Alan Modra [Tue, 31 Jul 2012 22:14:44 +0000 (07:44 +0930)]
re PR target/54131 (ICE building 416.gamess, reload_cse_simplify_operands)
PR target/54131
* config/rs6000/rs6000.c (mem_operand_gpr): Don't limit range
of lo_sum offsets. Comment. Assert mode at least word size
rather than bypassing powerpc64 word offset check.
Bill Schmidt [Tue, 31 Jul 2012 21:08:41 +0000 (21:08 +0000)]
rs6000.c (rs6000_builtin_vectorization_cost): Revise costs for vec_perm and vec_promote_demote down to more natural values.
2012-07-31 Bill Schmidt <wschmidt@linux.ibm.com>
* config/rs6000/rs6000.c (rs6000_builtin_vectorization_cost): Revise
costs for vec_perm and vec_promote_demote down to more natural values.
(struct _rs6000_cost_data): New data structure.
(rs6000_density_test): New function.
(rs6000_init_cost): Change to use rs6000_cost_data.
(rs6000_add_stmt_cost): Likewise.
(rs6000_finish_cost): Perform density test when vectorizing a loop.
Janus Weil [Tue, 31 Jul 2012 18:32:41 +0000 (20:32 +0200)]
re PR fortran/42418 (PROCEDURE: Rejects interface which is both specific and generic procedure)
2012-07-31 Janus Weil <janus@gcc.gnu.org>
PR fortran/42418
* decl.c (match_procedure_interface): Move some checks to
'resolve_procedure_interface'. Set flavor if appropriate.
* expr.c (gfc_check_pointer_assign): Cleanup of 'gfc_is_intrinsic'.
* intrinsic.c (gfc_is_intrinsic): Additional checks for attributes which
identify a procedure as being non-intrinsic.
* resolve.c (resolve_procedure_interface): Checks moved here from
'match_procedure_interface'. Minor cleanup.
(resolve_formal_arglist,resolve_symbol): Cleanup of
'resolve_procedure_interface'
(resolve_actual_arglist,is_external_proc): Cleanup of
'gfc_is_intrinsic'.
Steven Bosscher [Tue, 31 Jul 2012 18:24:40 +0000 (18:24 +0000)]
sched-vis.c (print_exp): Handle FMA explicitly.
* sched-vis.c (print_exp): Handle FMA explicitly. Simplify handling
of RTX_UNARY function-like operations. Print unknown expression
RTX classes as pseudo-functions.
(print_pattern): Simplify return pattern printing.
(dump_insn_slim): Print the value of notes as a pattern for proper
printing of CFA notes.
Steven Bosscher [Tue, 31 Jul 2012 11:48:56 +0000 (11:48 +0000)]
sched-vis.c (dump_insn_slim): Print print_rtx_head at the start of each new line.
* sched-vis.c (dump_insn_slim): Print print_rtx_head at the
start of each new line.
* final.c (final_scan_insn): If the final dump is requested as
slim RTL, dump slim RTL to the assembly file also.
Steven Bosscher [Tue, 31 Jul 2012 09:20:56 +0000 (09:20 +0000)]
re PR pch/53880 (compile time regression when generating precompiled headers for boost)
PR pch/53880
* gengtype.c (struct walk_type_data): Add have_this_obj field.
(walk_type): For functions that take a this_obj argument and
that process fields with a GTY((length)) argument, write the
test that write_types_local_process_field will write also at the
head of the loop, effectively unswitching the loop.
(write_func_for_structure, write_local_func_for_structure): Clear
have_this_obj before calling walk_type.
(write_local_func_for_structure): Set have_this_obj before walk_type.
(write_array): Set have_this_obj for output of local pointer walking
functions but not for marker functions.
(write_types_local_process_field): Assert have_this_obj is set.
* rtl.h (simplify_using_condition): Adjust prototype using bitmap
from coretypes.h.
Nick Clifton [Tue, 31 Jul 2012 09:09:29 +0000 (09:09 +0000)]
lib2funcs.c (__clrsbhi2): New function.
* config/stormy16/lib2funcs.c (__clrsbhi2): New function.
Implements __clrsb for an HImode argument.
* config/stormy16/clrsbhi2.c: New file:
* config/stormy16/t-stormy16 (LIB2ADD): Add clrsbhi2.c.
* expmed.h (NUM_MODE_VECTOR_INT): Define.
(struct expmed_op_cheap, struct expmed_op_costs): New structures.
(struct target_expmed): Convert x_mul_highpart_cost and
x_mul_widen_cost fields to be indexed by integer modes.
Convert x_sdiv_pow2_cheap and x_smod_pow2_cheap fields to be
of type struct expmed_op_cheap. Convert other cost fields to be
of type struct_expmed_op_costs.
(mul_widen_cost_ptr, mul_highpart_cost_ptr): Adjust for new
indexing of respective fields.
(expmed_op_cheap_ptr): New function.
(sdiv_pow2_cheap_ptr, smod_pow2_cheap_ptr): Call it.
(expmed_op_cost_ptr): New function.
(add_cost_ptr, neg_cost_ptr, shift_cost_ptr, shiftadd_cost_ptr,
shiftsub0_cost_ptr, shiftsub1_cost_ptr, mul_cost_ptr,
sdiv_cost_ptr, udiv_cost_ptr): Call it.
Janus Weil [Mon, 30 Jul 2012 19:55:41 +0000 (21:55 +0200)]
re PR fortran/51081 ([F03] Proc-pointer assignment: Rejects valid internal proc)
2012-07-30 Janus Weil <janus@gcc.gnu.org>
PR fortran/51081
* gfortran.h (gfc_resolve_intrinsic): Add prototype.
* expr.c (gfc_check_pointer_assign): Set INTRINSIC attribute if needed.
Check for invalid intrinsics.
* primary.c (gfc_match_rvalue): Check for intrinsics came too early.
Set procedure flavor if appropriate.
* resolve.c (resolve_intrinsic): Renamed to gfc_resolve_intrinsic.
(resolve_procedure_interface,resolve_procedure_expression,
resolve_function,resolve_fl_derived0,resolve_symbol): Ditto.
* gnatcmd.adb, makeutl.adb, makeutl.ads: Minor code reorganization.
2012-07-30 Vincent Pucci <pucci@adacore.com>
* exp_ch9.adb (Build_Lock_Free_Unprotected_Subprogram_Body): Minor
reformatting.
* sem_ch9.adb (Allows_Lock_Free_Implementation): Minor reformatting.
Capture the correct error message in case of a quantified expression.
2012-07-30 Thomas Quinot <quinot@adacore.com>
* g-socket.adb (Get_Socket_Option, Set_Socket_Option): On Windows, the
value is a milliseconds count in a DWORD, not a struct timeval.
* sem_ch12.adb (Earlier): Add local variables T1 and T2. Minor code
refactoring.
2012-07-30 Thomas Quinot <quinot@adacore.com>
* gnatcmd.adb, make.adb, makeutl.adb, makeutl.ads
(Test_If_Relative_Path): Rename to Ensure_Absolute_Path to better
reflect what this subprogram does. Rename argument Including_L_Switch
to For_Gnatbind, and also exempt -A from rewriting.
* bindusg.adb: Document optional =file argument to gnatbind -A.
2012-07-30 Ed Schonberg <schonberg@adacore.com>
* freeze.adb (Freeze_Entity): Do no apply restriction check on
storage pools to access to subprogram types.
* target.def (vector_alignment): New target hook.
* doc/tm.texi.in (TARGET_VECTOR_ALIGNMENT): Document new hook.
* doc/tm.texi: Regenerate.
* targhooks.c (default_vector_alignment): New function.
* targhooks.h (default_vector_alignment): Add prototype.
* stor-layout.c (layout_type): Use targetm.vector_alignment.
* config/arm/arm.c (arm_vector_alignment): New function.
(TARGET_VECTOR_ALIGNMENT): Define.
* tree-vect-data-refs.c (vect_update_misalignment_for_peel): Use
vector type alignment instead of size.
* tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Use
element type size directly instead of computing it from alignment.
Fix variable naming and comment.
testsuite/ChangeLog:
* lib/target-supports.exp
(check_effective_target_vect_natural_alignment): New function.
* gcc.dg/align-2.c: Only run on targets with natural alignment
of vector types.
* gcc.dg/vect/slp-25.c: Adjust tests for targets without natural
alignment of vector types.
Co-Authored-By: Richard Earnshaw <rearnsha@arm.com>
From-SVN: r189974
Richard Guenther [Mon, 30 Jul 2012 11:18:15 +0000 (11:18 +0000)]
tree-into-ssa.c (mark_def_sites): Also process virtual operands.
2012-07-30 Richard Guenther <rguenther@suse.de>
* tree-into-ssa.c (mark_def_sites): Also process virtual operands.
(rewrite_stmt): Likewise.
(rewrite_enter_block): Likewise.
(pass_build_ssa): Do not update virtual SSA form during TODO.
(mark_symbol_for_renaming): Do nothing if we are not in SSA form.
* lto-streamer-in.c (lto_read_body): Set in_ssa_p earlier.
gengtype.c (adjust_field_type): Diagnose duplicate "length" option applications and option being applied to...
gcc:
2012-07-27 Laurynas Biveinis <laurynas.biveinis@gmail.com>
Steven Bosscher <steven@gcc.gnu.org>
* gengtype.c (adjust_field_type): Diagnose duplicate "length"
option applications and option being applied to arrays of atomic
types.
(walk_type): Allow "atomic" option on strings too.
* dwarf2out.h (struct dw_vec_struct): Use the "atomic" GTY option
for the array field.
* vec.h: Describe the atomic object "A" type of the macros in
the header comment.
(VEC_T_GTY_ATOMIC, DEF_VEC_A, DEF_VEC_ALLOC_A): Define.
* emit-rtl.c (locations_locators_vals): use the atomic object
vector.
* doc/gty.texi: Clarify that GTY option "length" is only for
arrays of non-atomic objects. Fix typo in the description of the
"atomic" option.
François Dumont [Sun, 29 Jul 2012 16:44:18 +0000 (16:44 +0000)]
re PR libstdc++/54075 ([4.7.1] unordered_map insert still slower than 4.6.2)
2012-07-29 François Dumont <fdumont@gcc.gnu.org>
PR libstdc++/54075
* include/bits/hashtable_policy.h
(_Prime_rehash_policy::_M_next_bkt): Add a growth factor set to 2
to boost growth in the number of buckets.
* testsuite/performance/23_containers/insert/unordered_set.cc: New.
* config/i386/i386.c (ix86_avoid_lea_for_addr): Handle
zero-extended addresses. Return false if the address has less
than two components.
(ix86_split_lea_for_addr): Unconditionally convert target and
all address operands to requested mode.
* config/i386/i386.md (*lea<mode>): Recover operands from curr_insn.
Pass SImode to ix86_split_lea_for_addr when splitting zero-extended
address.
(zero-extended add splitter): New splitter to conditionally split
non-destructive adds.
(*zero_extendsidi2_rex64): Conditionally emit leal instead of movl.
Alan Modra [Sat, 28 Jul 2012 00:04:10 +0000 (09:34 +0930)]
re PR target/54093 (ICE in in extract_insn, at recog.c:2129)
PR target/54093
* config/rs6000/rs6000.c (rs6000_secondary_reload): Limit 32-bit
multi-gpr reload to cases where predicate passes. Do the same for
64-bit multi-gpr reload.
PR target/54089
* config/sh/sh.c (shiftcosts): Remove case where first operand
is a const_int. Move COSTS_N_INSNS usage into caller ...
(sh_rtx_costs) ... here. Return false when shiftcosts cannot be
calculated instead of MAX_COST.
Richard Guenther [Fri, 27 Jul 2012 12:31:49 +0000 (12:31 +0000)]
tree-into-ssa.c (def_blocks_p): New typedef.
2012-07-27 Richard Guenther <rguenther@suse.de>
* tree-into-ssa.c (def_blocks_p): New typedef.
(insert_phi_nodes_compare_def_blocks): New function.
(insert_phi_nodes): Do not walk over referenced vars, instead
walk over recorded def_blocks, record relevant ones and sort
them to avoid repeated hashtable lookups.
Nick Clifton [Thu, 26 Jul 2012 14:23:51 +0000 (14:23 +0000)]
mn10300.c (REG_SAVE_BYTES): Delete.
* config/mn10300/mn10300.c (REG_SAVE_BYTES): Delete.
(mn10300_get_live_callee_saved_regs): If requested return a count
of the number of bytes in the mask.
(mn10300_expand_prologue): Add argument to invocation of
mn10300_get_live_callee_regs.
(mn10300_expand_epilogue): Compute reg_save_bytes by calling
mn10300_get_live_callee_saved_regs.
(mn10300_initial_offset): Likewise.
* config/mn10300/mn10300-protos.h (mn10300_get_live_callee_saved_regs):
Update prototype.
* config/mn10300/mn10300.md (return_ret): Add argument to
invocation of mn10300_get_live_callee_saved_regs.
Steven Bosscher [Thu, 26 Jul 2012 12:02:54 +0000 (12:02 +0000)]
bitmap.h: Add explanation of sparse set as linked-list bitmap.
* bitmap.h: Add explanation of sparse set as linked-list bitmap.
* sbitmap.h: Add explanation about non-sparse sets as simple bitmap.
(TEST_BIT): Make a static inline function for stronger type checking.
(SET_BIT): Don't handle sbitmaps with popcount.
(RESET_BIT): Likewise.
(SET_BIT_WITH_POPCOUNT): New, like SET_BIT but with popcount.
(RESET_BIT_WITH_POPCOUNT): New, like RESET_BIT but with popcount.
* ebitmap.c (ebitmap_clear_bit): Use SET_BIT_WITH_POPCOUNT and
RESET_BIT_WITH_POPCOUNT on wordmask bitmaps.
(ebitmap_set_bit, ebitmap_and_into, ebitmap_and, ebitmap_ior_into,
ebitmap_and_compl_into, ebitmap_and_compl): Likewise.
* sparseset.h: Add explanation of sparse set representation.
Mikael Morin [Thu, 26 Jul 2012 08:53:56 +0000 (08:53 +0000)]
re PR fortran/44354 (implied do loop with its own variable name as upper bound)
fortran/
PR fortran/44354
* trans-array.c (gfc_trans_array_constructor_value):
Evaluate the iteration bounds before the inner variable shadows
the outer.
testsuite/
PR fortran/44354
* gfortran.dg/array_constructor_39.f90: New test.
Mikael Morin [Thu, 26 Jul 2012 08:47:33 +0000 (08:47 +0000)]
re PR fortran/44354 (implied do loop with its own variable name as upper bound)
fortran/
PR fortran/44354
* array.c (sought_symbol): New variable.
(expr_is_sought_symbol_ref, find_symbol_in_expr): New functions.
(resolve_array_list): Check for references to the induction
variable in the iteration bounds and issue a diagnostic if some
are found.
testsuite/
PR fortran/44354
* gfortran.dg/array_constructor_38.f90: New test.
re PR target/51244 ([SH] Inefficient conditional branch and code around T bit)
PR target/51244
* config/sh/sh.opt (mzdcbranch): New option.
* doc/invoke.texi: Document it.
* config/sh/sh.md (negsi_cond): Use TARGET_ZDCBRANCH as condition
instead of TARGET_HARD_SH4.
* config/sh/sh.c (sh_option_override): Set TARGET_ZDCBRANCH as default
for TARGET_HARD_SH4.
Eric Botcazou [Wed, 25 Jul 2012 19:57:51 +0000 (19:57 +0000)]
gimple-low.c (lower_try_catch): New function.
* gimple-low.c (lower_try_catch): New function.
(lower_stmt) <GIMPLE_TRY>: Use it to lower GIMPLE_TRY_CATCH.
<GIMPLE_CATCH>: Delete.
<GIMPLE_EH_FILTER>: Likewise.
Eric Botcazou [Wed, 25 Jul 2012 19:43:15 +0000 (19:43 +0000)]
expr.c (expand_expr_real_1): Do not expand operand #1 and #2 of BIT_FIELD_REF for ignored results.
* expr.c (expand_expr_real_1): Do not expand operand #1 and #2
of BIT_FIELD_REF for ignored results.
* fold-const.c (fold_ternary_loc) <BIT_FIELD_REF>: Check that the
CONSTRUCTOR has vector type before using vector accessors on it.
* gimplify.c (gimplify_compound_lval): Do not gimplify operand #1 and
#2 of BIT_FIELD_REF.
(gimplify_expr) <BIT_FIELD_REF>: Likewise.
* tree-nested.c (convert_nonlocal_reference_op) <BIT_FIELD_REF>: Do
not process operand #1 and #2.
(convert_local_reference_op): Likewise.
* tree-ssa-pre.c (create_component_ref_by_pieces_1): Remove superfluous
breaks throughout.
<BIT_FIELD_REF>: Reuse operand #1 and #2 directly.
* tree.c (stabilize_reference) <BIT_FIELD_REF>: Do not stabilize
operand #1 and #2.
(UPDATE_FLAGS): Do not consider operand #2 of BIT_FIELD_REF.
(build6_stat): Delete.
* tree.h (build6_stat, build6, build6_stat_loc, build6_loc): Likewise.
ada/
* gcc-interface/utils2.c (gnat_stabilize_reference) <BIT_FIELD_REF>: Do
not stabilize operand #1 and #2.
François Dumont [Wed, 25 Jul 2012 19:32:48 +0000 (19:32 +0000)]
re PR libstdc++/54075 ([4.7.1] unordered_map insert still slower than 4.6.2)
2012-07-25 François Dumont <fdumont@gcc.gnu.org>
PR libstdc++/54075
* include/bits/hashtable.h
(_Hashtable<>::_Hashtable(_InputIterator, _InputIterator,
size_type, ...): Remove std::max usage to guarantee that hashtable
state is consistent with hash policy state.
(_Hashtable<>::rehash): Likewise. Set _M_prev_resize to 0 to avoid
the hashtable to be shrinking on next insertion.
* testsuite/23_containers/unordered_set/modifiers/reserve.cc: New.
* testsuite/23_containers/unordered_multiset/modifiers/reserve.cc: New.
* testsuite/23_containers/unordered_map/modifiers/reserve.cc: New.
* testsuite/23_containers/unordered_multimap/modifiers/reserve.cc: New.
* final.c [ASSEMBLER_DIALECT](do_assembler_dialects): New
function to implement assembler dialects.
(output_asm_insn): Use do_assembler_dialects.
(asm_fprintf): Likewise.
re PR bootstrap/54092 (Bootstrap fails while building Ada at stage 1)
PR bootstrap/54092
* libfuncs.h: Don't include optabs.h.
(struct libfunc_entry): Use "int" for op, mode1, mode2 members.
* optabs.c (hash_libfunc): Don't cast members to int.
* Makefile.in (LIBFUNCS_H): Don't include OPTABS_H.
projects / thirdparty / gcc.git / log
