Bill Schmidt [Tue, 22 Oct 2013 17:31:17 +0000 (17:31 +0000)]
rs6000.c (altivec_expand_vec_perm_const): Reverse meaning of merge-high and merge-low masks for little endian...
gcc:
2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
* config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
meaning of merge-high and merge-low masks for little endian; avoid
use of vector-pack masks for little endian for mismatched modes.
gcc/testsuite:
2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
* gcc.target/powerpc/altivec-perm-1.c: Move the two vector pack
tests into...
* gcc.target/powerpc/altivec-perm-3.c: ...this new test, which is
restricted to big-endian targets.
Paul Thomas [Tue, 22 Oct 2013 04:40:57 +0000 (04:40 +0000)]
PR fortran 57893
2013-10-22 Paul Thomas <pault@gcc.gnu.org>
PR fortran 57893
* class.c : Include target-memory.h.
(gfc_find_intrinsic_vtab) Build a minimal expression so that
gfc_element_size can be used to obtain the storage size, rather
that the kind value.
2013-10-22 Paul Thomas <pault@gcc.gnu.org>
PR fortran 57893
* gfortran.dg/unlimited_polymorphic_13.f90 : New test.
Diego Novillo [Mon, 21 Oct 2013 19:36:37 +0000 (15:36 -0400)]
Re-factor inclusion of tree.h.
This moves tree.h out of every header. This exposes dependencies of
tree.h in files that should probably not need it after tree and gimple
are separated.
After this change, no header should include tree.h directly. It should
only be included by a .c file. Unfortunately, I did not find an
automatic way of forcing this.
Tested on x86_64 with all languages enabled and using
contrib/config-list.mk.
Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
and get_inner_reference returning different pmode for non-volatile
bit-field members dependent on flag_strict_volatile_bitfields.
* stor-layout.c (layout_decl): Remove special handling of
flag_strict_volatile_bitfields.
* expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
Jeff Law [Mon, 21 Oct 2013 15:25:09 +0000 (09:25 -0600)]
tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
* tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
Remove VISISTED as a local variable. When we have a threadable jump, verify
the destination of the jump has not been visised.
(thread_across_edge): Allocate VISITED bitmap once at function scope and
use it throughout. Make sure to set appropriate bits in VISITED for E (start
of jump thread path).
* tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading through
a joiner if any edge on the path has a recorded jump thread.
Jan Hubicka [Sun, 20 Oct 2013 11:18:12 +0000 (13:18 +0200)]
x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM, [...]): Keep enabled only for K8 and AMDFAM10.
* config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
(X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
Uros Bizjak [Sat, 19 Oct 2013 12:58:20 +0000 (14:58 +0200)]
re PR target/58792 (ICE at mode-switching.c:421 when compiling clang lib/AST/MicrosoftCXXABI.cpp)
PR target/58792
* config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
and SI_REG for 64bit SYSV ABI targets.
Uros Bizjak [Sat, 19 Oct 2013 12:32:25 +0000 (14:32 +0200)]
mode-switching.c (create_pre_exit): Rename maybe_builtin_apply to multi_reg_return.
* mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
to multi_reg_return. Clarify that we are skipping USEs of multiple
return registers. Use bool type where appropriate.
Eric Botcazou [Sat, 19 Oct 2013 11:11:03 +0000 (11:11 +0000)]
utils.c (scale_by_factor_of): New function.
* gcc-interface/utils.c (scale_by_factor_of): New function.
(rest_of_record_type_compilation): Use scale_by_factor_of in order to
scale the original offset for both rounding cases; in the second case,
take into accout the addend to compute the alignment. Tidy up.
Teresa Johnson [Fri, 18 Oct 2013 17:25:44 +0000 (17:25 +0000)]
predict.c (probably_never_executed): Compare frequency-based count to number of training runs.
2013-10-18 Teresa Johnson <tejohnson@google.com>
* predict.c (probably_never_executed): Compare frequency-based
count to number of training runs.
* params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
Tim Shen [Fri, 18 Oct 2013 16:13:07 +0000 (16:13 +0000)]
regex_scanner.tcc: (_Scanner<>::_M_scan_normal...
2013-10-18 Tim Shen <timshen91@gmail.com>
* include/bits/regex_scanner.tcc: (_Scanner<>::_M_scan_normal,
_Scanner<>::_M_eat_escape_ecma, _Scanner<>::_M_eat_escape_posix,
_Scanner<>::_M_eat_escape_awk): Narrow character before finding in maps.
* testsuite/28_regex/algorithms/regex_match/ecma/wchar_t/cjk_match.cc:
New.
Richard Biener [Fri, 18 Oct 2013 08:36:28 +0000 (08:36 +0000)]
stor-layout.c (layout_type): Do not change TYPE_PRECISION or TYPE_UNSIGNED of integral types.
2013-10-18 Richard Biener <rguenther@suse.de>
* stor-layout.c (layout_type): Do not change TYPE_PRECISION
or TYPE_UNSIGNED of integral types.
(set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
NULL_TREE for zero-precision integral types.
Michael Meissner [Thu, 17 Oct 2013 21:20:46 +0000 (21:20 +0000)]
p8vector-fp.c: New test for floating point scalar operations when...
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.
Charles Baylis [Thu, 17 Oct 2013 20:57:21 +0000 (22:57 +0200)]
builtin-apply2.c: Skip test on arm hardfloat ABI targets.
2013-10-17 Charles Bayis <charles.baylis@linaro.org>
* gcc.dg/builtin-apply2.c: Skip test on arm hardfloat ABI targets.
* gcc.dg/tls/pr42894.c: Remove dg-options for arm*-*-* targets.
* gcc.target/arm/thumb-ltu.c: Remove dg-skip-if and require
effective target arm_thumb1_ok.
* lib/target-supports.exp
(check_effective_target_arm_fp16_ok_nocache): Don't force
-mfloat-abi=soft when building for hardfloat target.
Michael Meissner [Thu, 17 Oct 2013 19:12:57 +0000 (19:12 +0000)]
rs6000.c (enum rs6000_reload_reg_type): Add new fields to the reg_addr array that describes the valid addressing mode...
2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
fields to the reg_addr array that describes the valid addressing
mode for any register, general purpose registers, floating point
registers, and Altivec registers.
(FIRST_RELOAD_REG_CLASS): Likewise.
(LAST_RELOAD_REG_CLASS): Likewise.
(struct reload_reg_map_type): Likewise.
(reload_reg_map_type): Likewise.
(RELOAD_REG_VALID): Likewise.
(RELOAD_REG_MULTIPLE): Likewise.
(RELOAD_REG_INDEXED): Likewise.
(RELOAD_REG_OFFSET): Likewise.
(RELOAD_REG_PRE_INCDEC): Likewise.
(RELOAD_REG_PRE_MODIFY): Likewise.
(reg_addr): Likewise.
(mode_supports_pre_incdec_p): New helper functions to say whether
a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
(mode_supports_pre_modify_p): Likewise.
(rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
print the valid address mode bits for each mode.
(rs6000_debug_print_mode): Likewise.
(rs6000_debug_reg_global): Likewise.
(rs6000_setup_reg_addr_masks): New function to set up the address
mask bits for each type.
(rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
Call rs6000_setup_reg_addr_masks to set up the address mask bits.
(rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
PRE_MODIFY are supported.
(rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
registers, instead of {src,dest}_av_p.
(rs6000_print_options_internal): Tweak the debug output slightly.
Michael Meissner [Thu, 17 Oct 2013 19:04:37 +0000 (19:04 +0000)]
rs6000.c (enum rs6000_reload_reg_type): Add new fields to the reg_addr array that describes the valid addressing mode...
2013-10-07 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
fields to the reg_addr array that describes the valid addressing
mode for any register, general purpose registers, floating point
registers, and Altivec registers.
(FIRST_RELOAD_REG_CLASS): Likewise.
(LAST_RELOAD_REG_CLASS): Likewise.
(struct reload_reg_map_type): Likewise.
(reload_reg_map_type): Likewise.
(RELOAD_REG_VALID): Likewise.
(RELOAD_REG_MULTIPLE): Likewise.
(RELOAD_REG_INDEXED): Likewise.
(RELOAD_REG_OFFSET): Likewise.
(RELOAD_REG_PRE_INCDEC): Likewise.
(RELOAD_REG_PRE_MODIFY): Likewise.
(reg_addr): Likewise.
(mode_supports_pre_incdec_p): New helper functions to say whether
a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
(mode_supports_pre_modify_p): Likewise.
(rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
print the valid address mode bits for each mode.
(rs6000_debug_print_mode): Likewise.
(rs6000_debug_reg_global): Likewise.
(rs6000_setup_reg_addr_masks): New function to set up the address
mask bits for each type.
(rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
Call rs6000_setup_reg_addr_masks to set up the address mask bits.
(rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
PRE_MODIFY are supported.
(rs6000_print_options_internal): Tweak the debug output slightly.
Andrew MacLeod [Thu, 17 Oct 2013 17:41:07 +0000 (17:41 +0000)]
tree-flow.h (struct omp_region): Move to omp-low.c.
* tree-flow.h (struct omp_region): Move to omp-low.c.
Remove omp_ prototypes and variables.
* gimple.h (omp_reduction_init): Move prototype to omp-low.h.
(copy_var_decl): Relocate prototype from tree-flow.h.
* gimple.c (copy_var_decl): Relocate from omp-low.c.
* tree.h: Move prototype to omp-low.h.
* omp-low.h: New File. Relocate prototypes here.
* omp-low.c (struct omp_region): Make local here.
(root_omp_region): Make static.
(copy_var_decl) Move to gimple.c.
(new_omp_region): Make static.
(make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
* tree-cfg.c: Include omp-low.h.
(make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
* gimplify.c: Include omp-low.h.
* tree-parloops.c: Likewise.
c
* c-parser.c: Include omp-low.h.
* c-typeck.c: Likewise.
cp
* parser.c: Include omp-low.h.
* semantics.c: Likewise.
Michael Meissner [Thu, 17 Oct 2013 17:07:49 +0000 (17:07 +0000)]
re PR target/58673 (ICE in final_scan_insn for movti_ppc64 with base+offset address)
[gcc]
2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/58673
* config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
restrict TImode addresses to single indirect registers if both
-mquad-memory and -mvsx-timode are used.
(rs6000_output_move_128bit): Use quad_load_store_p to determine if
we should emit load/store quad. Remove using %y for quad memory
addresses.
* config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
constraints to allow load/store quad on machines where TImode is
not allowed in VSX registers. Use 'n' instead of 'F' constraint
for TImode to load integer constants.
[gcc/testsuite]
2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/58673
* gcc.target/powerpc/pr58673-1.c: New file to test whether
-mquad-word + -mno-vsx-timode causes errors.
* gcc.target/powerpc/pr58673-2.c: Likewise.
Michael Meissner [Thu, 17 Oct 2013 17:06:24 +0000 (17:06 +0000)]
re PR target/58673 (ICE in final_scan_insn for movti_ppc64 with base+offset address)
[gcc]
2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/58673
* config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
restrict TImode addresses to single indirect registers if both
-mquad-memory and -mvsx-timode are used.
(rs6000_output_move_128bit): Use quad_load_store_p to determine if
we should emit load/store quad. Remove using %y for quad memory
addresses.
* config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
constraints to allow load/store quad on machines where TImode is
not allowed in VSX registers. Use 'n' instead of 'F' constraint
for TImode to load integer constants.
[gcc/testsuite]
2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/58673
* gcc.target/powerpc/pr58673-1.c: New file to test whether
-mquad-word + -mno-vsx-timode causes errors.
* gcc.target/powerpc/pr58673-2.c: Likewise.
* sem_prag.adb (Check_Dependency_Clause):
Recognize the scenario where successful clause matching has
depleted the available refinement items and the clause to match
technically refines to null => null.
2013-10-17 Tristan Gingold <gingold@adacore.com>
* exp_prag.adb (Expand_Pragma_Import_Or_Interface): Specify
External_Name instead of Link_Name for the RTTI declaration.
Arnaud Charlet [Thu, 17 Oct 2013 14:07:57 +0000 (16:07 +0200)]
[multiple changes]
2013-10-17 Yannick Moy <moy@adacore.com>
* sem_res.adb (Resolve_Short_Circuit): Only
generate expression-with-action when full expansion is set.
2013-10-17 Yannick Moy <moy@adacore.com>
* debug.adb Remove obsolete comment.
2013-10-17 Thomas Quinot <quinot@adacore.com>
* exp_ch4.adb (Process_Transient_Object.Find_Enclosing_Contexts):
Avoid late insertion when expanding an expression with action
nested within a transient block; Do not inconditionally generate
a finalization call if the generated object is from a specific
branch of a conditional expression.
2013-10-17 Pascal Obry <obry@adacore.com>
* g-arrspl.adb: Ensure Finalize call is idempotent.
* g-arrspl.adb (Finalize): Makes the call idempotent.
Arnaud Charlet [Thu, 17 Oct 2013 14:06:56 +0000 (16:06 +0200)]
[multiple changes]
2013-10-17 Robert Dewar <dewar@adacore.com>
* sem_prag.adb (Record_Possible_Body_Reference): Fix test for
being in body.
(Add_Constituent): Merged into Check_Refined_Global_Item.
(Check_Matching_Constituent): A constituent that has the proper Part_Of
option and comes from a private child or a sibling is now collected.
(Check_Matching_Modes): Merged into Check_Refined_Global_Item.
(Check_Refined_Global_Item): Code cleanup.
(Collect_Constituent): New routine.
(Inconsistent_Mode_Error): Moved out from Check_Matching_Modes.
2013-10-17 Ed Schonberg <schonberg@adacore.com>
* freeze.adb (Check_Current_Instance, Process): Add RM reference
and mention immutably limited types, when the current instance
is illegal in Ada 2012.
2013-10-17 Ed Schonberg <schonberg@adacore.com>
* sem_warn.adb (Check_Unused_Withs): If the main unit is a
subunit, apply the check to the units mentioned in its context
only. This provides additional warnings on with_clauses that
are superfluous.
* sem_ch3.adb (Analyze_Declarations): Emit an
error message concerning state refinement when the spec defines at
least one non-null abstract state and the body's SPARK mode is On.
(Requires_State_Refinement): New routine.
2013-10-17 Robert Dewar <dewar@adacore.com>
* sem_ch7.ads: Comment fixes.
2013-10-17 Robert Dewar <dewar@adacore.com>
* sem_ch7.adb (Analyze_Package_Specification): Remove circuit
for ensuring that a package spec requires a body for some other
reason than that it contains the declaration of an abstract state.
2013-10-17 Tristan Gingold <gingold@adacore.com>
* exp_ch11.adb (Expand_N_Raise_Expression): Fix call of
Possible_Local_Raise.
2013-10-17 Thomas Quinot <quinot@adacore.com>
* exp_pakd.adb (Expand_Bit_Packed_Element_Set): Unchecked
conversion of Or_Rhs to Etype of New_Rhs is required only when
the latter is the result of a byte swap operation.
2013-10-17 Thomas Quinot <quinot@adacore.com>
* exp_dist.adb (Build_To_Any_Function): For a type with opaque
representation that is not transmitted as an unconstrained value,
use 'Write, not 'Output, to generate the opaque representation.
2013-10-17 Yannick Moy <moy@adacore.com>
* sem_res.adb (Resolve_Short_Circuit): Only
generate expression-with-action when full expansion is set.
2013-10-17 Yannick Moy <moy@adacore.com>
* debug.adb Remove obsolete comment.
2013-10-17 Thomas Quinot <quinot@adacore.com>
* exp_ch4.adb (Process_Transient_Object.Find_Enclosing_Contexts):
Avoid late insertion when expanding an expression with action
nested within a transient block; Do not inconditionally generate
a finalization call if the generated object is from a specific
branch of a conditional expression.
2013-10-17 Pascal Obry <obry@adacore.com>
* g-arrspl.adb: Ensure Finalize call is idempotent.
* g-arrspl.adb (Finalize): Makes the call idempotent.
* aspects.adb, aspects.ads, sem_prag.ads: Remove all entries
for Refined_Pre from the various tables.
* par-prag.adb: Remove the entry for Refined_Pre from the list
of pragmas not needing special processing by the parser.
* sem_ch13.adb (Analyze_Aspect_Specifications):
Remove the processing for aspect Refined_Pre.
(Check_Aspect_At_Freeze_Point): Remove the entry for aspect
Refined_Pre.
* sem_prag.adb (Analyze_Pragma): Refined_Pre is no longer a
valid assertion kind. Remove the analysis of pragma Refined_Pre.
(Analyze_Refined_Pragma): Update the comment on usage.
(Find_Related_Subprogram_Or_Body): Update the comment on
usage. Pragma Refined_Pre is no longer processed by this routine.
(Is_Valid_Assertion_Kind): Refined_Pre is no longer a valid
assertion kind.
* snames.ads-tmpl: Remove predefined name Refined_Pre. Remove
the pragma id for Refined_Pre.
* exp_util.adb, exp_util.ads (Entity_Of): Moved to Sem_Util.
* sem_prag.adb (Analyze_Global_In_Decl_List): Mark a null
item list as being analyzed.
(Analyze_Global_List): Mark a
null global list and multiple global items as being analyzed.
(Analyze_Input_Item): Check the unit that defines the input
variable or state, not the reference to it.
* sem_util.ads, sem_util.adb (Entity_Of): Moved from Exp_Util. Ensure
that the input has an entity.
Thomas Quinot [Thu, 17 Oct 2013 13:58:39 +0000 (13:58 +0000)]
exp_util.adb (Get_Current_Value_Condition, [...]): Handle the case of expressions with actions * exp_util.adb (Insert_Actions):...
2013-10-17 Thomas Quinot <quinot@adacore.com>
* exp_util.adb (Get_Current_Value_Condition,
Set_Current_Value_Condition): Handle the case of expressions
with actions * exp_util.adb (Insert_Actions): Handle the case
of an expression with actions whose Actions list is empty.
* exp_util.adb (Remove_Side_Effects.Side_Effect_Free): An
expression with actions that has no Actions and whose Expression
is side effect free is itself side effect free.
* exp_util.adb (Remove_Side_Effects): Do not set an incorrect etype on
temporary 'R' (Def_Id), which is in general an access to Exp_Type, not
an Exp_Type.
* sem_res.adb (Resolve): For an expression with
actions, resolve the expression early. * sem_res.adb
(Resolve_Expression_With_Actions): Rewrite an expression with
actions whose value is compile time known and which has no
actions into just its expression, so that its constant value is
available downstream.
* sem_res.adb (Resolve_Short_Circuit):
Wrap the left operand in an expression with actions to contain
any required finalization actions.
* exp_ch4.adb (Expand_Expression_With_Actions): For an
expression with actions returning a Boolean expression, ensure
any finalization action is kept within the Actions list.
* sem_warn.adb (Check_References, Check_Unset_Reference): add
missing circuitry to handle expressions with actions.
* checks.adb (Ensure_Valid): For an expression with actions,
insert the validity check on the Expression.
* sem_ch13.adb (Build_Static_Predicate.Get_RList): An expression
with actions that has a non-empty Actions list is not static. An
expression with actions that has an empty Actions list has the
static ranges of its Expression.
* sem_util.adb (Has_No_Obvious_Side_Effects): An expression with
actions with an empty Actions list has no obvious side effects
if its Expression itsekf has no obvious side effects.
Ed Schonberg [Thu, 17 Oct 2013 13:54:29 +0000 (13:54 +0000)]
sem_aux.ads, [...] (Is_Immutably_Limited_Type): Make predicate compatible with Ada 2012 definition
2013-10-17 Ed Schonberg <schonberg@adacore.com>
* sem_aux.ads, sem_aux.adb (Is_Immutably_Limited_Type): Make
predicate compatible with Ada 2012 definition
(Is_Limited_View): New name for previous version of
Is_Immutably_Limited_Type. Predicate is true for an untagged
record type with a limited component.
* exp_ch7.adb, exp_ch6.adb, exp_ch4.adb, exp_ch3.adb, exp_aggr.adb,
sem_util.adb, sem_res.adb, sem_prag.adb, sem_attr.adb, sem_ch8.adb,
sem_ch6.adb, sem_ch3.adb, exp_util.adb: Use Is_Limited_View
* freeze.adb Use Is_Immutably_Limited_Type to check the legality
of references to the current instance, Is_Limited_View otherwise.
projects / thirdparty / gcc.git / log
