Arnaud Charlet [Thu, 10 Oct 2013 11:07:30 +0000 (13:07 +0200)]
[multiple changes]
2013-10-10 Robert Dewar <dewar@adacore.com>
* exp_ch3.adb (Expand_N_Variant_Part): Expand statically
predicated subtype which appears in Discrete_Choices list.
* exp_ch5.adb (Expand_N_Case_Statement): Expand statically
predicated subtype which appears in Discrete_Choices list of
case statement alternative.
* exp_util.ads, exp_util.adb (Expand_Static_Predicates_In_Choices): New
procedure.
* sem_case.adb: Minor reformatting (Analyze_Choices): Don't
expand out Discrete_Choices that are names of subtypes with
static predicates. This is now done in the analyzer so that the
-gnatct tree is properly formed for ASIS.
* sem_case.ads (Generic_Choices_Processing): Does not apply
to aggregates any more, so change doc accordingly, and remove
unneeded Get_Choices argument.
* sem_ch3.adb (Analyze_Variant_Part): Remove no
longer used Get_Choices argument in instantiation of
Generic_Choices_Processing.
* sem_ch4.adb (Analyze_Case_Expression): Remove no
longer used Get_Choices argument in instantiation of
Generic_Choices_Processing.
* sem_ch5.adb (Analyze_Case_Statement): Remove no
longer used Get_Choices argument in instantiation of
Generic_Choices_Processing.
* sinfo.ads: For N_Variant_Part, and N_Case_Statement_Alternative,
document that choices that are names of statically predicated
subtypes are expanded in the code generation tree passed to the
back end, but not in the ASIS tree generated for -gnatct.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch7.adb: Revert previous change.
2013-10-10 Gary Dismukes <dismukes@adacore.com>
* sem_ch13.adb (Analyze_Attribute_Definition_Clause): In the case where
the Storage_Pool aspect is specified by an aspect clause and a
renaming is used to capture the evaluation of the pool name,
insert the renaming in front of the aspect's associated entity
declaration rather than in front of the corresponding attribute
definition (which hasn't been appended to the declaration
list yet).
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Is_Interface_Conformant): The controlling type
of the interface operation is obtained from the ultimate alias
of the interface primitive parameter, because that may be in
fact an implicit inherited operation whose signature involves
the type extension and not the desired interface.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* par-ch13.adb (Aspect_Specifications_Present): In Ada 2012,
recognize an aspect specification with a misspelled name if it
is followed by a a comma or semicolon.
Arnaud Charlet [Thu, 10 Oct 2013 11:01:42 +0000 (13:01 +0200)]
[multiple changes]
2013-10-10 Robert Dewar <dewar@adacore.com>
* gnat_rm.texi: Minor fix.
2013-10-10 Robert Dewar <dewar@adacore.com>
* sem_ch13.adb (Analyze_Attribute_Definition_Clause, case
Address): Remove the Comes_From_Source test for the overlap
warning.
2013-10-10 Robert Dewar <dewar@adacore.com>
* sem_util.adb: Minor code reorganization (use Nkind_In).
* sem_warn.adb: Minor code reorganization (optimization in
Check_Unset_Reference).
* exp_ch9.adb, exp_ch4.adb, sinfo.ads: Minor reformatting.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch7.adb (Install_Parent_Private_Declarations): When
instantiating a child unit, do not install private declaration of
a non-generic ancestor of the generic that is also an ancestor
of the current unit: its private part will be installed when
private part of ancestor itself is analyzed.
2013-10-10 Thomas Quinot <quinot@adacore.com>
* freeze.adb (Check_Component_Storage_Order): Retrieve component
aliased status from type entities directly instead of going back
to original component definition.
* sem_ch7.adb: Minor reformatting.
Arnaud Charlet [Thu, 10 Oct 2013 10:59:13 +0000 (12:59 +0200)]
[multiple changes]
2013-10-10 Robert Dewar <dewar@adacore.com>
* sem_ch13.adb (Analyze_Aspect_Specifications): For Address
attribute, consider it to be set in source, because of aliasing
considerations.
(Analyze_Attribute_Definition_Clause): For the
purpose of warning on overlays, take into account the aspect case.
2013-10-10 Robert Dewar <dewar@adacore.com>
* a-cfdlli.ads, a-cfhase.ads, a-cforma.ads, a-cfhama.ads, a-cforse.ads,
a-cofove.ads: Minor reformatting.
2013-10-10 Arnaud Charlet <charlet@adacore.com>
* gnat_ugn.texi: Remove obsolete mention to -laddr2line.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* exp_ch4.adb (Expand_N_Case_Expression): Indicate that the
generated variable used as a target of the expression needs
no initialization.
2013-10-10 Jose Ruiz <ruiz@adacore.com>
* exp_util.adb (Corresponding_Runtime_Package): Remove the condition
related to No_Dynamic_Attachment which was wrong. Protected types
with interrupt handlers (when not using a restricted profile)
are always treated as protected types with entries, regardless
of the No_Dynamic_Attachment restriction.
* exp_ch9.adb (Expand_N_Protected_Type_Declaration): Simplify the code
using the result of Corresponding_Runtime_Package.
(Install_Private_Data_Declarations): When having
static handlers and a non restricted profile, we use the
type Static_Interrupt_Protection always, so we removed an
extra wrong condition looking at the No_Dynamic_Attachment
restriction. Simplify the code using the result of
Corresponding_Runtime_Package.
(Make_Initialize_Protection): Simplify the code using
the result of Corresponding_Runtime_Package.
(Install_Private_Data_Declaration): The No_Dynamic_Attachment
restriction has nothing to do with static handlers. Remove the extra
erroneous condition that was creating the wrong data type.
Arnaud Charlet [Thu, 10 Oct 2013 10:55:36 +0000 (12:55 +0200)]
[multiple changes]
2013-10-10 Yannick Moy <moy@adacore.com>
* gnat_rm.texi, a-cfdlli.ads, a-cfhase.ads, a-cforma.ads, a-cfhama.ads,
a-cforse.ads, a-cofove.ads: Update comment and doc of formal containers
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Analyze_Aspect_Specifications): For Pre/Post
conditions that apply to a subprogram body, preserve the placement
and order of the generated pragmas, which must appear before
other declarations in the body.
Arnaud Charlet [Thu, 10 Oct 2013 10:53:30 +0000 (12:53 +0200)]
[multiple changes]
2013-10-10 Bob Duff <duff@adacore.com>
* gnat_ugn.texi: Add gnat2xml doc.
2013-10-10 Doug Rupp <rupp@adacore.com>
* s-vxwork-arm.ads: Fix interface to FP_CONTEXT.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Analyze_Aspect_Specification): An aspect Import
on a variable need not have a convention specified, as long as
the implicit convention of the object, obtained from its type,
is Ada or Ada-related.
2013-10-10 Robert Dewar <dewar@adacore.com>
* cstand.adb (Standard_Unsigned_64): New internal type.
* gnat_rm.texi: Update documentation on To_Address.
* sem_attr.adb (Analyze_Attribute, case To_Address): Fix
problem with out of range static values given as literals or
named numbers.
* stand.ads (Standard_Unsigned_64): New internal type.
* stand.adb: Minor reformatting.
Arnaud Charlet [Thu, 10 Oct 2013 10:50:15 +0000 (12:50 +0200)]
[multiple changes]
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Analyze_Selected_Component,
Has_Mode_Conformant_Spec): If selected component may be an
indexing of a parameterless call to a protected function, and
expansion is disabled, this is a valid candidate interpretation.
Arnaud Charlet [Thu, 10 Oct 2013 10:47:59 +0000 (12:47 +0200)]
[multiple changes]
2013-10-10 Robert Dewar <dewar@adacore.com>
* gnatlink.adb: Minor reformatting.
2013-10-10 Yannick Moy <moy@adacore.com>
* debug.adb: Free flag d.E and change doc for flag d.K.
2013-10-10 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Check_Precondition_Postcondition): If the
pragma comes from an aspect spec, and the subprogram is a
library unit, treat as a ppc in a declarative part in ASIS mode,
so that expression in aspect is properly analyzed. In this case
there is no later point at which the aspect specification would
be examined.
2013-10-10 Bob Duff <duff@adacore.com>
* opt.ads: Minor comment fix.
2013-10-10 Vadim Godunko <godunko@adacore.com>
* a-coinho-shared.ads, a-coinho-shared.adb: New file.
* s-atocou.ads: Add procedure to initialize counter.
* s-atocou.adb: Likewise.
* s-atocou-builtin.adb: Likewise.
* s-atocou-x86.adb: Likewise.
* gcc-interface/Makefile.in: Select special version of
Indefinite_Holders package on platforms where atomic built-ins
are supported. Update tools target pairs for PikeOS.
Zhenqiang Chen [Wed, 9 Oct 2013 18:57:14 +0000 (18:57 +0000)]
tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
* tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
(operand_equal_for_value_replacement): New function, extracted from
value_replacement and enhanced to catch more cases.
(value_replacement): Use operand_equal_for_value_replacement.
Marek Polacek [Wed, 9 Oct 2013 14:51:28 +0000 (14:51 +0000)]
re PR c++/58635 ([c++11] ICE with __transaction_atomic and noexcept(false))
PR c++/58635
cp/
* semantics.c (finish_return_stmt): Return error_mark_node
when error_operand_p of the expr is true.
(build_transaction_expr): Check for EXPR_P before setting the
expr location.
testsuite/
* g++.dg/tm/pr58635-1.C: New test.
* g++.dg/tm/pr58635-2.C: New test.
Andrew MacLeod [Wed, 9 Oct 2013 13:42:17 +0000 (13:42 +0000)]
tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
* tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
(struct int_tree_map): Move to tree-hasher.h
(SCALE, LABEL, PERCENT): Move to gimple.h
* tree-flow-inline.h: Delete. Move functions to other files.
(unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
* gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
(gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
* gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
(SCALE, LABEL, PERCENT): Relocate from tree-flow.h
* tree-hasher.h: Don't include tree-flow.h.
(struct int_tree_map): Relocate from tree-flow.h.
* tree-sra.c (contains_view_convert_expr_p): Relocate from
tree-flow-inline.h and make static.
* tree-ssa-alias.h (ranges_overlap_p): Relocate from tree-flow-inline.h.
* tree-ssa-operands.c (gimple_ssa_operands): Relocate from
tree-flow-inline.h and make static.
* tree.h (is_global_var, may_be_aliased): Relocate from
tree-flow-inline.h.
* Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
* value-prof.c: No longer include tree-flow-inline.h.
* tree-switch-conversion.c: No longer include tree-flow-inline.h.
Andreas Krebbel [Wed, 9 Oct 2013 07:24:35 +0000 (07:24 +0000)]
s390.c (s390_register_info_stdarg_fpr): Remove packed stack special handling.
2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
* config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
packed stack special handling.
(s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
back to fixed stack slots for FPRs saved due to stdarg.
Andreas Krebbel [Wed, 9 Oct 2013 07:22:52 +0000 (07:22 +0000)]
s390.c (struct s390_frame_layout): New field gpr_save_slots.
2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
* config/s390/s390.c (struct s390_frame_layout): New field
gpr_save_slots.
(cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
(s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
regs_ever_clobbered to char*.
(s390_regs_ever_clobbered): Check crtl->saves_all_registers
instead of cfun->has_nonlocal_label. Ignore frame related restore
INSNs.
(s390_register_info): Enable FPR save slots. Move/Copy some
functionality into ...
(s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
(s390_register_info_stdarg_gpr, s390_optimize_register_info): New
function.
(s390_frame_info): Do gpr slot allocation here now. stdarg does
not imply a stack frame.
(s390_init_frame_layout): Remove variable clobbered_regs.
(s390_update_register_info): Remove function.
(s390_hard_regno_rename_ok): Call-saved regs without a save slot
cannot be used for register renaming.
(s390_hard_regno_scratch_ok): New function.
(TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
(s390_initial_elimination_offset): Change offset calculation of
the return address pointer.
(save_gprs): Deal with only r6 being saved from the call-saved
regs.
(restore_gprs): Set frame related flag.
(s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New
functions.
(s390_emit_prologue): Call s390_register_info instead of
s390_update_frame_layout. Call s390_save_gprs_to_fprs.
(s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
(s390_optimize_prologue): Call s390_optimize_register_info.
Try to remove also FPR slot save/restore INSNs. Remove frame
related flags from restore INSNs.
DJ Delorie [Wed, 9 Oct 2013 01:05:07 +0000 (21:05 -0400)]
rl78.c (rl78_print_operand_1): Change %c to %C to avoid conflict with the MI use of %c.
* config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
avoid conflict with the MI use of %c.
* config/rl78/rl78-real.md: change %c to %C throughout.
* config/rl78/rl78-virt.md: Likewise.
Jonathan Wakely [Tue, 8 Oct 2013 12:33:37 +0000 (12:33 +0000)]
re PR libstdc++/58659 (Construction of shared_ptr from unique_ptr mismatches new/delete and std::allocator for __shared_ptr_count)
PR libstdc++/58659
* include/bits/shared_ptr_base.h (__shared_count::__shared_count(P,D)):
Delegate to constructor taking allocator.
(__shared_count::_S_create_from_up): Inline into ...
(__shared_count::__shared_count(unique_ptr<Y,D>&&): Here. Use
std::conditional instead of constrained overloads. Allocate memory
using the allocator type that will be used for deallocation.
* testsuite/20_util/shared_ptr/cons/58659.cc: New.
* testsuite/20_util/shared_ptr/cons/43820_neg.cc: Adjust.
Andrew MacLeod [Tue, 8 Oct 2013 12:13:13 +0000 (12:13 +0000)]
tree-flow.h: Remove some prototypes.
* tree-flow.h: Remove some prototypes.
* tree.h: Remove some protypes, add a couple.
* tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
using_eh_for_cleanups_p): Add interface routines for front ends.
* tree-eh.h: New file. Add protoptyes.
* tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
(add_stmt_to_eh_lp_fn): Make static.
(lower_try_finally): Use new using_eh_for_cleanups_p.
* emit-rtl.c: Include tree-eh.h.
* gimple.h: Include tree-eh.h.
Tim Shen [Tue, 8 Oct 2013 03:41:14 +0000 (03:41 +0000)]
regex_executor.h: Add _TodoList class.
2013-10-08 Tim Shen <timshen91@gmail.com>
* include/bits/regex_executor.h: Add _TodoList class.
* include/bits/regex_executor.tcc (_BFSExecutor<>::_M_main): Add
_M_match_stack and _M_stack to make everything faster. Break if
_M_stack is empty, to reduce unnecessary idling.
* testsuite/performance/28_regex/split.cc: New.
Andreas Krebbel [Mon, 7 Oct 2013 07:50:25 +0000 (07:50 +0000)]
2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
* config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
constraint letters from expanders.
("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
retry count to general_operand.
("tabort"): Give operand 0 a mode.
("tabort_1"): Add mode and constraint letter for operand 0.
* doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
* gcc.target/s390/htm-1.c: Add more tests to cover different
operand types.
Paolo Carlini [Sun, 6 Oct 2013 21:41:19 +0000 (21:41 +0000)]
re PR c++/58126 (No diagnostic when inheriting an uninitialized const or reference member)
/cp
2013-10-06 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/58126
* class.c (check_bases): Propagate CLASSTYPE_READONLY_FIELDS_NEED_INIT
and CLASSTYPE_REF_FIELDS_NEED_INIT from bases to derived.
* init.c (diagnose_uninitialized_cst_or_ref_member_1): Extend error
messages about uninitialized const and references members to mention
the base class.
/testsuite
2013-10-06 Paolo Carlini <paolo.carlini@oracle.com>
Paolo Carlini [Fri, 4 Oct 2013 23:17:51 +0000 (23:17 +0000)]
re PR c++/58503 ([c++11] ICE with invalid range in range-based for-loop)
/cp
2013-10-04 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/58503
* parser.c (cp_parser_perform_range_for_lookup): If eventually
either *begin or *end is type-dependent, return NULL_TREE.
(do_range_for_auto_deduction): If cp_parser_perform_range_for_lookup
returns NULL_TREE, don't actually do_auto_deduction.
/testsuite
2013-10-04 Paolo Carlini <paolo.carlini@oracle.com>
Paolo Carlini [Thu, 3 Oct 2013 22:50:59 +0000 (22:50 +0000)]
re PR c++/58584 ([c++11] ICE with invalid argument for alignas)
/cp
2013-10-04 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/58584
* decl2.c (save_template_attributes): Handle error_mark_node as
*attr_p argument.
(cp_check_const_attributes): Likewise for attributes.
* parser.c (cp_parser_std_attribute_spec): When alignas_expr is an
error_mark_node call cp_parser_skip_to_end_of_statement.
/testsuite
2013-10-04 Paolo Carlini <paolo.carlini@oracle.com>
Jan Hubicka [Thu, 3 Oct 2013 17:27:32 +0000 (19:27 +0200)]
i386.c (ix86_issue_rate): Pentium4, Nocona has issue rate of 2.
* i386.c (ix86_issue_rate): Pentium4, Nocona has issue rate of 2.
Core2, Corei7 and Haswell has issue rate of 4.
(ix86_adjust_cost): Remove ATOM case; fix core2/corei7/Haswell case.
Wei Mi [Thu, 3 Oct 2013 17:18:26 +0000 (17:18 +0000)]
lra-constraints.c (insert_move_for_subreg): New function extracted from simplify_operand_subreg.
2013-10-03 Wei Mi <wmi@google.com>
* lra-constraints.c (insert_move_for_subreg): New function
extracted from simplify_operand_subreg.
(simplify_operand_subreg): Add reload for paradoxical subreg.
Rong Xu [Thu, 3 Oct 2013 17:15:56 +0000 (17:15 +0000)]
ipa-inline-analysis.c (find_foldable_builtin_expect): Find the candidate of builtin_expect such that we should fix the size/time...
* ipa-inline-analysis.c (find_foldable_builtin_expect): Find
the candidate of builtin_expect such that we should fix the
size/time estimation.
(estimate_function_body_sizes): Do the acutally size/time fix-up
for builtin_expect.
Rong Xu [Thu, 3 Oct 2013 17:13:50 +0000 (17:13 +0000)]
predict.c (tree_predict_by_opcode): Get the probability for builtin_expect from param builtin_expect_probability.
* predict.c (tree_predict_by_opcode): Get the probability
for builtin_expect from param builtin_expect_probability.
* params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
* predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
* doc/invoke.texi: Add documentation for builtin-expect-probability.
* gcc.target/i386/cold-attribute-2.c: Fix the test by using original
probability.
* gcc.dg/tree-ssa/ipa-split-5.c: Ditto.
* gcc.dg/tree-ssa/ipa-split-6.c: Ditto.
--This li (t)ene, and those below, will be ignored--
M gcc/params.def
M gcc/predict.def
M gcc/ChangeLog
M gcc/testsuite/gcc.dg/tree-ssa/ipa-split-5.c
M gcc/testsuite/gcc.dg/tree-ssa/ipa-split-6.c
M gcc/testsuite/gcc.target/i386/cold-attribute-2.c
M gcc/predict.c
M gcc/doc/invoke.texi
Marc Glisse [Thu, 3 Oct 2013 16:16:40 +0000 (18:16 +0200)]
del_op.cc (operator delete): Don't test for 0 before free.
2013-10-03 Marc Glisse <marc.glisse@inria.fr>
* libsupc++/del_op.cc (operator delete): Don't test for 0 before free.
* libsupc++/del_opnt.cc (free): Only declare if freestanding.
(operator delete): Qualify free with std::.
rs6000-builtin.def (XSRDPIM): Use floatdf2, ceildf2, btruncdf2, instead of vsx_* name.
[gcc]
2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
ceildf2, btruncdf2, instead of vsx_* name.
* config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
iterators to only do V2DF and V4SF here. Move the DF code to
rs6000.md where it is combined with SF mode. Replace <VSv> with
just 'v' since only vector operations are handled with these insns
after moving the DF support to rs6000.md.
(vsx_sub<mode>3): Likewise.
(vsx_mul<mode>3): Likewise.
(vsx_div<mode>3): Likewise.
(vsx_fre<mode>2): Likewise.
(vsx_neg<mode>2): Likewise.
(vsx_abs<mode>2): Likewise.
(vsx_nabs<mode>2): Likewise.
(vsx_smax<mode>3): Likewise.
(vsx_smin<mode>3): Likewise.
(vsx_sqrt<mode>2): Likewise.
(vsx_rsqrte<mode>2): Likewise.
(vsx_fms<mode>4): Likewise.
(vsx_nfma<mode>4): Likewise.
(vsx_copysign<mode>3): Likewise.
(vsx_btrunc<mode>2): Likewise.
(vsx_floor<mode>2): Likewise.
(vsx_ceil<mode>2): Likewise.
(vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
(vsx_sminsf3): Likewise.
(vsx_fmadf4): Likewise.
(vsx_fmsdf4): Likewise.
(vsx_nfmadf4): Likewise.
(vsx_nfmsdf4): Likewise.
(vsx_cmpdf_internal1): Likewise.
* config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
simpler to select whether a target has SPE or traditional floating
point support in iterators.
(TARGET_DF_SPE): Likewise.
(TARGET_SF_FPR): Likewise.
(TARGET_DF_FPR): Likewise.
(TARGET_SF_INSN): Macros to say whether floating point support
exists for a given operation for expanders.
(TARGET_DF_INSN): Likewise.
* config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
combining of SF/DF mode operations, using both traditional and VSX
registers.
(Fvsx): Likewise.
(Ff): Likewise.
(Fv): Likewise.
(Fs): Likewise.
(Ffre): Likewise.
(FFRE): Likewise.
(abs<mode>2): Combine SF/DF modes using traditional floating point
instructions. Add support for using the upper DF registers with
VSX support, and SF registers with power8-vector support. Update
expanders for operations supported by both the SPE and traditional
floating point units.
(abs<mode>2_fpr): Likewise.
(nabs<mode>2): Likewise.
(nabs<mode>2_fpr): Likewise.
(neg<mode>2): Likewise.
(neg<mode>2_fpr): Likewise.
(add<mode>3): Likewise.
(add<mode>3_fpr): Likewise.
(sub<mode>3): Likewise.
(sub<mode>3_fpr): Likewise.
(mul<mode>3): Likewise.
(mul<mode>3_fpr): Likewise.
(div<mode>3): Likewise.
(div<mode>3_fpr): Likewise.
(sqrt<mode>3): Likewise.
(sqrt<mode>3_fpr): Likewise.
(fre<Fs>): Likewise.
(rsqrt<mode>2): Likewise.
(cmp<mode>_fpr): Likewise.
(smax<mode>3): Likewise.
(smin<mode>3): Likewise.
(smax<mode>3_vsx): Likewise.
(smin<mode>3_vsx): Likewise.
(negsf2): Delete SF operations that are merged with DF.
(abssf2): Likewise.
(addsf3): Likewise.
(subsf3): Likewise.
(mulsf3): Likewise.
(divsf3): Likewise.
(fres): Likewise.
(fmasf4_fpr): Likewise.
(fmssf4_fpr): Likewise.
(nfmasf4_fpr): Likewise.
(nfmssf4_fpr): Likewise.
(sqrtsf2): Likewise.
(rsqrtsf_internal1): Likewise.
(smaxsf3): Likewise.
(sminsf3): Likewise.
(cmpsf_internal1): Likewise.
(copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
(negdf2): Delete DF operations that are merged with SF.
(absdf2): Likewise.
(nabsdf2): Likewise.
(adddf3): Likewise.
(subdf3): Likewise.
(muldf3): Likewise.
(divdf3): Likewise.
(fred): Likewise.
(rsqrtdf_internal1): Likewise.
(fmadf4_fpr): Likewise.
(fmsdf4_fpr): Likewise.
(nfmadf4_fpr): Likewise.
(nfmsdf4_fpr): Likewise.
(sqrtdf2): Likewise.
(smaxdf3): Likewise.
(smindf3): Likewise.
(cmpdf_internal1): Likewise.
(lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
(btrunc<mode>2): Delete separate expander, and combine with the
insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
(btrunc<mode>2_fpr): Likewise.
(ceil<mode>2): Likewise.
(ceil<mode>2_fpr): Likewise.
(floor<mode>2): Likewise.
(floor<mode>2_fpr): Likewise.
(fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
Add support for using the upper registers with VSX and
power8-vector. Move insns to be closer to the define_expands. On
VSX systems, prefer the traditional form of FMA over the VSX
version, since the traditional form allows the target not to
overlap with the inputs.
(fms<mode>4_fpr): Likewise.
(nfma<mode>4_fpr): Likewise.
(nfms<mode>4_fpr): Likewise.
[gcc/testsuite]
2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
* gcc.target/powerpc/p8vector-fp.c: New test for floating point
scalar operations when using -mupper-regs-sf and -mupper-regs-df.
* gcc.target/powerpc/ppc-target-1.c: Update tests to allow either
VSX scalar operations or the traditional floating point form of
the instruction.
* gcc.target/powerpc/ppc-target-2.c: Likewise.
* gcc.target/powerpc/recip-3.c: Likewise.
* gcc.target/powerpc/recip-5.c: Likewise.
* gcc.target/powerpc/pr72747.c: Likewise.
* gcc.target/powerpc/vsx-builtin-3.c: Likewise.
PR58460, the add and sub shifted register instruction forms in AArch64
do not permit the stack register. This patch removes k constraint
from the relevant patterns and adds reduced form of the test case.
lra-constraints.c (process_alt_operand): Calculate scratch_p and use it.
2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
* lra-constraints.c (process_alt_operand): Calculate scratch_p and
use it. Use smaller increase for scratch. Don't increase reject
for early clobber scratch.
* lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
setting eliminated regs except setting fp from hfp.
(lra_eliminate): Check lra_insn_recog_data on NULL.
re PR target/58587 (Bootstrap error with ada with -mcpu=power7)
2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/58587
* config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
setting -mvsx-timode by default until the underlying problem is
fixed.
(RS6000_CPU, power7 defaults): Likewise.