Likewise.
14741 (expand_omp_for_static_chunk): Likewise.
14742 (expand_omp_for): For the degenerate case fixup loops.
14743 (expand_omp_sections): Fix default bb placement in loops.
14744 (expand_omp_atomic_pipeline): Properly add loops.
14746 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14748 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14750 2013-04-29 Tom de Vries <tom@codesourcery.com>
14752 * tree-ssa-tail-merge.c: Update header comment.
14754 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14756 * config/aarch64/arm_neon.h
14757 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14758 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14759 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14760 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14761 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14762 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14763 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14764 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14766 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14768 * config/aarch64/aarch64-simd.md
14769 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14770 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14771 fix_trunc, fixuns_trunc.
14772 (ftrunc<VDQF:mode>2): New.
14773 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14774 (fix_trunc_optab): New.
14776 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14778 * config/aarch64/aarch64-builtins.c
14779 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14780 iceilf, lround, iroundf.
14782 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14785 * config/i386/i386.h (enum ix86_tune_indices)
14786 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14787 New, split from X86_TUNE_INTER_UNIT_MOVES.
14788 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14789 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14790 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14791 (TARGET_INTER_UNIT_MOVES): Remove.
14792 * config/i386/i386.c (initial_ix86_tune_features): Update.
14793 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14794 (ix86_expand_convert_uns_didf_sse): Use
14795 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14796 (ix86_expand_vector_init_one_nonzero): Ditto.
14797 (ix86_expand_vector_init_interleave): Ditto.
14798 (inline_secondary_memory_needed): Return true for moves from SSE class
14799 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14800 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14801 * config/i386/constraints.md (Yi, Ym): Depend on
14802 TARGET_INTER_UNIT_MOVES_TO_VEC.
14803 (Yj, Yn): New constraints.
14804 * config/i386/i386.md (*movdi_internal): Change constraints of
14805 operand 1 from Yi to Yj and from Ym to Yn.
14806 (*movsi_internal): Ditto.
14807 (*movdf_internal): Ditto.
14808 (*movsf_internal): Ditto.
14809 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14810 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14811 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14812 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14813 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14814 * config/i386/sse.md (movdi_to_sse): Ditto.
14815 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14816 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14817 TARGET_INTER_UNIT_MOVES.
14818 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14819 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14820 instead of TARGET_INTER_UNIT_MOVES.
14821 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14822 operand 1 from Yi to Yj and from Ym to Yn.
14824 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14826 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14827 (float_truncate_hi_): Likewise.
14828 (float_extend_lo_): Likewise.
14829 (float_truncate_lo_): Likewise.
14830 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14831 (aarch64_float_extend_lo_v2df): Likewise.
14832 (vec_unpacks_hi_v4sf): Likewise.
14833 (aarch64_float_truncate_lo_v2sf): Likewise.
14834 (aarch64_float_truncate_hi_v4sf): Likewise.
14835 (vec_pack_trunc_v2df): Likewise.
14836 (vec_pack_trunc_df): Likewise.
14838 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14840 * config/aarch64/aarch64-builtins.c
14841 (aarch64_fold_builtin): Fold float conversions.
14842 * config/aarch64/aarch64-simd-builtins.def
14843 (floatv2si, floatv4si, floatv2di): New.
14844 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14845 * config/aarch64/aarch64-simd.md
14846 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14847 * config/aarch64/iterators.md (FLOATUORS): New.
14848 (optab): Add float, floatuns.
14849 (su_optab): Likewise.
14851 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14853 * config/aarch64/aarch64-builtins.c
14854 (aarch64_builtin_vectorized_function): Use new names for
14856 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14857 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14858 (fcvtzu): Split as...
14859 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14860 (fcvtas): Split as...
14861 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14862 (fcvtau): Split as...
14863 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14864 (fcvtps): Split as...
14865 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14866 (fcvtpu): Split as...
14867 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14868 (fcvtms): Split as...
14869 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14870 (fcvtmu): Split as...
14871 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14872 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14873 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14874 (lfrintnusf, lfrintnudf): Likewise.
14875 * config/aarch64/aarch64-simd.md
14876 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14878 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14879 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14880 (fcvt_pattern): Likewise.
14882 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14884 * config/aarch64/aarch64-simd.md
14885 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14886 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14888 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14890 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14891 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14892 (vrnd<a,m,n,p>_f32): Implement using builtins.
14893 (vrnd<i,x><q>_f<32, 64>): New.
14895 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14897 * config/aarch64/aarch64-builtins.c
14898 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14899 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14900 (frintz): Rename to...
14902 (frintp): Rename to...
14904 (frintm): Rename to...
14906 (frinti): Rename to...
14907 (nearbyint): ...this.
14908 (frintx): Rename to...
14910 (frinta): Rename to...
14912 * config/aarch64/aarch64-simd.md
14913 (aarch64_frint<frint_suffix><mode>): Delete.
14914 (<frint_pattern><mode>2): Convert to insn.
14915 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14916 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14917 (frint_pattern): Likewise.
14918 (frint_suffix): Likewise.
14920 2013-04-29 Richard Biener <rguenther@suse.de>
14922 PR tree-optimization/57081
14923 * loop-init.c: Include tree-flow.h.
14924 (loop_optimizer_finalize): Free number of iteration estimates.
14925 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14927 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14929 PR tree-optimization/57083
14930 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14931 non-singleton shift count range, zero extend low_bound for uns case.
14933 * config/i386/predicates.md (general_vector_operand): New predicate.
14934 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14935 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14936 if they aren't nonimmediate operands. If their original values
14937 satisfy const_vector_equal_evenodd_p, don't shift them.
14938 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14939 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14940 if not nonimmediate_operand.
14941 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14942 instead of register_operand.
14943 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14945 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14947 * stor-layout.c (finalize_size_functions): Allocate a structure and
14948 reset cfun before dumping the functions.
14950 2013-04-27 Jakub Jelinek <jakub@redhat.com>
14952 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
14955 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
14956 use xop_pmacsdqh if uns_p.
14957 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
14958 the immediate rotate count.
14960 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
14962 * rtl.h (struct rtx_def): Add comment for field jump.
14963 (LRA_SUBREG_P): New macro.
14964 * recog.c (register_operand): Check LRA_SUBREG_P.
14965 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14967 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14968 (lra_final_code_change): Skip subreg change for operators.
14969 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14970 if there are no operand changes.
14971 * lra-constraints.c (curr_insn_set): New.
14972 (match_reload): Set LRA_SUBREG_P.
14973 (emit_spill_move): Ditto.
14974 (check_and_process_move): Use curr_insn_set. Process only single
14975 set insns. Don't initialize sec_mem_p and change_p.
14976 (simplify_operand_subreg): Use LRA_SUBREG_P.
14977 (reg_in_class_p): New function.
14978 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14979 of #ifdef. Add code to remove cycling.
14980 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14981 non-null disp. Reload inner instead of disp when base and index
14982 are null. Try to put lo_sum into register.
14983 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14984 (check_and_process_move): Move code for move cost check to
14985 simple_move_p. Remove equiv_substitution.
14986 (simple_move_p): New function.
14987 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14988 curr_insn_set. Call check_and_process_move only for single set
14989 insns. Use the new function. Move call of check_and_process_move
14990 after operand equiv substitution and address process.
14992 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14995 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
14996 with nonlocal goto receivers or returns twice calls, ignore
14997 unininitialized values from abnormal edges to nl goto receiver
14998 or returns twice call.
15000 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15002 PR tree-optimization/57051
15003 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15004 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15007 2013-04-26 Richard Biener <rguenther@suse.de>
15009 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15010 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15011 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15012 (fixup_loop_arrays_after_move): New function.
15013 (move_sese_region_to_fn): Properly outline the loop tree parts
15014 of the SESE region.
15016 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15018 * config/i386/i386.md (type, unit): Fix long lines.
15020 2013-04-26 Richard Biener <rguenther@suse.de>
15022 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15023 (lto-streamer-out.o): Likewise.
15024 * cfgloop.c (init_loops_structure): Export, add struct function
15025 argument and adjust.
15026 (flow_loops_find): Adjust.
15027 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15028 (init_loops_structure): Declare.
15029 * lto-streamer-in.c: Include cfgloop.h.
15030 (input_cfg): Input the loop tree.
15031 * lto-streamer-out.c: Include cfgloop.h.
15032 (output_cfg): Output the loop tree.
15033 (output_struct_function_base): Do not drop PROP_loops.
15035 2013-03-26 Richard Biener <rguenther@suse.de>
15037 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15038 (pass_build_cfg): Provide PROP_loops.
15039 (move_sese_region_to_fn): Remove loops that are outlined into fn
15041 * tree-inline.c: Include cfgloop.h.
15042 (initialize_cfun): Do not drop PROP_loops.
15043 (copy_loops): New function.
15044 (copy_cfg_body): Copy loop structure.
15045 (tree_function_versioning): Initialize destination loop tree.
15046 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15047 (pass_parallelize_loops): Do IL verification.
15048 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15049 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15050 the CFG make sure we fixup loops as well.
15051 * tree-ssa-tail-merge.c: Include cfgloop.h.
15052 (replace_block_by): When merging loop latches mark loops for fixup.
15053 * lto-streamer-out.c (output_struct_function_base): Drop
15054 PROP_loops for now.
15055 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15056 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15057 * ipa-split.c: Include cfgloop.h.
15058 (split_function): Add the new return block to the loop tree root.
15059 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15060 whether we have removed the forwarder block.
15061 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15062 * cfgloop.h (place_new_loop): Declare.
15063 * cfgloopmanip.c (place_new_loop): Export.
15064 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15065 (tree-switch-conversion.o): Likewise.
15066 (tree-complex.o): Likewise.
15067 (tree-inline.o): Likewise.
15068 (tree-ssa-tailmerge.o): Likewise.
15069 (ipa-split.o): Likewise.
15070 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15071 (tree-ssa-copy.o): Likewise.
15072 * tree-switch-conversion.c: Include cfgloop.h
15073 (process_switch): If we emit a bit-test cascade, schedule loops
15075 * tree-complex.c: Include cfgloop.h.
15076 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15077 * asan.c: Include cfgloop.h.
15078 (create_cond_insert_point): Properly add new basic-blocks to
15079 loops, schedule loop fixup.
15080 * cfgloop.c (verify_loop_structure): Check that looks are not
15082 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15084 (expand_omp_for_generic): Likewise.
15085 (expand_omp_sections): Likewise.
15086 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15087 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15088 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15089 is initialized, not when loops are present.
15090 * tree-parloops.c (parallelize_loops): Remove checking here.
15091 * passes.c (init_optimization_passes): Schedule a copy-propagation
15092 pass before complete unrolling of inner loops.
15094 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15096 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15097 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15098 (colorize_init): Add argument to _WIN32 version.
15099 * toplev.c: Include diagnostic-color.h.
15100 (process_options): Default to -fdiagnostics-color=auto if
15101 GCC_COLORS env var is in the environment.
15102 * common.opt (fdiagnostics-color=): Add Var and Init.
15103 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15104 env var is in the environment, the default is auto rather than never.
15106 * diagnostic.h (file_name_as_prefix): Add context argument.
15107 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15108 the string as locus.
15109 * langhooks.c (lhd_print_error_function): Adjust caller.
15111 2013-04-25 Lawrence Crowl <crowl@google.com>
15113 * var-tracking.c (shared_hash_def::htab):
15114 Change type to hash_table. Update dependent calls and types.
15116 2013-04-25 Lawrence Crowl <crowl@google.com>
15118 * Makefile.in: Update as needed below.
15120 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15121 Move declaration to after the type's method definitons.
15123 * attribs.c (htab_t scoped_attributes::attribute_hash):
15124 Change type to hash_table. Update dependent calls and types.
15126 * bitmap.c (htab_t bitmap_desc_hash):
15127 Change type to hash_table. Update dependent calls and types.
15129 * cselib.c (htab_t cselib_hash_table):
15130 Change type to hash_table. Update dependent calls and types.
15132 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15133 (hash_string_slot_node): Move implementation into lto-streamer.h
15134 struct string_slot_hasher.
15135 (eq_string_slot_node): Likewise.
15137 * data-streamer-out.c: Update output_block::string_hash_table
15138 dependent calls and types.
15140 * dwarf2cfi.c (htab_t trace_index):
15141 Change type to hash_table. Update dependent calls and types.
15143 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15144 Change type to hash_table. Update dependent calls and types.
15145 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15146 (htab_t optimize_external_refs::map): Likewise.
15147 (htab_t output_comp_unit::extern_map): Likewise.
15148 (htab_t output_comdat_type_unit::extern_map): Likewise.
15149 (htab_t output_macinfo::macinfo_htab): Likewise.
15150 (htab_t optimize_location_lists::htab): Likewise.
15151 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15153 * except.c (htab_t ehspec_hash_type):
15154 Change type to hash_table. Update dependent calls and types.
15155 (assign_filter_values::ttypes): Likewise.
15156 (assign_filter_values::ehspec): Likewise.
15157 (sjlj_assign_call_site_values::ar_hash): Likewise.
15158 (convert_to_eh_region_ranges::ar_hash): Likewise.
15160 * gcse.c (htab_t pre_ldst_table):
15161 Change type to hash_table. Update dependent calls and types.
15163 * ggc-common.c (htab_t saving_htab):
15164 Change type to hash_table. Update dependent calls and types.
15165 (htab_t loc_hash): Likewise.
15166 (htab_t ptr_hash): Likewise.
15167 (call_count): Rename ggc_call_count.
15168 (call_alloc): Rename ggc_call_alloc.
15169 (loc_descriptor): Rename make_loc_descriptor.
15170 (add_statistics): Rename ggc_add_statistics.
15172 * ggc-common.c (saving_htab):
15173 Change type to hash_table. Update dependent calls and types.
15175 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15176 (push_gimplify_context): Likewise.
15177 (pop_gimplify_context): Likewise.
15178 (struct gimple_temp_hash_elt): Added.
15179 (struct gimplify_hasher): Likewise.
15180 (struct gimplify_ctx.temp_htab):
15181 Change type to hash_table. Update dependent calls and types.
15183 * gimple-fold.c: Include gimplify-ctx.h.
15185 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15186 Change type to hash_table. Update dependent calls and types.
15187 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15188 avoid potential global name collision.
15190 * gimplify.c: Include gimplify-ctx.h.
15191 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15192 (htab_t gimplify_ctx::temp_htab):
15193 Update dependent calls and types for new type hash_table.
15194 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15195 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15197 * gimplify-ctx.h: New.
15198 (struct gimple_temp_hash_elt): Move from gimplify.c.
15199 (class gimplify_hasher): New.
15200 (struct gimplify_ctx): Move from gimple.h.
15201 (htab_t gimplify_ctx::temp_htab):
15202 Change type to hash_table. Update dependent calls and types.
15204 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15205 (htab_t ivs_params::newivs_index):
15206 Change type to hash_table. Update dependent calls and types.
15207 (htab_t ivs_params::params_index): Likewise.
15208 (htab_t print_generated_program::params_index): Likewise.
15209 (htab_t gloog::newivs_index): Likewise.
15210 (htab_t gloog::params_index): Likewise.
15212 * graphite.c: Include graphite-htab.h.
15213 4htab_t graphite_transform_loops::bb_pbb_mapping):
15214 Change type to hash_table. Update dependent calls and types.
15216 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15217 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15218 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15220 * graphite-dependences.c: Include graphite-htab.h.
15221 (loop_is_parallel_p): Change hash table type of parameter.
15223 * graphite-htab.h: New.
15224 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15225 (extern find_pbb_via_hash): Move from graphite-poly.h.
15226 (extern loop_is_parallel_p): Move from graphite-poly.h.
15227 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15229 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15230 (extern loop_is_parallel_p): Move to graphite-htab.h.
15231 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15233 * haifa-sched.c (htab_t delay_htab):
15234 Change type to hash_table. Update dependent calls and types.
15235 (htab_t delay_htab_i2): Likewise.
15237 * ira-color.c (htab_t allocno_hard_regs_htab):
15238 Change type to hash_table. Update dependent calls and types.
15240 * ira-costs.c (htab_t cost_classes_htab):
15241 Change type to hash_table. Update dependent calls and types.
15243 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15244 Change type to hash_table. Update dependent calls and types.
15246 * loop-iv.c (htab_t bivs):
15247 Change type to hash_table. Update dependent calls and types.
15249 * loop-unroll.c (htab_t opt_info::insns_to_split):
15250 Change type to hash_table. Update dependent calls and types.
15251 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15253 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15254 (struct string_slot_hasher): New.
15255 (htab_t output_block::string_hash_table):
15256 Change type to hash_table. Update dependent calls and types.
15258 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15259 (htab_t file_name_hash_table):
15260 Change type to hash_table. Update dependent calls and types.
15262 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15265 * lto-streamer.c (htab_t tree_htab):
15266 Change type to hash_table. Update dependent calls and types.
15268 * omp-low.c: Include gimplify-ctx.h.
15270 * passes.c (htab_t name_to_pass_map):
15271 Change type to hash_table. Update dependent calls and types.
15272 (pass_traverse): Rename to passes_pass_traverse.
15274 * plugin.c (htab_t event_tab):
15275 Change type to hash_table. Update dependent calls and types.
15277 * postreload-gcse.c (htab_t expr_table):
15278 Change type to hash_table. Update dependent calls and types.
15279 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15281 * sese.c (debug_rename_map_1): Make extern.
15282 (htab_t copy_bb_and_scalar_dependences::rename_map):
15283 Change type to hash_table. Update dependent calls and types.
15285 * sese.h (extern debug_rename_map): Move to .c file.
15287 * store-motion.c (htab_t store_motion_mems_table):
15288 Change type to hash_table. Update dependent calls and types.
15290 * trans-mem.c (htab_t tm_new_mem_hash):
15291 Change type to hash_table. Update dependent calls and types.
15293 * tree-browser.c (htab_t TB_up_ht):
15294 Change type to hash_table. Update dependent calls and types.
15296 * tree-cfg.c (htab_t discriminator_per_locus):
15297 Change type to hash_table. Update dependent calls and types.
15299 * tree-complex.c: Include tree-hasher.h
15300 (htab_t complex_variable_components):
15301 Change type to hash_table. Update dependent calls and types.
15303 * tree-eh.c (htab_t finally_tree):
15304 Change type to hash_table. Update dependent calls and types.
15306 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15307 struct int_tree_hasher.
15308 (extern int_tree_map_eq): Likewise.
15309 (uid_decl_map_hash): Removed.
15310 (extern decl_tree_map_eq): Likewise.
15312 * tree-hasher.h: New.
15313 (struct int_tree_hasher): New.
15314 (typedef int_tree_htab_type): New.
15316 * tree-inline.c: Include gimplify-ctx.h.
15318 * tree-mudflap.c: Include gimplify-ctx.h.
15320 * tree-parloops.c: Include tree-hasher.h.
15321 (htab_t eliminate_local_variables_stmt::decl_address):
15322 Change type to hash_table. Update dependent calls and types.
15323 (htab_t separate_decls_in_region::decl_copies): Likewise.
15325 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15326 Change type to hash_table. Update dependent calls and types.
15328 * tree-sra.c (candidates):
15329 Change type to hash_table. Update dependent calls and types.
15331 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15333 (int_tree_map_hash): Likewise.
15335 * tree-ssa-dom.c (htab_t avail_exprs):
15336 Change type to hash_table. Update dependent calls and types.
15338 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15339 Change type to hash_table. Update dependent calls and types.
15341 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15342 Change type to hash_table. Update dependent calls and types.
15344 * tree-ssa-phiopt.c (seen_ssa_names):
15345 Change type to hash_table. Update dependent calls and types.
15347 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15348 Change type to hash_table. Update dependent calls and types.
15350 * tree-ssa-uncprop.c (equiv):
15351 Change type to hash_table. Update dependent calls and types.
15353 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15355 PR rtl-optimization/57003
15356 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15357 call note_stores with kill_clobbered_value callback again after
15358 killing regs_invalidated_by_call.
15360 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15362 * config/aarch64/aarch64-simd.md
15363 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15364 (aarch64_simd_bsl<mode>): Likewise.
15365 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15367 2013-04-25 Marek Polacek <polacek@redhat.com>
15369 PR tree-optimization/57066
15370 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15372 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15374 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15376 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15378 * config/aarch64/aarch64-builtins.c
15379 (aarch64_fold_builtin): New.
15380 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15381 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15382 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15383 * config/aarch64/arm_neon.h
15384 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15386 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15387 Tejas Belagod <tejas.belagod@arm.com>
15389 * config/aarch64/aarch64-builtins.c
15390 (aarch64_gimple_fold_builtin): New.
15391 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15392 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15393 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15394 (addvv4sf): Update.
15395 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15397 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15399 * config/aarch64/aarch64.md
15400 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15402 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15404 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15405 (*ngcsi_uxtw): New pattern.
15407 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15408 Julian Brown <julian@codesourcery.com>
15410 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15411 (TB_DREG): Add T_V4HF.
15412 (v4hf_UP): New macro.
15413 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15414 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15415 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15417 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15418 (arm_vector_mode_supported_p): Handle V4HF.
15419 (arm_mangle_map): Handle V4HFmode.
15420 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15421 * config/arm/arm_neon_builtins.def: Add entries for
15422 vcvtv4hfv4sf, vcvtv4sfv4hf.
15423 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15424 (neon_vcvtv4hfv4sf): Likewise.
15425 * config/arm/neon-gen.ml: Handle half-precision floating point
15427 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15428 * config/arm/arm_neon.h: Regenerate.
15429 * config/arm/neon.ml (type elts): Add F16.
15430 (type vectype): Add T_float16x4, T_floatHF.
15431 (type vecmode): Add V4HF.
15432 (type features): Add Requires_FP_bit feature.
15433 (elt_width): Handle F16.
15434 (elt_class): Likewise.
15435 (elt_of_class_width): Likewise.
15436 (mode_of_elt): Refactor.
15437 (type_for_elt): Handle F16, fix error messages.
15438 (vectype_size): Handle T_float16x4.
15439 (vcvt_sh): New function.
15440 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15441 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15442 (string_of_mode): Handle V4HF.
15443 * doc/arm-neon-intrinsics.texi: Regenerate.
15445 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15447 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15448 format specifier in 'X' case.
15450 2013-04-25 Alan Modra <amodra@gmail.com>
15453 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15454 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15455 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15456 Repeat for many other rotate/shift and mask patterns using subregs.
15457 Name lshiftrt insns.
15458 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15459 on WORDS_BIG_ENDIAN.
15461 2013-04-25 Alan Modra <amodra@gmail.com>
15463 * config.gcc: Support little-endian powerpc-linux targets.
15464 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15465 (LINK_OS_LINUX_SPEC): Define.
15466 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15467 Preserve MASK_LITTLE_ENDIAN.
15468 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15469 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15470 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15471 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15472 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15473 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15474 Correct fp word order for little-endian. Don't shift toc entries
15475 smaller than a word for little-endian.
15476 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15477 (bswapdi2 splits): Correct low-part subreg for little-endian.
15478 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15479 low/high where such is correct only for be.
15480 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15481 little-endian for -mcall-aixdesc.
15483 2013-04-25 Alan Modra <amodra@gmail.com>
15485 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15486 replace_equiv_address_nv.
15488 2013-04-25 Alan Modra <amodra@gmail.com>
15490 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15492 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15495 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15496 * rtl.h (struct rtx_def): ...
15498 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15500 PR rtl-optimizations/57046
15501 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15502 for multi-reg splits.
15504 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15506 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15508 2013-04-24 Sterling Augustine <saugustine@google.com>
15510 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15511 (comp_dir_string, debug_str_dwo_section): New.
15512 (DEBUG_STR_DWO_SECTION): Rename to ...
15513 (DEBUG_DWO_STR_SECTION): ... this.
15514 (DEBUG_NORM_STR_SECTION): Delete.
15515 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15516 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15517 (find_AT_string): Move most logic to ...
15518 (find_AT_string_in_table): ... here. New.
15519 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15520 add_skeleton_AT_string. Delete logic.
15521 (output_skeleton_debug_sections): Remove call to
15522 add_top_level_skeleton_die_attrs.
15523 (add_comp_dir_attribute): Move logic to comp_dir_string.
15524 (dwarf2out_init): Initialize debug_str_dwo_section.
15525 (output_indirect_string): Call find_string_form.
15526 (output_indirect_strings): Rewrite.
15527 (prune_unused_types): Empty skeleton_debug_str_hash.
15528 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15529 (dwarf2out_finish): Call output_indirect_strings.
15531 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15533 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15535 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15537 * rtl.h (struct rtx_def): Add comment for field jump.
15538 (LRA_SUBREG_P): New macro.
15539 * recog.c (register_operand): Check LRA_SUBREG_P.
15540 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15542 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15543 (lra_final_code_change): Skip subreg change for operators.
15544 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15545 if there are no operand changes.
15546 * lra-constraints.c (curr_insn_set): New.
15547 (match_reload): Set LRA_SUBREG_P.
15548 (emit_spill_move): Ditto.
15549 (check_and_process_move): Use curr_insn_set. Process only single
15550 set insns. Don't initialize sec_mem_p and change_p.
15551 (simplify_operand_subreg): Use LRA_SUBREG_P.
15552 (reg_in_class_p): New function.
15553 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15554 of #ifdef. Add code to remove cycling.
15555 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15556 non-null disp. Reload inner instead of disp when base and index
15557 are null. Try to put lo_sum into register.
15558 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15559 (check_and_process_move): Move code for move cost check to
15560 simple_move_p. Remove equiv_substitution.
15561 (simple_move_p): New function.
15562 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15563 curr_insn_set. Call check_and_process_move only for single set
15564 insns. Use the new function. Move call of check_and_process_move
15565 after operand equiv substitution and address process.
15567 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15569 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15570 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15571 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15573 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15575 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15577 2013-04-24 Marek Polacek <polacek@redhat.com>
15579 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15580 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15581 (select_loops_exit_conditions): Likewise.
15582 (number_of_iterations_for_all_loops): Likewise.
15583 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15584 (scev_analysis): Likewise.
15586 2013-04-02 Catherine Moore <clm@codesourcery.com>
15587 Chao-ying Fu <fu@mips.com>
15589 * config/mips/micromips.md (jraddiusp): New pattern.
15590 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15591 instruction if possible.
15593 2013-04-24 Alan Modra <amodra@gmail.com>
15595 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15597 2013-04-24 Julian Brown <julian@codesourcery.com>
15598 Chung-Lin Tang <cltang@codesourcery.com>
15600 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15601 dependency behavior in enumeration type DIE generation. Add TODO note
15602 to comments about future DW_FORM_sdata/udata re-work of related code.
15604 2013-04-23 Lawrence Crowl <crowl@google.com>
15606 * Makefile.in: Update as needed below.
15608 * hash-table.h (class hash_table):
15609 Correct many methods with parameter types compare_type to the correct
15610 value_type. (Correct code was unlikely to notice the change.)
15611 (hash_table::elements_with_deleted) New.
15612 (class hashtable::iterator): New.
15613 (hashtable::begin()): New.
15614 (hashtable::end()): New.
15615 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15617 * statistics.c (statistics_hashes):
15618 Change type to hash_table. Update dependent calls and types.
15620 * tree-into-ssa.c (var_infos):
15621 Change type to hash_table. Update dependent calls and types.
15623 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15624 Change type to hash_table. Update dependent calls and types.
15626 * tree-ssa-loop-im.c (struct mem_ref.refs):
15627 Change type to hash_table. Update dependent calls and types.
15629 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15630 Change type to hash_table. Update dependent calls and types.
15632 * tree-ssa-sccvn.c (vn_tables_s::nary):
15633 Change type to hash_table. Update dependent calls and types.
15634 (vn_tables_s::phis): Likewise.
15635 (vn_tables_s::references): Likewise.
15637 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15638 (vn_reference_eq): Update parameter and return types.
15640 * tree-ssa-structalias.c (pointer_equiv_class_table):
15641 Change type to hash_table. Update dependent calls and types.
15642 (location_equiv_class_table): Likewise.
15644 * tree-vect-data-refs.c: Consequential changes for making
15645 peeling a hash_table.
15647 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15648 (destroy_loop_vec_info): Dependent hash_table update.
15650 * tree-vectorizer.h (peeling_htab):
15651 Change type to hash_table. Update dependent calls and types.
15653 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15655 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15656 to check the register content is equal or not.
15657 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15658 to assign register content record.
15659 * lra-eliminations.c (update_reg_eliminate): Use
15660 lra_update_reg_val_offset to update register content offset.
15661 * lra-int.h (struct lra_reg): Add offset member.
15662 (lra_reg_val_equal_p): New static inline function.
15663 (lra_update_reg_val_offset): New static inline function.
15664 (lra_assign_reg_val): New static inline function.
15665 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15666 to assign register content record.
15667 (initialize_lra_reg_info_element): Initial offset to zero.
15669 2013-04-23 Catherine Moore <clm@codesourcery.com>
15671 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15672 operands. Record compression.
15674 2013-04-23 Xinliang David Li <davidxl@google.com>
15676 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15678 2013-04-23 Richard Biener <rguenther@suse.de>
15680 PR middle-end/57036
15681 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15682 parameter, only add abnormal goto edges from the copied body
15683 if the call could perform abnormal gotos.
15684 (copy_cfg_body): Adjust.
15686 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15688 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15690 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15692 * coretypes.h (gimple_stmt_iterator): Add struct to make
15695 2013-04-23 Richard Biener <rguenther@suse.de>
15697 PR tree-optimization/57026
15698 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15699 from SSA names occuring in abnormal PHI nodes.
15701 2013-04-22 Andi Kleen <ak@linux.intel.com>
15703 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15705 2013-04-22 Andi Kleen <ak@linux.intel.com>
15707 * lto/lto.c (print_lto_report_1): Declare early.
15708 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15710 2013-04-22 Andi Kleen <ak@linux.intel.com>
15712 * common.opt (-flto-report-wpa): Add.
15713 * doc/invoke.texi (-flto-report-wpa): Add.
15714 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15717 2013-04-22 Xinliang David Li <davidxl@google.com>
15719 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15720 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15721 * Makefile.in: New dependency
15723 David Daney <ddaney.cavm@gmail.com>
15725 * configure.ac (gcc_cv_as_micromips_support): Use the
15726 --fatal-warnings option.
15727 * configure: Regenerate.
15729 2013-04-22 Marek Polacek <polacek@redhat.com>
15732 * tsan.c (instrument_expr): Don't instrument expression
15733 in case its size is zero.
15735 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15739 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15741 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15743 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15745 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15746 (gimple_stmt_iterator): New typedef.
15747 * gimple.h (gimple_stmt_iterator): Rename to...
15748 (gimple_stmt_iterator_d): ... This.
15749 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15750 trees be valid for GIMPLE and GENERIC.
15751 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15752 * gimple-fold.c (gimple_fold_call): Call target hook
15753 gimple_fold_builtin.
15754 * hooks.c (hook_bool_gsiptr_false): New.
15755 * hooks.h (hook_bool_gsiptr_false): New.
15756 * target.def (fold_stmt): New.
15757 * doc/tm.texi: Regenerate.
15759 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15762 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15763 a set sp if no stack realignment.
15765 2013-04-22 Nick Clifton <nickc@redhat.com>
15767 * config.gcc (tilegx-linux): Extend extra_objs rather than
15769 (tilepro-linux): Likewise.
15771 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15773 * config/aarch64/aarch64-builtins.c
15775 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15776 (VAR<1-12>): Add MAP parameter.
15777 (BUILTIN_*): Likewise.
15778 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15779 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15780 (aarch64_ushl_n<mode>): Likewise.
15781 (aarch64_sshr_n<mode>): Likewise.
15782 (aarch64_ushr_n<mode>): Likewise.
15783 (aarch64_<maxmin><mode>): Likewise.
15784 (aarch64_sqrt<mode>): Likewise.
15785 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15786 (vshr<q>_n_*): Likewise.
15788 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15790 * config/aarch64/aarch64-builtins.c
15791 (aarch64_simd_builtin_type_mode): Handle SF types.
15793 (BUILTIN_GPF): Define.
15794 (aarch64_init_simd_builtins): Handle SF types.
15795 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15796 (frecps): Likewise.
15797 (frecpx): Likewise.
15798 * config/aarch64/aarch64-simd.md
15799 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15800 (aarch64_frecpe<mode>): New.
15801 (aarch64_frecps<mode>): Likewise.
15802 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15803 (v8type): Add frecp<esx>.
15804 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15805 (aarch64_frecps<mode>): Likewise.
15806 * config/aarch64/iterators.md (FRECP): New.
15807 (frecp_suffix): Likewise.
15808 * config/aarch64/arm_neon.h
15809 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15811 2013-04-22 Christian Bruel <christian.bruel@st.com>
15814 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15815 (REG_CLASS_NAMES): Idem.
15816 (REG_CLASS_CONTENTS): Idem.
15817 (REGCLASS_HAS_FP_REG): Idem.
15818 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15819 (sh_conditional_register_usage): Idem.
15821 2013-04-21 Jeff Law <law@redhat.com>
15823 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15824 (ssa_forward_propagate_and_combine): Use it.
15826 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15828 * lra.c: Update the flow chart diagram.
15830 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15832 PR rtl-optimization/56847
15833 * lra-constraints.c (process_alt_operands): Discourage alternative
15834 with non-matche doffsettable memory constraint fro memory with
15837 2013-04-19 Richard Biener <rguenther@suse.de>
15839 PR tree-optimization/56982
15840 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15842 * gimplify.c (gimplify_call_expr): Notice special calls.
15843 (gimplify_modify_expr): Likewise.
15844 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15845 abnormal control flow receivers.
15846 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15847 in the same way as cfun->has_nonlocal_labels.
15848 (gimple_purge_dead_abnormal_call_edges): Likewise.
15849 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15850 receivers start a basic-block.
15852 2013-04-19 Richard Biener <rguenther@suse.de>
15854 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15856 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15857 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15858 (SLP_TREE_LOAD_PERMUTATION): Add.
15859 (vect_transform_slp_perm_load): Adjust prototype.
15860 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15861 (vect_free_slp_instance): Likewise.
15862 (vect_create_new_slp_node): Likewise.
15863 (vect_supported_slp_permutation_p): Remove.
15864 (vect_slp_rearrange_stmts): Adjust.
15865 (vect_supported_load_permutation_p): Likewise. Inline
15866 vect_supported_slp_permutation_p here.
15867 (vect_analyze_slp_instance): Compute load permutations per
15868 slp node instead of per instance.
15869 (vect_get_slp_defs): Adjust.
15870 (vect_transform_slp_perm_load): Likewise.
15871 (vect_schedule_slp_instance): Remove redundant code.
15872 (vect_schedule_slp): Remove hack for PR56270, add it ...
15873 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15874 CSE loads for SLP. Adjust.
15876 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15878 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15879 spelling in two comments.
15881 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15884 * config/arm/arm.c (load_multiple_sequence): Require SP
15885 as base register for loads if SP is in the register list.
15887 2013-04-19 Martin Jambor <mjambor@suse.cz>
15889 PR tree-optimization/56718
15890 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15891 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15892 and made public. Adjusted all callers.
15893 (ipa_intraprocedural_devirtualization): New function.
15894 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15895 (ipa_intraprocedural_devirtualization): Likewise.
15896 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15898 2013-04-19 Richard Biener <rguenther@suse.de>
15900 PR tree-optimization/57000
15901 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15903 2013-04-19 Terry Guo <terry.guo@arm.com>
15905 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15907 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15908 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15909 (cortex_m4_fmacs): Use new reservations.
15910 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15912 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15914 PR rtl-optimization/56999
15915 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15917 (lra_coalesce): Remove split_origin_bitmap and related code.
15918 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15919 ranges if necessary.
15921 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15923 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15925 (ix86_expand_call): Remove clobbered_registers array and use
15926 x86_64_ms_sysv_extra_clobbered_registers instead.
15927 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15929 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15931 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15932 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15933 (*call_value_rex64_ms_sysv): Ditto.
15935 2013-04-18 Cary Coutant <ccoutant@google.com>
15937 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15938 parent instead of die_mark.
15940 2013-04-18 Diego Novillo <dnovillo@google.com>
15942 * gimple.c (create_gimple_tmp): New.
15943 (get_expr_type): New.
15944 (build_assign): New.
15945 (build_type_cast): New.
15946 * gimple.h (enum ssa_mode): Define.
15947 (gimple_seq_set_location): New.
15948 * asan.c (build_check_stmt): Change some gimple_build_* calls
15949 to use build_assign and build_type_cast.
15951 2013-04-18 Richard Biener <rguenther@suse.de>
15953 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
15954 handle negative step. Remove redundant checks.
15955 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
15956 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
15957 for negative step and grouped loads fail to vectorize.
15959 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15961 * emit-rtl.c (reset_insn_used_flags): New function.
15962 (reset_all_used_flags): Use it.
15963 (verify_insn_sharing): New function.
15964 (verify_rtl_sharing): Fix verification for SEQUENCEs.
15966 2013-04-18 Jakub Jelinek <jakub@redhat.com>
15968 PR tree-optimization/56984
15969 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
15970 and (x >> M) >= N don't register any assertion if N << M is the
15973 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15975 * lower-subreg.c (resolve_simple_move): If called self-recursive,
15976 do not delete_insn insns that have not yet been emitted, only
15977 unlink them with remove_insn.
15978 * df-scan.c (df_insn_delete): Revert r197492.
15980 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
15982 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
15983 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
15985 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15987 * config/arm/arm.md (movsicc_insn): Convert define_insn into
15988 define_insn_and_split.
15989 (and_scc,ior_scc,negscc): Likewise.
15990 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
15992 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15994 * config/arm/arm.c (use_return_insn): Return 0 for targets that
15995 can benefit from using a sequence of LDRD instructions in epilogue
15996 instead of a single LDM instruction.
15998 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16001 * doc/extend.texi: Fix typo.
16003 2013-04-17 Richard Biener <rguenther@suse.de>
16005 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16006 (vect_build_slp_tree): ... here.
16007 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16008 match. Remove special-casing of mismatched complex loads.
16009 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16010 re-try the match with swapped commutative operands.
16011 (vect_supported_load_permutation_p): Remove special-casing of
16012 mismatched complex loads.
16013 (vect_analyze_slp_instance): Adjust.
16015 2013-04-17 Richard Biener <rguenther@suse.de>
16017 PR rtl-optimization/56921
16018 * cfgloop.h (struct loop): Add simple_loop_desc member.
16019 (struct niter_desc): Mark with GTY(()).
16020 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16021 * loop-iv.c (get_simple_loop_desc): Likewise.
16022 (free_simple_loop_desc): Likewise.
16025 2013-04-16 Richard Biener <rguenther@suse.de>
16027 PR rtl-optimization/56921
16028 * loop-init.c (pass_rtl_move_loop_invariants): Add
16029 TODO_do_not_ggc_collect to todo_flags_finish.
16030 (pass_rtl_unswitch): Same.
16031 (pass_rtl_unroll_and_peel_loops): Same.
16032 (pass_rtl_doloop): Same.
16034 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16036 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16037 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16038 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16040 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16041 * tree-streamer.c (record_common_node): Adjust reference in comment.
16043 2013-04-17 Terry Guo <terry.guo@arm.com>
16045 * config/arm/cortex-m4.md: Add a new bypass.
16047 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16049 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16051 (*subs_<optab><mode>_multp2): New pattern.
16052 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16053 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16055 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16057 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16058 (*subs_mul_imm_<mode>): New pattern.
16060 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16063 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16064 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16065 (vsx_movti_32bit): Change j->wa to O->wa.
16067 2013-04-16 Richard Biener <rguenther@suse.de>
16069 PR rtl-optimization/56921
16070 * loop-init.c (pass_rtl_move_loop_invariants): Add
16071 TODO_do_not_ggc_collect to todo_flags_finish.
16072 (pass_rtl_unswitch): Same.
16073 (pass_rtl_unroll_and_peel_loops): Same.
16074 (pass_rtl_doloop): Same.
16076 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16078 * config/arm/arm.c (emit_multi_reg_push): New declaration
16079 for an existing function.
16080 (arm_emit_strd_push): New function.
16081 (arm_expand_prologue): Used here.
16082 (arm_emit_ldrd_pop): New function.
16083 (arm_expand_epilogue): Used here.
16084 (arm_get_frame_offsets): Update condition.
16085 (arm_emit_multi_reg_pop): Add a special case for load of a single
16086 register with writeback.
16088 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16090 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16093 2013-04-16 Richard Biener <rguenther@suse.de>
16095 PR tree-optimization/56756
16096 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16097 (first_mem_ref_loc): New.
16098 (execute_sm): Place the load temporarily before a previous
16099 access instead of in the latch edge to ensure its SSA dependencies
16100 are defined at points dominating the load.
16102 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16104 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16105 correct fix by moving header and footer insn to the footer of
16106 the merged basic block. Clear BB_END of the merged-away block.
16108 PR middle-end/43631
16109 * emit-rtl.c (make_note_raw): New function.
16110 (link_insn_into_chain): New static inline function.
16111 (add_insn): Use it.
16112 (add_insn_before, add_insn_after): Factor insn chain linking code...
16113 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16114 using link_insn_into_chain.
16115 (note_outside_basic_block_p): New helper function for emit_note_after
16116 and emit_note_before.
16117 (emit_note_after): Use nobb variant of add_insn_after if the note
16118 should not be contained in a basic block.
16119 (emit_note_before): Use nobb variant of add_insn_before if the note
16120 should not be contained in a basic block.
16121 (emit_note_copy): Use make_note_raw.
16122 (emit_note): Likewise.
16123 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16124 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16125 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16126 the moved barrier the tail of the basic block it follows.
16127 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16129 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16131 PR tree-optimization/56962
16132 * gimple-ssa-strength-reduction.c (record_increment): Only set
16133 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16134 either rhs1 or rhs2 is equal to c->base_expr.
16136 2013-04-15 Richard Biener <rguenther@suse.de>
16138 PR tree-optimization/56933
16139 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16141 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16142 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16143 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16144 dependence check ...
16145 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16147 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16148 GROUP_READ_WRITE_DEPENDENCE.
16150 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16152 * emit-rtl.c (reset_all_used_flags): New function.
16153 (verify_rtl_sharing): Call reset_all_used_flags before and after
16154 performing the checks.
16156 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16158 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16159 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16160 * config/arm/constraints.md (De): New constraint.
16161 * config/arm/neon.md (anddi3_neon): Delete.
16162 (neon_vand<mode>): Expand to standard anddi3 pattern.
16163 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16164 Move earlier in the file.
16165 (neon_inv_logic_op2): Likewise.
16166 (arm_anddi_operand_neon): New predicate.
16168 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16170 * configure.ac (gcc_cv_ld_as_needed): Set
16171 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16172 Use -z ignore, -z record on *-*-solaris2*.
16173 (HAVE_LD_AS_NEEDED): Update comment.
16174 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16175 * configure: Regenerate.
16176 * config.in: Regenerate.
16177 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16178 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16179 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16180 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16181 equivalents. Fix markup.
16182 * doc/tm.texi: Regenerate.
16184 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16186 * config/i386/i386.opt: New option mstack-protector-guard=.
16187 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16188 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16189 TARGET_SSP_TLS_GUARD.
16190 * config/i386/i386.c (ix86_option_override_internal): Set
16191 ix86_stack_protector_guard.
16192 * config/i386/i386.md (stack_protect_set): Enable for
16193 TARGET_SSP_TLS_GUARD only.
16194 (stack_protect_set_<mode>): Ditto.
16195 (stack_protect_test): Ditto.
16196 (stack_protect_test_<mode>): Ditto.
16197 * doc/invoke.texi (i386 Option): Document.
16199 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16202 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16203 (S_MODES): Set H_MODE bit.
16204 (SF_MODES): Set only S_MODE and SF_MODE bits.
16205 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16206 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16207 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16208 <MODE_FLOAT>: Likewise.
16210 2013-04-15 Joey Ye <joey.ye@arm.com>
16212 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16214 2013-04-15 Joey Ye <joey.ye@arm.com>
16216 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16218 (thumb_far_jump_used_p): Count instruction size and set
16221 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16223 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16224 * resource.c (mark_target_live_regs): Reformat conditional block.
16226 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16228 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16229 notes, they are emitted only just before final.
16230 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16232 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16234 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16235 * cfgrtl.c (delete_insn): Call it here instead.
16236 * lra-spills.c (lra_final_code_change): Use delete_insn.
16237 * haifa-sched.c (sched_remove_insn): Likewise.
16238 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16239 returning to the nop pool.
16240 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16241 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16243 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16245 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16246 * doc/tm.texi: Regenerated.
16248 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16250 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16253 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16255 * df-core.c (df_find_def): Compare register numbers.
16256 (df_find_use): Likewise.
16258 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16261 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16262 lra_in_progress for return.
16264 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16266 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16267 define_insn into define_insn_and_split and emit movsicc patterns.
16269 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16271 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16273 2013-04-12 Richard Biener <rguenther@suse.de>
16275 * tree-pass.h (TODO_do_not_ggc_collect): New.
16276 * passes.c (execute_one_ipa_transform_pass): Honor
16277 TODO_do_not_ggc_collect.
16278 (execute_one_pass): Likewise.
16281 2013-04-10 Richard Biener <rguenther@suse.de>
16283 * passes.c (init_optimization_passes): Remove reload pass.
16284 * ira.c (do_reload): Merge into ...
16286 (rest_of_handle_reload): Remove.
16287 (pass_reload): Likewise.
16288 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16289 of reload for vzeroupper pass placement.
16291 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16293 PR tree-optimization/56918
16294 PR tree-optimization/56920
16295 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16296 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16297 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16298 use wide_mul_with_sign method.
16300 2013-04-12 Richard Biener <rguenther@suse.de>
16302 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16303 not be considered a gimple constant.
16305 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16307 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16308 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16311 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16312 Jakub Jelinek <jakub@redhat.com>
16314 * opts.c: Include diagnostic-color.h.
16315 (common_handle_option): Handle OPT_fdiagnostics_color_.
16316 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16317 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16318 (diagnostic-color.o): New.
16319 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16320 (diagnostic_color_rule): New enum.
16321 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16322 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16323 the location string.
16324 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16325 either NULL, or color kind.
16326 * diagnostic-color.c: New file.
16327 * diagnostic-color.h: New file.
16328 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16330 * doc/invoke.texi (-fdiagnostics-color): Document.
16331 * pretty-print.h (pp_show_color): Define.
16332 (struct pretty_print_info): Add show_color field.
16333 * diagnostic.c: Include diagnostic-color.h.
16334 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16335 macros. Colorize error:, warning: etc. strings and also the location
16337 (diagnostic_show_locus): Colorize the caret line.
16338 * pretty-print.c: Include diagnostic-color.h.
16339 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16340 inside of %< %> quotes or quoted through q format modifier.
16342 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16344 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16346 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16348 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16349 code in CC_NZ mode.
16350 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16353 2013-04-11 Marek Polacek <polacek@redhat.com>
16355 PR tree-optimization/48184
16356 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16358 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16360 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16361 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16362 (skip_simple_arithmetic): Tidy up.
16363 * tree.h (skip_simple_constant_arithmetic): Declare.
16365 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16367 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16369 2013-04-11 Richard Biener <rguenther@suse.de>
16371 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16372 generate vector constants.
16374 2013-04-11 Richard Biener <rguenther@suse.de>
16376 PR tree-optimization/56878
16377 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16378 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16380 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16381 Prefer to align the DR with the most invariant base address.
16383 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16385 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16388 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16390 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16391 floating-point vector comparisons against 0.
16393 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16395 PR tree-optimization/56899
16396 * fold-const.c (extract_muldiv_1): Apply distributive law
16397 only if TYPE_OVERFLOW_WRAPS (ctype).
16399 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16402 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16403 loading has side effect.
16405 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16407 * config/sparc/sparc.c: Include tree-pass.h.
16408 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16409 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16410 head of file. Change return type. Split off gate function.
16411 (sparc_gate_work_around_errata): New function.
16412 (pass_work_around_errata): New pass definition.
16413 (insert_pass_work_around_errata) New pass insert definition to
16414 insert pass_work_around_errata just after delayed-branch scheduling.
16415 (sparc_option_override): Insert the pass.
16416 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16418 2013-04-10 David S. Miller <davem@davemloft.net>
16420 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16421 or -mcpu=hypersparc.
16423 * target.def (cstore_mode): New hook.
16424 * target.h: Include insn-codes.h
16425 * targhooks.c: Likewise.
16426 (default_cstore_mode): New function.
16427 * targhooks.h: Declare it.
16428 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16429 * doc/tm.texi: Rebuild.
16430 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16431 target hook, rather than inspecting the insn_data.
16432 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16433 (TARGET_CSTORE_MODE): Redefine.
16434 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16436 * config/sparc/predicates.md (cstore_result_operand): New special
16438 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16439 Use it for operand 0.
16440 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16441 (*snesi_special): Likewise.
16442 (*snesi_zero): Likewise.
16443 (*seqsi_zero): Likewise.
16444 (*sltu_insn): Likewise.
16445 (*sgeu_insn): Likewise.
16446 (*seqdi_special): Make operand 0 and comparison operation be of
16448 (*snedi_special): Likewise.
16449 (*snedi_special_vis3): Likewise.
16450 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16451 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16452 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16453 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16454 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16455 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16456 (*sltu_extend_sp64): Likewise.
16457 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16458 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16459 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16460 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16461 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16463 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16465 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16466 (aarch64_start_file): Use the new function.
16468 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16469 Jason Merrill <jason@redhat.com>
16471 * common.opt: Add -gdwarf.
16472 * opts.c (common_handle_option): Handle it.
16473 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16475 2013-04-10 Richard Biener <rguenther@suse.de>
16477 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16478 (execute_one_ipa_transform_pass): But unconditionally here.
16479 (execute_one_pass): And here.
16480 (init_optimization_passes): Remove reload pass.
16481 * tree-pass.h (TODO_ggc_collect): Remove.
16482 (pass_reload): Likewise.
16483 * ira.c (do_reload): Merge into ...
16485 (rest_of_handle_reload): Remove.
16486 (pass_reload): Likewise.
16487 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16488 of reload for vzeroupper pass placement.
16489 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16490 and todo_flags_finish of all passes.
16492 2013-04-10 Richard Biener <rguenther@suse.de>
16494 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16495 first_const_oprnd field, rename first_def_type to first_op_type.
16496 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16497 (vect_get_and_check_slp_defs): Always use the type of the
16498 operand. Allow mixed vect_external_def, vect_constant_def types.
16499 (vect_get_constant_vectors): Handle mixed vect_external_def,
16500 vect_constant_def types.
16502 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16504 PR tree-optimization/55524
16505 * tree-ssa-math-opts.c
16506 (convert_mult_to_fma): Don't use an fms construct
16507 when we don't have an fms operation, but fnma, and it looks
16508 likely that we'll be able to use the latter.
16510 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16512 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16514 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16515 inline fail caused by overwritable functions.
16517 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16519 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16520 unnecessary bits in the constant power of two case.
16522 2013-04-10 Richard Biener <rguenther@suse.de>
16524 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16525 broken code swapping operands.
16526 (vect_build_slp_tree): Do not compute load permutations here.
16527 (vect_analyze_slp_instance): Compute load permutations here,
16528 after building the SLP tree.
16530 2013-04-09 Christian Bruel <christian.bruel@st.com>
16532 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16533 of next/prev_real_insn.
16535 2013-04-09 Jan Hubicka <jh@suse.cz>
16537 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16538 Drop aliased parameter.
16539 (function_and_variable_visibility): Do not handle alias pairs.
16540 * cgraph.c (varpool_externally_visible_p): Update prototype.
16541 * varpool.c (varpool_add_new_variable): Update.
16543 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16545 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16547 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16549 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16551 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16553 2013-04-09 Marek Polacek <polacek@redhat.com>
16555 PR tree-optimization/48762
16556 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16558 2013-04-09 Richard Biener <rguenther@suse.de>
16560 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16562 (vect_build_slp_tree): Likewise.
16563 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16564 calculating the cost of a SLP instance.
16565 (vect_analyze_slp_instance): Use it from here, after building
16568 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16570 PR middle-end/56883
16571 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16572 expand_omp_for_static_chunk): Use simple_p = true in
16573 force_gimple_operand_gsi calls when assigning to addressable decls.
16575 2013-04-09 Jeff Law <law@redhat.com>
16577 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16578 when the boolean was created by converting a wider object which
16579 had a boolean range.
16581 2013-04-09 Richard Biener <rguenther@suse.de>
16583 * tree-vectorizer.h (slp_void_p): Remove.
16584 (slp_tree): Typedef before _slp_tree declaration.
16585 (struct _slp_tree): Use a vector of slp_tree as children.
16586 (vect_get_place_in_interleaving_chain): Remove.
16587 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16589 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16591 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16592 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16593 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16594 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16595 Use slp_node instead of slp_void_p and adjust.
16597 2013-04-09 Richard Biener <rguenther@suse.de>
16599 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16600 work that is not necessary.
16602 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16604 PR tree-optimization/56854
16605 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16606 forward into clobber stmts if it would change MEM_REF lhs into
16609 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16611 * tree.c (type_hash_lookup, type_hash_add): Make static.
16612 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16614 2013-04-09 Richard Biener <rguenther@suse.de>
16616 * tree.h (unsave_expr_now): Remove.
16617 * tree-inline.c (mark_local_for_remap_r): Remove.
16618 (unsave_expr_1): Likewise.
16619 (unsave_r): Likewise.
16620 (unsave_expr_now): Likewise.
16621 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16622 (propagate_tree_value): Likewise.
16624 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16626 * doc/rtl.texi (sequence): Rewrite documentation to match the
16627 current use of SEQUENCE rtl objects.
16628 * rtl.def (SEQUENCE): Likewise.
16630 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16631 Update documentation.
16632 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16633 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16635 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16637 2013-04-08 Teresa Johnson <tejohnson@google.com>
16639 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16640 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16642 (estimate_edge_size_and_time): Add comment to suggest using rounding
16644 (estimate_node_size_and_time): Ditto.
16645 (remap_edge_change_prob): Use helper rounding divide methods.
16646 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16647 (gimple_mod_pow2_value_transform): Ditto.
16648 (gimple_mod_subtract_transform): Ditto.
16649 (gimple_ic_transform): Ditto.
16650 (gimple_stringops_transform): Ditto.
16651 * stmt.c (conditional_probability): Ditto.
16652 (emit_case_dispatch_table): Ditto.
16653 * lto-cgraph.c (merge_profile_summaries): Ditto.
16654 * tree-optimize.c (execute_fixup_cfg): Ditto.
16655 * cfgcleanup.c (try_forward_edges): Ditto.
16656 * cfgloopmanip.c (scale_loop_profile): Ditto.
16658 (duplicate_loop_to_header_edge): Ditto.
16659 (lv_adjust_loop_entry_edge): Ditto.
16660 * tree-vect-loop.c (vect_transform_loop): Ditto.
16661 * profile.c (compute_branch_probabilities): Ditto.
16662 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16663 * lto-streamer-in.c (input_cfg): Ditto.
16664 * gimple-streamer-in.c (input_bb): Ditto.
16665 * ipa-cp.c (update_profiling_info): Ditto.
16666 (update_specialized_profile): Ditto.
16667 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16668 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16670 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16671 (compute_trg_info): Ditto.
16672 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16673 (purge_dead_edges): Ditto.
16674 * loop-unswitch.c (unswitch_loop): Ditto.
16675 * cgraphclones.c (cgraph_clone_edge): Ditto.
16676 (cgraph_clone_node): Ditto.
16677 * tree-inline.c (copy_bb): Ditto.
16678 (copy_edges_for_bb): Ditto.
16679 (initialize_cfun): Ditto.
16680 (copy_cfg_body): Ditto.
16681 (expand_call_inline): Ditto.
16683 2013-04-08 Kai Tietz <ktietz@redhat.com>
16685 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16686 TARGET_CYGWIN64 by TARGET_64BIT.
16688 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16690 * config/epiphany/epiphany.md (GPR_1): New constant.
16691 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16692 * config/epiphany/epiphany.c (gen_compare_reg):
16693 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16694 is already in place.
16695 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16696 Don't require being called during rtl expansion; If y operlaps r0,
16698 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16699 (epiphany_expand_epilogue): Likewise.
16701 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16702 Don't use CC_FPmode for ORDERED / UNORDERED.
16703 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16705 * config/epiphany/constraints.md (CnL): New constraint.
16706 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16707 * config/epiphany/predicates.md (add_operand): Allow 1024.
16709 * config/epiphany/epiphany.md (logical_op): New code iterator.
16710 (op_mnc): New code attribute.
16711 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16712 (mov_f+1, mov_f+2): New peephole2 patterns.
16714 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16715 (cstoresi4): Also allow re-use of zero result when doing a NE
16716 comparison to a non-zero operand.
16717 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16719 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16720 Use gen_addsi3_i / gen_subsi3_i.
16722 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16726 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16727 contain anything but clobbers, at most one __builtin_stack_restore,
16728 optionally debug stmts and final resx, and if it has at least one
16729 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16730 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16731 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16732 which isn't defaut definition, remove them.
16733 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16734 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16735 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16736 with MEM_REF LHS with SSA_NAME address.
16738 2013-04-08 Jeff Law <law@redhat.com>
16740 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16742 2013-04-08 Richard Biener <rguenther@suse.de>
16744 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16746 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16747 determined vector type.
16748 (vect_analyze_data_refs): Likewise.
16749 (vect_get_new_vect_var): Adjust.
16750 (vect_create_destination_var): Preserve SSA name versions.
16751 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16752 not dump anything here.
16754 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16756 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16757 Add member lr_slot_known.
16758 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16760 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16761 Remove code that sets lr_slot_offset according to what a previous
16762 version of epiphany_emit_save_restore used to do.
16763 (epiphany_emit_save_restore): When doing an lr save or restore,
16764 set/verify lr_slot_known and lr_slot_offset.
16766 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16769 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16772 2013-04-08 Richard Biener <rguenther@suse.de>
16774 * alias.c (find_base_term): Fix thinko in previous change.
16776 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16778 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16779 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16780 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16781 if possible to compute val.
16782 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16783 For QImode integers don't require anything about precision. Use
16784 const_with_all_bytes_same to find out if the constant doesn't have
16785 repeated bytes in it.
16787 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16789 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16792 2013-04-08 Marek Polacek <polacek@redhat.com>
16794 PR rtl-optimization/48182
16795 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16798 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16801 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16802 nuses, make sure we have a label.
16804 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16807 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16808 (rs6000_emit_swdiv_low_precision): Remove.
16809 (rs6000_emit_swdiv): Rewrite to handle between one and four
16810 iterations of Newton-Raphson generally; modify required number of
16811 iterations for some cases.
16812 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16814 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16816 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16817 set-but-unused variable.
16819 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16820 basic blocks of released function bodies garbage-collectable.
16822 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16823 (struct rtl_opt_pass): Add TODO_df_finish.
16825 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16827 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16829 * config/arm/constraints.md (q): New constraint.
16830 * config/arm/ldrdstrd.md: New file.
16831 * config/arm/arm.md (ldrdstrd.md) New include.
16832 (arm_movdi): Use "q" instead of "r" constraint
16833 for double-word memory access.
16834 (movdf_soft_insn): Likewise.
16835 * config/arm/vfp.md (movdi_vfp): Likewise.
16836 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16837 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16838 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16839 (mem_ok_for_ldrd_strd): Likewise.
16840 (output_move_double): Update assertion.
16842 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16844 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16846 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16848 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16849 define_insn_and_split.
16850 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16852 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16854 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16855 define_insn_and_split.
16856 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16857 (shiftsi3_compare): New pattern.
16858 (rrx): New pattern.
16859 * config/arm/unspecs.md (UNSPEC_RRX): New.
16861 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16863 * config/arm/arm.md (negdi_extendsidi): New pattern.
16864 (negdi_zero_extendsidi): Likewise.
16866 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16868 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16869 define_insn_and_split.
16870 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16871 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16873 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16875 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16876 define_insn_and_split.
16877 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16878 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16880 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16882 * config/arm/arm.md (subsi3_carryin): New pattern.
16883 (subsi3_carryin_const): Likewise.
16884 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16885 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16887 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16889 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16891 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16893 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16894 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16896 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16898 * config/arm/arm.c (arm_expand_builtin): Change fcode
16899 type to unsigned int.
16901 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16903 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16905 2013-04-04 Ian Lance Taylor <iant@google.com>
16907 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16910 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16912 PR middle-end/56729
16913 * df-scan.c (df_insn_delete): Disable failing assert.
16915 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16917 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16918 New function prototype.
16919 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16920 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16921 (arm_builtin_vectorized_function): New function.
16923 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16925 * config/arm/arm_neon_builtins.def: New file.
16926 * config/arm/arm.c (neon_builtin_data): Move contents to
16927 arm_neon_builtins.def.
16928 (enum arm_builtins): Include neon builtin definitions.
16929 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16930 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16932 2013-04-04 Marek Polacek <polacek@redhat.com>
16934 PR tree-optimization/48186
16935 * predict.c (maybe_hot_frequency_p): Return false if
16936 HOT_BB_FREQUENCY_FRACTION is 0.
16937 (cgraph_maybe_hot_edge_p): Likewise.
16939 2013-04-04 Richard Biener <rguenther@suse.de>
16941 PR tree-optimization/56826
16942 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16945 2013-04-04 Richard Biener <rguenther@suse.de>
16947 PR tree-optimization/56213
16948 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16949 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16951 2013-04-04 Richard Biener <rguenther@suse.de>
16953 PR tree-optimization/56837
16954 * tree-loop-distribution.c (classify_partition): For non-zero
16955 values require that the value has the same precision as its
16956 mode to be useful as memset value.
16958 2013-04-03 Nick Clifton <nickc@redhat.com>
16960 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
16961 (fmssf4): Use fmsf.s on E3V5 architectures.
16962 (fnmasf4): Use fnmaf.s on E3V5 architectures.
16963 (fnmssf4): Use fnmsf.s on E3V5 architectures.
16965 2013-04-03 Jeff Law <law@redhat.com>
16967 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
16968 (lra-eliminations.o): Likewise.
16970 2013-04-03 Teresa Johnson <tejohnson@google.com>
16972 * gcov-io.c (compute_working_sets): Moved most of body of old
16973 compute_working_sets here from profile.c.
16974 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
16975 (gcov_working_set_t): Moved typedef here from basic-block.h
16976 (compute_working_set): Declare.
16977 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
16978 (get_working_sets): Renamed from compute_working_set,
16979 replace most of body with call to new compute_working_sets.
16980 (get_exec_counts): Replace call to compute_working_sets
16981 to get_working_sets.
16982 * profile.h (get_working_sets): Renamed from compute_working_set.
16983 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
16984 to get_working_sets.
16985 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
16986 * gcov-dump.c (dump_working_sets): New function.
16988 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
16990 * hwint.c (sext_hwi, zext_hwi): New functions.
16991 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
16992 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
16993 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
16994 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
16995 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
16996 (sext_hwi, zext_hwi): New functions.
16998 2013-04-03 Jeff Law <law@redhat.com>
17000 PR tree-optimization/56799
17001 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17002 back test for widening conversion erroneously dropped in prior change.
17004 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17007 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17008 instead of next_real_insn.
17010 2013-04-03 Marek Polacek <polacek@redhat.com>
17013 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17015 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17018 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17020 (thumb1_output_casesi): Likewise.
17021 (thumb2_output_casesi): Likewise.
17023 2013-04-03 Richard Biener <rguenther@suse.de>
17025 PR tree-optimization/56817
17026 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17028 (tree_unroll_loops_completely_1): ... new function to manually
17029 walk the loop tree, properly defering outer loops of unrolled
17030 loops to later iterations.
17032 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17034 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17035 (vectorizable_load): Likewise.
17036 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17037 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17039 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17041 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17044 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17046 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17048 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17050 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17052 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17054 PR tree-optimization/56790
17055 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17058 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17060 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17062 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17063 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17066 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17069 * tree.h (force_folding_builtin_constant_p): New decl.
17070 * builtins.c (force_folding_builtin_constant_p): New variable.
17071 (fold_builtin_constant_p): Fold immediately also if
17072 force_folding_builtin_constant_p.
17074 2013-04-03 Richard Biener <rguenther@suse.de>
17076 PR tree-optimization/56812
17077 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17078 DRs of the same interleaving chain are independent.
17080 2013-04-02 Jason Merrill <jason@redhat.com>
17082 * gdbinit.in (pbb): Use debug fn.
17084 2013-04-02 Lawrence Crowl <crowl@google.com>
17086 * sese.h (struct ivtype_map_elt_s): Remove unused.
17087 (extern debug_ivtype_map): Remove unused.
17088 (extern eq_ivtype_map_elts): Remove unused.
17089 * sese.c (debug_ivtype_map): Removed unused.
17090 (debug_ivtype_map_1): Removed unused.
17091 (debug_ivtype_elt): Remove unused.
17092 (eq_ivtype_map_elts): Remove unused.
17095 2013-04-02 Kai Tietz <ktietz@redhat.com>
17098 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17099 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17100 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17102 (legitimize_pe_coff_symbol): Likewise.
17103 (is_imported_p): New helper-function.
17104 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17105 for Windows x64 targets.
17106 (ix86_expand_prologue): Optimize for pe-coff targets.
17107 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17108 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17109 medium/large code-model.
17110 (legitimize_pic_address): Likewise.
17111 (legitimize_tls_address): Likewise.
17112 (ix86_expand_call): Likewise.
17113 (x86_output_mi_thunk): Likewise.
17114 (get_dllimport_decl): Add new beimport argument.
17115 (construct_plt_address): Don't assert for x64 pe-coff targets.
17116 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17118 (SYMBOL_FLAG_STUBVAR): New macro.
17119 (SYMBOL_REF_STUBVAR_P): Likewise.
17120 * config/i386/winnt.c (stub_list): New structure.
17121 (stub_head): New local variable.
17122 (i386_pe_record_stub): New function.
17123 (i386_pe_file_end): Emit refptr-stubs.
17125 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17127 PR rtl-optimization/56745
17128 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17129 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17132 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17133 and both of them are MEM_REFs, just compare first argument for
17134 equality and attempt to deal even with differing offsets.
17137 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17138 of gimple_clobber_p to be MEM_REF.
17139 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17140 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17141 after gimplification.
17142 * asan.c (get_mem_ref_of_assignment): Don't instrument
17143 gimple_clobber_p stmts.
17144 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17145 gimple_clobber_p stmt if they have MEM_REF lhs and
17146 are dead because of another gimple_clobber_p stmt.
17147 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17148 gimple_clobber_p stmts like debug stmts.
17149 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17150 that refer to unused VAR_DECLs or uninitialized values.
17151 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17152 gimple_clobber_p stmts if they refer to removed parameters.
17153 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17156 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17158 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17159 using SWI48 mode attribute.
17161 2013-04-02 Wei Mi <wmi@google.com>
17163 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17164 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17165 *<rotate_insn><mode>3_mask in i386.md.
17167 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17169 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17171 2013-04-02 Richard Biener <rguenther@suse.de>
17173 PR tree-optimization/56778
17174 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17175 Runtime alias tests are not supported for gather loads.
17176 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17177 stmts referenced from SSA operands before updating SSA form.
17179 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17180 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17182 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17183 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17184 * config/arm/cortex-a53.md: New file.
17185 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17186 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17187 * config/arm/arm.c (arm_issue_rate): Likewise.
17188 * config/arm/arm-tune.md: Regenerate
17189 * config/arm/arm-tables.opt: Regenerate.
17190 * config/arm/arm-cores.def: Add cortex-a53.
17192 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17194 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17197 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17199 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17200 scalar load/store operations using B/H registers.
17201 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17203 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17205 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17207 * config/aarch64/aarch64.c
17208 (aarch64_simd_scalar_immediate_valid_for_move): New.
17209 * config/aarch64/aarch64-protos.h
17210 (aarch64_simd_scalar_immediate_valid_for_move): New.
17211 * config/aarch64/constraints.md (Dh, Dq): New.
17212 * config/aarch64/iterators.md (hq): New.
17214 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17216 * reorg.c (get_branch_condition): Deal with conditional returns.
17217 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17219 2013-04-01 Wei Mi <wmi@google.com>
17221 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17222 Truncate operand 2 using %b asm operand modifier.
17223 (*<shift_insn><mode>3_mask): Ditto.
17224 (*<rotate_insn><mode>3_mask): Ditto.
17226 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17228 PR middle-end/56798
17229 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17231 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17233 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17235 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17237 2013-03-30 Lawrence Crowl <crowl@google.com>
17239 * dse.c (clear_alias_sets): Remove never set.
17240 (disqualified_clear_alias_sets): Remove never set.
17241 (clear_alias_mode_pool): Remove never set.
17242 (dse_step0): Remove condition that is never true.
17243 (canon_address): Remove condition that is never true.
17244 (dse_step7): Remove condition that is never true.
17245 (rest_of_handle_dse): Remove condition that is never true.
17246 (rest_of_handle_dse::did_global): Remove never read from above.
17247 (dse_step2_spill): Remove never called from above.
17248 (dse_step5_spill): Remove never called from above.
17250 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17252 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17253 JUMP_TABLE_DATA changes.
17254 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17255 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17256 (Insns) <jump_table_data>: New entry.
17257 * doc/tm.texi: Regenerate.
17259 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17261 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17262 for table jump at the end of a basic block using tablejump_p.
17263 * targhooks.c (default_invalid_within_doloop): Likewise.
17264 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17265 target hook implementation that is identical to the default hook.
17266 (rs6000_invalid_within_doloop): Remove.
17268 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17269 unused variable from tablejump_p call.
17271 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17272 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17273 (INSN_DELETED_P): Likewise.
17274 (emit_jump_table_data): New prototype.
17275 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17276 after 4th as unused.
17277 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17278 * sched-vis.c (print_insn): Likewise.
17279 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17280 insn for compatibility with back ends that use next_active_insn to
17281 identify jump table data.
17282 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17283 (remove_insn): Likewise.
17284 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17286 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17287 (emit_jump_table_data): New function.
17289 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17290 basic block, a JUMP_TABLE_DATA never is.
17291 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17292 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17293 off from code handling real insns.
17294 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17295 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17296 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17297 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17298 is not a NONDEBUG_INSN_P.
17299 * ira-costs.c (scan_one_insn): Likewise.
17300 * jump.c (mark_all_labels): Likewise.
17301 (mark_jump_label_1): Likewise.
17302 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17303 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17305 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17306 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17307 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17308 JUMP_TABLE_DATA_P insns.
17309 (calculate_elim_costs_all_insns): Likewise.
17310 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17311 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17312 (delete_output_reload): Code style fixups.
17313 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17314 insn flags on this non-insn.
17315 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17316 as scheduling barriers, for pre-change compatibility.
17317 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17318 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17320 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17321 redundant JUMP_TABLE_DATA_P test.
17322 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17323 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17324 (frv_for_each_packet): Likewise.
17325 * config/i386/i386.c (min_insn_size): Likewise.
17326 (ix86_avoid_jump_mispredicts): Likewise.
17327 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17328 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17329 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17330 (mips16_insn_length): Robustify.
17331 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17332 (mips16_split_long_branches): Likewise.
17333 * config/pa/pa.c (pa_combine_instructions): Likewise.
17334 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17335 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17336 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17337 as contributing to pool range lengths.
17338 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17339 Remove redundant JUMP_TABLE_DATA_P test.
17340 (sh_loop_align): Likewise.
17341 (split_branches): Likewise.
17342 (sh_insn_length_adjustment): Likewise.
17343 * config/spu/spu.c (get_branch_target): Likewise.
17345 2013-03-29 Jan Hubicka <jh@suse.cz>
17347 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17348 gcov streaming; stream hot bb threshold to ltrans.
17349 * predict.c (get_hot_bb_threshold): Break out from ....
17350 (maybe_hot_count_p): ... here.
17351 (set_hot_bb_threshold): New function.
17352 * lto-section-in.c (lto_section_name): Add profile.
17353 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17354 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17355 and data-streamer.h
17356 (histogram_entry): New structure.
17357 (histogram, histogram_pool): New global vars.
17358 (histogram_hash): New structure.
17359 (histogram_hash::hash): New method.
17360 (histogram_hash::equal): Likewise.
17361 (account_time_size): New function.
17362 (cmp_counts): New function.
17363 (dump_histogram): New function.
17364 (ipa_profile_generate_summary): New function.
17365 (ipa_profile_write_summary): New function.
17366 (ipa_profile_read_summary): New function.
17367 (ipa_profile): Decide on threshold.
17368 (pass_ipa_profile): Add ipa_profile_write_summary and
17369 ipa_profile_read_summary.
17370 * Makefile.in (ipa.o): Update dependencies.
17371 * lto-streamer.h (LTO_section_ipa_profile): New section.
17373 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17375 * tree.h (VAR_P): New.
17377 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17380 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17382 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17384 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17385 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17386 (control_flow_insn_p): Likewise.
17387 * cfgrtl.c (duplicate_insn_chain): Likewise.
17388 * final.c (get_attr_length_1): Likewise.
17389 (shorten_branches): Likewise.
17390 (final_scan_insn): Likewise.
17391 * function.c (instantiate_virtual_regs): Likewise.
17392 * gcse.c (insert_insn_end_basic_block): Likewise.
17393 * ira-costs.c (scan_one_insn): Likewise.
17394 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17395 * lra.c (check_rtl): Likewise.
17396 * reload1.c (elimination_costs_in_insn): Likewise.
17397 * reorg.c (follow_jumps): Likewise.
17399 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17400 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17401 (thumb_far_jump_used_p): Likewise.
17402 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17403 (workaround_speculation): Likewise.
17404 (add_sched_insns_for_speculation): Likewise.
17405 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17406 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17407 (frv_for_each_packet): Likewise.
17408 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17409 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17410 (final_emit_insn_group_barriers): Likewise.
17411 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17412 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17413 (mips16_insn_length): Likewise.
17414 * config/pa/pa.c (pa_reorg): Likewise.
17415 (pa_combine_instructions): Likewise.
17416 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17417 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17418 (sh_reorg): Likewise.
17419 (split_branches): Likewise.
17420 * config/spu/spu.c (get_branch_target): Likewise.
17422 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17425 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17427 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17428 Fix declaration name.
17430 2013-03-28 Lawrence Crowl <crowl@google.com>
17432 * graphds.h (struct graph.indicies): Remove unused.
17433 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17434 (SCOP_ORIGINAL_PDDRS): Remove unused.
17435 * sese.h (extern insert_loop_close_phis): Removed unused.
17436 (extern insert_guard_phis): Removed unused.
17437 (extern ivtype_map_elt_info): Removed unused.
17438 (new_ivtype_map_elt): Removed unused.
17439 * sese.c (ivtype_map_elt_info): Removed unused.
17441 2013-03-28 Lawrence Crowl <crowl@google.com>
17443 * Makefile.in: Add several missing include dependences.
17445 (test-dump.o): New. This object is not added to any executable,
17446 but is present for ad-hoc testing.
17448 (debug (const bitmap_head_def &)): New.
17449 (debug (const bitmap_head_def *)): New.
17451 (extern debug (const bitmap_head_def &)): New.
17452 (extern debug (const bitmap_head_def *)): New.
17454 (debug (edge_def &)): New.
17455 (debug (edge_def *)): New.
17457 (debug (basic_block_def &)): New.
17458 (debug (basic_block_def *)): New.
17460 (dump_node (const_tree, int, FILE *)): Correct source file.
17462 (debug (die_struct &)): New.
17463 (debug (die_struct *)): New.
17465 (extern debug (die_struct &)): New.
17466 (extern debug (die_struct *)): New.
17467 * gimple-pretty-print.c
17468 (debug (gimple_statement_d &)): New.
17469 (debug (gimple_statement_d *)): New.
17470 * gimple-pretty-print.h
17471 (extern debug (gimple_statement_d &)): New.
17472 (extern debug (gimple_statement_d *)): New.
17474 (debug (ira_allocno_copy &)): New.
17475 (debug (ira_allocno_copy *)): New.
17476 (debug (ira_allocno &)): New.
17477 (debug (ira_allocno *)): New.
17479 (extern debug (ira_allocno_copy &)): New.
17480 (extern debug (ira_allocno_copy *)): New.
17481 (extern debug (ira_allocno &)): New.
17482 (extern debug (ira_allocno *)): New.
17484 (debug (live_range &)): New.
17485 (debug (live_range *)): New.
17487 (debug (lra_live_range &)): New.
17488 (debug (lra_live_range *)): New.
17490 (debug (lra_live_range &)): New.
17491 (debug (lra_live_range *)): New.
17493 (debug (omega_pb_d &)): New.
17494 (debug (omega_pb_d *)): New.
17496 (extern debug (omega_pb_d &)): New.
17497 (extern debug (omega_pb_d *)): New.
17499 (debug (const rtx_def &)): New.
17500 (debug (const rtx_def *)): New.
17502 (debug_tree (tree): Move within file.
17503 (debug_raw (const tree_node &)): New.
17504 (debug_raw (const tree_node *)): New.
17505 (dump_tree_via_hooks (const tree_node *, int)): New.
17506 (debug (const tree_node &)): New.
17507 (debug (const tree_node *)): New.
17508 (debug_verbose (const tree_node &)): New.
17509 (debug_verbose (const tree_node *)): New.
17510 (debug_head (const tree_node &)): New.
17511 (debug_head (const tree_node *)): New.
17512 (debug_body (const tree_node &)): New.
17513 (debug_body (const tree_node *)): New.
17514 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17515 (debug (vec<tree, va_gc> &)): New.
17516 (debug (vec<tree, va_gc> *)): New.
17518 (extern debug (const rtx_def &)): New.
17519 (extern debug (const rtx_def *)): New.
17521 (debug_raw (simple_bitmap_def &)): New.
17522 (debug_raw (simple_bitmap_def *)): New.
17523 (debug (simple_bitmap_def &)): New.
17524 (debug (simple_bitmap_def *)): New.
17526 (extern debug (simple_bitmap_def &)): New.
17527 (extern debug (simple_bitmap_def *)): New.
17528 (extern debug_raw (simple_bitmap_def &)): New.
17529 (extern debug_raw (simple_bitmap_def *)): New.
17531 (debug (vinsn_def &)): New.
17532 (debug (vinsn_def *)): New.
17533 (debug_verbose (vinsn_def &)): New.
17534 (debug_verbose (vinsn_def *)): New.
17535 (debug (expr_def &)): New.
17536 (debug (expr_def *)): New.
17537 (debug_verbose (expr_def &)): New.
17538 (debug_verbose (expr_def *)): New.
17539 (debug (vec<rtx> &)): New.
17540 (debug (vec<rtx> *)): New.
17542 (extern debug (vinsn_def &)): New.
17543 (extern debug (vinsn_def *)): New.
17544 (extern debug_verbose (vinsn_def &)): New.
17545 (extern debug_verbose (vinsn_def *)): New.
17546 (extern debug (expr_def &)): New.
17547 (extern debug (expr_def *)): New.
17548 (extern debug_verbose (expr_def &)): New.
17549 (extern debug_verbose (expr_def *)): New.
17550 (extern debug (vec<rtx> &)): New.
17551 (extern debug (vec<rtx> *)): New.
17553 (_list_iter_cond_expr): Make inline instead of static.
17555 (debug (sreal &)): New.
17556 (debug (sreal *)): New.
17558 (extern debug (sreal &)): New.
17559 (extern debug (sreal *)): New.
17561 (extern debug_raw (const tree_node &)): New.
17562 (extern debug_raw (const tree_node *)): New.
17563 (extern debug (const tree_node &)): New.
17564 (extern debug (const tree_node *)): New.
17565 (extern debug_verbose (const tree_node &)): New.
17566 (extern debug_verbose (const tree_node *)): New.
17567 (extern debug_head (const tree_node &)): New.
17568 (extern debug_head (const tree_node *)): New.
17569 (extern debug_body (const tree_node &)): New.
17570 (extern debug_body (const tree_node *)): New.
17571 (extern debug (vec<tree, va_gc> &)): New.
17572 (extern debug (vec<tree, va_gc> *)): New.
17574 (debug (struct loop &)): New.
17575 (debug (struct loop *)): New.
17576 (debug_verbose (struct loop &)): New.
17577 (debug_verbose (struct loop *)): New.
17578 * tree-dump.c: Add header dependence.
17580 (extern debug (struct loop &)): New.
17581 (extern debug (struct loop *)): New.
17582 (extern debug_verbose (struct loop &)): New.
17583 (extern debug_verbose (struct loop *)): New.
17585 (debug (data_reference &)): New.
17586 (debug (data_reference *)): New.
17587 (debug (vec<data_reference_p> &)): New.
17588 (debug (vec<data_reference_p> *)): New.
17589 (debug (vec<ddr_p> &)): New.
17590 (debug (vec<ddr_p> *)): New.
17592 (extern debug (data_reference &)): New.
17593 (extern debug (data_reference *)): New.
17594 (extern debug (vec<data_reference_p> &)): New.
17595 (extern debug (vec<data_reference_p> *)): New.
17596 (extern debug (vec<ddr_p> &)): New.
17597 (extern debug (vec<ddr_p> *)): New.
17599 (debug (pt_solution &)): New.
17600 (debug (pt_solution *)): New.
17602 (extern debug (pt_solution &)): New.
17603 (extern debug (pt_solution *)): New.
17605 (debug (_var_map &)): New.
17606 (debug (_var_map *)): New.
17607 (debug (tree_live_info_d &)): New.
17608 (debug (tree_live_info_d *)): New.
17610 (extern debug (_var_map &)): New.
17611 (extern debug (_var_map *)): New.
17612 (extern debug (tree_live_info_d &)): New.
17613 (extern debug (tree_live_info_d *)): New.
17615 2013-03-28 Jan Hubicka <jh@suse.cz>
17617 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17619 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17621 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17622 record only when desired or required.
17624 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17626 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17627 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17629 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17631 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17632 (*andsi3_compare0_uxtw): New pattern.
17633 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17634 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17636 2013-03-28 Jan Hubicka <jh@suse.cz>
17638 * data-streamer-in.c (streamer_read_gcov_count): New function.
17639 * gimple-streamer-out.c: Include value-prof.h.
17640 (output_gimple_stmt): Output histogram.
17641 (output_bb): Use streamer_write_gcov_count.
17642 * value-prof.c: Include data-streamer.h
17643 (dump_histogram_value): Add HIST_TYPE_MAX.
17644 (stream_out_histogram_value): New function.
17645 (stream_in_histogram_value): New function.
17646 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17647 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17648 * data-streamer-out.c (streamer_write_gcov_count): New function.
17649 (streamer_write_gcov_count_stream): New function.
17650 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17651 (lto_output_node): Likewise.
17652 (input_node, input_edge): Likewise.
17653 * lto-streamer-out.c (output_cfg): Update streaming.
17654 * lto-streamer-in.c (input_cfg): Likewise.
17655 * data-streamer.h (streamer_write_gcov_count,
17656 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17657 * gimple-streamer-in.c: Include value-prof.h
17658 (input_gimple_stmt): Input histograms.
17659 (input_bb): Update profile streaming.
17661 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17663 * genmodes.c (emit_max_int): New function.
17664 (emit_insn_modes_h): Added call to emit_max_function.
17665 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17667 * machmode.def: Fixed comment.
17669 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17671 * combine.c (try_combine): Removed useless assert.
17672 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17674 2013-03-28 Marek Polacek <polacek@redhat.com>
17675 Richard Biener <rguenther@suse.de>
17677 PR tree-optimization/56695
17678 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17679 build signed result of a vector comparison.
17680 * tree-cfg.c (verify_gimple_comparison): Check that a result
17681 of a vector comparison has signed type.
17683 2013-03-28 Richard Biener <rguenther@suse.de>
17685 PR tree-optimization/37021
17686 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17687 do not restrict gaps between groups.
17688 * tree-vect-stmts.c (vectorizable_load): Properly account for
17689 a gap between groups.
17691 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17693 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17694 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17697 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17699 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17700 * doc/extend.texi (Named Address Spaces): Ditto.
17701 (Variable Attributes): Ditto.
17703 2013-03-27 Kai Tietz <ktietz@redhat.com>
17705 * config.build: Add support for cygwin x64 target.
17706 * config.gcc: Likewise.
17707 * config.host: Likewise.
17708 * configure.ac: Likewise
17709 * configure: Regenerated.
17711 2013-03-27 Kai Tietz <ktietz@redhat.com>
17713 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17714 * config/i386/t-cygwin-w64: New file.
17715 * config/i386/cygwin-w64.h: New file.
17716 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17717 and add support for x64-cygwin target.
17718 (CPP_SPEC): Likewise.
17719 (CXX_WRAP_SPEC_LIST): Undefine before define.
17720 (LIBGCJ_SONAME): Use 15 as version.
17722 2013-03-27 Richard Biener <rguenther@suse.de>
17724 PR tree-optimization/56716
17725 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17726 dumping for ref nodes.
17728 2013-03-27 Martin Jambor <mjambor@suse.cz>
17730 PR tree-optimization/55334
17731 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17732 restricted pointers to arrays.
17734 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17736 * Makefile.in (.SUFFIXES): Add .cc.
17737 (.c.o): Apply same recipe for implicit rule .cc.o.
17739 2013-03-27 Richard Biener <rguenther@suse.de>
17741 PR tree-optimization/37021
17742 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17743 REALPART/IMAGPART_EXPRs around the supported refs.
17744 * tree-ssa-structalias.c (find_func_aliases): Assume that
17745 floating-point values are not used to transfer pointers.
17747 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17749 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17750 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17751 * doc/tm.texi: Regenerate.
17752 * targhooks.h (default_has_ifunc_p): New.
17753 * targhooks.c (default_has_ifunc_p): Ditto.
17754 * config/linux-protos.h: New file.
17755 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17756 hook for linux which disables support of indirect functions in android.
17757 * config/linux-android.c: New file.
17758 * config/t-linux-android.c: Ditto.
17759 * config.gcc: Added new object file linux-android.o.
17760 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17761 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17762 * varasm.c (do_assemble_alias): Likewise.
17763 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17764 doesn't support indirect functions.
17765 * configure: Regenerate.
17767 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17770 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17771 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17774 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17776 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17778 2013-03-27 Terry Guo <terry.guo@arm.com>
17780 * config/arm/arm-cores.def: Added core cortex-r7.
17781 * config/arm/arm-tune.md: Regenerated.
17782 * config/arm/arm-tables.opt: Regenerated.
17783 * doc/invoke.texi: Added entry for core cortex-r7.
17785 2013-03-27 Walter Lee <walt@tilera.com>
17787 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17788 double-decrement of next_scratch_regno.
17790 2013-03-27 Walter Lee <walt@tilera.com>
17792 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17794 (insn_v1mulus): Ditto.
17795 (insn_v2muls): Ditto.
17797 2013-03-27 Walter Lee <walt@tilera.com>
17799 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17800 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17802 2013-03-27 Walter Lee <walt@tilera.com>
17804 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17805 (*sibcall_value): Ditto.
17807 2013-03-27 Walter Lee <walt@tilera.com>
17809 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17810 (insn_mnz_v8qi): ... this ...
17811 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17813 (insn_v<n>mnz): Replaced by ...
17814 (insn_v1mnz): ... this ...
17815 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17817 (insn_mz_<mode>): Replaced by ...
17818 (insn_mz_v8qi): ... this ...
17819 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17821 (insn_v<n>mz): Replaced by ...
17822 (insn_v1mz): ... this ...
17823 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17826 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17828 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17830 2013-03-26 Roland McGrath <mcgrathr@google.com>
17832 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17833 than fprintf with a non-constant, non-format string.
17835 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17837 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17838 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17839 operand 0 predicate.
17840 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17841 attribute. Use general_x64nomem_operand as operand 1 predicate.
17842 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17843 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17844 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17845 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17846 (mov<mode>_insv_1): Remove expander. Merge insn with
17847 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17848 Use general_x64nomem_operand as operand 1 predicate.
17849 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17850 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17851 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17852 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17853 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17854 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17855 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17856 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17857 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17858 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17859 (general_x64nomem_operand): Ditto.
17861 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17863 * config/rtems.opt: Add -pthread option.
17865 2013-03-26 Richard Biener <rguenther@suse.de>
17867 * alias.c (find_base_term): Avoid redundant and not used recursion.
17868 (base_alias_check): Get the initial base term from the caller.
17869 (true_dependence_1): Compute and pass base terms to base_alias_check.
17870 (write_dependence_p): Likewise.
17871 (may_alias_p): Likewise.
17873 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17875 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17876 PC-relative load in SI modes and above only.
17878 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17880 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17881 * config/arm/iwmmxt.md (WCGR0): Update.
17882 (WCGR1, WCGR2, WCGR3): Likewise.
17884 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17886 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17887 Use x64 and nox64 isa attributes.
17889 2013-03-26 Richard Biener <rguenther@suse.de>
17891 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17892 alignment computations and rely on get_object_alignment_1
17893 for the !TYPE_P case.
17894 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17896 2013-03-26 Walter Lee <walt@tilera.com>
17898 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17899 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17901 2013-03-25 Jeff Law <law@redhat.com>
17903 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17904 check for INTEGRAL_TYPE_P that was missing due to checking in
17905 wrong version of prior patch.
17907 2013-03-25 Walter Lee <walt@tilera.com>
17909 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17910 TILEGX_INSN_SHUFFLEBYTES1.
17911 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17913 (tilegx_builtins): Ditto.
17914 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17916 2013-03-25 Walter Lee <walt@tilera.com>
17918 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17919 (floatunssisf2): New pattern.
17920 (floatsidf2): New pattern.
17921 (floatunssidf2): New pattern.
17923 2013-03-25 Walter Lee <walt@tilera.com>
17925 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17926 tests for constraint J, K, N, P.
17928 2013-03-25 Walter Lee <walt@tilera.com>
17930 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17931 Use indirect/pcrel encoding.
17932 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17935 2013-03-25 Steve Ellcey <sellcey@mips.com>
17937 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17938 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17939 * config/mips/mips.c (mips_option_override): Set IMADD default.
17940 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17941 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17942 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17943 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17944 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17946 2013-03-25 Jeff Law <law@redhat.com>
17948 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17949 slightly to avoid creating and folding useless trees. Simplify
17950 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
17952 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
17954 * config/i386/i386.md (*zero_extendsidi2): Merge with
17955 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
17956 * config/i386/predicates.md (x86_64_zext_operand): Rename from
17957 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
17958 targets. Clarify comment.
17960 2013-03-25 Martin Jambor <mjambor@suse.cz>
17962 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
17963 pass-through jump functions differently.
17964 (ipa_read_jump_function): Likewise. Also use setter functions to set
17967 2013-03-25 Martin Jambor <mjambor@suse.cz>
17969 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
17970 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
17972 (ipa_get_indirect_edge_target): New function.
17973 (devirtualization_time_bonus): New parameter known_aggs, pass it to
17974 ipa_get_indirect_edge_target. Update all callers.
17975 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
17976 ipa_get_indirect_edge_target_1 instead of calling
17977 ipa_get_indirect_edge_target.
17978 (create_specialized_node): Pass aggvlas to
17979 ipcp_discover_new_direct_edges.
17981 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17983 * config/arm/arm.md (f_sels, f_seld): New types.
17984 (*cmov<mode>): New pattern.
17985 * config/arm/predicates.md (arm_vsel_comparison_operator): New
17988 2013-03-25 Kai Tietz <ktietz@redhat.com>
17990 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
17991 POSIX-printf for mingw-hosted builds.
17993 2013-03-25 Richard Biener <rguenther@suse.de>
17995 PR middle-end/56694
17996 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
17997 must-not-throw stmt location.
17999 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18001 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18002 Emit load-acquire versions when acq is true.
18003 (arm_emit_store_exclusive): Add rel parameter.
18004 Emit store-release versions when rel is true.
18005 (arm_split_compare_and_swap): Use acquire-release instructions
18007 of barriers when appropriate.
18008 (arm_split_atomic_op): Likewise.
18009 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18010 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18011 (VUNSPEC_SLX): Likewise.
18012 (VUNSPEC_LDA): Likewise.
18013 (VUNSPEC_STL): Likewise.
18014 * config/arm/sync.md (atomic_load<mode>): New pattern.
18015 (atomic_store<mode>): Likewise.
18016 (arm_load_acquire_exclusive<mode>): Likewise.
18017 (arm_load_acquire_exclusivesi): Likewise.
18018 (arm_load_acquire_exclusivedi): Likewise.
18019 (arm_store_release_exclusive<mode>): Likewise.
18021 2013-03-25 Catherine Moore <clm@codesourcery.com>
18023 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18024 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18025 * config/mip/predicates.md (lwsp_swsp_operand,
18026 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18027 sb16_operand, db4_operand, db7_operand, ib3_operand,
18028 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18029 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18030 andi16_operand): New predicates.
18031 * config/mips/mips.md (compression): New attribute.
18032 (enabled): New attribute.
18033 (length): Consider compression in computing length.
18034 (shift_compression): New code attribute.
18035 (*add<mode>3): New operands. Record compression.
18036 (sub<mode>3): Likewise.
18037 (one_cmpl<mode>2): Likewise.
18038 (*and<mode>3): Likewise.
18039 (*ior<mode>3): Likewise.
18040 (unnamed pattern for xor): Likewise.
18041 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18042 (*<optab><mode>3): Likewise.
18043 (*mov<mode>_internal: Likewise.
18044 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18045 (mips_unsigned_immediate_p): New.
18046 (umips_lwsp_swsp_address_p): New.
18047 (m16_based_address_p): New.
18048 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18049 (mips_unsigned_immediate_p): New prototype.
18050 (lwsp_swsp_address_p): New prototype.
18051 (m16_based_address_p): New prototype.
18052 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18053 (mips_signed_immediate_p): New function.
18054 (m16_based_address_p): New function.
18055 (lwsp_swsp_address_p): New function.
18056 (mips_print_operand_punctuation): Recognize short delay slot insns
18057 for microMIPS.add<mode>3"
18059 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18062 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18063 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18065 2013-03-25 Richard Biener <rguenther@suse.de>
18067 PR tree-optimization/56689
18068 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18071 2013-03-25 Richard Biener <rguenther@suse.de>
18073 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18075 (memory_references): Likewise.
18076 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18077 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18078 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18079 (gather_mem_refs_in_loops): Fold into ...
18080 (analyze_memory_references): ... this. Move initialization
18081 to tree_ssa_lim_initialize.
18082 (fill_always_executed_in): Rename to ...
18083 (fill_always_executed_in_1): ... this.
18084 (fill_always_executed_in): Move contains_call computation to
18085 this new function from ...
18086 (tree_ssa_lim_initialize): ... here.
18087 (tree_ssa_lim): Call fill_always_executed_in.
18089 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18091 * postreload.c (reload_combine): Fix code detecting returns.
18093 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18095 * function.c (emit_use_return_register_into_block): On cc0 targets,
18096 do not emit the sequence between cc0 setter and user.
18098 2013-03-25 Kai Tietz <ktietz@redhat.com>
18100 * config/i386/predicates.md (local_symbolic_operand): Interpret
18101 dll-imported symbols as none-local.
18103 2013-03-25 Richard Biener <rguenther@suse.de>
18105 * tree-ssa-loop-im.c (struct depend): Remove.
18106 (struct lim_aux_data): Make depends a vec of gimples.
18107 (free_lim_aux_data): Adjust.
18108 (add_dependency): Likewise.
18109 (set_level): Likewise.
18111 2013-03-25 Richard Biener <rguenther@suse.de>
18113 PR middle-end/56434
18114 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18115 the pointer returned by calls with ECF_MALLOC set.
18117 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18119 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18121 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18123 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18124 using MMXMODE mode iterator.
18125 (*move<mode>_internal): Merge with *movv2sf_internal and
18126 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18128 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18130 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18131 (record_last_mem_set_info): Likewise.
18133 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18134 of XNEWVEC followed by memset.
18135 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18137 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18139 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18140 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18141 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18142 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18143 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18144 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18145 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18146 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18147 BARRIER_P instead of GET_CODE.
18149 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18151 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18152 inaccuracy in the probing code.
18154 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18155 (ctrapdi4): Likewise.
18157 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18159 * calls.c (expand_call): Add missing guard to code handling return
18160 of non-BLKmode structures in MSB.
18161 * function.c (expand_function_end): Likewise.
18163 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18165 * combine.c (try_combine): Adjust comment. Do not add the set of
18166 insn #0 if the destination indirectly is set or dies in insn #2.
18167 Tidy up code to distribute a new note.
18169 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18171 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18172 also for alternatives 16 and 17.
18174 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18176 * config/i386/sse.md (*mov<mode>_internal): Merge with
18177 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18178 Emit insn template depending on type attribute. Use
18179 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18180 movd instead of movq mnemonic for interunit moves. Rewrite mode
18181 attribute calculation. Remove unit attribute calculation.
18182 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18183 Set prefix_data16 attribute for DImode ssemov types.
18184 Use Ym instead of y for SSE-MMX conversion alternatives.
18185 Reorder operand constraints.
18187 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18189 * df.h (df_insn_delete): Adjust prototype.
18190 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18191 and let it decide whether mark the basic block dirty.
18192 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18193 * df-scan.c (df_insn_info_delete): New helper function, split
18194 off from df_insn_delete.
18195 (df_scan_free_bb_info): Use it.
18196 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18198 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18199 that the insn is actually an insn and it has a non-NULL basic block.
18200 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18202 2013-03-22 Richard Biener <rguenther@suse.de>
18204 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18206 (mem_ref_alloc): Do not allocate them.
18207 (refs_independent_p): Do not query or maintain a cache.
18209 2013-03-22 Richard Biener <rguenther@suse.de>
18211 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18212 (gather_mem_refs_in_loops): Do not compute it.
18213 (analyze_memory_references): Do not allocate it.
18214 (tree_ssa_lim_finalize): Do not free it.
18215 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18217 2013-03-22 Richard Biener <rguenther@suse.de>
18219 * is-a.h (as_a): Use gcc_checking_assert.
18221 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18223 * config/aarch64/aarch64.c (aarch64_print_operand): New
18224 format specifier for printing a constant in hex.
18225 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18226 format specifier for printing second operand.
18228 2013-03-22 Richard Biener <rguenther@suse.de>
18230 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18232 (gather_mem_refs_in_loops): Perform store accumulation here.
18233 (create_vop_ref_mapping_loop): Remove.
18234 (create_vop_ref_mapping): Likewise.
18235 (analyze_memory_references): Initialize refs_stored_in_loop.
18236 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18237 (record_indep_loop): Remove.
18238 (record_dep_loop): New function.
18239 (ref_indep_loop_p_1): Adjust to only walk over references
18240 in the loop, not its subloops.
18241 (ref_indep_loop_p): Rename to ...
18242 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18243 maintaining a more fine-grained cache.
18244 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18245 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18247 2013-03-22 Richard Biener <rguenther@suse.de>
18249 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18250 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18251 aggregate mem_ref_loc.
18252 (free_mem_ref_locs): Inline into ...
18253 (memref_free): ... this and adjust.
18254 (mem_ref_alloc): Adjust.
18255 (mem_ref_locs_alloc): Remove.
18256 (record_mem_ref_loc): Adjust.
18257 (get_all_locs_in_loop): Rewrite into ...
18258 (for_all_locs_in_loop): ... this iterator.
18259 (rewrite_mem_ref_loc): New functor.
18260 (rewrite_mem_refs): Use for_all_locs_in_loop.
18261 (sm_set_flag_if_changed): New functor.
18262 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18263 (ref_always_accessed): New functor.
18264 (ref_always_accessed_p): Use for_all_locs_in_loop.
18266 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18268 * tree-pass.h (PROP_gimple_lvec): New.
18269 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18270 (init_optimization_passes): Move pass_lower_vector.
18271 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18273 (pass_lower_vector): Provide PROP_gimple_lvec.
18274 (pass_lower_vector_ssa): Likewise.
18275 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18277 2013-03-21 Mark Wielaard <mjw@redhat.com>
18279 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18281 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18283 * config/i386/i386.md (*movdi_internal): Disparage slightly
18284 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18285 conversion alternatives.
18287 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18289 PR middle-end/48087
18290 * diagnostic.def (DK_WERROR): New kind.
18291 * diagnostic.h (werrorcount): Define.
18292 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18293 promoted to DK_ERROR, increment DK_WERROR counter instead of
18295 * toplev.c (toplev_main): Call print_ignored_options even if
18296 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18297 even if just werrorcount is non-zero.
18300 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18302 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18303 (string_cst_pool_decl): New function.
18304 (optimize_one_addr_into_implicit_ptr): New function.
18305 (resolve_addr_in_expr): Optimize DWARF location expression
18306 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18307 which doesn't live in memory, but has DW_AT_location or
18308 DW_AT_const_value, or refers to a string literal, into
18309 DW_OP_GNU_implicit_pointer.
18310 (optimize_location_into_implicit_ptr): New function.
18311 (resolve_addr): If removing DW_AT_location of a variable because
18312 it was DW_OP_addr of address of the variable, but the variable doesn't
18313 live in memory, try to emit const value attribute for the initializer.
18315 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18317 * tree.h (VECTOR_TYPE_P): New macro.
18318 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18319 TYPE_MODE): Use it.
18320 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18321 VEC_COND_EXPR cannot be lvalues.
18322 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18324 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18326 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18327 Restrict the transformation to equal modes.
18329 2013-03-21 Richard Biener <rguenther@suse.de>
18331 PR tree-optimization/39326
18332 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18333 (MEM_ANALYZABLE): Adjust.
18334 (record_mem_ref_loc): Move bitmap ops ...
18335 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18336 unanalyzable refs, do not record locations for it.
18337 (analyze_memory_references): Allocate ref zero as shared
18339 (refs_independent_p): Do not test for unanalyzed mems here.
18340 (ref_indep_loop_p_1): Special-case disambiguation against
18341 the unanalyzed ref.
18342 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18344 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18346 * config/arm/arm-protos.h (tune_params): Add
18347 prefer_neon_for_64bits field.
18348 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18349 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18350 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18351 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18352 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18353 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18354 (arm_option_override): Handle -mneon-for-64bits new option.
18355 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18356 (prefer_neon_for_64bits): Declare new variable.
18357 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18358 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18359 (arch_enabled): Handle new arch types. Remove support for onlya8
18361 (one_cmpldi2): Use new arch names.
18362 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18363 * config/arm/arm.opt (mneon-for-64bits): Add option.
18364 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18365 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18366 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18368 * doc/invoke.texi (-mneon-for-64bits): Document.
18370 2013-03-21 Richard Biener <rguenther@suse.de>
18372 PR tree-optimization/39326
18373 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18374 (sort_bbs_in_loop_postorder_cmp): New function.
18375 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18377 2013-03-21 Richard Biener <rguenther@suse.de>
18379 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18380 (vect_insert_into_interleaving_chain): Likewise.
18381 (vect_drs_dependent_in_basic_block): Inline ...
18382 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18384 (vect_analyze_data_ref_dependence): ... here. Simplify.
18385 (vect_check_interleaving): Simplify.
18386 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18387 (vect_slp_analyze_data_ref_dependences): ... this new function.
18388 (dr_group_sort_cmp): New function.
18389 (vect_analyze_data_ref_accesses): Compute data-reference groups
18390 here instead of in vect_analyze_data_ref_dependence. Use
18391 a more efficient algorithm.
18392 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18393 vect_slp_analyze_data_ref_dependences. Call
18394 vect_analyze_data_ref_accesses earlier.
18395 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18396 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18397 (vect_slp_analyze_data_ref_dependences): New prototype.
18399 2013-03-21 Richard Biener <rguenther@suse.de>
18401 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18402 ref is stored in the loop.
18403 (find_refs_for_sm): Walk only over all stores.
18404 (store_motion_loop): Allocate from lim_bitmap_obstack.
18405 (store_motion): Likewise.
18407 2013-03-21 Richard Biener <rguenther@suse.de>
18409 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18410 Update virtual SSA form.
18412 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18414 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18415 * configure: Regenerate.
18416 * config.in: Regenerate.
18417 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18418 if !HAVE_LD_EH_FRAME_CIEV3.
18420 2013-03-21 Richard Biener <rguenther@suse.de>
18422 * tree-cfg.c (verify_expr_no_block): New function.
18423 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18424 nor DECL_VALUE_EXPR have locations with associated blocks.
18425 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18426 (clear_unused_block_pointer): Remove code dealing with
18427 blocks in DECL_DEBUG_EXPR locations.
18429 2013-03-21 Richard Biener <rguenther@suse.de>
18431 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18432 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18433 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18434 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18435 instead of DECL_DEBUG_EXPR_IS_FROM.
18436 * gimplify.c (gimplify_modify_expr): Likewise.
18437 * tree-cfg.c (verify_expr_location_1): Likewise.
18438 * tree-complex.c (create_one_component_var): Likewise.
18439 * tree-sra.c (create_access_replacement): Likewise.
18440 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18441 (clear_unused_block_pointer): Likewise.
18442 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18443 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18444 * var-tracking.c (var_debug_decl): Likewise.
18445 (track_expr_p): Likewise.
18446 * tree-inline.c (add_local_variables): Likewise. Set
18447 DECL_HAS_DEBUG_EXPR_P after copying it.
18448 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18449 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18451 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18454 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18455 * configure: Regenerate.
18456 * config.in: Regenerate.
18457 * config/i386/i386.md (*movdf_internal): Use
18458 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18459 movd instead of movq mnemonic for interunit moves.
18460 (*movdi_internal): Ditto.
18462 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18464 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18465 (abd<mode>_3): New pattern.
18466 (aba<mode>_3): New pattern.
18467 (fabd<mode>_3): New pattern.
18469 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18471 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18472 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18473 occurrence of REGISTER_PREFIX as its empty string.
18475 2013-03-20 Jeff Law <law@redhat.com>
18477 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18478 addititional equivalences for equality comparisons between an SSA_NAME
18479 and a constant where the SSA_NAME was set from a widening conversion.
18481 2013-03-20 Walter Lee <walt@tilera.com>
18483 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18485 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18487 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18488 depending on type attribute.
18489 (*movti_internal): Ditto.
18490 (*movtf_internal): Ditto.
18491 (*movxf_internal): Ditto.
18492 (*movdf_internal): Ditto.
18493 (*movsf_internal): Ditto.
18495 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18497 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18498 maybe_vex for sselog1 and ssemov types.
18499 (*movdi_internal): Reorder operand constraints.
18500 (*movsi_internal): Ditto. Set prefix attribute to
18501 maybe_vex for sselog1 and ssemov types.
18502 (*movtf_internal): Set prefix attribute to maybe_vex
18503 for sselog1 and ssemov types.
18504 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18505 DImode ssemov types. Reorder operand constraints.
18506 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18507 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18508 attribute for SImode ssemov types. Reorder operand constraints.
18510 2013-03-20 Martin Jambor <mjambor@suse.cz>
18512 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18513 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18515 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18517 * config/rs6000/predicates.md (indexed_address, update_address_mem
18518 update_indexed_address_mem): New predicates.
18519 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18520 attribute for load/store instructions.
18521 * config/rs6000/dfp.md (movsd_store): Likewise.
18522 (movsd_load): Likewise.
18523 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18524 (unnamed HI->DI extend define_insn): Likewise.
18525 (unnamed SI->DI extend define_insn): Likewise.
18526 (unnamed QI->SI extend define_insn): Likewise.
18527 (unnamed QI->HI extend define_insn): Likewise.
18528 (unnamed HI->SI extend define_insn): Likewise.
18529 (unnamed HI->SI extend define_insn): Likewise.
18530 (extendsfdf2_fpr): Likewise.
18531 (movsi_internal1): Likewise.
18532 (movsi_internal1_single): Likewise.
18533 (movhi_internal): Likewise.
18534 (movqi_internal): Likewise.
18535 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18536 attribute for load/store instructions.
18537 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18539 (mov<mode>_softfloat): Likewise.
18540 (mov<mode>_hardfloat32): Likewise.
18541 (mov<mode>_hardfloat64): Likewise.
18542 (mov<mode>_softfloat64): Likewise.
18543 (movdi_internal32): Likewise.
18544 (movdi_internal64): Likewise.
18545 (probe_stack_<mode>): Likewise.
18547 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18549 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18550 floating point, and decimal floating point to reload iterator.
18552 * config/rs6000/constraints.md (wl constraint): New constraints to
18553 return FLOAT_REGS if certain options are used to reduce the number
18554 of separate patterns that exist in the file.
18555 (wx constraint): Likewise.
18556 (wz constraint): Likewise.
18558 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18559 -mdebug=reg, print wg, wl, wx, and wz constraints.
18560 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18561 Initialize the reload functions for 64-bit binary/decimal floating
18563 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18564 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18565 create the buffer on the stack to overcome not having a 32-bit
18567 (rs6000_emit_move): Likewise.
18568 (rs6000_secondary_memory_needed_rtx): Likewise.
18569 (rs6000_alloc_sdmode_stack_slot): Likewise.
18570 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18571 via xxlxor, just like DFmode 0.0.
18573 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18574 define as 1 if we are running on a power7 or newer.
18575 (enum r6000_reg_class_enum): Add new constraints.
18577 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18578 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18579 with other moves by using conditional constraits (wg). Use LFIWZX
18580 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18581 (movsd splitter): Likewise.
18582 (movsd_hardfloat): Likewise.
18583 (movsd_softfloat): Likewise.
18585 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18586 binary and decimal floating point moves.
18587 (fmove_ok): New attributes to combine binary and decimal floating
18588 point moves, and to combine power6x (mfpgpr) moves along normal
18590 (real_value_to_target): Likewise.
18591 (f32_lr): Likewise.
18592 (f32_lm): Likewise.
18593 (f32_li): Likewise.
18594 (f32_sr): Likewise.
18595 (f32_sm): Likewise.
18596 (f32_si): Likewise.
18597 (movsf): Combine binary and decimal floating point moves. Combine
18598 power6x (mfpgpr) moves with other moves by using conditional
18599 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18600 (mov<mode> for SFmode/SDmode); Likewise.
18601 (SFmode/SDmode splitters): Likewise.
18602 (movsf_hardfloat): Likewise.
18603 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18604 (movsf_softfloat): Likewise.
18605 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18607 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18608 wx and wz constraints.
18610 * config/rs6000/constraints.md (wg constraint): New constraint to
18611 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18613 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18616 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18617 -mdebug=reg, print wg, wl, wx, and wz constraints.
18618 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18619 Initialize the reload functions for 64-bit binary/decimal floating
18621 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18622 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18623 create the buffer on the stack to overcome not having a 32-bit
18625 (rs6000_emit_move): Likewise.
18626 (rs6000_secondary_memory_needed_rtx): Likewise.
18627 (rs6000_alloc_sdmode_stack_slot): Likewise.
18628 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18629 via xxlxor, just like DFmode 0.0.
18631 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18632 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18633 with other moves by using conditional constraits (wg). Use LFIWZX
18634 and STFIWX for loading SDmode on power7.
18635 (movdd splitters): Likewise.
18636 (movdd_hardfloat32): Likewise.
18637 (movdd_softfloat32): Likewise.
18638 (movdd_hardfloat64_mfpgpr): Likewise.
18639 (movdd_hardfloat64): Likewise.
18640 (movdd_softfloat64): Likewise.
18642 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18643 64-bit binary and decimal floating point moves.
18644 (FMOVE64X): Likewise.
18645 (movdf): Combine 64-bit binary and decimal floating point moves.
18646 Combine power6x (mfpgpr) moves with other moves by using
18647 conditional constraits (wg).
18648 (mov<mode> for DFmode/DDmode): Likewise.
18649 (DFmode/DDmode splitters): Likewise.
18650 (movdf_hardfloat32): Likewise.
18651 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18652 (movdf_softfloat32): Likewise.
18653 (movdf_hardfloat64_mfpgpr): Likewise.
18654 (movdf_hardfloat64): Likewise.
18655 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18656 (movdf_softfloat64): Likewise.
18657 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18658 (reload_<mode>_load): Move to later in the file so they aren't in
18659 the middle of the floating point move insns.
18660 (reload_<mode>_store): Likewise.
18662 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18665 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18666 constraint if -mdebug=reg.
18667 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18668 Enable using dd reload support if needed.
18670 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18671 binary and decimal floating point moves in rs6000.md.
18672 (movtd_internal): Likewise.
18674 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18675 decimal floating point moves.
18677 (movtf_internal): Likewise.
18678 (mov<mode>_internal, TDmode/TFmode): Likewise.
18679 (movtf_softfloat): Likewise.
18680 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18682 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18683 movdi_internal64, using wg constraint for move direct operations.
18684 (movdi_internal64): Likewise.
18686 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18687 MODES_TIEABLE_P for selected modes. Print the numerical value of
18688 the various virtual registers. Use GPR/FPR first/last values,
18689 instead of hard coding the register numbers. Print which modes
18690 have reload functions registered.
18691 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18692 settings before/after setting cpu, target and subtarget settings.
18693 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18694 and for secondary reload failures in rs6000_secondary_reload_inner.
18695 (rs6000_secondary_reload_fail): Likewise.
18696 (rs6000_secondary_reload_inner): Likewise.
18698 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18699 macros for first/last GPR and FPR registers.
18700 (LAST_GPR_REGNO): Likewise.
18701 (FIRST_FPR_REGNO): Likewise.
18702 (LAST_FPR_REGNO): Likewise.
18704 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18705 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18706 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18707 (vcond<mode><mode>): Likewise.
18708 (vcondu<mode><mode>): Likewise.
18709 (vector_gtu<mode>): Likewise.
18710 (vector_gte<mode>): Likewise.
18711 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18712 to prevent the compiler from converting DImode operations to TImode.
18713 (ior<mode>3): Likewise.
18714 (and<mode>3): Likewise.
18715 (one_cmpl<mode>2): Likewise.
18716 (nor<mode>3): Likewise.
18717 (andc<mode>3): Likewise.
18719 * config/rs6000/constraints.md (wt constraint): New constraint
18720 that returns VSX_REGS if TImode is allowed in VSX registers.
18722 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18723 constant under VSX.
18725 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18726 similar to TImode, but it is restricted to being in the GPRs.
18728 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18729 TImode to occupy a single VSX register.
18731 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18732 -mvsx-timode for power7/power8.
18733 (power7 cpu): Likewise.
18734 (power8 cpu): Likewise.
18736 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18737 sure that TFmode/TDmode take up two registers if they are ever
18738 allowed in the upper VSX registers.
18739 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18741 (rs6000_init_hard_regno_mode_ok): Likewise.
18742 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18743 constraint. Print if LRA is turned on.
18744 (rs6000_option_override_internal): Give an error if -mvsx-timode
18745 and VSX is not enabled.
18746 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18747 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18748 to reg+offset addressing. Use PTImode when checking offset
18749 addresses for validity.
18750 (reg_offset_addressing_ok_p): Likewise.
18751 (rs6000_legitimate_offset_address_p): Likewise.
18752 (rs6000_legitimize_address): Likewise.
18753 (rs6000_legitimize_reload_address): Likewise.
18754 (rs6000_legitimate_address_p): Likewise.
18755 (rs6000_eliminate_indexed_memrefs): Likewise.
18756 (rs6000_emit_move): Likewise.
18757 (rs6000_secondary_reload): Likewise.
18758 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18759 reloads to fpr registers to continue to use reg+offset addressing,
18760 but 64-bit reloads to altivec registers need reg+reg addressing.
18761 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18762 it. Treat LO_SUM like a PLUS operation.
18763 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18764 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18765 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18766 registers to share a register with a smaller sized type, since VSX
18767 puts scalars in the upper 64-bits.
18768 (print_operand): Add support for PTImode.
18769 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18770 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18771 registers, but don't have arithmetic support.
18772 (rs6000_memory_move_cost): Add test for VSX.
18773 (rs6000_opt_masks): Add -mvsx-timode.
18775 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18778 (VSr): Use wt constraint for TImode.
18779 (VSv): Drop TImode support.
18780 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18781 (vsx_movti_64bit): Likewise.
18782 (vsx_movti_32bit): Likewise.
18783 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18784 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18785 one '?' on the appropriate output constraint. Do not allow TImode
18786 logical operations on 32-bit systems.
18787 (vsx_ior<mode>3): Likewise.
18788 (vsx_xor<mode>3): Likewise.
18789 (vsx_one_cmpl<mode>2): Likewise.
18790 (vsx_nor<mode>3): Likewise.
18791 (vsx_andc<mode>3): Likewise.
18792 (vsx_concat_<mode>): Likewise.
18793 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18795 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18796 OPTION_MASK_VSX_TIMODE.
18797 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18798 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18800 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18801 (TI2 iterator): New iterator for TImode, PTImode.
18802 (wd mode attribute): Add values for vector types.
18803 (movti_string): Replace TI move operations with operations for TImode
18804 and PTImode. Add support for TImode being allowed in VSX registers.
18805 (mov<mode>_string, TImode/PTImode): Likewise.
18806 (movti_ppc64): Likewise.
18807 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18808 (TI mode splitters): Likewise.
18810 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18813 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18815 PR tree-optimization/56355
18816 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18817 Also handle integers with undefined overflow.
18819 2013-03-20 Catherine Moore <clm@codesourcery.com>
18820 Maciej W. Rozycki <macro@codesourcery.com>
18821 Tom de Vries <tom@codesourcery.com>
18822 Nathan Sidwell <nathan@codesourcery.com>
18823 Iain Sandoe <iain@codesourcery.com>
18824 Nathan Froyd <froydnj@codesourcery.com>
18825 Chao-ying Fu <fu@mips.com>
18827 * doc/extend.texi: (micromips, nomicromips, nocompression):
18828 Document new function attributes.
18829 * doc/invoke.texi (minterlink-compressed, mmicromips,
18830 m14k, m14ke, m14kec): Document new options.
18831 (minterlink-mips16): Update documentation.
18832 * doc/md.texi (ZC, ZD): Document new constraints.
18833 * configure.ac (gcc_cv_as_micromips): Check if linker
18834 supports the .set micromips directive.
18835 * configure: Regenerate.
18836 * config.in: Regenerate.
18837 * config/mips/mips-tables.opt: Regenerate.
18838 * config/mips/micromips.md: New file.
18839 * constraints.md (ZC, ZD): New constraints.
18840 * config/mips/predicates.md (movep_src_register): New predicate.
18841 (movep_src_operand): New predicate.
18842 (non_volatile_mem_operand): New predicate.
18843 * config/mips/mips.md (multimem): New type.
18844 (length): Differentiate between 17-bit and 18-bit branch offsets.
18845 (MOVEP1, MOVEP2): New mode iterator.
18846 (mov_<load>l): Use ZC constraint.
18847 (mov_<load>r): Likewise.
18848 (mov_<store>l): Likewise.
18849 (mov_<store>r): Likewise.
18850 (*branch_equality<mode>_inverted): Add microMIPS support.
18851 (*branch_equality<mode>): Likewise.
18852 (*jump_absolute): Likewise.
18853 (indirect_jump_<mode>): Likewise.
18854 (tablejump_<mode>): Likewise.
18855 (<optab>_internal): Likewise.
18856 (sibcall_internal): Likewise.
18857 (sibcall_value_internal): Likewise.
18858 (prefetch): Use constraint ZD.
18859 * config/mips/mips.opt (minterlink-compressed): New option.
18860 (minterlink-mips16): Now an alias for minterlink-compressed.
18861 (mmicromips): New option.
18862 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18863 (compare_and_swap_12): Likewise.
18864 (sync_add<mode>): Likewise.
18865 (sync_<optab>_12): Likewise.
18866 (sync_old_<optab>_12): Likewise.
18867 (sync_new_<optab>_12): Likewise.
18868 (sync_nand_12): Likewise.
18869 (sync_old_nand_12): Likewise.
18870 (sync_new_nand_12): Likewise.
18871 (sync_sub<mode>): Likewise.
18872 (sync_old_add<mode>): Likewise.
18873 (sync_old_sub<mode>): Likewise.
18874 (sync_new_add<mode>): Likewise.
18875 (sync_new_sub<mode>): Likewise.
18876 (sync_<optab><mode>): Likewise.
18877 (sync_old_<optab><mode>): Likewise.
18878 (sync_new_<optab><mode>): Likewise.
18879 (sync_nand<mode>): Likewise.
18880 (sync_old_nand<mode>): Likewise.
18881 (sync_new_nand<mode>): Likewise.
18882 (sync_lock_test_and_set<mode>): Likewise.
18883 (test_and_set_12): Likewise.
18884 (atomic_compare_and_swap<mode>): Likewise.
18885 (atomic_exchange<mode>_llsc): Likewise.
18886 (atomic_fetch_add<mode>_llsc): Likewise.
18887 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18888 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18889 (umips_save_restore_pattern_p): Likewise.
18890 (umips_load_store_pair_p): Likewise.
18891 (umips_output_load_store_pair): Likewise.
18892 (umips_movep_target_p): Likewise.
18893 (umips_12bit_offset_address_p): Likewise.
18894 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18895 (mips_base_mips16): Rename this...
18896 (mips_base_compression_flags): ...to this. Update all uses.
18897 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18898 (mips_mips16_decl_p): Delete.
18899 (mips_nomips16_decl_p): Delete.
18900 (mips_get_compress_on_flags): New function.
18901 (mips_get_compress_off_flags): New function.
18902 (mips_get_compress_mode): New function.
18903 (mips_get_compress_on_name): New function.
18904 (mips_get_compress_off_name): New function.
18905 (mips_insert_attributes): Support multiple compression types.
18906 (mips_merge_decl_attributes): Likewise.
18907 (umips_12bit_offset_address_p): New function.
18908 (mips_start_function_definition): Emit .set micromips directive.
18909 (mips_call_may_need_jalx_p): New function.
18910 (mips_function_ok_for_sibcall): Add microMIPS support.
18911 (mips_print_operand_punctuation): Support short delay slots and
18913 (umips_swm_mask, umips_swm_encoding): New.
18914 (umips_build_save_restore): New function.
18915 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18916 (was_mips16_p): Remove.
18917 (old_compression_mode): New.
18918 (mips_set_compression_mode): New function.
18919 (mips_set_current_function): Add microMIPS support.
18920 (mips_option_override): Likewise.
18921 (umips_save_restore_pattern_p): New function.
18922 (umips_output_save_restore): New function.
18923 (umips_load_store_pair_p_1): New function.
18924 (umips_load_store_pair_p): New function.
18925 (umips_output_load_store_pair_1): New function.
18926 (umips_output_load_store_pair): New function.
18927 (umips_movep_target_p) New function.
18928 (mips_prepare_pch_save): Add microMIPS support.
18929 * config/mips/mips.h (TARGET_COMPRESSION): New.
18930 (TARGET_CPU_CPP_BUILTINS): Update macro
18931 to use new compression flags and to support microMIPS.
18932 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18933 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18934 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18935 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18936 (ASM_SPEC): Support mmicromips and mno-micromips.
18937 (M16STORE_REG_P): New macro.
18938 (MIPS_CALL): Support TARGET_MICROMIPS.
18939 (MICROMIPS_J): New macro.
18940 (mips_base_mips16): Rename this...
18941 (mips_base_compression_flags): ...to this.
18942 (UMIPS_12BIT_OFFSET_P): New macro.
18943 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18944 (MULTILIB_DIRNAMES): Likewise.
18945 2013-03-20 Richard Biener <rguenther@suse.de>
18947 PR tree-optimization/56661
18948 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18949 the result does not have to be distinct.
18951 2013-03-20 Richard Biener <rguenther@suse.de>
18953 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
18956 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18957 Steven Bosscher <steven@gcc.gnu.org>
18959 PR rtl-optimization/56605
18960 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
18962 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18965 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
18966 that require movd instead of movq.
18968 2013-03-20 Richard Biener <rguenther@suse.de>
18970 * tree-ssa-structalias.c (struct variable_info): Add pointer
18971 to the first field of an aggregate with sub-vars. Make
18972 this and the pointer to the next subfield its ID.
18973 (vi_next): New function.
18974 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
18975 storedanything_id, integer_id): Increment by one.
18976 (new_var_info, get_call_vi, lookup_call_clobber_vi,
18977 get_call_clobber_vi): Adjust.
18978 (solution_set_expand): Simplify and speedup.
18979 (solution_set_add): Inline into ...
18980 (set_union_with_increment): ... this. Adjust accordingly.
18981 (do_sd_constraint): Likewise.
18982 (do_ds_constraint): Likewise.
18983 (do_complex_constraint): Simplify.
18984 (build_pred_graph): Adjust.
18985 (solve_graph): Likewise. Simplify and speedup.
18986 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
18987 get_constraint_for_component_ref, get_constraint_for_1,
18988 first_vi_for_offset, first_or_preceding_vi_for_offset,
18989 create_function_info_for, create_variable_info_for_1,
18990 create_variable_info_for, intra_create_variable_infos): Adjust.
18991 (init_base_vars): Push NULL for ID zero.
18992 (compute_points_to_sets): Adjust.
18994 2013-03-20 Richard Biener <rguenther@suse.de>
18996 * cfgloop.c (verify_loop_structure): Streamline and avoid
18997 ICEing on corrupt loop tree.
18998 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19001 2013-03-20 Richard Biener <rguenther@suse.de>
19003 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19004 check whether an SSA update is needed.
19006 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19008 * config/mips/constraints.md (T): Rename to...
19012 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19013 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19015 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19017 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19018 (*subsi3_carryin_uxtw): Likewise.
19020 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19022 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19023 (*rorsi3_insn_uxtw): Likewise.
19025 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19027 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19028 (*extrsi5_insn_uxtw): Likewise.
19030 2013-03-19 Richard Biener <rguenther@suse.de>
19032 PR tree-optimization/56273
19033 * passes.c (init_optimization_passes): Move second VRP after DOM.
19035 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19037 * config/i386/i386.md (*movti_internal): Merge from
19038 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19039 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19040 nox64 isa attributes.
19042 2013-03-18 Richard Biener <rguenther@suse.de>
19044 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19046 (merge_node_constraints): Likewise.
19047 (build_succ_graph): Likewise.
19048 (valid_graph_edge): Inline into single caller.
19049 (unify_nodes): Likewise. Use bitmap_set_bit return value
19051 (scc_visit): Fix formatting and variable use.
19052 (do_sd_constraint): Use gcc_checking_assert.
19053 (do_ds_constraint): Likewise.
19054 (do_complex_constraint): Likewise.
19055 (condense_visit): Likewise. Cleanup.
19056 (dump_pred_graph): New function.
19057 (perform_var_substitution): Dump the pred-graph before
19058 variable substitution.
19059 (find_equivalent_node): Use gcc_checking_assert.
19060 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19062 2013-03-18 Richard Biener <rguenther@suse.de>
19064 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19065 Remove cond_expr_stmt_list argument and do not gimplify the
19067 (vect_loop_versioning): Adjust.
19068 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19069 Cleanup to use less temporaries.
19070 (vect_create_data_ref_ptr): Cleanup.
19072 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19074 PR tree-optimization/56635
19075 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19076 require types_compatible_p types.
19078 2013-03-18 Nick Clifton <nickc@redhat.com>
19080 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19081 spurious backslash.
19083 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19084 Add missing line to comment describing function.
19086 2013-03-18 Richard Biener <rguenther@suse.de>
19088 PR tree-optimization/56210
19089 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19090 Handle string / character search functions.
19091 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19093 2013-03-18 Richard Biener <rguenther@suse.de>
19095 PR middle-end/56483
19096 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19097 and implement properly.
19098 * gimple.h (gimple_cond_single_var_p): Remove.
19100 2013-03-18 Richard Biener <rguenther@suse.de>
19102 * tree-data-ref.h (find_data_references_in_loop): Declare.
19103 * tree-data-ref.c (get_references_in_stmt): Use a stack
19104 vector pre-allocated in the callers.
19105 (find_data_references_in_stmt): Adjust.
19106 (graphite_find_data_references_in_stmt): Likewise.
19107 (create_rdg_vertices): Likewise.
19108 (find_data_references_in_loop): Export.
19109 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19110 Compute dependences here...
19111 (vect_analyze_data_refs): ...not here. When we encounter
19112 a non-vectorizable data reference in basic-block vectorization
19113 truncate the data reference vector. Do not bother to
19114 fixup data-dependence information for gather loads.
19115 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19116 of data references, as reported.
19118 2013-03-18 Richard Biener <rguenther@suse.de>
19120 PR tree-optimization/3713
19121 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19122 has_constants and expr.
19123 (stmt_has_constants): Properly valueize SSA names when deciding
19124 whether the stmt has constants.
19126 2013-03-18 Richard Biener <rguenther@suse.de>
19128 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19129 whole function when there is nothing to do.
19130 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19131 * tree-vectorizer.c (vectorize_loops): Update virtual and
19132 loop-closed SSA once.
19133 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19135 2013-03-18 Richard Biener <rguenther@suse.de>
19137 PR middle-end/56113
19138 * domwalk.c (bb_postorder): New global static.
19139 (cmp_bb_postorder): New function.
19140 (walk_dominator_tree): Replace scheme imposing an order for
19141 visiting dominator sons by one sorting them at the time they
19142 are pushed on the stack.
19144 2013-03-18 Richard Biener <rguenther@suse.de>
19146 PR tree-optimization/39326
19147 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19148 (struct mem_ref): Replace mem member with ao_ref typed member.
19149 (MEM_ANALYZABLE): Adjust.
19150 (memref_eq): Likewise.
19151 (mem_ref_alloc): Likewise.
19152 (gather_mem_refs_stmt): Likewise.
19153 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19154 (execute_sm_if_changed_flag_set): Adjust.
19155 (execute_sm): Likewise.
19156 (ref_always_accessed_p): Likewise.
19157 (refs_independent_p): Likewise.
19158 (can_sm_ref_p): Likewise.
19160 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19163 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19164 return 1 even for !unsignedp.
19166 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19168 * config/i386/i386.md (isa): Add x64 and nox64.
19169 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19170 (*pushtf): Enable *roF alternative for x64 isa only.
19171 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19172 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19173 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19174 (*movtf_internal): Merge from *movtf_internal_rex64 and
19175 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19176 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19177 nox64 isa attributes.
19178 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19179 nox64 isa attributes.
19180 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19182 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19184 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19186 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19189 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19190 class names. Remove trailing comma after "ALL_REGS".
19192 2013-03-16 Jan Hubicka <jh@suse.cz>
19194 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19195 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19196 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19197 of cgraph_get_create_node.
19198 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19200 2013-03-16 Jason Merrill <jason@redhat.com>
19203 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19204 with DW_AT_default_value.
19206 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19208 * BASE-VER: Set to 4.9.0.
19210 2013-03-14 Andi Kleen <ak@linux.intel.com>
19213 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19214 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19215 Document _x* TSX intrinsics.
19217 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19218 David Holsgrove <david.holsgrove@xilinx.com>
19220 * configure.ac: Add MicroBlaze TLS support detection.
19221 * configure: Regenerate.
19222 * config/microblaze/microblaze-protos.h
19223 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19224 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19225 * config/microblaze/microblaze.c (microblaze_address_type): Add
19226 ADDRESS_TLS and tls_reloc address types.
19227 (microblaze_address_info): Add tls_reloc.
19228 (TARGET_HAVE_TLS): Define.
19229 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19230 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19231 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19232 load_tls_operand, microblaze_call_tls_get_addr,
19233 microblaze_legitimize_tls_address): New functions.
19234 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19235 (get_base_reg): Use microblaze_tls_symbol_p.
19236 (microblaze_classify_address): Handle TLS.
19237 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19238 label_mentioned_p and microblaze_tls_referenced_p.
19239 (microblaze_legitimize_address): Handle TLS.
19240 (microblaze_address_insns): Handle ADDRESS_TLS.
19241 (pic_address_needs_scratch): Handle TLS.
19242 (print_operand_address): Handle TLS.
19243 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19244 (microblaze_expand_move): Handle TLS.
19245 (microblaze_legitimate_constant_p): Check
19246 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19247 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19248 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19249 (PIC_OFFSET_TABLE_REGNUM): Set.
19250 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19251 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19252 (addsi3, movsi_internal2, movdf_internal): Update constraints
19253 * config/microblaze/predicates.md (arith_plus_operand): Define
19254 (move_operand): Redefine as move_src_operand,
19255 check microblaze_tls_referenced_p.
19257 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19259 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19260 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19262 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19264 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19267 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19269 PR tree-optimization/53265
19270 * common.opt (Waggressive-loop-optimizations): New option.
19271 * tree-ssa-loop-niter.c: Include tree-pass.h.
19272 (do_warn_aggressive_loop_optimizations): New function.
19273 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19274 if number_of_latch_executions returned constant.
19275 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19276 early. If number_of_latch_executions returned constant, set
19277 nb_iterations_upper_bound back to it.
19278 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19280 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19281 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19283 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19284 (MULTILIB_OSDIRNAMES): Set.
19285 * genmultilib: If defaultosdirname doesn't start with :: , set
19286 defaultosdirname2 instead, clear it and emit two . multilib_raw
19287 entries instead of just one.
19289 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19291 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19292 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19293 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19294 (SUBTARGET_OVERRIDE_OPTIONS): New.
19296 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19299 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19300 (musermode): Convert to Var(TARGET_USERMODE).
19301 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19302 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19303 * config/sh/sh.c (sh_option_override): Use
19304 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19305 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19307 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19309 (udivsi3_i4_single, divsi3_i4_single): Use
19310 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19312 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19314 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19317 2013-03-13 Richard Biener <rguenther@suse.de>
19319 PR tree-optimization/56608
19320 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19321 calls when vectorizing basic-blocks.
19323 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19326 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19329 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19331 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19333 2013-03-11 Jan Hubicka <jh@suse.cz>
19336 * lto-streamer-out.c (output_symbol_p): Skip references from
19337 constructors of external variables.
19339 2013-03-11 Jan Hubicka <jh@suse.cz>
19341 PR middle-end/56571
19342 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19344 * emit-rtl.c (verify_rtx_sharing): Likewise.
19345 (copy_insn_1): Likewise.
19346 * rtl.c (copy_rtx): Likewise.
19348 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19351 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19352 output_operand_lossage message.
19354 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19357 * arm.c (shift_op): Validate RTL pattern on the fly.
19358 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19361 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19364 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19365 new call patterns that clobber %fr12.
19366 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19367 split and postreload patterns.
19368 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19369 registers %fr12 and %fr12R as call used.
19371 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19373 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19374 (canon_address, record_store, replace_read, check_mem_read_rtx,
19375 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19376 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19377 rest_of_handle_dse): Likewise.
19379 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19381 PR middle-end/56524
19382 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19384 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19385 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19386 (save_optabs_if_changed): Replace with...
19387 (init_tree_optimization_optabs): ...this.
19388 * optabs.c (save_optabs_if_changed): Rename to...
19389 (init_tree_optimization_optabs): ...this. Take the optimization node
19390 as argument. Do nothing if the base optabs are already correct.
19391 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19392 to recompute optabs.
19393 * function.h (function): Remove optabs field.
19394 * function.c (invoke_set_current_function_hook): Call
19395 init_tree_optimization_optabs. Use the result to initialize
19398 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19400 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19401 if GTMA_HAS_NO_INSTRUMENTATION.
19402 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19403 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19404 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19405 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19406 GTMA_HAS_NO_INSTRUMENTATION.
19408 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19410 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19413 2013-03-08 Marek Polacek <polacek@redhat.com>
19414 Jakub Jelinek <jakub@redhat.com>
19416 PR tree-optimization/56478
19417 * predict.c (is_comparison_with_loop_invariant_p): Change the
19418 type of loop_step to tree.
19419 (predict_loops): Adjust.
19420 (predict_iv_comparison): Perform the computations on double_ints.
19422 2013-03-08 Richard Biener <rguenther@suse.de>
19424 PR tree-optimization/56570
19425 * tree-cfg.c (verify_expr_location_1): Verify locations for
19427 * tree-sra.c (create_access_replacement): Strip locations
19428 from DECL_DEBUG_EXPRs.
19430 2013-03-08 Richard Biener <rguenther@suse.de>
19432 * tree-inline.c (expand_call_inline): Do not associate
19433 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19434 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19436 2013-03-08 Richard Biener <rguenther@suse.de>
19438 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19439 or block changes with -Og. Fix for location / block encoding
19440 changes and PHI arguments with locations.
19442 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19444 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19446 (struct output_info): Likewise.
19447 (register_overhead): Remove bad gcc_assert.
19448 (bitmap_find_bit): If there is only a single bitmap element, do not
19449 count a miss as a search.
19450 (print_statistics): Update for counter type changes.
19451 (dump_bitmap_statistics): Likewise. Print headers such that they
19452 are properly lined up with the printed counters.
19454 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19456 PR tree-optimization/56559
19457 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19458 check that it has only a single use.
19460 2013-03-07 Richard Biener <rguenther@suse.de>
19462 * doc/invoke.texi (fwhole-program): Discourage use in combination
19465 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19467 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19469 PR tree-optimization/56539
19470 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19471 instead of GSI_CONTINUE_LINKING as last argument to
19472 force_gimple_operand_gsi. Adjust function comment.
19474 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19477 PR middle-end/56548
19478 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19479 promoted mode, convert the result back to the original mode.
19481 2013-03-06 Richard Biener <rguenther@suse.de>
19483 PR middle-end/56294
19484 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19485 (insert_updated_phi_nodes_compare_uids): New function.
19486 (update_ssa): Sort symbols_to_rename after UID before
19487 traversing it to insert PHI nodes.
19489 2013-03-06 Richard Biener <rguenther@suse.de>
19491 PR middle-end/50494
19492 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19493 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19496 2013-02-13 Richard Biener <rguenther@suse.de>
19499 * varasm.c (output_constant_def_1): Get the decl representing
19500 the constant as argument.
19501 (output_constant_def): Wrap output_constant_def_1.
19502 (make_decl_rtl): Use output_constant_def_1 with the decl
19503 representing the constant.
19504 (build_constant_desc): Optionally re-use a decl already
19505 representing the constant.
19506 (tree_output_constant_def): Adjust.
19508 2013-03-06 Joey Ye <joey.ye@arm.com>
19511 * gcc.c (convert_white_space): New function.
19512 (main): Handles white space in function name.
19514 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19517 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19518 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19519 to SH_DIV_CALL_TABLE for TARGET_SH2.
19520 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19522 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19523 call-table options.
19525 2013-03-05 Sterling Augustine <saugustine@google.com>
19526 Cary Coutant <ccoutant@google.com>
19529 * dwarf2out.c (resolve_addr): Don't call
19530 remove_loc_list_addr_table_entries a second time for the same
19533 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19536 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19537 (avoid_complex_debug_insns): New function.
19538 (expand_debug_locations): Call it.
19540 PR rtl-optimization/56484
19541 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19542 lifetimes of hard registers on small register class machines.
19544 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19546 * config/microblaze/microblaze-protos.h: Rename
19547 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19548 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19550 (microblaze_fast_interrupt_function_p): New function.
19551 (microblaze_is_interrupt_handler): Rename to
19552 microblaze_is_interrupt_variant and add fast_interrupt check.
19553 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19554 (save_restore_insns): Likewise.
19555 (compute_frame_size): Likewise.
19556 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19557 (microblaze_globalize_label): Likewise.
19558 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19559 * config/microblaze/microblaze.md: Use wrapper
19560 microblaze_is_interrupt_variant.
19562 2013-03-05 Kai Tietz <ktietz@redhat.com>
19564 * sdbout.c (sdbout_one_type): Switch to current function's section
19565 supporting cold/hot.
19567 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19569 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19572 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19574 PR middle-end/56461
19575 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19576 if VALGRIND_GET_VBITS is defined, temporarily make object
19577 memory all defined, and restore previous valgrind addressability
19578 and definability afterwards. Free this_object at the end.
19580 PR middle-end/56461
19581 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19582 right before calling lra_create_live_ranges, also call it
19583 when clearing live_p. Only call lra_clear_live_ranges
19584 at the end if live_p.
19586 PR middle-end/56461
19587 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19589 2013-03-05 Richard Biener <rguenther@suse.de>
19591 PR tree-optimization/56521
19592 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19595 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19598 * except.h (remove_unreachable_eh_regions): New prototype.
19599 * except.c (remove_eh_handler_splicer): New function, split out
19600 of remove_eh_handler.
19601 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19602 warning about running it on many EH regions one at a time.
19603 (remove_unreachable_eh_regions_worker): New function, walk the
19604 EH tree in depth-first order and remove non-marked regions.
19605 (remove_unreachable_eh_regions): New function.
19606 * tree-eh.c (mark_reachable_handlers): New function, split out
19607 from remove_unreachable_handlers.
19608 (remove_unreachable_handlers): Use mark_reachable_handlers and
19609 remove_unreachable_eh_regions.
19610 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19611 and remove_unreachable_eh_regions.
19613 2013-03-05 Richard Biener <rguenther@suse.de>
19615 PR middle-end/56525
19616 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19617 not freeing them until the end.
19619 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19621 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19623 2013-03-05 Richard Biener <rguenther@suse.de>
19625 PR tree-optimization/56270
19626 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19627 of loads after scheduling an SLP instance.
19629 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19631 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19633 (check_gcc_parallelize): Run guality.exp as a separate job from
19634 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19635 struct-layout-1.exp with stackalign.exp.
19637 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19639 PR middle-end/56461
19640 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19641 load_index sbitmap even if some bit in it isn't set.
19643 PR middle-end/56461
19644 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19645 (discover_iteration_bound_by_body_walk): Change queues to
19646 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19647 spelling in comment. Call safe_push on queues[bound_index] directly.
19648 Release queues[queue_index] in every iteration unconditionally.
19649 Release bounds vector.
19651 PR middle-end/56461
19652 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19653 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19654 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19657 2013-03-05 Richard Biener <rguenther@suse.de>
19660 * tree-inline.c (remap_blocks_to_null): New function.
19661 (expand_call_inline): When expanding a call stmt without
19662 an associated block inline remap all callee blocks to NULL.
19664 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19666 PR rtl-optimization/56494
19667 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19668 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19669 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19671 PR middle-end/56461
19672 * sel-sched-ir.c (free_sched_pools): Release
19673 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19674 if succs_info_pool.max_top isn't -1.
19677 * opts.c (opts_obstack, opts_concat): Moved to...
19678 * opts-common.c (opts_obstack, opts_concat): ... here.
19680 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19682 PR middle-end/56461
19683 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19685 2013-03-04 Martin Jambor <mjambor@suse.cz>
19687 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19688 all appropriate places.
19690 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19692 PR tree-optimization/56424
19693 * ipa-split.c (split_function): Do not set the RSO flag if result is
19694 not by reference and its type is a register type.
19696 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19698 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19699 (microblaze_legitimate_pic_operand): Likewise
19700 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19701 new function microblaze_legitimate_pic_operand
19702 * config/microblaze/microblaze-protos.h
19703 (microblaze_legitimate_pic_operand): Declare.
19705 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19707 * config/microblaze/predicates.md (call_insn_simple_operand):
19708 New predicate for supported rtx code types.
19709 * config/microblaze/microblaze.md (call_internal1): Use
19710 call_insn_simple_operand predicate.
19712 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19714 PR middle-end/56461
19715 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19716 partitions.ordered_remove.
19718 PR middle-end/56461
19719 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19720 vec_oprnds0.create (1) for modifier == NONE.
19722 PR middle-end/56461
19723 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19724 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19725 vec_oprnds1 right before pushing anything to it for
19728 PR middle-end/56461
19729 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19730 set nbbs to 0 instead of having separate code path.
19731 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19732 instead of false as last argument if returning NULL.
19734 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19736 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19737 the attribute is now called "target" instead of "option".
19738 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19739 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19740 attribute/pragma name for TARGET_OPTION_VALID_P and
19741 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19742 * doc/tm.texi: Regenerated.
19744 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19746 * config/microblaze/microblaze.c:
19747 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19748 * config/microblaze/microblaze.h: Add -mxl-reorder to
19750 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19751 instructions emitted if TARGET_REORDER.
19752 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19753 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19756 2013-03-01 Xinliang David Li <davidxl@google.com>
19758 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19761 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19763 PR middle-end/56461
19764 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19765 vector even when returning true. Fix up function comment formatting.
19767 PR middle-end/56461
19768 * ira-build.c (ira_loop_nodes_count): New variable.
19769 (create_loop_tree_nodes): Initialize it.
19770 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19772 PR middle-end/56461
19773 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19774 method on dr_chain and result_chain.
19775 * tree-vect-stmts.c (vectorizable_store): Only call
19776 result_chain.create if j == 0.
19778 PR middle-end/56461
19779 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19780 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19781 before overwriting it.
19783 2013-03-01 Tobias Burnus <burnus@net-b.de>
19785 * doc/extended.texi (C Extensions): Change order in @menu
19787 (Other MIPS Built-in Functions): Move last MIPS entry before
19788 "picoChip Built-in Functions".
19789 (SH Built-in Functions): Move after RX Built-in Functions.
19790 * doc/gcc.texi (Introduction): Change order in @menu
19792 * doc/md.texi (Constraints): Ditto.
19793 * gty.texi (Type Information): Ditto.
19794 (User-provided marking routines for template types): Make
19796 * doc/invoke.texi (AArch64 Options): Move before
19797 "Adapteva Epiphany Options".
19799 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19800 Jakub Jelinek <jakub@redhat.com>
19803 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19804 no_address_safety_analysis attribute.
19805 * doc/extend.texi (no_address_safety_attribute): Rename to
19806 no_sanitize_address attribute, mention no_address_safety_analysis
19807 attribute as deprecated alias.
19809 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19811 PR middle-end/56461
19812 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19813 type to vec<vec<tree> > *.
19814 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19815 to be vec<tree> instead of vec<tree> *, set vec_defs
19816 to vNULL and call vec_defs.create (number_of_vects), adjust other
19818 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19819 vectorizable_condition): Adjust vect_get_slp_defs callers.
19821 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19823 * config/aarch64/aarch64.c
19824 (aarch64_float_const_representable): Remove unused variable.
19826 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19828 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19830 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19832 * config/aarch64/aarch64-builtins.c
19833 (aarch64_init_simd_builtins): Make static.
19835 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19837 * config/aarch64/aarch64.c
19838 (aarch64_simd_make_constant): Make static.
19840 2013-02-28 Martin Jambor <mjambor@suse.cz>
19842 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19843 with no initialization to the RHS of debug statements.
19845 2013-02-28 Martin Jambor <mjambor@suse.cz>
19847 PR tree-optimization/56294
19848 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19850 (get_access_replacement): Do not call create_access_replacement.
19851 Assert a replacement exists.
19852 (get_repl_default_def_ssa_name): Create the replacement declaration
19855 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19857 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19858 final_end_function.
19860 2013-02-28 Marek Polacek <polacek@redhat.com>
19862 PR rtl-optimization/56466
19863 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19864 if we're changing a loop.
19865 (peel_loops_completely): Likewise.
19867 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19870 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19872 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19875 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19876 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19877 INTX_FTYPE_FX, FX_FTYPE_INTX.
19878 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19880 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19882 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19883 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19884 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19885 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19886 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19887 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19888 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19889 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19890 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19891 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19892 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19893 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19894 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19895 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19896 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19897 (avrxmega6): Increase max flash segments from 5 to 6.
19898 * config/avr/t-multilib: Regenerate.
19899 * config/avr/avr-tables.opt: Regenerate.
19900 * doc/avr-mmcu.texi: Regenerate.
19902 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19904 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19905 (avr_device_to_arch): Rename to avr_device_to_ld.
19906 (avr_device_to_as): New prototype.
19907 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19908 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19909 * config/avr/driver-avr.c (avr_device_to_as): New.
19910 (avr_device_to_arch): Rename to avr_device_to_ld.
19912 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19914 PR middle-end/56461
19915 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19916 method on dr_chain and result_chain.
19918 PR middle-end/56461
19919 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19920 pointer_set_destroy on not_executed_last_iteration.
19922 PR middle-end/56461
19923 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19925 PR middle-end/56461
19926 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19927 FOR_EACH_DEFINED_FUNCTION when freeing state.
19929 PR middle-end/56461
19930 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19932 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19935 PR middle-end/56461
19936 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19937 known_aggs[i].items and release known_aggs vector.
19939 PR middle-end/56461
19940 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19942 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19944 * config/microblaze/microblaze.c (microblaze_emit_compare):
19945 Use xor for EQ/NE comparisions.
19946 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19947 (cbranchsf4): Adjust operator to comparison_operator.
19949 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19951 PR middle-end/56461
19952 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
19954 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
19955 vec_safe_push, always update *slot.
19956 (redirect_edge_var_map_clear): Use vec_free.
19957 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
19958 (free_var_map_entry): Use vec_free.
19959 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
19960 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
19962 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
19964 PR middle-end/45472
19965 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
19966 when the may_trap_p bit of the exprs being merged differs.
19967 Reorder tests for speculativeness in the logical and operator.
19969 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19971 * incpath.c (add_standard_paths): Use reconcat instead of concat
19972 where appropriate and avoid leaking memory.
19974 * opts.h: Include obstack.h.
19975 (opts_concat): New prototype.
19976 (opts_obstack): New declaration.
19977 * opts.c (opts_concat): New function.
19978 (opts_obstack): New variable.
19979 (init_options_struct): Call gcc_init_obstack on opts_obstack.
19980 (finish_options): Use opts_concat instead of concat
19981 and XOBNEWVEC instead of XNEWVEC.
19982 * opts-common.c (generate_canonical_option, decode_cmdline_option,
19983 generate_option): Likewise.
19984 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
19985 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
19988 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
19989 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
19991 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19993 PR middle-end/56461
19994 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
19996 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19998 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
19999 (arm_block_move_unaligned_straight): Likewise.
20000 (arm_adjust_block_mem): Likewise.
20002 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20005 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20006 temp, cond and label.
20007 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20010 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20011 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20014 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20015 of prologue/epilogue functions.
20018 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20019 Remove unused variable cfa_offset.
20020 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20023 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20024 type promotion to unsigned.
20027 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20028 for HOST_WIDE_INT of 32 bit / same size as int.
20029 (arm_block_move_unaligned_straight): Likewise.
20030 (arm_adjust_block_mem): Likewise.
20033 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20036 2013-02-26 Marek Polacek <polacek@redhat.com>
20038 PR tree-optimization/56426
20039 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20041 2013-02-26 Richard Biener <rguenther@suse.de>
20044 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20045 unused variable loops.
20047 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20049 PR tree-optimization/56448
20050 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20051 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20052 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20053 later operands of the references, or even first operand for
20054 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20056 PR tree-optimization/56443
20057 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20058 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20059 to type_for_mode langhook.
20061 2013-02-25 Matt Turner <mattst88@gmail.com>
20063 * doc/invoke.texi: Document r4700.
20065 2013-02-25 Richard Biener <rguenther@suse.de>
20067 PR tree-optimization/56175
20068 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20070 (simplify_bitwise_binary): ... here. Also guard the conversion
20071 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20073 2013-02-25 Catherine Moore <clm@codesourcery.com>
20076 2013-02-24 Catherine Moore <clm@codesourcery.com>
20077 Maciej W. Rozycki <macro@codesourcery.com>
20078 Tom de Vries <tom@codesourcery.com>
20079 Nathan Sidwell <nathan@codesourcery.com>
20080 Iain Sandoe <iain@codesourcery.com>
20081 Nathan Froyd <froydnj@codesourcery.com>
20082 Chao-ying Fu <fu@mips.com>
20084 * doc/extend.texi: (micromips, nomicromips, nocompression):
20085 Document new function attributes.
20086 * doc/invoke.texi (minterlink-compressed, mmicromips,
20087 m14k, m14ke, m14kec): Document new options.
20088 (minterlink-mips16): Update documentation.
20089 * doc/md.texi (ZC, ZD): Document new constraints.
20090 * configure.ac (gcc_cv_as_micromips): Check if linker
20091 supports the .set micromips directive.
20092 * configure: Regenerate.
20093 * config.in: Regenerate.
20094 * config/mips/mips-tables.opt: Regenerate.
20095 * config/mips/micromips.md: New file.
20096 * constraints.md (ZC, AD): New constraints.
20097 * config/mips/predicates.md (movep_src_register): New predicate.
20098 (movep_src_operand): New predicate.
20099 (non_volatile_mem_operand): New predicate.
20100 * config/mips/mips.md (multimem): New type.
20101 (length): Differentiate between 17-bit and 18-bit branch offsets.
20102 (MOVEP1, MOVEP2): New mode iterator.
20103 (mov_<load>l): Use ZC constraint.
20104 (mov_<load>r): Likewise.
20105 (mov_<store>l): Likewise.
20106 (mov_<store>r): Likewise.
20107 (*branch_equality<mode>_inverted): Add microMIPS support.
20108 (*branch_equality<mode>): Likewise.
20109 (*jump_absolute): Likewise.
20110 (indirect_jump_<mode>): Likewise.
20111 (tablejump_<mode>): Likewise.
20112 (<optab>_internal): Likewise.
20113 (sibcall_internal): Likewise.
20114 (sibcall_value_internal): Likewise.
20115 (prefetch): Use constraint ZD.
20116 * config/mips/mips.opt (minterlink-compressed): New option.
20117 (minterlink-mips16): Now an alias for minterlink-compressed.
20118 (mmicromips): New option.
20119 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20120 (compare_and_swap_12): Likewise.
20121 (sync_add<mode>): Likewise.
20122 (sync_<optab>_12): Likewise.
20123 (sync_old_<optab>_12): Likewise.
20124 (sync_new_<optab>_12): Likewise.
20125 (sync_nand_12): Likewise.
20126 (sync_old_nand_12): Likewise.
20127 (sync_new_nand_12): Likewise.
20128 (sync_sub<mode>): Likewise.
20129 (sync_old_add<mode>): Likewise.
20130 (sync_old_sub<mode>): Likewise.
20131 (sync_new_add<mode>): Likewise.
20132 (sync_new_sub<mode>): Likewise.
20133 (sync_<optab><mode>): Likewise.
20134 (sync_old_<optab><mode>): Likewise.
20135 (sync_new_<optab><mode>): Likewise.
20136 (sync_nand<mode>): Likewise.
20137 (sync_old_nand<mode>): Likewise.
20138 (sync_new_nand<mode>): Likewise.
20139 (sync_lock_test_and_set<mode>): Likewise.
20140 (test_and_set_12): Likewise.
20141 (atomic_compare_and_swap<mode>): Likewise.
20142 (atomic_exchange<mode>_llsc): Likewise.
20143 (atomic_fetch_add<mode>_llsc): Likewise.
20144 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20145 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20146 (umips_save_restore_pattern_p): Likewise.
20147 (umips_load_store_pair_p): Likewise.
20148 (umips_output_load_store_pair): Likewise.
20149 (umips_movep_target_p): Likewise.
20150 (umips_12bit_offset_address_p): Likewise.
20151 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20152 (mips_base_mips16): Rename this...
20153 (mips_base_compression_flags): ...to this. Update all uses.
20154 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20155 (mips_mips16_decl_p): Delete.
20156 (mips_nomips16_decl_p): Delete.
20157 (mips_get_compress_on_flags): New function.
20158 (mips_get_compress_off_flags): New function.
20159 (mips_get_compress_mode): New function.
20160 (mips_get_compress_on_name): New function.
20161 (mips_get_compress_off_name): New function.
20162 (mips_insert_attributes): Support multiple compression types.
20163 (mips_merge_decl_attributes): Likewise.
20164 (umips_12bit_offset_address_p): New function.
20165 (mips_start_function_definition): Emit .set micromips directive.
20166 (mips_call_may_need_jalx_p): New function.
20167 (mips_function_ok_for_sibcall): Add microMIPS support.
20168 (mips_print_operand_punctuation): Support short delay slots and
20170 (umips_swm_mask, umips_swm_encoding): New.
20171 (umips_build_save_restore): New function.
20172 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20173 (was_mips16_p): Remove.
20174 (old_compression_mode): New.
20175 (mips_set_compression_mode): New function.
20176 (mips_set_current_function): Add microMIPS support.
20177 (mips_option_override): Likewise.
20178 (umips_save_restore_pattern_p): New function.
20179 (umips_output_save_restore): New function.
20180 (umips_load_store_pair_p_1): New function.
20181 (umips_load_store_pair_p): New function.
20182 (umips_output_load_store_pair_1): New function.
20183 (umips_output_load_store_pair): New function.
20184 (umips_movep_target_p) New function.
20185 (mips_prepare_pch_save): Add microMIPS support.
20186 * config/mips/mips.h (TARGET_COMPRESSION): New.
20187 (TARGET_CPU_CPP_BUILTINS): Update macro
20188 to use new compression flags and to support microMIPS.
20189 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20190 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20191 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20192 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20193 (ASM_SPEC): Support mmicromips and mno-micromips.
20194 (M16STORE_REG_P): New macro.
20195 (MIPS_CALL): Support TARGET_MICROMIPS.
20196 (MICROMIPS_J): New macro.
20197 (mips_base_mips16): Rename this...
20198 (mips_base_compression_flags): ...to this.
20199 (UMIPS_12BIT_OFFSET_P): New macro.
20200 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20201 (MULTILIB_DIRNAMES): Likewise.
20203 2013-02-25 Tom de Vries <tom@codesourcery.com>
20205 PR rtl-optimization/56131
20206 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20207 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20208 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20210 2013-02-25 Tobias Burnus <burnus@net-b.de>
20212 * doc/invoke.texi (-fsanitize=): Move from optimization
20213 to debugging options.
20215 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20217 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20219 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20220 Alexander Monakov <amonakov@ispras.ru>
20222 PR middle-end/56077
20223 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20224 flush pending lists also on non-jumps. Adjust comment.
20226 2013-02-24 Catherine Moore <clm@codesourcery.com>
20227 Maciej W. Rozycki <macro@codesourcery.com>
20228 Tom de Vries <tom@codesourcery.com>
20229 Nathan Sidwell <nathan@codesourcery.com>
20230 Iain Sandoe <iain@codesourcery.com>
20231 Nathan Froyd <froydnj@codesourcery.com>
20232 Chao-ying Fu <fu@mips.com>
20234 * doc/extend.texi: (micromips, nomicromips, nocompression):
20235 Document new function attributes.
20236 * doc/invoke.texi (minterlink-compressed, mmicromips,
20237 m14k, m14ke, m14kec): Document new options.
20238 (minterlink-mips16): Update documentation.
20239 * doc/md.texi (ZC, ZD): Document new constraints.
20240 * configure.ac (gcc_cv_as_micromips): Check if linker
20241 supports the .set micromips directive.
20242 * configure: Regenerate.
20243 * config.in: Regenerate.
20244 * config/mips/mips-tables.opt: Regenerate.
20245 * config/mips/micromips.md: New file.
20246 * constraints.md (ZC, AD): New constraints.
20247 * config/mips/predicates.md (movep_src_register): New predicate.
20248 (movep_src_operand): New predicate.
20249 (non_volatile_mem_operand): New predicate.
20250 * config/mips/mips.md (multimem): New type.
20251 (length): Differentiate between 17-bit and 18-bit branch offsets.
20252 (MOVEP1, MOVEP2): New mode iterator.
20253 (mov_<load>l): Use ZC constraint.
20254 (mov_<load>r): Likewise.
20255 (mov_<store>l): Likewise.
20256 (mov_<store>r): Likewise.
20257 (*branch_equality<mode>_inverted): Add microMIPS support.
20258 (*branch_equality<mode>): Likewise.
20259 (*jump_absolute): Likewise.
20260 (indirect_jump_<mode>): Likewise.
20261 (tablejump_<mode>): Likewise.
20262 (<optab>_internal): Likewise.
20263 (sibcall_internal): Likewise.
20264 (sibcall_value_internal): Likewise.
20265 (prefetch): Use constraint ZD.
20266 * config/mips/mips.opt (minterlink-compressed): New option.
20267 (minterlink-mips16): Now an alias for minterlink-compressed.
20268 (mmicromips): New option.
20269 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20270 (compare_and_swap_12): Likewise.
20271 (sync_add<mode>): Likewise.
20272 (sync_<optab>_12): Likewise.
20273 (sync_old_<optab>_12): Likewise.
20274 (sync_new_<optab>_12): Likewise.
20275 (sync_nand_12): Likewise.
20276 (sync_old_nand_12): Likewise.
20277 (sync_new_nand_12): Likewise.
20278 (sync_sub<mode>): Likewise.
20279 (sync_old_add<mode>): Likewise.
20280 (sync_old_sub<mode>): Likewise.
20281 (sync_new_add<mode>): Likewise.
20282 (sync_new_sub<mode>): Likewise.
20283 (sync_<optab><mode>): Likewise.
20284 (sync_old_<optab><mode>): Likewise.
20285 (sync_new_<optab><mode>): Likewise.
20286 (sync_nand<mode>): Likewise.
20287 (sync_old_nand<mode>): Likewise.
20288 (sync_new_nand<mode>): Likewise.
20289 (sync_lock_test_and_set<mode>): Likewise.
20290 (test_and_set_12): Likewise.
20291 (atomic_compare_and_swap<mode>): Likewise.
20292 (atomic_exchange<mode>_llsc): Likewise.
20293 (atomic_fetch_add<mode>_llsc): Likewise.
20294 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20295 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20296 (umips_save_restore_pattern_p): Likewise.
20297 (umips_load_store_pair_p): Likewise.
20298 (umips_output_load_store_pair): Likewise.
20299 (umips_movep_target_p): Likewise.
20300 (umips_12bit_offset_address_p): Likewise.
20301 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20302 (mips_base_mips16): Rename this...
20303 (mips_base_compression_flags): ...to this. Update all uses.
20304 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20305 (mips_mips16_decl_p): Delete.
20306 (mips_nomips16_decl_p): Delete.
20307 (mips_get_compress_on_flags): New function.
20308 (mips_get_compress_off_flags): New function.
20309 (mips_get_compress_mode): New function.
20310 (mips_get_compress_on_name): New function.
20311 (mips_get_compress_off_name): New function.
20312 (mips_insert_attributes): Support multiple compression types.
20313 (mips_merge_decl_attributes): Likewise.
20314 (umips_12bit_offset_address_p): New function.
20315 (mips_start_function_definition): Emit .set micromips directive.
20316 (mips_call_may_need_jalx_p): New function.
20317 (mips_function_ok_for_sibcall): Add microMIPS support.
20318 (mips_print_operand_punctuation): Support short delay slots and
20320 (umips_swm_mask, umips_swm_encoding): New.
20321 (umips_build_save_restore): New function.
20322 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20323 (was_mips16_p): Remove.
20324 (old_compression_mode): New.
20325 (mips_set_compression_mode): New function.
20326 (mips_set_current_function): Add microMIPS support.
20327 (mips_option_override): Likewise.
20328 (umips_save_restore_pattern_p): New function.
20329 (umips_output_save_restore): New function.
20330 (umips_load_store_pair_p_1): New function.
20331 (umips_load_store_pair_p): New function.
20332 (umips_output_load_store_pair_1): New function.
20333 (umips_output_load_store_pair): New function.
20334 (umips_movep_target_p) New function.
20335 (mips_prepare_pch_save): Add microMIPS support.
20336 * config/mips/mips.h (TARGET_COMPRESSION): New.
20337 (TARGET_CPU_CPP_BUILTINS): Update macro
20338 to use new compression flags and to support microMIPS.
20339 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20340 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20341 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20342 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20343 (ASM_SPEC): Support mmicromips and mno-micromips.
20344 (M16STORE_REG_P): New macro.
20345 (MIPS_CALL): Support TARGET_MICROMIPS.
20346 (MICROMIPS_J): New macro.
20347 (mips_base_mips16): Rename this...
20348 (mips_base_compression_flags): ...to this.
20349 (UMIPS_12BIT_OFFSET_P): New macro.
20350 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20351 (MULTILIB_DIRNAMES): Likewise.
20353 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20356 * target-globals.c (save_target_globals): For init_reg_sets and
20357 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20359 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20361 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20362 * config/aarch64/t-aarch64
20363 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20365 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20367 PR inline-asm/56148
20368 * lra-constraints.c (process_alt_operands): Reload operand
20369 conflicting with earlier clobber only if no more other conflicting
20372 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20375 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20376 if not linking a shared library.
20378 2013-02-22 Seth LaForge <sethml@google.com>
20380 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20382 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20384 * config/arm/arm.md (split for extendsidi): Update condition.
20385 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20386 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20387 (qhs_zextenddi_cstr): Likewise.
20389 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20391 PR middle-end/56420
20392 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20393 avoid signed wrapping.
20394 (expand_mult): Handle properly multiplication by
20395 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20396 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20397 in the compiler if coeff is HOST_WIDE_INT_MIN.
20398 (expand_divmod): Don't make ext_op1 static, change it's type to
20399 uhwi. Avoid undefined behavior in -INTVAL (op1).
20401 PR rtl-optimization/50339
20402 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20404 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20405 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20406 into splitting_ashiftrt field.
20407 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20409 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20412 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20414 PR middle-end/56108
20415 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20416 are sure to go irrevocable.
20418 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20420 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20421 scalars are valid operands.
20423 2013-02-21 Martin Jambor <mjambor@suse.cz>
20425 PR tree-optimization/56310
20426 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20427 only matching indices and non-negative final offsets.
20428 (intersect_aggregates_with_edge): Pass src_idx to
20429 agg_replacements_to_vector. Pass src_idx insstead of index to
20430 intersect_with_agg_replacements.
20432 2013-02-21 Martin Jambor <mjambor@suse.cz>
20434 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20435 instead of hard-wired defaults.
20437 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20439 * doc/invoke.texi (MIPS Options): Update documentation of the
20440 floating-point multiply-accumulate instruction restrictions.
20442 2013-02-21 Kostya Serebryany <kcc@google.com>
20444 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20445 asan_shadow_offset on x86_64 linux.
20447 2013-02-21 Richard Biener <rguenther@suse.de>
20449 PR tree-optimization/56415
20451 2013-02-11 Richard Biener <rguenther@suse.de>
20453 PR tree-optimization/56273
20454 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20457 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20460 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20463 PR inline-asm/56405
20464 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20465 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20467 2013-02-20 Jan Hubicka <jh@suse.cz>
20469 PR tree-optimization/56265
20470 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20471 when target is referenced for first time.
20473 2013-02-20 Richard Biener <rguenther@suse.de>
20475 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20476 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20477 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20478 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20479 not return anything.
20480 (rename_ssa_copies): Do not remove unused locals.
20481 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20482 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20483 * passes.c (execute_function_todo): Do not schedule unused locals
20484 removal if cleanup_tree_cfg did something.
20485 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20486 about the number of removed locals.
20488 2013-02-20 Richard Biener <rguenther@suse.de>
20490 PR tree-optimization/56398
20491 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20493 2013-02-20 Martin Jambor <mjambor@suse.cz>
20495 PR tree-optimization/55334
20496 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20497 restricted pointers to arrays.
20499 2013-02-20 Richard Biener <rguenther@suse.de>
20500 Jakub Jelinek <jakub@redhat.com>
20502 PR tree-optimization/56396
20503 * tree-ssa-ccp.c (n_const_val): New static variable.
20504 (get_value): Return NULL for SSA names we don't have a lattice
20506 (ccp_initialize): Initialize n_const_val.
20507 * tree-ssa-copy.c (n_copy_of): New static variable.
20508 (init_copy_prop): Initialize n_copy_of.
20509 (get_value): Return NULL_TREE for SSA names we don't have a
20512 2013-02-20 Martin Jambor <mjambor@suse.cz>
20514 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20516 2013-02-20 Richard Biener <rguenther@suse.de>
20518 * genpreds.c (write_lookup_constraint): Do not compare first
20519 letter of the constraint again.
20521 2013-02-20 Richard Biener <rguenther@suse.de>
20523 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20525 (get_use_iv_cost): Terminate hashtable walk when coming across
20528 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20530 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20531 reassociation for avx2 targets.
20533 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20535 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20536 Add version check for v8.10.a to enable microblaze_has_clz
20537 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20538 version and TARGET_PATTERN_COMPARE check
20539 * config/microblaze/microblaze.md: New clzsi2 instruction
20541 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20543 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20544 function before branching.
20546 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20548 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20550 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20552 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20554 PR middle-end/55889
20555 * sel-sched.c: Include ira.h.
20556 (implicit_clobber_conflict_p): New function.
20557 (moveup_expr): Use it.
20558 * Makefile.in (sel-sched.o): Depend on ira.h.
20560 2013-02-19 Richard Biener <rguenther@suse.de>
20562 PR tree-optimization/56384
20563 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20564 (vn_hash_type): Split out from ...
20565 (vn_hash_constant_with_type): ... here.
20566 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20567 (vn_phi_eq): Compare types from vn_phi_s structure.
20568 (vn_phi_lookup): Populate vn_phi_s type.
20569 (vn_phi_insert): Likewise.
20571 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20573 PR tree-optimization/56350
20574 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20575 if haven't found reduction or nested cycle operand, rather than
20576 asserting we must find it.
20578 PR tree-optimization/56381
20579 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20582 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20583 Jakub Jelinek <jakub@redhat.com>
20586 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20587 (swap_optab_enable): Same.
20588 (init_all_optabs): Use argument instead of global.
20589 * tree.h (struct tree_optimization_option): New field target_optabs.
20590 * expr.h (init_all_optabs): Add argument to prototype.
20591 (TREE_OPTIMIZATION_OPTABS): New.
20592 (save_optabs_if_changed): Protoize.
20593 * optabs.h: Declare this_fn_optabs.
20594 * optabs.c (save_optabs_if_changed): New.
20595 Declare this_fn_optabs.
20596 (init_optabs): Add argument to init_all_optabs() call.
20597 * function.c (invoke_set_current_function_hook): Handle per
20599 * function.h (struct function): New field optabs.
20600 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20601 optimization_current_node has changed.
20602 * target-globals.h (save_target_globals_default_opts): Protoize.
20603 * target-globals.c (save_target_globals_default_opts): New.
20605 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20608 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20609 registers %fr12 and %fr12R as call used.
20612 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20613 and HImode, require all displacements to be an integer multiple of
20615 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20616 only allow QImode and HImode when reload is in progress and strict is
20617 true. Likewise for symbolic addresses. Use base14_operand to check
20618 displacements in REG+BASE addresses.
20620 2013-02-18 Richard Biener <rguenther@suse.de>
20622 PR tree-optimization/56366
20623 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20624 handle sign-conversion of outer-loop initial induction value.
20626 2013-02-18 Richard Biener <rguenther@suse.de>
20628 PR middle-end/56349
20629 * cfghooks.c (merge_blocks): If we merge a latch into another
20630 block adjust references to it.
20631 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20632 (verify_loop_structure): Verify that a recorded latch is in fact
20635 2013-02-18 Richard Biener <rguenther@suse.de>
20637 PR tree-optimization/56321
20638 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20639 order SSA name release and virtual operand unlinking.
20641 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20643 * config/microblaze/microblaze.md (save_stack_block): Define.
20644 (restore_stack_block): Likewise.
20646 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20648 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20649 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20650 * config/microblaze/microblaze.c (microblaze_option_override):
20651 Bail out early for PIC modes when target does not support PIC.
20653 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20655 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20656 Replace with a microblaze version.
20657 (microblaze_trampoline_init): Adapt for microblaze.
20658 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20661 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20662 Dodji Seketeli <dodji@redhat.com>
20665 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20666 (instrument_mem_region_access): Do not forget to always put
20667 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20668 0) statement, even for cases where either 'base' or 'base + len'
20669 are not instrumented -- because they have been previously
20670 instrumented. Simplify the logic by putting all the statements
20671 instrument 'base + len' inside a sequence, and then insert that
20672 sequence right before the current insertion point. Then, to
20673 instrument 'base + len', just get an iterator on that statement.
20674 And do not forget to update the pointer to iterator the function
20675 received as argument.
20677 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20679 PR rtl-optimization/56348
20680 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20682 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20684 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20685 (clean_graph_dump_file): Pass base to start_graph_dump.
20687 2013-02-14 Richard Henderson <rth@redhat.com>
20690 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20692 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20694 * collect2-aix.h: Define F_LOADONLY.
20696 2013-02-14 Richard Biener <rguenther@suse.de>
20699 * varasm.c (output_constant_def_1): Get the decl representing
20700 the constant as argument.
20701 (output_constant_def): Wrap output_constant_def_1.
20702 (make_decl_rtl): Use output_constant_def_1 with the decl
20703 representing the constant.
20704 (build_constant_desc): Optionally re-use a decl already
20705 representing the constant.
20706 (tree_output_constant_def): Adjust.
20708 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20711 * asan.c (instrument_builtin_call): Really put the length of the
20712 second source argument into src1_len.
20714 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20716 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20717 argument. If it is false, don't create edge from then_bb to
20719 (insert_if_then_before_iter): Pass true to it.
20720 (build_check_stmt): Pass false to it.
20721 (transform_statements): Flush hash table only on extended basic
20722 block boundaries, rather than at the beginning of every bb.
20723 Don't flush hash table on nonfreeing_call_p calls.
20724 * tree-flow.h (nonfreeing_call_p): New prototype.
20725 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20727 2013-02-13 David S. Miller <davem@davemloft.net>
20729 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20731 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20734 * ira.c (max_regno_before_ira): Move from ...
20736 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20739 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20741 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20743 2013-02-13 Richard Biener <rguenther@suse.de>
20746 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20747 globals in MEM_REFs.
20749 2013-02-13 Richard Biener <rguenther@suse.de>
20751 * loop-init.c (loop_optimizer_init): Clear loop state when
20752 re-initializing preserved loops.
20753 * loop-unswitch.c (unswitch_single_loop): Return whether
20754 we unswitched the loop. Do not verify loop state here.
20755 (unswitch_loops): When we unswitched a loop discover new loops.
20757 2013-02-13 Kostya Serebryany <kcc@google.com>
20759 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20761 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20763 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20765 Avoid instrumenting duplicated memory access in the same basic block
20766 * Makefile.in (asan.o): Add new dependency on hash-table.h
20767 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20768 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20769 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20770 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20771 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20772 (get_mem_ref_of_assignment): New functions.
20773 (get_mem_refs_of_builtin_call): Extract from
20774 instrument_builtin_call and tweak a little bit to make it fit with
20776 (instrument_builtin_call): Use the new
20777 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20778 of is_gimple_builtin_call.
20779 (instrument_derefs, instrument_mem_region_access): Insert the
20780 instrumented memory reference into the hash table.
20781 (maybe_instrument_assignment): Renamed instrument_assignment into
20782 this, and change it to advance the iterator when instrumentation
20783 actually happened and return true in that case. This makes it
20784 homogeneous with maybe_instrument_assignment, and thus give a
20785 chance to callers to be more 'regular'.
20786 (transform_statements): Clear the memory reference hash table
20787 whenever we enter a new BB, when we cross a function call, or when
20788 we are done transforming statements. Use
20789 maybe_instrument_assignment instead of instrumentation. No more
20790 need to special case maybe_instrument_assignment and advance the
20791 iterator after calling it; it's now handled just like
20792 maybe_instrument_call. Update comment.
20794 2013-02-13 Richard Biener <rguenther@suse.de>
20796 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20797 Fix loop discovery code.
20799 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20801 PR inline-asm/56148
20802 * lra-constraints.c (process_alt_operands): Match early clobber
20803 operand with itself. Check conflicts with earlyclobber only if
20804 the operand is not reloaded. Prefer to reload conflicting operand
20805 if earlyclobber and matching operands are the same.
20807 2013-02-12 Richard Biener <rguenther@suse.de>
20810 * lto-streamer-out.c (write_symbol): Do not output symbols
20811 for hard register variables.
20813 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20816 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20817 (umulsidi3_insn, mulsidi3_insn): New insns.
20819 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20821 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20822 (struct tune_params): Add vec_costs field.
20823 * config/arm/arm.c (arm_builtin_vectorization_cost)
20824 (arm_add_stmt_cost): New functions.
20825 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20826 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20827 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20828 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20829 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20830 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20831 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20833 2013-02-12 Richard Biener <rguenther@suse.de>
20836 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20837 decls again if possible.
20839 2013-02-12 Richard Biener <rguenther@suse.de>
20841 PR middle-end/56288
20842 * tree-ssa.c (verify_ssa_name): Fix check, move
20843 SSA_NAME_IN_FREE_LIST check up.
20845 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20846 Steven Bosscher <steven@gcc.gnu.org>
20848 PR rtl-optimization/56151
20849 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20850 equal to op0 or op1, and last_insn pattern is CODE operation
20851 with MEM dest and one of the operands matches that MEM.
20853 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20855 * doc/extend.texi: Document Function Multiversioning and "default"
20856 parameter string to target attribute.
20857 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20858 target attribute parameter is "default".
20859 (ix86_compare_version_priority): Remove checks for target attribute.
20860 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20861 Remove check for target attribute equal to NULL. Add assert.
20862 (ix86_generate_version_dispatcher_body): Change error to sorry.
20864 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20865 Jack Howarth <howarth@bromo.med.uc.edu>
20866 Patrick Marlier <patrick.marlier@gmail.com>
20869 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20870 define ENDFILE_SPEC as TM_DESTRUCTOR.
20871 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20873 2013-02-11 Alexander Potapenko <glider@google.com>
20874 Jack Howarth <howarth@bromo.med.uc.edu>
20875 Jakub Jelinek <jakub@redhat.com>
20878 * config/darwin.c (cdtor_record): Rename ctor_record.
20879 (sort_cdtor_records): Rename sort_ctor_records.
20880 (finalize_dtors): New routine to sort destructors by
20881 priority before use in assemble_integer.
20882 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20884 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20886 PR rtl-optimization/56275
20887 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20888 offset is non-negative and less than cmode size before
20889 calling simplify_subreg.
20891 2013-02-11 Richard Biener <rguenther@suse.de>
20893 PR tree-optimization/56264
20894 * cfgloop.h (fix_loop_structure): Adjust prototype.
20895 * loop-init.c (fix_loop_structure): Return the number of
20896 newly discovered loops.
20897 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20898 are discovered, do a full loop-closed SSA rewrite.
20900 2013-02-11 Richard Biener <rguenther@suse.de>
20902 PR tree-optimization/56273
20903 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20905 (check_array_ref): Fix missing newline in dumps.
20906 (search_for_addr_array): Likewise.
20908 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20910 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20912 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20915 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20917 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20919 PR rtl-optimization/56246
20920 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20922 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20923 constraints are satisfied.
20925 2013-02-08 Jeff Law <law@redhat.com>
20928 * emit-rtl.c (reg_is_parm_p): New function.
20929 * regs.h (reg_is_parm_p): New prototype.
20930 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20931 callee-clobbered registers.
20933 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20936 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20937 If there is no implicit builtin declaration, just return NULL.
20939 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20941 * config/i386/sse.md (FMAMODEM): New mode iterator.
20942 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20943 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20945 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20947 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20948 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20949 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20951 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20953 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
20954 (microblaze*-*-elf): Likewise.
20955 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
20957 * config/microblaze/microblaze-c.c: Add builtin defines for
20958 _LITTLE_ENDIAN and _BIG_ENDIAN.
20959 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
20960 add to TARGET_DEFAULT flags.
20961 Expand ASM_SPEC and LINK_SPEC.
20962 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
20963 * config/microblaze/microblaze.md: Update extendsidi2 and
20964 movdi_internal instructions to use low-order / high-order reg
20966 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
20967 options and inversemask / mask of LITTLE_ENDIAN.
20968 * config/microblaze/t-microblaze: Expand multilib options to
20969 include mlittle-endian (le) and update exceptions patterns.
20971 2013-02-08 Jakub Jelinek <jakub@redhat.com>
20973 PR rtl-optimization/56195
20974 * lra-constraints.c (get_reload_reg): Don't reuse regs
20975 if they have smaller mode than requested, if they have
20976 wider mode than requested, try to return a SUBREG.
20978 PR tree-optimization/56250
20979 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
20980 if type is unsigned and code isn't MULT_EXPR.
20982 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20984 PR tree-optimization/56064
20985 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
20986 bits according to mode.
20987 * fixed-value.h (fixed_from_double_int)
20988 (const_fixed_from_double_int): Adjust comments.
20990 2013-02-08 Richard Biener <rguenther@suse.de>
20993 * lto-streamer.h (struct data_in): Remove current_file, current_line
20994 and current_col members.
20995 * lto-streamer-out.c (lto_output_location): Stream changed bits
20996 en-block for efficiency.
20997 * lto-streamer-in.c (clear_line_info): Remove.
20998 (lto_input_location): Cache current file, line and column
20999 globally via local statics. Read changed bits en-block.
21000 (input_function): Do not call clear_line_info.
21001 (lto_read_body): Likewise.
21002 (lto_input_toplevel_asms): Likewise.
21004 2013-02-08 Michael Matz <matz@suse.de>
21006 PR tree-optimization/52448
21007 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21008 (nt_call_phase): New static.
21009 (add_or_mark_expr): Only mark accesses with newer phase than any
21011 (nonfreeing_call_p): New.
21012 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21013 (nt_fini_block): Keep blocks marked as visited.
21014 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21016 2013-02-08 Richard Biener <rguenther@suse.de>
21018 * ira.c (ira): Free broken dominator information.
21020 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21022 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21024 2013-02-08 Marek Polacek <polacek@redhat.com>
21026 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21028 2013-02-08 Richard Biener <rguenther@suse.de>
21030 PR middle-end/56181
21031 * cfgloop.h (flow_loops_find): Adjust.
21032 (bb_loop_header_p): Declare.
21033 * cfgloop.c (bb_loop_header_p): New function split out from ...
21034 (flow_loops_find): ... here. Adjust function signature,
21035 support incremental loop structure update.
21036 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21037 * cfgloopmanip.c (fix_loop_structure): Move ...
21038 * loop-init.c (fix_loop_structure): ... here.
21039 (apply_loop_flags): Split out from ...
21040 (loop_optimizer_init): ... here.
21041 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21042 in incremental mode, only remove dead loops here.
21044 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21047 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21048 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21049 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21050 (*round<mode>3.libgcc): New insns for fixed-modes.
21051 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21052 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21053 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21054 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21055 implementations. Define to __builtin_avr_absFX,
21056 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21057 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21058 __builtin_avr_countlsFX, respectively.
21059 * config/avr/avr-c.c (target.h): Include it.
21060 (enum avr_builtin_id): New enum.
21061 (avr_resolve_overloaded_builtin): New static function.
21062 (avr_register_target_pragmas): Use it to set
21063 targetm.resolve_overloaded_builtin.
21064 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21065 tree nodes used by DEF_BUILTIN.
21066 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21067 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21068 <AVR_BUILTIN_xxBITS>: Same.
21070 2013-02-08 Richard Biener <rguenther@suse.de>
21072 * cfgloop.c (verify_loop_structure): Properly handle
21073 a loop exiting to another loop header.
21074 * ira-int.h (ira_loops): Remove.
21075 * ira.c (ira_loops): Remove.
21076 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21077 (do_reload): Use loop_optimizer_finalize.
21078 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21079 number_of_loops to access the loop tree.
21080 (more_one_region_p): Likewise.
21081 (finish_loop_tree_nodes): Likewise.
21082 (rebuild_regno_allocno_maps): Likewise.
21083 (mark_loops_for_removal): Likewise.
21084 (mark_all_loops_for_removal): Likewise.
21085 (remove_unnecessary_regions): Likewise.
21086 (ira_build): Likewise.
21087 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21089 2013-02-08 Richard Biener <rguenther@suse.de>
21091 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21092 * ipa-pure-const.c (analyze_function): Avoid calling
21093 mark_irreducible_loops twice.
21094 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21096 2013-02-07 David S. Miller <davem@davemloft.net>
21098 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21100 * var-tracking.c (vt_add_function_parameter): Test the presence of
21101 HAVE_window_save properly and do not remap argument registers when
21102 we have a leaf function.
21104 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21107 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21109 * config/i386/i386.c (ix86_print_operand): Ditto.
21111 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21113 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21115 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21117 PR rtl-optimization/56225
21118 * lra-constraints.c (process_alt_operands): Check that reload hard
21119 reg can hold value for strict_low_part.
21121 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21124 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21125 dwarf2out_end_function.
21126 (in_first_function_p, maybe_at_text_label_p,
21127 first_loclabel_num_not_at_text_label): New variables.
21128 (dwarf2out_var_location): In the first function find out
21129 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21130 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21133 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21135 PR rtl-optimization/56178
21136 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21137 SUBREG of a register. Tidy up related block of code.
21138 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21139 note if the source is a register or a SUBREG of a register.
21141 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21144 * config/rs6000/rs6000.md (ptrm): New mode attr.
21145 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21146 call_value_indirect_aix<pttrsize>,
21147 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21150 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21152 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21153 if -bnortl. Convert to strcmp and strncmp.
21155 2013-02-07 Alan Modra <amodra@gmail.com>
21158 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21159 addresses won't wrap when offsetting.
21160 (rs6000_secondary_reload): Provide secondary reloads needed for
21161 wrapping LO_SUM addresses.
21163 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21165 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21166 MACH, just __MACH__.
21168 2013-02-06 Richard Biener <rguenther@suse.de>
21170 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21171 instead of calling fix_loop_structure.
21173 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21175 PR middle-end/56217
21176 * omp-low.c (use_pointer_for_field): Return false if
21177 lower_send_shared_vars doesn't generate any copy-out code.
21179 2013-02-06 Tom de Vries <tom@codesourcery.com>
21181 PR rtl-optimization/56131
21182 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21183 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21184 of the label is NULL. Add comment.
21186 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21188 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21191 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21192 (STATIC_LIBTSAN_LIBS): Likewise.
21193 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21194 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21195 is defined, don't add anything else beyond that.
21196 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21197 (LINK_COMMAND_SPEC): Use them.
21199 PR tree-optimization/56205
21200 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21201 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21202 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21204 2013-02-05 Richard Biener <rguenther@suse.de>
21206 PR tree-optimization/53342
21207 PR tree-optimization/53185
21208 * tree-vectorizer.h (vect_check_strided_load): Remove.
21209 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21210 not disallow peeling for vectorized strided loads.
21211 (vect_check_strided_load): Make static and simplify.
21212 (vect_analyze_data_refs): Adjust.
21213 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21214 correctly when vectorizing strided loads.
21216 2013-02-05 Richard Biener <rguenther@suse.de>
21218 * doc/install.texi: Refer to ISL, not PPL.
21220 2013-02-05 Jan Hubicka <jh@suse.cz>
21222 PR tree-optimization/55789
21223 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21225 2013-02-05 Jan Hubicka <jh@suse.cz>
21227 PR tree-optimization/55789
21228 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21229 the dead call anyway.
21231 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21234 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21236 2013-02-04 Alexander Potapenko <glider@google.com>
21237 Jack Howarth <howarth@bromo.med.uc.edu>
21238 Jakub Jelinek <jakub@redhat.com>
21241 * config/darwin.c (sort_ctor_records): Stabilized qsort
21242 on constructor priority by using original position.
21243 (finalize_ctors): New routine to sort constructors by
21244 priority before use in assemble_integer.
21245 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21247 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21250 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21251 about visibility on artificial decls.
21252 * config/sol2.c (solaris_assemble_visibility): Likewise.
21254 2013-02-04 Kai Tietz <ktietz@redhat.com>
21257 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21258 argument and improve checking of return-argument types for 16-byte
21260 (ix86_function_value_1): Add additional valtype argument on call
21261 of function_value_64.
21262 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21263 handling infunction_value_64 function.
21265 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21267 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21269 2013-02-04 Richard Biener <rguenther@suse.de>
21271 PR tree-optimization/56188
21272 * tree-ssa-structalias.c (label_visit): Consider case with
21273 initially non-empty points-to set.
21274 (perform_var_substitution): Dump node mapping and clean up.
21276 2013-02-04 Richard Guenther <rguenther@suse.de>
21279 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21280 node prevail as last resort.
21281 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21282 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21284 2013-02-04 Richard Biener <rguenther@suse.de>
21286 PR tree-optimization/56113
21287 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21289 (equiv_class_lookup_or_add): ... this.
21290 (label_visit): Adjust and fix error in previous patch.
21291 (perform_var_substitution): Adjust.
21293 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21295 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21296 * config/sh/sh4-300.md: Likewise.
21297 * config/sh/sh4a.md: Likewise.
21298 * config/sh/constraints.md: Likewise.
21299 * config/sh/sh.md: Likewise.
21300 * config/sh/netbsd-elf.h: Likewise.
21301 * config/sh/predicates.md: Likewise.
21302 * config/sh/sh-protos.h: Likewise.
21303 * config/sh/ushmedia.h: Likewise.
21304 * config/sh/linux.h: Likewise.
21305 * config/sh/sh.c: Likewise.
21306 * config/sh/superh.h: Likewise.
21307 * config/sh/elf.h: Likewise.
21308 * config/sh/sh4.md: Likewise.
21309 * config/sh/sh.h: Likewise.
21311 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21313 * config/pa/constraints.md: Adjust unused letters. Change "T"
21314 constraint to match_test floating_point_store_memory_operand().
21315 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21316 (base14_operand): New.
21317 (floating_point_store_memory_operand): New.
21318 (integer_store_memory_operand): Revise to use base14_operand and
21319 reg_plus_base_memory_operand.
21320 (move_dest_operand): Allow symbolic_memory_operands.
21321 (symbolic_memory_operand): Check for LO_SOM.
21322 (symbolic_operand): Change default case to break.
21323 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21324 CONST_DOUBLE values to be reloaded by putting them into memory when
21325 the destination is a floating point register.
21326 (movdf): Remove code to handle CONST_DOUBLE.
21328 (reload_indf_r1): New.
21329 (reload_insf_r1): New.
21330 Consistently use "Q" and "T" constraints with integer and floating
21331 point move instructions, respectively.
21332 (movdi): Remove FAIL.
21333 Change predicate for source operand unamed DImode move from
21334 general_operand to move_src_operand.
21335 (umulsidi3): Change predicate for destination operand to
21337 Likewise for similar unamed patterns.
21338 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21339 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21340 (hppa_legitimize_address): Simplify mask calculation.
21341 (pa_emit_move_sequence): Revised handling of secondary reloads from
21342 REG+D addresses for floating point loads and stores. Directly handle
21343 loading CONST0_RTX (mode) to a floating point register.
21344 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21345 to floating point registers. Don't restrict secondary reloads to
21346 floating point registers to integer modes. Revise some comments and
21348 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21349 (pa_legitimate_address_p): New.
21350 (pa_legitimize_reload_address): New.
21351 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21352 (STRICT_REG_OK_FOR_BASE_P): New.
21353 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21354 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21356 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21357 Andrew Dixie <andrewd@gentrack.com>
21359 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21362 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21364 * expmed.c (extract_bit_field_1): Pass the full width of the
21365 structure to get_best_reg_extraction_insn.
21367 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21370 * configure.ac (use_cxa_atexit): Add AIX.
21371 * configure: Regenerate.
21373 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21375 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21378 * final.c (need_profile_function): New variable.
21379 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21380 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21381 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21382 notes, targetm.asm_out.function_prologue doesn't emit anything,
21383 HAVE_prologue and profiler should be emitted before prologue,
21384 set need_profile_function instead of emitting it.
21385 (final_scan_insn): If need_profile_function, emit
21386 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21387 NOTE_INSN_FUNCTION_BEG note.
21389 2013-02-01 Richard Henderson <rth@redhat.com>
21391 * config/rs6000/rs6000.md (smulditi3): New.
21394 * config/alpha/alpha.md (umulditi3): New.
21396 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21398 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21399 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21401 2013-02-01 Richard Biener <rguenther@suse.de>
21403 PR tree-optimization/56113
21404 * tree-ssa-structalias.c (label_visit): Reduce work for
21405 single-predecessor nodes.
21407 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21409 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21410 range isn't testing for zero.
21412 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21414 PR middle-end/56113
21415 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21417 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21418 Nick Clifton <nickc@redhat.com>
21420 * config/v850/constraints.md (Q): Define as a memory constraint.
21421 * config/v850/predicates.md (label_ref_operand): New predicate.
21422 (e3v5_shift_operand): New predicate.
21423 (ior_operator): New predicate.
21424 * config/v850/t-v850: Add e3v5 multilib.
21425 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21426 (v850_gen_movdi): Prototype.
21427 * config/v850/v850.c: Add support for e3v5 architecture.
21428 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21430 (construct_save_jarl): Add e3v5 long JARL support.
21431 (v850_adjust_insn_length): New function. Adjust length of call
21432 insns when using e3v5 instructions.
21433 (v850_gen_movdi): New function: Generate instructions to move a
21435 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21436 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21437 (TARGET_USE_FPU): Enable for e3v5.
21438 (CONST_OK_FOR_W): New macro.
21439 (ADJUST_INSN_LENGTH): Define.
21440 * config/v850/v850.md (UNSPEC_LOOP): Define.
21441 (attr cpu): Add v850e3v5.
21442 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21443 (movdi): New pattern.
21444 (movdi_internal): New pattern.
21445 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21446 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21447 (cstoresf4): Likewise.
21448 (cstoredf4): Likewise.
21449 (insv): New pattern.
21450 (rotlso3_a): New pattern.
21451 (rotlsi3_b): New pattern
21452 (rotlsi3_v850e3v5): New pattern.
21453 (doloop_begin): New pattern.
21454 (fix_loop_counter): New pattern.
21455 (doloop_end): New pattern.
21456 (branch_normal): Add e3v5 long branch support.
21457 (branch_invert): Likewise.
21458 (branch_z_normal): Likewise.
21459 (branch_z_invert): Likewise.
21460 (branch_nz_normal): Likewise.
21461 (branch_nz_invert): Likewise.
21462 (call_internal_short): Add e3v5 register-indirect JARL support.
21463 (call_internal_long): Likewise.
21464 (call_value_internal_short): Likewise.
21465 (call_value_internal_long): Likewise.
21466 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21467 (mloop): New option.
21468 * config.gcc: Add support for configuring v840e3v5 target.
21469 * doc/invoke.texi: Document new v850 specific command line options.
21471 2013-01-31 Paul Koning <ni1d@arrl.net>
21475 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21476 children if parent is a class.
21477 (prune_unused_types_prune): Don't add DW_AT_declaration.
21479 2013-01-31 Richard Biener <rguenther@suse.de>
21481 PR tree-optimization/56157
21482 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21483 match up operand with SLP child.
21485 2013-01-31 Jason Merrill <jason@redhat.com>
21488 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21489 parameters the first time.
21490 (gen_scheduled_generic_parms_dies): Check completeness here.
21492 2013-01-31 Richard Biener <rguenther@suse.de>
21494 PR middle-end/53073
21495 * common.opt (faggressive-loop-optimizations): New flag,
21496 enabled by default.
21497 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21498 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21499 infer_loop_bounds_from_undefined by it.
21501 2013-01-31 Richard Biener <rguenther@suse.de>
21503 PR tree-optimization/56150
21504 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21505 visit virtual operands.
21506 (find_uses_to_rename_bb): Likewise.
21508 2013-01-31 Richard Biener <rguenther@suse.de>
21510 PR tree-optimization/56150
21511 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21512 mixed store non-store stmts.
21514 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21517 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21518 LIBASAN_EARLY_SPEC is defined.
21519 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21520 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21522 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21525 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21526 invalid args instead of ICEing on it.
21527 (ix86_valid_target_attribute_tree): Return error_mark_node if
21528 ix86_valid_target_attribute_inner_p failed.
21529 (ix86_valid_target_attribute_p): Return false only if
21530 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21531 target("default") attribute.
21532 (sorted_attr_string): Change argument from const char * to tree,
21533 merge in all target attribute arguments rather than just one.
21534 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21535 instead of free. Avoid using strcat.
21536 (ix86_mangle_function_version_assembler_name): Mangle
21537 target("default") as if no target attribute is present. Adjust
21538 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21539 instead of xmalloc and XDELETEVEC instead of free.
21540 (ix86_function_versions): Don't return true if one of the decls
21541 doesn't have target attribute. If they don't and one of the decls
21542 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21543 sorted_attr_string caller. Use XDELETEVEC instead of free.
21544 (ix86_supports_function_versions): Remove.
21545 (make_name): Fix up formatting.
21546 (make_dispatcher_decl): Remove resolver_name and its initialization.
21547 Avoid leaking memory.
21548 (is_function_default_version): Return true if there is
21549 target("default") attribute rather than no target attribute at all.
21550 (make_resolver_func): Avoid leaking memory.
21551 (ix86_generate_version_dispatcher_body): Likewise.
21552 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21553 * target.def (supports_function_versions): Remove.
21554 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21555 * doc/tm.texi: Regenerated.
21557 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21559 PR rtl-optimization/56144
21560 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21561 for values with side effects.
21563 2013-01-30 Richard Biener <rguenther@suse.de>
21565 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21566 (sparseset_pop): Likewise.
21567 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21568 to be able to use quick_push in the worker loop.
21570 2013-01-30 Marek Polacek <polacek@redhat.com>
21572 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21574 2013-01-30 Richard Biener <rguenther@suse.de>
21577 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21579 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21581 PR tree-optimization/56064
21582 * fixed-value.c (fixed_from_double_int): New function.
21583 * fixed-value.h (fixed_from_double_int): New prototype.
21584 (const_fixed_from_double_int): New static inline function.
21585 * fold-const.c (native_interpret_fixed): New static function.
21586 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21587 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21588 (native_encode_fixed): New static function.
21589 (native_encode_expr) <FIXED_CST>: Use it.
21590 (native_interpret_int): Move double_int worker code to...
21591 * double-int.c (double_int::from_buffer): ...this new static method.
21592 * double-int.h (double_int::from_buffer): Prototype it.
21594 2013-01-30 Richard Biener <rguenther@suse.de>
21596 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21597 New pointer-map and obstack.
21598 (init_alias_vars): Allocate pointer-map and obstack.
21599 (delete_points_to_sets): Free them.
21600 (find_what_var_points_to): Cache result.
21601 (find_what_p_points_to): Adjust for changed interface of
21602 find_what_var_points_to.
21603 (compute_points_to_sets): Likewise.
21604 (ipa_pta_execute): Likewise.
21606 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21608 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21609 * configure: Regenerate.
21610 * config.in: Regenerate.
21611 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21612 #nobits/#progbits if supported.
21614 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21617 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21618 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21619 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21621 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21623 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21624 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21626 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21628 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21629 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21631 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21633 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21635 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21636 * config/arm/cortex-a7.md: New bypasses using
21637 arm_mac_accumulator_is_result.
21639 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21641 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21642 (cortex_a7_neon_mla): Likewise.
21643 (cortex_a7_fpfmad): New reservation.
21644 (cortex_a7_fpmacs): Use ffmas and update required units.
21645 (cortex_a7_fpmuld): Update required units and latency.
21646 (cortex_a7_fpmacd): Likewise.
21647 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21648 (cortex_a7_neon). Likewise.
21649 (bypass) Update participating units.
21651 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21653 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21654 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21656 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21657 (vfp_fmul): Use ffmad.
21658 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21659 (cortex_r4_fmacd): Use ffmad.
21660 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21661 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21662 (cortex_a9_fmacd): Use ffmad.
21663 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21664 (cortex_a8_vfp_macd): Use ffmad.
21665 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21666 (cortex_a5_fpmacd): Use ffmad.
21667 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21668 (cortex_a15_vfp_macd): Use ffmad.
21669 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21671 2013-01-29 Jason Merrill <jason@redhat.com>
21674 * varasm.c (default_assemble_visibility): Don't warn about
21675 visibility on artificial decls.
21677 2013-01-29 Richard Biener <rguenther@suse.de>
21679 PR tree-optimization/56113
21680 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21682 (label_visit): Free duplicate bitmaps and record the leader instead.
21683 (perform_var_substitution): Adjust.
21685 2013-01-29 Richard Biener <rguenther@suse.de>
21687 PR tree-optimization/55270
21688 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21689 the CFG, schedule loops for fixup.
21691 2013-01-29 Nick Clifton <nickc@redhat.com>
21693 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21696 2013-01-28 Leif Ekblad <leif@rdos.net>
21698 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21699 * config/i386/i386.h (TARGET_RDOS): New macro.
21700 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21701 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21702 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21703 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21704 DEFAULT_LARGE_SECTION_THRESHOLD.
21705 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21706 * config/i386/rdos.h: New file.
21707 * config/i386/rdos64.h: New file.
21709 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21712 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21715 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21717 PR rtl-optimization/56117
21718 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21719 call cselib_lookup_from_insn on the MEM before calling
21720 add_insn_mem_dependence.
21722 2013-01-28 Richard Biener <rguenther@suse.de>
21724 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21725 to a stmt that didn't have one.
21726 (copy_phis_for_bb): Likewise for PHI arguments.
21727 (copy_debug_stmt): Likewise for debug stmts.
21729 2013-01-28 Richard Biener <rguenther@suse.de>
21731 PR tree-optimization/56034
21732 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21733 (partition_builtin_p): Adjust.
21734 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21735 it is the last partition.
21736 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21737 up the vertex for the definition.
21738 (classify_partition): Classify whether a partition is a
21739 PKIND_REDUCTION, thus has uses outside of the loop.
21740 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21741 Merge all PKIND_REDUCTION partitions into the last partition.
21742 (tree_loop_distribution): Seed partitions from reductions as well.
21744 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21746 PR tree-optimization/56125
21747 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21748 pow(x,c) into sqrt(x) * powi(x, n/2) or
21749 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21750 optimizing for size.
21751 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21752 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21755 PR tree-optimization/56094
21756 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21757 to UNKNOWN_LOCATION while gimplifying expr.
21759 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21762 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21763 operand 0 in movabs insn template for -masm=intel asm alternative.
21764 (*movabs<mode>_2): Ditto for operand 1.
21766 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21769 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21772 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21774 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21777 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21779 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21780 Undef to avoid warning.
21782 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21784 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21785 * configure: Regenerate.
21787 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21789 PR tree-optimization/56098
21790 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21791 for stmts with volatile ops.
21792 (cond_store_replacement): Don't optimize if assign has volatile ops.
21793 (cond_if_else_store_replacement_1): Don't optimize if either
21794 then_assign or else_assign have volatile ops.
21795 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21798 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21800 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21802 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21804 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21805 missing ':' in asm example.
21807 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21809 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21810 entries into lane and laneq entries.
21811 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21812 Remove AdvSIMD scalar modes.
21813 (aarch64_sq<r>dmulh_laneq<mode>): New.
21814 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21816 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21817 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21818 * config/aarch64/iterators.md (VCOND): New.
21821 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21824 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21825 Add NULL LIBNAME argument to existing definitions.
21826 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21827 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21828 * config/avr/avr.c (DEF_BUILTIN): Same.
21829 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21830 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21831 implementation is available (DECL_ASSEMBLER_NAME is set).
21832 (avr_fold_absfx): New static function.
21833 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21834 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21835 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21836 AVR_BUILTIN_ABSLLK.
21837 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21838 (abshk, absk, abslk, absllk): Provide as static inline functions.
21840 2013-01-25 Marek Polacek <polacek@redhat.com>
21842 PR tree-optimization/56035
21843 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21845 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21847 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21848 (*movtf_internal_rex64): Add (!o,C) alternative
21849 (*movxf_internal_rex64): Ditto.
21850 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21852 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21854 * doc/invoke.texi: fix typo.
21855 * doc/objc.texi: fix typo.
21857 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21859 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21860 for the first two alternatives.
21862 2013-01-24 Diego Novillo <dnovillo@google.com>
21864 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21865 (ggc-zone.o): Remove.
21866 * configure.ac: Remove option --with-gc.
21867 * configure: Re-generate.
21868 * doc/install.texi: Remove documentation for --with-gc.
21869 * gengtype.c (write_enum_defn): Remove. Update all users.
21870 (write_Types_process_field): Remove generation of gt_e_* argument.
21871 (output_type_enum): Remove. Update all users.
21872 (write_enum_defn): Remove. Update all users.
21873 (enum alloc_zone): Remove. Update all users.
21874 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21875 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21876 Update all callers.
21877 (struct ptr_data): Remove field TYPE. Update all users.
21878 (gt_pch_note_object): Remove argument TYPE. Update all users.
21879 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21881 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21882 (struct alloc_zone): Remove.
21883 (ggc_internal_alloc_zone_stat): Remove.
21884 (ggc_internal_cleared_alloc_zone_stat): Remove.
21885 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21886 (ggc_pch_count_object): Remove last argument. Update all users.
21887 (ggc_pch_alloc_object): Remove last argument. Update all users.
21888 (struct alloc_zone): Remove.
21889 * ggc-zone.c: Remove.
21890 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21891 (struct alloc_zone): Remove.
21892 (ggc_alloc_typed_stat): Remove.
21893 (ggc_alloc_typed): Remove.
21894 (ggc_splay_alloc): Remove first argument.
21895 (rtl_zone): Remove. Update all users.
21896 (tree_zone): Remove. Update all users.
21897 (tree_id_zone): Remove. Update all users.
21898 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21899 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21900 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21901 * tree-ssanames.c: Remove references to zone allocator in comments.
21903 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21905 * config/avr/avr.c (avr_out_fract): Make register numbers that
21906 might be outside of source operand signed.
21908 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21910 * config/i386/constraints.md (Yf): New constraint.
21911 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21912 of f constraint to conditionaly disable x87 register preferences.
21913 (*movdf_internal): Ditto.
21914 (*movsf_internal): Ditto.
21916 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21918 PR inline-asm/55934
21919 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21920 that have operands with impossible constraints.
21921 Add a FIXME for a speed-up opportunity.
21922 * lra-constraints.c (process_alt_operands): Verify that a class
21923 selected from constraints on asms is valid for the operand mode.
21924 (curr_insn_transform): Remove incorrect comment.
21926 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21928 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21929 TOC operand is a valid symbol ref in the constant pool.
21931 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21933 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21935 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21938 * config/avr/stdfix.h: New file.
21939 * t-avr (stdfix-gcc.h): New rule to build it.
21940 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21942 2013-01-23 Kostya Serebryany <kcc@google.com>
21944 * config/darwin.h: remove dependency on
21945 CoreFoundation (asan on Mac OS).
21947 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21950 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21951 instead of cmpdi_operand for first comparison operand.
21952 Don't assert that comparison operands aren't both constants.
21954 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21956 * doc/install.texi (Downloading the Source): Update references to
21957 downloading separate components.
21959 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21961 * doc/extend.texi (__int128): Improve grammar.
21963 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
21966 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
21967 alternative to (o,r).
21968 (*movdi_internal_rex64): Remove (!o,n) alternative.
21969 (DImode immediate->memory splitter): Remove.
21970 (DImode immediate->memory peephole2): Remove.
21971 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
21972 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
21973 alternative to (!o,*r).
21974 (*movtf_internal_sse): New pattern.
21975 (*movxf_internal_rex64): New pattern.
21976 (*movxf_internal): Disable for TARGET_64BIT.
21977 (*movdf_internal_rex64): Remove (!o,F) alternative.
21979 2013-01-22 Jakub Jelinek <jakub@redhat.com>
21981 PR middle-end/56074
21982 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
21983 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
21984 * tree-vect-loop-manip.c (find_loop_location): Also ignore
21985 stmt locations where LOCATION_LOCUS of the stmt location is
21986 UNKNOWN_LOCATION or BUILTINS_LOCATION.
21989 * config/i386/i386.md (UNSPEC_STOS): New.
21990 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
21991 *strsetqi_1): Add UNSPEC_STOS.
21993 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
21996 * doc/invoke.texi: Remove left over -Wsynth example.
21998 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22000 PR tree-optimization/56051
22001 * fold-const.c (fold_binary_loc): Don't fold
22002 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22003 a narrowing conversion, or widening conversion from signed
22006 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22008 PR rtl-optimization/56023
22009 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22010 dependent on debug instruction.
22012 2013-01-21 Martin Jambor <mjambor@suse.cz>
22014 PR middle-end/56022
22015 * function.c (allocate_struct_function): Call
22016 invoke_set_current_function_hook earlier.
22018 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22020 * reload1.c (init_reload): Only initialize reload_obstack
22021 during the first call.
22023 2013-01-21 Marek Polacek <polacek@redhat.com>
22025 * cfgloop.c (verify_loop_structure): Fix up grammar.
22027 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22029 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22030 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22032 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22035 * config/arm/marvell-pj4.md: Update copyright year.
22036 Fix up use of alu to alu_reg and simple_alu_imm.
22038 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22040 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22042 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22045 * lra-constraints.c (curr_insn_transform): Don't reuse original
22046 insn for secondary memory move when memory mode should be different.
22048 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22050 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22051 atomic_storedi_1): New patterns.
22053 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22055 btver2 pipeline descriptions.
22056 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22058 * config/i386/i386.md (btver2_decode): New type attributes.
22059 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22061 * config/i386/btver2.md: New file describing btver2 pipelines.
22063 2013-01-19 Andrew Pinski <apinski@cavium.com>
22065 PR tree-optimization/52631
22066 * tree-ssa-sccvn (visit_use): Before looking up the original
22067 statement, try looking up the simplified expression.
22069 2013-01-19 Anthony Green <green@moxielogic.com>
22071 * config/moxie/moxie.c (moxie_expand_prologue): Set
22072 current_function_static_stack_size.
22074 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22076 PR tree-optimization/56029
22077 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22078 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22080 2013-01-18 Sharad Singhai <singhai@google.com>
22082 PR tree-optimization/55995
22083 * dumpfile.c (dump_loc): Print location only if available.
22084 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22086 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22089 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22090 secondary memory move.
22091 (inherit_reload_reg): Use rclass instead of cl for
22092 check_secondary_memory_needed_p.
22094 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22096 PR middle-end/56015
22097 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22098 the case where writing real complex part of target modifies op1.
22100 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22102 * config/aarch64/aarch64-simd.md
22103 (aarch64_vcond_internal<mode>): Handle unordered cases.
22104 * config/aarch64/iterators.md (v_cmp_result): New.
22106 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22107 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22109 * config/arm/marvell-pj4.md: New file.
22110 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22111 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22112 (generic_vfp): Likewise.
22113 * config/arm/arm-cores.def: Add marvell-pj4.
22114 * config/arm/arm-tune.md: Regenerate.
22115 * config/arm/arm-tables.opt: Regenerate.
22116 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22117 * doc/invoke.texi: Document marvell-pj4.
22119 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22121 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22123 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22128 * var-tracking.c (negative_power_of_two_p): New.
22129 (global_get_addr_cache, local_get_addr_cache): New.
22130 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22131 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22133 (vt_stack_offset_p): Remove.
22134 (vt_canon_true_dep): Always canonicalize loc's address.
22135 (clobber_overlapping_mems): Make sure we have a MEM.
22136 (local_get_addr_clear_given_value): New.
22137 (val_reset): Clear local cached entries.
22138 (compute_bb_dataflow): Create and release the local cache.
22139 Disable duplicate MEMs clobbering.
22140 (emit_notes_in_bb): Clobber MEMs likewise.
22141 (vt_emit_notes): Create and release the local cache.
22142 (vt_initialize, vt_finalize): Create and release the global
22143 cache, respectively.
22144 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22146 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22148 PR libmudflap/53359
22149 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22150 not found in the symtab.
22152 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22155 PR rtl-optimization/55547
22156 PR rtl-optimization/53827
22159 * alias.c (offset_overlap_p): New, factored out of...
22160 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22161 the conservative special case for symbolic constants. Don't
22162 adjust zero sizes on alignment.
22164 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22166 PR rtl-optimization/52573
22167 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22168 REG_UNUSED for the same register.
22170 2013-01-17 Richard Biener <rguenther@suse.de>
22171 Marek Polacek <polacek@redhat.com>
22173 PR rtl-optimization/55833
22174 * loop-unswitch.c (unswitch_loops): Move loop verification...
22175 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22176 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22177 Set it to true when we're removing a loop from hierarchy tree in
22178 an irreducible region.
22179 (fix_bb_placements): Adjust caller.
22180 (fix_loop_placements): Likewise.
22182 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22184 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22185 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22186 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22187 Remove ID. Adjust comments.
22188 * config/avr/avr-c.c (avr_builtin_name): Remove.
22189 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22190 * config/avr/avr.c (avr_tolower): New static function.
22191 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22192 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22193 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22196 2013-01-17 Jan Hubicka <jh@suse.cz>
22198 PR tree-optimization/55273
22199 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22201 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22204 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22205 store through atomic_store<mode>_1.
22206 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22208 2013-01-17 Martin Jambor <mjambor@suse.cz>
22210 PR tree-optimizations/55264
22211 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22212 for virtual methods.
22213 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22214 virtual methods before inlining is over.
22215 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22217 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22220 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22222 PR rtl-optimization/56005
22223 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22224 pending reads for prefetch.
22226 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22228 * config/aarch64/aarch64.md
22229 (*cstoresi_neg_uxtw): New pattern.
22230 (*cmovsi_insn_uxtw): New pattern.
22231 (*<optab>si3_uxtw): New pattern.
22232 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22233 (*<optab>si3_insn_uxtw): New pattern.
22234 (*bswapsi2_uxtw): New pattern.
22236 2013-01-16 Richard Biener <rguenther@suse.de>
22238 * tree-inline.c (tree_function_versioning): Remove set but
22239 never used variable.
22241 2013-01-16 Richard Biener <rguenther@suse.de>
22243 PR tree-optimization/55964
22244 * tree-flow.h (rename_variables_in_loop): Remove.
22245 (rename_variables_in_bb): Likewise.
22246 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22247 (copy_loop_before): Adjust and delete update-ssa status.
22248 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22249 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22250 (rename_variables_in_loop): Remove.
22251 (slpeel_update_phis_for_duplicate_loop): Likewise.
22252 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22253 use available cfg machinery instead of duplicating it.
22254 Update PHI nodes and perform poor-mans SSA update here.
22255 (slpeel_tree_peel_loop_to_edge): Adjust.
22257 2013-01-16 Richard Biener <rguenther@suse.de>
22259 PR tree-optimization/54767
22260 PR tree-optimization/53465
22261 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22262 (vrp_visit_phi_node): For PHI arguments coming via backedges
22263 drop all symbolical range information.
22264 (execute_vrp): Compute backedges.
22266 2013-01-16 Richard Biener <rguenther@suse.de>
22268 * doc/install.texi: Update CLooG and ISL requirements to
22271 2013-01-16 Christian Bruel <christian.bruel@st.com>
22274 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22275 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22276 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22278 2013-01-16 DJ Delorie <dj@redhat.com>
22280 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22281 (UNSPECV_SP_SWITCH_E): New.
22282 (sp_switch_1): Change to an unspec.
22283 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22286 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22288 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22289 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22290 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22291 (expand_mem_thread_fence): Ditto.
22292 (expand_mem_signal_fence): Ditto.
22293 (expand_atomic_load): Ditto.
22294 (expand_atomic_store): Ditto.
22296 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22298 PR rtl-optimization/55547
22299 PR rtl-optimization/53827
22302 * alias.c (memrefs_conflict_p): Set sizes to negative after
22305 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22308 * function.c (thread_prologue_and_epilogue_insns): Always
22309 add crtl->drap_reg to set_up_by_prologue.set, even if
22310 stack_realign_drap is false.
22312 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22314 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22315 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22316 *call): Fix indention.
22318 2013-01-15 Tom de Vries <tom@codesourcery.com>
22321 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22324 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22326 PR rtl-optimization/55153
22327 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22329 2013-01-15 Martin Jambor <mjambor@suse.cz>
22331 PR tree-optimization/55920
22332 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22333 accesses as grp_to_be_debug_replaced.
22335 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22337 PR tree-optimization/55920
22338 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22339 there is non-useless type conversion needed from debug rhs to lhs,
22340 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22342 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22343 Mikael Pettersson <mikpe@it.uu.se>
22346 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22348 (ASM_OUTPUT_CASE_LABEL): Remove.
22349 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22350 * final.c (shorten_branches): Update alignment of labels before
22351 jump tables if CASE_VECTOR_SHORTEN_MODE.
22353 2013-01-15 Richard Biener <rguenther@suse.de>
22356 * system.h: Do not include gmp.h for building host tools.
22358 2013-01-15 Richard Biener <rguenther@suse.de>
22360 PR middle-end/55882
22361 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22362 account for bitpos when computing alignment.
22364 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22366 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22367 (ix86_target_macros_internal): Likewise.
22369 * config/i386/i386.c (m_CORE2I7): Removed.
22370 (m_CORE_HASWELL): New macro.
22371 (m_CORE_ALL): Likewise.
22372 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22373 (initial_ix86_arch_features): Likewise.
22374 (processor_target_table): Initializations for Core avx2.
22375 (cpu_names): New names "core-avx2".
22376 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22377 PROCESSOR_CORE_HASWELL.
22378 (ix86_issue_rate): New case.
22379 (ia32_multipass_dfa_lookahead): Likewise.
22380 (ix86_sched_init_global): Likewise.
22382 * config/i386/i386.h (TARGET_HASWELL): New macro.
22383 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22384 (processor_type): New PROCESSOR_HASWELL.
22386 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22388 PR tree-optimization/55955
22389 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22390 *SHIFT_EXPR and *ROTATE_EXPR codes.
22392 PR tree-optimization/48766
22393 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22394 -ftrapv disable -fwrapv.
22396 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22399 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22400 etc. to 1 and not to __flash.
22401 Use LL suffix for __INT24_MAX__ with -mint8.
22402 Use ULL suffix for __UINT24_MAX__ with -mint8.
22404 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22406 * config/avr/avr-arch.h
22407 (struct base_arch_s): Use typedef avr_arch_t instead.
22408 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22409 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22410 * config/avr/avr.c: Same.
22411 * config/avr/avr-devices.c: Same.
22412 * config/avr/driver-avr.c: Same.
22413 * config/avr/gen-avr-mmcu-texi.c: Same.
22414 * config/avr/avr-mcus.def: Adjust comment.
22416 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22418 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22419 * config/aarch64/iterators.md (VALLDI): New.
22421 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22422 Andi Kleen <ak@linux.intel.com>
22425 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22426 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22429 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22431 * config/avr/avr-stdint.h: Remove trailing blanks.
22432 * config/avr/avr-log.h: Same.
22433 * config/avr/avr-arch.h: Same.
22434 * config/avr/avr-devices.c: Same.
22435 * config/avr/avr-dimode.md: Same.
22436 * config/avr/predicates.md: Same.
22437 * config/avr/avr-c.c: Same. And fix typo.
22439 * config/avr/avr-protos.h: Same. And:
22440 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22441 (init_cumulative_args): Rename to avr_init_cumulative_args.
22442 (expand_prologue): Rename to avr_expand_prologue.
22443 (expand_epilogue): Rename to avr_expand_epilogue.
22444 (adjust_insn_length): Rename to avr_adjust_insn_length.
22445 (notice_update_cc): Rename to avr_notice_update_cc.
22446 (final_prescan_insn): Rename to avr_final_prescan_insn.
22447 * config/avr/avr.c: Same.
22448 * config/avr/avr.h: Same.
22449 * config/avr/avr.md: Remove trailing blanks.
22450 (prologue): Use avr_expand_prologue.
22451 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22453 2013-01-14 Richard Biener <rguenther@suse.de>
22455 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22456 verify_location, collect_subblocks): New functions.
22457 (verify_gimple_in_cfg): Verify that locations only reference
22458 BLOCKs in the functions BLOCK tree.
22460 2013-01-14 Richard Biener <rguenther@suse.de>
22462 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22464 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22466 (insert_out_of_ssa_copy_on_edge): Likewise.
22467 (rewrite_close_phi_out_of_ssa): Likewise.
22468 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22470 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22471 propagated constants.
22472 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22475 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22477 * config/avr/avr-modes.def: Add GPL copyright notice.
22479 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22481 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22482 MEMMODEL_MASK to determine memory model.
22483 (atomic_store<mode>): Ditto from operands[2].
22484 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22486 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22489 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22490 (fold_gimple_assign): Don't call unshare_expr here.
22491 (fold_ctor_reference): Call unshare_expr.
22493 2013-01-13 Terry Guo <terry.guo@arm.com>
22495 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22496 * doc/fragments.texi: Document MULTILIB_REUSE.
22497 * gcc.c (multilib_reuse): New internal spec.
22498 (set_multilib_dir): Also search multilib from multilib_reuse.
22499 * genmultilib (tmpmultilib3): Refactor code.
22500 (tmpmultilib4): Ditto.
22501 (multilib_reuse): New multilib argument.
22503 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22505 * Makefile.in: Update copyright.
22507 2013-01-12 Tom de Vries <tom@codesourcery.com>
22509 PR middle-end/55890
22510 * calls.c (expand_call): Check if arg_nr is valid.
22512 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22514 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22515 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22516 documentation. Add missing '__' in front of
22517 __builtin_ia32_packssdw256.
22519 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22522 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22523 NO_REGS for larl operands.
22524 (s390_reload_larl_operand): Use s390_load_address instead of
22527 2013-01-11 Richard Biener <rguenther@suse.de>
22529 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22530 (verify_node_sharing): ... here.
22531 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22533 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22535 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22536 Substitute TREECHECKING.
22537 * configure: Regenerate.
22538 * Makefile.in (TREECHECKING): New.
22540 2013-01-11 Richard Guenther <rguenther@suse.de>
22542 PR tree-optimization/44061
22543 * tree-vrp.c (extract_range_basic): Compute zero as
22544 value-range for __builtin_constant_p of function parameters.
22546 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22548 Update copyright years.
22550 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22552 PR rtl-optimization/55672
22553 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22554 const to be eliminable.
22556 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22558 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22559 * configure: Regenerate.
22561 2013-01-10 Richard Biener <rguenther@suse.de>
22563 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22565 2013-01-10 Richard Biener <rguenther@suse.de>
22568 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22569 locations for virtual PHI arguments.
22570 (rewrite_update_phi_arguments): Likewise.
22572 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22574 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22577 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22579 PR tree-optimization/55921
22580 * tree-complex.c (expand_complex_asm): New function.
22581 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22583 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22586 * config/s390/s390.c (s390_symref_operand_p)
22587 (s390_loadrelative_operand_p): Merge the two functions.
22588 (s390_check_qrst_address, print_operand_address): Add parameters
22589 to s390_loadrelative_operand_p invokation.
22590 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22591 (s390_reload_larl_operand, s390_secondary_reload): Use
22592 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22593 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22595 2013-01-09 Mike Stump <mikestump@comcast.net>
22597 * dse.c (record_store): Remove unnecessary assert.
22599 2013-01-09 Jan Hubicka <jh@suse.cz>
22601 PR tree-optimization/55569
22602 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22603 * cfgloop.h (scale_loop_profile): Likewise.
22605 2013-01-09 Jan Hubicka <jh@suse.cz>
22608 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22610 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22612 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22614 PR middle-end/55114
22615 * expr.h (maybe_emit_group_store): Declare.
22616 * expr.c (maybe_emit_group_store): New function.
22617 * builtins.c (expand_builtin_int_roundingfn): Call it.
22618 (expand_builtin_int_roundingfn_2): Likewise.
22620 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22622 PR rtl-optimization/55829
22623 * lra-constraints.c (match_reload): Add code for absent output.
22624 (curr_insn_transform): Add code for reloads of matched inputs
22627 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22629 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22630 attribute of movddup insn to DF.
22631 (*vec_interleave_lowv2df): Ditto.
22632 (vec_dupv2df): Ditto.
22634 2013-01-09 Jan Hubicka <jh@suse.cz>
22636 PR tree-optimiation/55875
22637 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22638 EVERY_ITERATION parameter.
22639 (number_of_iterations_exit): Check if exit is executed every iteration.
22640 (idx_infer_loop_bounds): Similarly here.
22641 (n_of_executions_at_most): Simplify
22642 to only test for cases where statement is dominated by the
22643 particular bound; handle correctly the "postdominance" test.
22644 (scev_probably_wraps_p): Use max loop iterations info
22645 as a global bound first.
22647 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22648 Nick Clifton <nickc@redhat.com>
22650 * config/v850/v850.md (cbranchsf4): New pattern.
22651 (cstoresf4): New pattern.
22652 (cbranchdf4): New pattern.
22653 (cstoredf4): New pattern.
22654 (movsicc): Disallow floating point comparisons.
22655 (cmpsf_le_insn): Fix order of operators.
22656 (cmpsf_lt_insn): Likewise.
22657 (cmpsf_eq_insn): Likewise.
22658 (cmpdf_le_insn): Likewise.
22659 (cmpdf_lt_insn): Likewise.
22660 (cmpdf_eq_insn): Likewise.
22661 (cmpsf_ge_insn): Use LE comparison.
22662 (cmpdf_ge_insn): Likewise.
22663 (cmpsf_gt_insn): Use LT comparison.
22664 (cmpdf_gt_insn): Likewise.
22665 (cmpsf_ne_insn): Delete pattern.
22666 (cmpdf_ne_insn): Delete pattern.
22667 * config/v850/v850.c (v850_gen_float_compare): Use
22668 gen_cmpdf_eq_insn for NE comparison.
22669 (v850_float_z_comparison_operator)
22670 (v850_float_nz_comparison_operator): Move from here ...
22671 * config/v850/predicates.md: ... to here. Move GT and GE
22672 comparisons into v850_float_z_comparison_operator.
22673 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22675 (v850_float_nz_comparison_operator): Likewise.
22677 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22679 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22680 with calls to gen_insvsi/gen_insvdi.
22682 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22684 * config/i386/i386.c (initial_ix86_tune_features): Set up
22685 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22687 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22688 Jakub Jelinek <jakub@redhat.com>
22690 PR tree-optimization/48189
22691 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22692 If nitercst is 0, don't predict the exit edge.
22694 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22696 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22697 in asm_fprintf with reg_names.
22698 (aarch64_print_operand_address): Likewise.
22699 (aarch64_return_addr): Likewise.
22700 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22702 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22704 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22705 (INT_U6_BITS): Likewise.
22706 * config/pa/predicates.md (uint6_operand): New predicate.
22707 (shift5_operand, shift6_operand): Likewise.
22708 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22710 (lshrdi3): Use shift6_operand.
22711 (shrpsi4, shrpdi4): New insn patterns.
22712 (extzv): Delete expander.
22713 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22714 predicates in unamed zero extract patterns. Tighten common constraint.
22715 (extv): Delete expander.
22716 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22717 predicates in unamed sign extract patterns. Tighten common constraint.
22718 (insv): Delete expander.
22719 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22720 predicates in unamed insert patterns. Tighten common constraint.
22721 Change uint32_operand predicate to uint6_operand predicate in unamed
22722 DImode pattern to insert constant values of type 1...1xxxx.
22724 2013-01-04 Jan Hubicka <jh@suse.cz>
22726 PR tree-optimization/55823
22727 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22730 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22731 Uros Bizjak <ubizjak@gmail.com>
22733 PR rtl-optimization/55845
22734 * df-problems.c (can_move_insns_across): Stop scanning at
22735 volatile_insn_p source instruction or give up if
22736 across_from .. across_to range contains any volatile_insn_p
22739 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22741 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22742 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22744 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22745 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22747 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22750 * asan.c (asan_clear_shadow): New function.
22751 (asan_emit_stack_protection): Use it.
22753 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22755 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22756 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22757 with tab instead of space.
22759 2013-01-08 Nick Clifton <nickc@redhat.com>
22761 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22762 register bank 0 at the start of an interrupt handler.
22763 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22766 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22768 * config/aarch64/aarch64-simd.md
22769 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22770 (aarch64_simd_bsl): Likewise.
22771 (aarch64_vcond_internal<mode>): Likewise.
22772 (vcond<mode><mode>): Likewise.
22773 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22774 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22776 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22778 * config/aarch64/aarch64-builtins.c
22779 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22781 2013-01-08 Martin Jambor <mjambor@suse.cz>
22784 * tree-sra.c (analyze_access_subtree): Return true also after
22785 potentially creating a debug-only replacement.
22787 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22789 PR middle-end/55890
22790 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22792 PR tree-optimization/54120
22793 * tree-vrp.c (range_fits_type_p): Don't allow
22794 src_precision < precision from signed vr to unsigned_p
22795 if vr->min or vr->max is negative.
22796 (simplify_float_conversion_using_ranges): Test can_float_p
22797 against CODE_FOR_nothing.
22799 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22800 Richard Biener <rguenther@suse.de>
22802 PR middle-end/55851
22803 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22804 types instead of just INTEGER_TYPE types.
22806 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22808 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22811 2013-01-07 Steve Ellcey <sellcey@mips.com>
22814 * config/mips/mips.opt: Change mad to mmad to match documentation.
22816 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22819 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22820 .progmemx.data now.
22822 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22825 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22826 (avr_addrspace_t): Add .section_name field.
22827 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22829 (avr_addrspace): Same. Initialize .section_name. Remove last
22830 NULL entry. Put __memx into .progmemx.data.
22831 (progmem_section_prefix): Remove.
22832 (avr_asm_init_sections): No need to initialize progmem_section.
22833 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22834 section name prefix.
22835 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22836 retrieve the progmem section.
22837 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22838 boundary to run over avr_addrspace[].
22839 (avr_register_target_pragmas): Ditto.
22841 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22843 * varasm.c (output_constant_def_contents): For asan_protect_global
22844 protected strings, adjust DECL_ALIGN if needed, before testing for
22846 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22847 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22849 (output_object_block): For asan protected decls, emit asan padding
22850 after their contents.
22851 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22852 (asan_finish_file): Test it here instead.
22854 2013-01-07 Nick Clifton <nickc@redhat.com>
22855 Matthias Klose <doko@debian.org>
22856 Doug Kwan <dougkwan@google.com>
22857 H.J. Lu <hongjiu.lu@intel.com>
22860 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22862 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22864 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22866 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22868 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22870 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22873 * doc/install.texi (Cross-Compiler-Specific Options): Document
22876 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22878 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22879 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22880 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22881 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22882 vqmovun_high_s64): Fix source operand number and update copyright.
22884 2013-01-07 Richard Biener <rguenther@suse.de>
22886 PR middle-end/55890
22887 * gimple.h (gimple_call_builtin_p): New overload.
22888 * gimple.c (validate_call): New function.
22889 (gimple_call_builtin_p): Likewise.
22890 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22891 Use gimple_call_builtin_p.
22892 (find_func_clobbers): Likewise.
22893 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22894 (strlen_optimize_stmt): Likewise.
22896 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22898 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22899 (vld1q_dup_*): Likewise.
22900 (vld1_*): Likewise.
22901 (vld1q_*): Likewise.
22902 (vld1_lane_*): Likewise.
22903 (vld1q_lane_*): Likewise.
22905 2013-01-07 Richard Biener <rguenther@suse.de>
22907 * lto-streamer.h (LTO_minor_version): Bump to 2.
22909 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22911 * config/aarch64/aarch64-protos.h
22912 (aarch64_const_double_zero_rtx_p): Rename to...
22913 (aarch64_float_const_zero_rtx_p): ...this.
22914 (aarch64_float_const_representable_p): New.
22915 (aarch64_output_simd_mov_immediate): Likewise.
22916 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22917 move immediate case.
22918 * config/aarch64/aarch64.c
22919 (aarch64_const_double_zero_rtx_p): Rename to...
22920 (aarch64_float_const_zero_rtx_p): ...this.
22921 (aarch64_print_operand): Allow printing of new constants.
22922 (aarch64_valid_floating_const): New.
22923 (aarch64_legitimate_constant_p): Check for valid floating-point
22925 (aarch64_simd_valid_immediate): Likewise.
22926 (aarch64_vect_float_const_representable_p): New.
22927 (aarch64_float_const_representable_p): Likewise.
22928 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22929 (aarch64_output_simd_mov_immediate): New.
22930 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22931 (*movdf_aarch64): Likewise.
22932 * config/aarch64/constraints.md (Ufc): New.
22933 (Y): call aarch64_float_const_zero_rtx.
22934 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22936 2013-01-07 Richard Biener <rguenther@suse.de>
22938 PR tree-optimization/55888
22939 PR tree-optimization/55862
22940 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22941 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22942 not if it is contained therein.
22944 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22946 * config/avr/t-avr: Typo.
22948 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22951 * config/avr/t-avr: Don't automatically rebuild
22952 $(srcdir)/config/avr/t-multilib
22953 $(srcdir)/config/avr/avr-tables.opt
22954 $(srcdir)/doc/avr-mmcu.texi
22955 (avr-mcus): New phony target to build them on request.
22956 (s-avr-mlib, s-avr-mmcu-texi): Remove.
22957 * avr/avr-mcus.def: Adjust comments.
22959 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
22961 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
22963 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22965 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
22967 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22969 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
22971 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
22973 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
22974 to generate profiling.
22975 * config/rs6000/aix64.h (LIB_SPEC): Same.
22977 2013-01-04 Andrew Pinski <apinski@cavium.com>
22979 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
22981 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
22983 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
22985 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
22987 (ix86_expand_move): Ditto.
22988 (ix86_zero_extend_to_Pmode): Ditto.
22989 (ix86_expand_call): Ditto.
22990 (ix86_expand_special_args_builtin): Ditto.
22991 (ix86_expand_builtin): Ditto.
22993 2013-01-04 Richard Biener <rguenther@suse.de>
22995 PR tree-optimization/55862
22996 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
22997 translating them through PHI nodes.
22999 2013-01-04 Martin Jambor <mjambor@suse.cz>
23001 PR tree-optimization/55755
23002 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23003 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23005 2013-01-04 Marek Polacek <polacek@redhat.com>
23007 PR middle-end/55859
23008 * opts.c (default_options_optimization): Clarify error message.
23010 2013-01-04 Richard Biener <rguenther@suse.de>
23012 PR middle-end/55863
23013 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23016 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23019 * config/pa/pa.md (movsi): Revert previous change.
23020 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23023 2013-01-03 Richard Henderson <rth@redhat.com>
23025 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23026 after eliminating TLS symbols.
23028 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23031 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23032 * graphite-poly.c (debug_gmp_value): Likewise.
23034 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23037 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23038 selected code model, define __code_mode_small__, __code_model_medium__,
23039 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23040 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23041 xchg temporary register with %k. Declare temporary register as
23043 [__x86_64__]: For medium and large code models, preserve %rbx register.
23045 2013-01-03 Richard Biener <rguenther@suse.de>
23047 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23048 (dump_subscript): Adjust.
23049 (finalize_ddr_dependent): Do not dump redundant info.
23050 (analyze_siv_subscript): Adjust.
23051 (subscript_dependence_tester): Likewise.
23052 (compute_affine_dependence): Likewise.
23054 2013-01-03 Richard Biener <rguenther@suse.de>
23057 2013-01-03 Richard Biener <rguenther@suse.de>
23059 PR tree-optimization/55857
23060 * tree-vect-stmts.c (vectorizable_load): Do not setup
23061 re-alignment for invariant loads.
23063 2013-01-02 Richard Biener <rguenther@suse.de>
23065 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23066 invariant load do not generate a vector load from the scalar location.
23068 2013-01-03 Richard Biener <rguenther@suse.de>
23070 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23071 for not vectorizing.
23072 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23073 not build INDIRECT_REFs, call get_name once only.
23074 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23075 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23077 2013-01-03 Richard Biener <rguenther@suse.de>
23079 PR tree-optimization/55857
23080 * tree-vect-stmts.c (vectorizable_load): Do not setup
23081 re-alignment for invariant loads.
23083 2013-01-03 Richard Biener <rguenther@suse.de>
23086 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23087 prefer a built-in decl.
23089 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23091 * gcc.c (process_command): Update copyright notice dates.
23092 * gcov.c (print_version): Likewise.
23093 * gcov-dump.c (print_version): Likewise.
23095 PR rtl-optimization/55838
23096 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23097 iv0.step, iv1.step and step.
23099 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23100 Marc Glisse <marc.glisse@inria.fr>
23102 PR tree-optimization/55832
23103 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23104 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23105 integer_{one,zero}_node.
23107 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23110 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23111 * var-tracking.c (reverse_op): Don't add reverse ops to
23112 VALUEs that have already
23113 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23115 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23117 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23119 2013-01-02 Teresa Johnson <tejohnson@google.com>
23121 * dumpfile.c (dump_loc): Print filename with location.
23122 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23123 new location_t parameter to emit complete unroll message with
23124 new dump framework.
23125 (canonicalize_loop_induction_variables): Compute loops location
23126 and pass to try_unroll_loop_completely.
23127 * loop-unroll.c (report_unroll_peel): New function.
23128 (peel_loops_completely): Use new dump format with location
23129 for main dumpfile message, and invoke report_unroll_peel on success.
23130 (decide_unrolling_and_peeling): Ditto.
23131 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23132 by report_unroll_peel.
23133 (decide_peel_completely): Ditto.
23134 (decide_unroll_constant_iterations): Ditto.
23135 (decide_unroll_runtime_iterations): Ditto.
23136 (decide_peel_simple): Ditto.
23137 (decide_unroll_stupid): Ditto.
23138 * cfgloop.c (get_loop_location): New function.
23139 * cfgloop.h (get_loop_location): Declare.
23141 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23143 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23146 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23148 PR middle-end/55198
23149 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23150 BLKmode objects when EXPAND_MEMORY is specified.
23152 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23154 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23156 (fold_builtin_cpu): Do not share cpu model decls across statements.
23158 2013-01-02 Jason Merrill <jason@redhat.com>
23161 * tree.c (build_array_type_1): Revert earlier change.
23163 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23165 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23167 * config/aarch64/aarch64-tune.md: Re-generate.
23169 2013-01-02 Richard Biener <rguenther@suse.de>
23171 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23172 invariant load do not generate a vector load from the scalar location.
23174 2013-01-02 Richard Biener <rguenther@suse.de>
23177 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23178 * configure: Regenerate.
23180 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23182 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23183 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23184 (expand_builtin_int_roundingfn_2): Keep the original target around
23185 for the fallback case.
23187 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23189 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23190 to be clear for sign changes.
23192 2013-01-01 Jan Hubicka <jh@suse.cz>
23194 * ipa-inline-analysis.c: Fix formatting.
23196 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23198 PR tree-optimization/55831
23199 * tree-vect-loop.c (get_initial_def_for_induction): Use
23200 gsi_after_labels instead of gsi_start_bb.
23202 Copyright (C) 2013 Free Software Foundation, Inc.
23204 Copying and distribution of this file, with or without modification,
23205 are permitted in any medium without royalty provided the copyright
23206 notice and this notice are preserved.