2013-05-10 Richard Biener <rguenther@suse.de>
* tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
disable peeling when we version for aliasing.
(vector_alignment_reachable_p): Honor explicit user alignment.
(vect_supportable_dr_alignment): Likewise.
* tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
* tree-vect-loop.c (vect_transform_loop): First apply versioning,
then peeling to arrange for the cost-model check to come first.
* gcc.target/i386/avx256-unaligned-load-2.c: Make well-defined.
* gcc.target/i386/l_fma_double_1.c: Adjust.
* gcc.target/i386/l_fma_double_2.c: Likewise.
* gcc.target/i386/l_fma_double_3.c: Likewise.
* gcc.target/i386/l_fma_double_4.c: Likewise.
* gcc.target/i386/l_fma_double_5.c: Likewise.
* gcc.target/i386/l_fma_double_6.c: Likewise.
* gcc.target/i386/l_fma_float_1.c: Likewise.
* gcc.target/i386/l_fma_float_2.c: Likewise.
* gcc.target/i386/l_fma_float_3.c: Likewise.
* gcc.target/i386/l_fma_float_4.c: Likewise.
* gcc.target/i386/l_fma_float_5.c: Likewise.
* gcc.target/i386/l_fma_float_6.c: Likewise.
From-SVN: r198767
* config/epiphany/epiphany.c (epiphany_expand_prologue):
When using gen_stack_adjust_str with a register offset, add a
REG_FRAME_RELATED_EXPR note.
From-SVN: r198753
* decl.c (grokdeclarator): Complain about reference, pointer, or
typedef to VLA.
(create_array_type_for_decl): Complain about array of VLA.
* pt.c (tsubst): Likewise.
* rtti.c (get_tinfo_decl): Talk about "array of runtime bound".
* semantics.c (finish_decltype_type): Complain about decltype of VLA.
* typeck.c (cp_build_addr_expr_1): Complain about VLA.
(cxx_sizeof_or_alignof_type): Likewise.
From-SVN: r198746
Richard Biener <rguenther@suse.de>
PR lto/54095
* symtab.c (symtab_make_decl_local): Do not add private names.
* lto.c (lto_register_var_decl_in_symtab): Don't do renaming.
(lto_register_var_decl_in_symtab): Likewise.
(lto_main): Promote statics.
* lto-partition.c (privatize_symbol_name): New function.
(promote_symbol): Use it.
(may_need_named_section_p): New predicate.
(rename_statics): New functions.
(lto_promote_cross_file_statics): Simplify; do renaming.
(lto_promote_statics_nonwpa): New function.
* lto-partition.h (lto_promote_statics_nonwpa): New function.
Co-Authored-By: Richard Biener <rguenther@suse.de>
From-SVN: r198741
When a non constant memory model is passed to __atomic_*
gcc falls back to seq_cst. This drops any HLE acquire or release bits.
This can happen when <atomic> is used with -O0
as the member functions are not always inlined then and the memory
argument passed in ends up being non-constant.
v2: Use _GLIBCXX_ALWAYS_INLINE
libstdc++-v3/:
2013-05-08 Andi Kleen <ak@linux.intel.com>
PR target/55947
* libstdc++-v3/include/bits/atomic_base.h
(_GLIBCXX_ALWAYS_INLINE): Add new macro.
(atomic_thread_fence, atomic_signal_fence, test_and_set,
clear, store, load, exchange, compare_exchange_weak)
compare_exchange_strong, fetch_add, fetch_sub, fetch_and,
fetch_or, fetch_xor): Mark _GLIBCXX_ALWAYS_INLINE.
From-SVN: r198733
in array new size calculation.
libstdc++-v3/
* libsupc++/new: Add std::bad_array_new_length.
* libsupc++/bad_array_new.cc: New.
* libsupc++/eh_aux_runtime.cc: Add __cxa_throw_bad_array_new_length.
* libsupc++/Makefile.in: Build them.
* config/abi/pre/gnu.ver: Add new symbols.
* config/abi/pre/gnu-versioned-namespace.ver: Add new symbols.
gcc/cp/
* init.c (throw_bad_array_new_length): New.
(build_new_1): Use it. Don't warn about braced-init-list.
(build_vec_init): Use it.
* call.c (build_operator_new_call): Use it.
From-SVN: r198731
2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
* gimple-ssa-strength-reduction.c (count_candidates): Change
return value to int.
(analyze_candidates_and_replace): Change type of length to int.
From-SVN: r198715
2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
PR tree-optimization/57200
* tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
Only call inform if the preceding warning_at returns true.
From-SVN: r198704
2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
* ira.c (update_equiv_regs): Add insn having equiv memory even if
it is not lhs of the insn.
(setup_reg_equiv): Remove insn having equiv memory which it is not
lhs of the insn.
* lra-constraints.c (process_address): Try to improve generation
code for address base + disp.
(lra_constraints): Make correct the code for checking insn setting
up backward equivalence. Remove insn only if it is in the init
insn list.
* lra-eliminations.c (update_reg_eliminate): Change return value.
(lra_eliminate): Use the result.
From-SVN: r198695
* config/i386/sse.md (ssescalarnummask): New mode attribute.
(PEXTR_MODE, PEXTR_MODEx): New mode iterators.
(*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
*sse4_1_pextrw_memory. Handle register target operands.
(*vec_extractv8hi_sse2): New pattern.
(*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
(*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
(*vec_extract<mode>_mem): New insn and split pattern.
From-SVN: r198691
2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
* gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
(incr_vec_index): Return -1 if increment not found.
(create_add_on_incoming_edge): Assert if increment not found.
(record_increment): Limit number of increments recorded.
(all_phi_incrs_profitable): Return false if an increment not found.
(replace_profitable_candidates): Don't process increments that were
not recorded.
(analyze_candidates_and_replace): Limit size of incr_vec.
From-SVN: r198682
2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
* gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
phi def as possibly hiding a basis for a CAND_ADD whose operands
have been commuted in the analysis.
(alloc_cand_and_find_basis): Add parms to call to find_phi_def.
From-SVN: r198678
2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
* config/aarch64/aarch64.md
(cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
shift value between 0-4.
From-SVN: r198677
2013-05-07 Richard Biener <rguenther@suse.de>
* double-int.h (rshift): New overload.
* double-int.c (rshift): New function.
* tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
(create_reference_ops_from_ref): Remove.
(vn_reference_insert): Use shared ops for constructing the
reference and copy it.
From-SVN: r198676
