Richard Biener [Mon, 9 Sep 2013 13:18:38 +0000 (13:18 +0000)]
re PR tree-optimization/58326 (ICE in check_loop_closed_ssa_use, at tree-ssa-loop-manip.c:551)
2013-09-09 Richard Biener <rguenther@suse.de>
PR middle-end/58326
* cfgloopmanip.c (fix_bb_placements): When fixing the placement
of a subloop record all its block as affecting loop-closed SSA form.
* gcc.dg/torture/pr58326-1.c: New testcase.
* gcc.dg/torture/pr58326-2.c: Likewise.
Jan Hubicka [Mon, 9 Sep 2013 07:37:53 +0000 (09:37 +0200)]
ipa-profile.c: Add toplevel comment.
* ipa-profile.c: Add toplevel comment.
(ipa_propagate_frequency_1): Be more conservative when profile is read.
(contains_hot_call_p): New function.
(ipa_propagate_frequency): Set frequencies based on counts when
profile is read.
* predict.c (compute_function_frequency): Use PROFILE_READ gueard for
profile; do not tamper with profile after inlining if it is read.
Fix handling of very long asm statements in inliner
An auto generated program with a 6.4mio line asm statement gave
with 4.7 and 4.8:
xxx.c:6400017:1: internal compiler error: in account_size_time, at
ipa-inline-analysis.c:601
The problem is that the inliner counts the number of lines in the asm
statement and multiplies that with a weight. With the weight this
overflows 32bit signed int, and triggers an assert for negative time.
Fix this by limiting the number of lines to 1000 for asm cost
estimation. The RTL backend also does similar multiplications for
jump shortening. I haven't tried to address this, but presumably
it's less likely to result in a failure.
gcc/:
2013-09-08 Andi Kleen <ak@linux.intel.com>
* tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
Fix PR 58300: Re-order events with -fvtable-verify=preinit
flag, so constructor init function is not written to assembly
file until after call to cgraph_process_new_functions.
ira.c (update_equiv_regs): Only call set_paradoxical_subreg for non-debug insns.
gcc/
* ira.c (update_equiv_regs): Only call set_paradoxical_subreg
for non-debug insns.
* lra.c (new_insn_reg): Take the containing insn as a parameter.
Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
(collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
accordingly.
Jan Hubicka [Sun, 8 Sep 2013 16:42:21 +0000 (18:42 +0200)]
devirt-11.C: Update template.
* testsuite/g++.dg/ipa/devirt-11.C: Update template.
* testsuite/g++.dg/ipa/devirt-16.C: New testcase.
* testsuite/g++.dg/ipa/devirt-17.C: New testcase.
* testsuite/g++.dg/ipa/devirt-18.C: New testcase.
* cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
targets and devirtualize to BUILT_IN_UNREACHABLE.
* timevar.def (TV_IPA_UNREACHABLE): New timevar.
* ipa.c (walk_polymorphic_call_targets): New function.
(symtab_remove_unreachable_nodes): Use it; do not keep all virtual
functions; use the new timevar.
* ipa-devirt.c (maybe_record_node): Do not insert static nodes that
was removed from the program.
(record_binfo): If BINFO corresponds to an anonymous namespace, we may
not consider it in the walk when its vtable is dead.
(possible_polymorphic_call_targets_1): Pass anonymous flag to
record_binfo.
(devirt_variable_node_removal_hook): New function.
(possible_polymorphic_call_targets): Also register
devirt_variable_node_removal_hook.
(ipa_devirt): Do not do non-speculative devirtualization.
(gate_ipa_devirt): One execute if devirtualizing speculatively.
Jan Hubicka [Sat, 7 Sep 2013 07:48:59 +0000 (09:48 +0200)]
cgraphunit.c (expand_thunk): Get body before touching arguments.
* cgraphunit.c (expand_thunk): Get body before touching arguments.
* lto-streamer-out.c: Stream thunks, too.
* lto-streamer-in.c (input_function): Pop cfun here
(lto_read_body): Instead of here.
Jan Hubicka [Fri, 6 Sep 2013 15:02:03 +0000 (17:02 +0200)]
re PR middle-end/58094 (IPA devirt testsuite errors)
PR middle-end/58094
* ipa-inline.c (has_caller_p): New function.
(want_inline_function_to_all_callers_p): Use it.
(sum_callers, inline_to_all_callers): Break out from ...
(ipa_inline): ... here.
[AArch64] Use neon_<ldm,stm>_2 where appropriate as "type".
gcc/
* config/aarch64/aarch64.md
(*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
is fpsimd_<load/store>2.
(load_pair<mode>): Likewise.
(store_pair<mode>): Likewise.
gcc/
* config/arm/types.md (type): Add "mrs" type.
* config/aarch64/aarch64.md
(aarch64_load_tp_hard): Make type "mrs".
* config/arm/arm.md
(load_tp_hard): Make type "mrs".
* config/arm/cortex-a15.md: Update with new attributes.
* config/arm/cortex-a5.md: Update with new attributes.
* config/arm/cortex-a53.md: Update with new attributes.
* config/arm/cortex-a7.md: Update with new attributes.
* config/arm/cortex-a8.md: Update with new attributes.
* config/arm/cortex-a9.md: Update with new attributes.
* config/arm/cortex-m4.md: Update with new attributes.
* config/arm/cortex-r4.md: Update with new attributes.
* config/arm/fa526.md: Update with new attributes.
* config/arm/fa606te.md: Update with new attributes.
* config/arm/fa626te.md: Update with new attributes.
* config/arm/fa726te.md: Update with new attributes.
[AArch64, ARM] Use "multiple" for type, where more than one instruction is used for a move
gcc/
* config/aarch64/aarch64.md
(*movti_aarch64): Use "multiple" for type where v8type is "move2".
(*movtf_aarch64): Likewise.
* config/arm/arm.md
(thumb1_movdi_insn): Use "multiple" for type where more than one
instruction is used for a move.
(*arm32_movhf): Likewise.
(*thumb_movdf_insn): Likewise.
gcc/
* config/arm/types.md (type): Rename fcpys to fmov.
* config/arm/vfp.md
(*arm_movsi_vfp): Rename type fcpys as fmov.
(*thumb2_movsi_vfp): Likewise
(*movhf_vfp_neon): Likewise
(*movhf_vfp): Likewise
(*movsf_vfp): Likewise
(*thumb2_movsf_vfp): Likewise
(*movsfcc_vfp): Likewise
(*thumb2_movsfcc_vfp): Likewise
* config/aarch64/aarch64-simd.md
(move_lo_quad_<mode>): Replace type mov_reg with fmovs.
* config/aarch64/aarch64.md
(*movsi_aarch64): Replace type mov_reg with fmovs.
(*movdi_aarch64): Likewise
(*movsf_aarch64): Likewise
(*movdf_aarch64): Likewise
* config/arm/arm.c
(cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
* config/arm/iwmmxt.md
(*iwmmxt_movsi_insn): Rename type fcpys as fmov.
* config/arm/arm1020e.md: Update with new attributes.
* config/arm/cortex-a15-neon.md: Update with new attributes.
* config/arm/cortex-a5.md: Update with new attributes.
* config/arm/cortex-a53.md: Update with new attributes.
* config/arm/cortex-a7.md: Update with new attributes.
* config/arm/cortex-a8-neon.md: Update with new attributes.
* config/arm/cortex-a9.md: Update with new attributes.
* config/arm/cortex-m4-fpu.md: Update with new attributes.
* config/arm/cortex-r4f.md: Update with new attributes.
* config/arm/marvell-pj4.md: Update with new attributes.
* config/arm/vfp11.md: Update with new attributes.
gcc/
* config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
* config/arm/arm.md (core_cycles): Remove fdiv.
* config/arm/vfp.md:
(*sqrtsf2_vfp): Update for attribute changes.
(*sqrtdf2_vfp): Likewise.
* config/aarch64/aarch64.md:
(sqrt<mode>2): Update for attribute changes.
* config/arm/arm1020e.md: Update with new attributes.
* config/arm/cortex-a15-neon.md: Update with new attributes.
* config/arm/cortex-a5.md: Update with new attributes.
* config/arm/cortex-a53.md: Update with new attributes.
* config/arm/cortex-a7.md: Update with new attributes.
* config/arm/cortex-a8-neon.md: Update with new attributes.
* config/arm/cortex-a9.md: Update with new attributes.
* config/arm/cortex-m4-fpu.md: Update with new attributes.
* config/arm/cortex-r4f.md: Update with new attributes.
* config/arm/marvell-pj4.md: Update with new attributes.
* config/arm/vfp11.md: Update with new attributes.
[ARM,AARCH64] Insn type reclassification. Split f_cvt type.
gcc/
* config/arm/types.md
(type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
* config/aarch64/aarch64.md
(l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
new attributes.
(fix_trunc<GPF:mode><GPI:mode>2): Likewise.
(fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
(float<GPI:mode><GPF:mode>2): Likewise.
* config/arm/vfp.md
(*truncsisf2_vfp): Update with new attributes.
(*truncsidf2_vfp): Likewise.
(fixuns_truncsfsi2): Likewise.
(fixuns_truncdfsi2): Likewise.
(*floatsisf2_vfp): Likewise.
(*floatsidf2_vfp): Likewise.
(floatunssisf2): Likewise.
(floatunssidf2): Likewise.
(*combine_vcvt_f32_<FCVTI32typename>): Likewise.
(*combine_vcvt_f64_<FCVTI32typename>): Likewise.
* config/arm/arm1020e.md: Update with new attributes.
* config/arm/cortex-a15-neon.md: Update with new attributes.
* config/arm/cortex-a5.md: Update with new attributes.
* config/arm/cortex-a53.md: Update with new attributes.
* config/arm/cortex-a7.md: Update with new attributes.
* config/arm/cortex-a8-neon.md: Update with new attributes.
* config/arm/cortex-a9.md: Update with new attributes.
* config/arm/cortex-m4-fpu.md: Update with new attributes.
* config/arm/cortex-r4f.md: Update with new attributes.
* config/arm/marvell-pj4.md: Update with new attributes.
* config/arm/vfp11.md: Update with new attributes.
gcc/
* config/arm/types.md: Add "no_insn", "multiple" and "untyped"
types.
* config/arm/arm-fixed.md: Add type attribute to all insn
patterns.
* config/arm/vfp.md: Add type attribute to all insn patterns.
* config/arm/arm.md: Add type attribute to all insn patterns.
* config/arm/thumb2.md: Add type attribute to all insn patterns.
* config/arm/arm1020e.md: Update with new attributes.
* config/arm/arm1026ejs.md: Update with new attributes.
* config/arm/arm1136jfs.md: Update with new attributes.
* config/arm/arm926ejs.md: Update with new attributes.
* config/arm/cortex-a15.md: Update with new attributes.
* config/arm/cortex-a5.md: Update with new attributes.
* config/arm/cortex-a53.md: Update with new attributes.
* config/arm/cortex-a7.md: Update with new attributes.
* config/arm/cortex-a8.md: Update with new attributes.
* config/arm/cortex-a9.md: Update with new attributes.
* config/arm/cortex-m4.md: Update with new attributes.
* config/arm/cortex-r4.md: Update with new attributes.
* config/arm/fa526.md: Update with new attributes.
* config/arm/fa606te.md: Update with new attributes.
* config/arm/fa626te.md: Update with new attributes.
* config/arm/fa726te.md: Update with new attributes.
Jan Hubicka [Fri, 6 Sep 2013 10:36:44 +0000 (10:36 +0000)]
Makefile.in (tree-sra.o): Update dependencies.
* Makefile.in (tree-sra.o): Update dependencies.
* tree-sra.c: Include ipa-utils.h
(scan_function): Use recursive_call_p.
(has_caller_p): New function.
(cgraph_for_node_and_aliases): Count also callers of aliases.
Andreas Krebbel [Fri, 6 Sep 2013 07:54:05 +0000 (07:54 +0000)]
s390.md: Add "bcr_flush" value to mnemonic attribute.
2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
* config/s390/s390.md: Add "bcr_flush" value to mnemonic
attribute.
("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
attribute to "z196_alone".
* config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
"zEC12_simple".
* gcc.c-torture/execute/20101011-1.c [__arc__] (DO_TEST): Define as 0.
* testsuite/gcc.target/arc: New directory.
* gcc.dg/torture/pr37868.c: Also skip for arc*-*-*.
* gcc.dg/stack-usage-1.c [__arc__] (SIZE): Define.
* testsuite/gcc.dg/torture/stackalign/builtin-apply-2.c
[__arc__] (STACK_ARGUMENTS_SIZE): Set to 0.
* testsuite/gcc.dg/builtin-apply2.c
[__arc__] (STACK_ARGUMENTS_SIZE): Set to 0.
Jeff Law [Thu, 5 Sep 2013 20:28:36 +0000 (14:28 -0600)]
tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed from thread_around_empty_block.
* tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
from thread_around_empty_block. Record threading path into PATH.
Recurse if threading through the initial block is successful.
(thread_across_edge): Corresponding changes to slightly simplify.
* config/aarch64/aarch64.md
(type): Remove frecpe, frecps, frecpx.
(aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
fix to be a TARGET_SIMD instruction.
(aarch64_frecps): Remove.
* config/aarch64/aarch64-simd.md
(aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
(aarch64_frecps<mode>): Handle all float/vector of float modes.
Richard Biener [Thu, 5 Sep 2013 12:45:20 +0000 (12:45 +0000)]
re PR tree-optimization/58137 ([trunk, ICE] full unroll + AVX2 vectorization)
2013-09-05 Richard Biener <rguenther@suse.de>
PR tree-optimization/58137
* tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
Do not create vectors of pointers.
* tree-vect-loop.c (get_initial_def_for_induction): Use proper
types for the components of the vector initializer.
* tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
Martin Jambor [Thu, 5 Sep 2013 12:41:16 +0000 (14:41 +0200)]
ipa-prop.c (remove_described_reference): Accept missing references, return false if that hppens, otherwise return true.
2013-09-05 Martin Jambor <mjambor@suse.cz>
* ipa-prop.c (remove_described_reference): Accept missing references,
return false if that hppens, otherwise return true.
(cgraph_node_for_jfunc): New function.
(try_decrement_rdesc_refcount): Likewise.
(try_make_edge_direct_simple_call): Use them.
(ipa_edge_removal_hook): Remove references from rdescs.
(ipa_edge_duplication_hook): Clone rdescs and their references
when the new edge has the same caller as the old one.
* cgraph.c (cgraph_resolve_speculation): Remove speculative
reference before removing any edges.
testsuite/
* g++.dg/ipa/remref-1.C: New test.
* g++.dg/ipa/remref-2.C: Likewise.
re PR tree-optimization/57370 (compiler hangs in reassoc)
2013-09-04 Easwaran Raman <eraman@google.com>
PR middle-end/57370
PR tree-optimization/58011
* tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
(build_and_add_sum): Use it.
(appears_later_in_bb): Simplify code.
Jan Hubicka [Wed, 4 Sep 2013 14:14:28 +0000 (16:14 +0200)]
Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
* Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
* ipa-devirt.c: Include diganostic.h
(odr_type_d): Add types and types_set.
(hash_type_name): Work for types with vtables during LTO.
(odr_hasher::remove): Fix comment; destroy types_set.
(add_type_duplicate): New function,
(get_odr_type): Use it.
(dump_type_inheritance_graph): Dump type duplicates.
* ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
graph.
* tree.c (types_same_for_odr): Give exact answers on types with
virtual tables.
c-pretty-print.h (c_pretty_printer::simple_type_specifier): Now a virtual member function.
c-family/
* c-pretty-print.h (c_pretty_printer::simple_type_specifier): Now
a virtual member function.
(pp_simple_type_specifier): Remove.
(pp_c_type_specifier): Likewise.
* c-pretty-print.c (c_pretty_printer::simple_type_specifier):
Rename from pp_c_type_specifier. Adjust.
(c_pretty_printer::c_pretty_printer): Do not assign to
simple_type_specifier.
cp/
* cxx-pretty-print.h (cxx_pretty_printer::simple_type_specifier):
Declare as overrider.
* cxx-pretty-print.c (cxx_pretty_printer::simple_type_specifier):
Rename from pp_cxx_simple_type_specifier.
(cxx_pretty_printer::cxx_pretty_printer): Do not assign to
simple_type_specifier.