Alan Modra [Mon, 19 Oct 2015 23:50:30 +0000 (10:20 +1030)]
PR66870 PowerPC64 Enable gold linker with split stack
A powerpc-linux/powerpc64-linux biarch compiler can default to either
-m32 or -m64, and we need to notice both -m32 and -m64 on the gccgo
command line. It's also possible to build a -m64 only compiler, so in
that case we can define TARGET_CAN_SPLIT_STACK.
gcc/
PR go/66870
* config/rs6000/sysv4.h (TARGET_CAN_SPLIT_STACK_64BIT): Don't define.
* config/rs6000/linux64.h (TARGET_CAN_SPLIT_STACK): Define.
(TARGET_CAN_SPLIT_STACK_64BIT): Define.
gcc/go/
PR go/66870
* gospec.c (saw_opt_m32): Rename to..
(is_m64): ..this, initialised by TARGET_CAN_SPLIT_STACK_64BIT.
Update uses.
(lang_specific_driver): Set is_m64 if OPT_m64, clear if OPT_m32.
PR rtl-optimization/66790
* df.h (DF_MIR): New macro.
(DF_LAST_PROBLEM_PLUS1): Update to be past DF_MIR
(DF_MIR_INFO_BB): New macro.
(DF_MIR_IN, DF_MIR_OUT): New macros.
(struct df_mir_bb_info): New.
(df_mir): New macro.
(df_mir_add_problem, df_mir_simulate_one_insn): New forward
declarations.
(df_mir_get_bb_info): New.
* df-problems.c (struct df_mir_problem_data): New.
(df_mir_free_bb_info, df_mir_alloc, df_mir_reset,
df_mir_bb_local_compute, df_mir_local_compute, df_mir_init,
df_mir_confluence_0, df_mir_confluence_n,
df_mir_transfer_function, df_mir_free, df_mir_top_dump,
df_mir_bottom_dump, df_mir_verify_solution_start,
df_mir_verify_solution_end): New.
(problem_MIR): New.
(df_mir_add_problem, df_mir_simulate_one_insn): New.
* timevar.def (TV_DF_MIR): New.
* ree.c: Include bitmap.h
(add_removable_extension): Add an INIT_REGS parameter. Use it
to skip zero-extensions that may get an uninitialized register.
(find_removable_extensions): Compute must-initialized registers
using the MIR dataflow problem. Update the call to
add_removable_extension.
(find_and_remove_re): Call df_mir_add_problem.
gcc/testsuite/ChangeLog:
* gnat.dg/opt50.adb: New test.
* gnat.dg/opt50_pkg.adb: New helper.
* gnat.dg/opt50_pkg.ads: New helper.
* common/config/mn10300/mn10300-common.c
(mn10300_option_optimization_table) <OPT_freorder_blocks_algorithm_>:
Use REORDER_BLOCKS_ALGORITHM_STC at -Os and up.
Richard Biener [Mon, 19 Oct 2015 14:00:28 +0000 (14:00 +0000)]
re PR tree-optimization/67975 (Failure to optimise equality between two call sequences)
2015-10-19 Richard Biener <rguenther@suse.de>
PR tree-optimization/67975
* tree-cfg.h (extract_true_false_controlled_edges): Declare.
* tree-cfg.c (extract_true_false_controlled_edges): Split out
core worker from ...
* tree-ssa-loop-im.c (extract_true_false_args_from_phi): ... here.
* tree-ssa-sccvn.c (vn_phi_compute_hash): Hash number of args
instead of block number for PHIs with two or one args.
(vn_phi_eq): Compare edge predicates of PHIs that are in different
blocks.
H.J. Lu [Mon, 19 Oct 2015 11:41:03 +0000 (11:41 +0000)]
Use GET_MODE_BITSIZE to get vector natural alignment
Since GET_MODE_ALIGNMENT is defined by psABI and the biggest alignment
is 4 byte for IA MCU psABI, we should use GET_MODE_BITSIZE for IA MCU
psABI to get vector natural alignment to check misaligned vector move.
* config/i386/i386.c (ix86_expand_vector_move): Use
GET_MODE_BITSIZE for IA MCU psABI to get vector natural
alignment.
H.J. Lu [Mon, 19 Oct 2015 11:18:14 +0000 (11:18 +0000)]
Don't leak ISA to __attribute__ ((target("arch=XXX")))
When processing __attribute__ ((target("arch=XXX"))), we should clear
the ISA bits in x_ix86_isa_flags first to avoid leaking ISA from
command line.
gcc/
PR target/67995
* config/i386/i386.c (ix86_valid_target_attribute_tree): If
arch= is set, clear all bits in x_ix86_isa_flags, except for
ISA_64BIT, ABI_64, ABI_X32, and CODE16.
PR middle-end/68002
* common.opt (fkeep-static-functions): New option.
* doc/invoke.texi: Document it.
* cgraphunit.c (cgraph_node::finalize_function): Use it.
if (cos((y*=2, -fabs(tan(x/-y)))) != cos((y*=2,tan(x/y))))
link_error ();
which is undefined behaviour. The test expected that y had the same
value in x/y and x/-y, but gimplification actually implements the
"obvious" interpretation, multiplying y by 2, using it for one cos call,
then multiplying it by 2 again and using it for the other cos call.
The file has other (valid) tests that side-effects don't block
optimisation, such as:
if (cosf((y*=3, -x)) != cosf((y*=3,x)))
link_error ();
so this patch simply removes this instance.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi.
Iain Sandoe [Sun, 18 Oct 2015 18:33:05 +0000 (18:33 +0000)]
Revise Darwin's sysroot handling to be compatible with the clang toolchain.
gcc/
* config/darwin.h (TARGET_SYSTEM_ROOT): Remove this from here,
(HAVE_LD_SYSROOT): New. (SYSROOT_SPEC): New.
(LINK_SYSROOT_SPEC): Revise to remove the default for target sysroot.
(STANDARD_STARTFILE_PREFIX_1): New.
(STANDARD_STARTFILE_PREFIX_2): New.
Mikael Morin [Sun, 18 Oct 2015 17:17:21 +0000 (17:17 +0000)]
Fix common-related error recovery ICE.
Fix an inconsistent state, between the in_common attribute
and the common_block pointer.
- adding a symbol to a common block list in gfc_match_common is delayed
after the call to gfc_add_in_common.
- gfc_restore_latest_undo_checkpoint is changed to check the common_block
pointer directly instead of the in_common attribute.
- gfc_restore_old_symbol is changed to also restore
the common-related pointers. This is done using a new function created
to factor the related memory management.
- In gfc_restore_last_undo_checkpoint, when a symbol has been removed
from the common block linked list, its common_next pointer is cleared.
PR fortran/67758
gcc/fortran/
* gfortran.h (gfc_symbol): Expand comment.
* match.c (gfc_match_common): Delay adding the symbol to
the common_block after the gfc_add_in_common call.
* symbol.c (gfc_free_symbol): Move common block memory handling...
(gfc_set_symbol_common_block): ... here as a new function.
(restore_old_symbol): Restore common block fields.
(gfc_restore_last_undo_checkpoint):
Check the common_block pointer instead of the in_common attribute.
When a symbol has been removed from the common block linked list,
clear its common_next pointer.
gcc/testsuite/
* gfortran.dg/common_25.f90: New file.
Iain Sandoe [Sun, 18 Oct 2015 14:31:25 +0000 (14:31 +0000)]
Darwin - Tidy mmacosx-version-min code and specs.
gcc/
* config/darwin-driver.c (darwin_default_min_version): Refactor code.
(darwin_driver_init): Note a version-min when provided on the c/l.
* config/darwin.h (%darwin_minversion): Remove.
* config/i386/darwin.h: Likewise.
* config/rs6000/darwin.h: Likewise.
* config/darwin.opt (mmacosx-version-min=): Use the configured default, rather than
an arbitrary constant.
Roland McGrath [Sun, 18 Oct 2015 10:33:37 +0000 (10:33 +0000)]
Fix PR63758 by using the _NSGetEnviron() API on Darwin
include/
Roland McGrath <roland@gnu.org>
PR other/63758
* environ.h: New file.
libiberty/
Roland McGrath <roland@gnu.org>
Iain Sandoe <iain@codesourcery.com>
PR other/63758
* pex-unix.c: Obtain the environment interface from settings in environ.h
rather than in-line code. Update copyright date.
* setenv.c: Likewise.
* xmalloc.c: Likewise.
Iain Sandoe [Sun, 18 Oct 2015 10:11:03 +0000 (10:11 +0000)]
Improve handling of -arch flag for Darwin
gcc/
* config/darwin-driver.c (darwin_driver_init): Handle '-arch' for
PPC, detect conflicts between -arch and multilib settings. Detect
and warn about conflicts between multiple -arch definitions.
Paul Thomas [Sun, 18 Oct 2015 09:31:21 +0000 (09:31 +0000)]
re PR fortran/67177 (MOVE_ALLOC not automatically allocating deferred character arrays in derived types)
2015-10-18 Paul Thomas <pault@gcc.gnu.org>
PR fortran/67177
PR fortran/67977
* primary.c (match_substring): Add an argument 'deferred' to
flag that a substring reference with null start and end should
not be optimized away for deferred length strings.
(match_string_constant, gfc_match_rvalue): Set the argument.
* trans-expr.c (alloc_scalar_allocatable_for_assignment): If
there is a substring reference return.
* trans-intrinsic.c (conv_intrinsic_move_alloc): For deferred
characters, assign the 'from' string length to the 'to' string
length. If the 'from' expression is deferred, set its string
length to zero. If the 'to' expression has allocatable
components, deallocate them.
2015-10-18 Paul Thomas <pault@gcc.gnu.org>
PR fortran/67177
* gfortran.dg/move_alloc_15.f90: New test
* gfortran.dg/move_alloc_16.f90: New test
PR fortran/67977
* gfortran.dg/deferred_character_assignment_1.f90: New test
Steven G. Kargl [Sat, 17 Oct 2015 16:50:47 +0000 (16:50 +0000)]
re PR fortran/67987 (ICE on declaring and initializing character with negative len)
2015-10-17 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/67987
* decl.c (char_len_param_value): Unwrap unlong line. If LEN < 0,
force it to zero per the Fortran 90, 95, 2003, and 2008 Standards.
* resolve.c (gfc_resolve_substring_charlen): Unwrap unlong line.
If 'start' is larger than 'end', length of substring is negative,
so explicitly set it to zero.
(resolve_charlen): Remove -Wsurprising warning. Update comment to
reflect that the text is from the F2008 standard.
2015-10-17 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/67987
* gfortran.df/pr67987.f90: New test.
* gfortran.dg/char_length_2.f90: Update testcase.
Kaushik Phatak [Fri, 16 Oct 2015 21:47:07 +0000 (21:47 +0000)]
divmodqi.S: Return 0x00 by default for div by 0.
* config/rl78/divmodqi.S: Return 0x00 by default for div by 0.
* config/rl78/divmodsi.S: Update return register to r8.
* config/rl78/divmodhi.S: Update return register to r8,r9.
Branch to main_loop_done_himode to pop registers before return.
* aspects.adb Add an entry for Constant_After_Elaboration in
table Canonical_Aspect.
* aspects.ads Add entries for Constant_After_Elaboration in
tables Aspect_Argument, Aspect_Delay, Aspect_Id, Aspect_Names
and Implementation_Defined_Aspect.
* par-prag.adb Pragma Constant_After_Elaboration does not require
special processing by the parser.
* sem_ch13.adb Add an entry for Constant_After_Elaboration
in table Sig_Flags.
(Analyze_Aspect_Specifications):
Add processing for aspect Constant_After_Elaboration.
(Check_Aspect_At_Freeze_Point): Aspect Constant_After_Elaboration
does not require special processing at freeze time.
* sem_prag.adb (Analyze_Pragma): Add processing for pragma
Constant_After_Elaboration. Use routine Find_Related_Context to
retrieve the context of pragma Part_Of.
(Duplication_Error): Update comment on usage.
(Find_Related_Context): New routine.
* sem_prag.ads Add an entry for Constant_After_Elaboration
in table Aspect_Specifying_Pragma.
(Analyze_Contract_Cases_In_Decl_Part): Update the comment on usage.
* sem_util.adb (Add_Contract_Item): Add processing for pragma
Constant_After_Elaboration.
* sem_util.ads (Add_Contract_Item): Update the comment on usage.
* snames.ads-tmpl Add new predefined name and aspect id for
Constant_After_Elaboration.
2015-10-16 Vincent Celier <celier@adacore.com>
* prj-pp.adb (Pretty_Print.Print): Correctly display extending
packages, instead of making them renamed packages.
* sem_ch12.adb (Analyze_Package_Instantiation):
Treat a missing SPARK_Mode annotation as having mode "Off".
(Analyze_Subprogram_Instantiation): Treat a missing SPARK_Mode
annotation as having mode "Off".
(Instantiate_Package_Body): Code
reformatting. Treat a missing SPARK_Mode annotation as having mode
"Off".
(Instantiate_Subprogram_Body): Code reformatting. Treat
a missing SPARK_Mode annotation as having mode "Off".
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* exp_ch5.adb: Code clean up.
* sem_ch13.adb: Minor fix in comment.
2015-10-16 Bob Duff <duff@adacore.com>
* a-exexda.adb: Change format of Exception_Information to be
more like what we print for unhandled exceptions.
* a-exstat.adb: Parse new format.
* a-except-2005.adb, a-except.adb: Document new format.
Arnaud Charlet [Fri, 16 Oct 2015 13:41:29 +0000 (15:41 +0200)]
[multiple changes]
2015-10-16 Javier Miranda <miranda@adacore.com>
* sem_ch5.adb (Analyze_Iterator_Specification): Associate a
transient scope with the renaming object declararation.
* exp_util.adb (Insert_Actions): if the enclosing interator
loop is marked as requiring the secondary stack then attach the
actions to the transient scope.
2015-10-16 Bob Duff <duff@adacore.com>
* exp_ch7.adb: Minor spelling fixes.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_ch3.adb (Replace_Anonymous_Access_To_Protected_Subprogram):
If anonymous type is component type of array type declaration,
analyze its declaration in the current scope, not the enclosing
one.
Arnaud Charlet [Fri, 16 Oct 2015 13:25:00 +0000 (15:25 +0200)]
[multiple changes]
2015-10-16 Gary Dismukes <dismukes@adacore.com>
* prj.adb, sem_util.adb, exp_ch6.adb: Minor reformatting.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* exp_ch5.adb (Expand_Formal_Container_Element_Loop): Modify
expansion to allow element iteration over formal containers
whose elements are indefinite types.
2015-10-16 Doug Rupp <rupp@adacore.com>
* s-taprop-linux.adb (Monotonic_Clock): Call clock_gettime
instead of gettimeofday.
* s-osinte-linux.ads (clock_gettime): New imported subprogram.
* exp_ch6.adb (Make_Build_In_Place_Call_In_Object_Declaration):
Indefinite or limited library level objects are now returned on
the heap.
* exp_ch7.adb (Build_Finalization_Master): Add formal
parameter For_Lib_Level. Add context specific insertion for a
finalization master created for an access result type related
to a build-in-place function call used to initialize a library
level object.
* exp_ch7.ads (Build_Finalization_Master): Add formal parameter
For_Lib_Level. Update the comment on usage.
* sem_util.adb (Mark_Coextensions): Code cleanup.
2015-10-16 Emmanuel Briot <briot@adacore.com>
* prj.adb (For_Every_Project_Imported_Context): Fix handling
of aggregated projects with duplicate names.
* a-ngelfu.ads: Minor whitespace fix.
Arnaud Charlet [Fri, 16 Oct 2015 13:11:18 +0000 (15:11 +0200)]
[multiple changes]
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Build_Predicate_Functions): The expression for
the predicate is side-effect free if it does not contain any
variable references.
2015-10-16 Bob Duff <duff@adacore.com>
* a-convec.adb ("="): Previous version depended
on "=" composing, but that doesn't quite work -- we want the "="
operator passed in to the generic. So we need a loop after all.
2015-10-16 Yannick Moy <moy@adacore.com>
* sem_util.adb (Is_Object_Reference): Attribute 'Loop_Entry produces
an object.
* sem_ch6.adb: Minor fix in comment.
Arnaud Charlet [Fri, 16 Oct 2015 13:08:04 +0000 (15:08 +0200)]
[multiple changes]
2015-10-16 Bob Duff <duff@adacore.com>
* a-contai.ads: Add two check names: Container_Checks and
Tampering_Check. Move the tampering check machinery from
Ada.Containers.Vectors to Ada.Containers. Later we can share it
with other containers.
Disable the tampering machinery in the presence of
Suppress(Tampering_Check).
Simplify the implementation of tampering checks. E.g. use RAII
to make incrementing/decrementing of the counts more concise.
* a-contai.adb: New package body, implementing the above.
* a-convec.ads, a-convec.adb: Use tampering check machinery
in Ada.Containers.
Disable all checking code when checks are suppressed.
Simplify many of the operations. Implement "&" in terms of Append,
rather than "by hand".
Remove: function "=" (L, R : Elements_Array) return Boolean is
abstract; so we can call the predefined "=" on Elements_Array.
For "=" on Vectors: Previously, we returned True immediately if
Left'Address = Right'Address. That seems like a non-optimization
("if X = X" is unusual), so removed that. Simplify by using
slice comparison ("=" on Element_Array will automatically call
"=" on the components, even if user defined).
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Chek_Record_Representation_Clause): When
iterating over components, skip anonymous subtypes created for
constrained array components.
Arnaud Charlet [Fri, 16 Oct 2015 12:53:03 +0000 (14:53 +0200)]
[multiple changes]
2015-10-16 Bob Duff <duff@adacore.com>
* adadecode.h, adadecode.c (ada_demangle): Remove
ada_demangle, no longer used.
* a-except-2005.adb: Bring System.Traceback.Symbolic into the
closure.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_util.adb, sem_util.ads (Get_Reference_Discriminant): Utility to
locate the access discriminant that supports implicit dereference on a
record type.
(Is_OK_Variable_For_Out_Parameter): Reject other illegal uses
of Implicit_Dereference on an access_to_constant when actual
parameter is a rewritten variable or function call.
Arnaud Charlet [Fri, 16 Oct 2015 12:47:43 +0000 (14:47 +0200)]
[multiple changes]
2015-10-16 Bob Duff <duff@adacore.com>
* a-tags.adb, s-trasym.adb, s-trasym.ads: Make sure we don't get elaboration circularities when polling is turned on.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_ch3.adb (Replace_Anonymous_Access_To_Protected_Subprogram):
When creating a full declaration for a formal parameter, the
entity is labelled as an Itype, so set Associated_Node_For_Itype
accordingly.
* einfo.ads: Clarify use of Associated_Node_For_Itype.
2015-10-16 Bob Duff <duff@adacore.com>
* bindgen: Move pragmas Warnings earlier, otherwise
we can get warnings on with_clauses.
Arnaud Charlet [Fri, 16 Oct 2015 12:43:21 +0000 (14:43 +0200)]
[multiple changes]
2015-10-16 Arnaud Charlet <charlet@adacore.com>
* s-osprim-mingw.adb, s-osprim-x32.adb, s-taprop-mingw.adb,
s-taprop-posix.adb (Monotonic_Clock): Removed, not used.
Remove remaining references to OS_Primitives.Monotonic_Clock.
Keep it only on windows which is a special case.
Arnaud Charlet [Fri, 16 Oct 2015 12:41:13 +0000 (14:41 +0200)]
[multiple changes]
2015-10-16 Eric Botcazou <ebotcazou@adacore.com>
* inline.adb (Subp_Info): Remove Listed component.
(Add_Inlined_Subprogram): Take an entity instead of an index.
Do not set Listed component to True.
(New_Entry): Do not initialize Listed component to False.
(Analyze_Inlined_Bodies): Do not test Listed component
(Must_Inline): Add calls not in the main unit only
if they are in a subprogram that can be inlined outside its unit.
(Add_Inlined_Body): Move test around and add comment.
* sem_ch4.adb (Try_Object_Operation, Try_One_Interpretation):
Do not reset the Obj_Type of the prefix if an interpretation
involves an untagged type, to prevent a crash when analyzing an
illegal program in All_Errors mode.
* exp_ch4.adb (Expand_N_Expression_With_Actions):
Force the evaluation of the expression when its type is Boolean.
(Force_Boolean_Evaluation): New routine.
2015-10-16 Bob Duff <duff@adacore.com>
* sem_util.adb (Has_Discrim_Dep_Array): Remove
this function, and the call. No longer needed now that the back
end can handle such things. Should result in further speedups
in some cases.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Build_Predicate_Functions): If expression for
predicate is side-effect free, indicate that the predicate
function is pure, to allow for optimization of redundant
predicate checks.
Gary Dismukes [Fri, 16 Oct 2015 12:22:22 +0000 (12:22 +0000)]
exp_ch9.adb (Build_Simple_Entry_Call): Set_Is_Internal on the temporary object used for a by-copy entry parameter...
2015-10-16 Gary Dismukes <dismukes@adacore.com>
* exp_ch9.adb (Build_Simple_Entry_Call): Set_Is_Internal on
the temporary object used for a by-copy entry parameter, to
ensure that the object doesn't get its No_Initialization flag
reset later in Default_Initialize_Object. Also, generate the
assignment of the actual to the temporary in the additional case
of a scalar out parameter whose type has a Default_Value aspect.
* exp_ch3.adb: Fix minor typo in comment.
* exp_util.adb, sem_util.ads, sem_ch6.adb, sem_ch7.adb, sem_ch8.adb:
Minor reformatting.
* sem_ch12.adb (Analyze_Formal_Package_Declaration): Set the
SPARK_Mode from the context.
2015-10-16 Bob Duff <duff@adacore.com>
* sem_util.adb (Requires_Transient_Scope):
If Typ is a generic formal incomplete type, look at the actual
type. Otherwise, we don't notice that the actual type is tagged,
has a variant part, etc, causing a mismatch of calling conventions
between caller and callee.
* testsuite/25_algorithms/random_shuffle/moveable.cc: Change variable
N from const int N = 200000 to const unsigned int N = 10000.
Delete useless fill_ascending function call.
Arnaud Charlet [Fri, 16 Oct 2015 11:00:05 +0000 (11:00 +0000)]
sem_ch12.adb: Minor punctuation fix in comment
2015-10-16 Arnaud Charlet <charlet@adacore.com>
* sem_ch12.adb: Minor punctuation fix in comment
* s-rident.ads: Minor consistency fix in comment
* exp_attr.adb, g-spipat.ads: punctuation fixes in comments.
* restrict.ads: Style fix in comments.
* sem_prag.adb: Minor grammar fix in comment
* exp_ch3.adb (Expand_N_Full_Type_Declaration): Do not capture,
set and restore the Ghost mode.
(Expand_N_Object_Declaration): Do not capture, set and restore the
Ghost mode.
(Freeze_Type): Redo the capture and restore of the Ghost mode.
(Restore_Globals): Removed.
* exp_ch5.adb (Expand_N_Assignment_Statement): Redo the capture
and restore of the Ghost mode.
(Restore_Globals): Removed.
* exp_ch6.adb (Expand_N_Procedure_Call_Statement):
Redo the capture and restore of the Ghost mode.
(Expand_N_Subprogram_Body): Redo the capture, set and restore
of the Ghost mode.
(Expand_N_Subprogram_Declaration): Do not
capture, set and restore the Ghost mode.
(Restore_Globals): Removed.
* exp_ch7.adb (Expand_N_Package_Body): Redo the capture, set
and restore of the Ghost mode.
(Expand_N_Package_Declaration): Do not capture, set and restore the
Ghost mode.
* exp_ch8.adb (Expand_N_Exception_Renaming_Declaration):
Redo the capture and restore of the Ghost mode.
(Expand_N_Object_Renaming_Declaration): Redo
the capture and restore of the Ghost mode.
(Expand_N_Package_Renaming_Declaration):
Redo the capture and restore of the Ghost mode.
(Expand_N_Subprogram_Renaming_Declaration): Redo the capture
and restore of the Ghost mode.
* exp_ch11.adb Remove with and use clauses for Ghost.
(Expand_N_Exception_Declaration): Do not capture, set and restore
the Ghost mode.
* exp_disp.adb (Make_DT): Redo the capture and restore of the
Ghost mode.
(Restore_Globals): Removed.
* exp_prag.adb (Expand_Pragma_Check): Do not capture, set
and restore the Ghost mode.
(Expand_Pragma_Contract_Cases):
Redo the capture and restore of the Ghost mode. Preserve the
original context of contract cases by setting / resetting the
In_Assertion_Expr counter.
(Expand_Pragma_Initial_Condition):
Redo the capture and restore of the Ghost mode.
(Expand_Pragma_Loop_Variant): Redo the capture and restore of
the Ghost mode.
(Restore_Globals): Removed.
* exp_util.adb (Make_Predicate_Call): Redo the capture and
restore of the Ghost mode.
(Restore_Globals): Removed.
* freeze.adb (Freeze_Entity): Redo the capture and restore of
the Ghost mode.
(Restore_Globals): Removed.
* ghost.adb (Check_Ghost_Context): Remove the RM reference from
the error message.
(Is_OK_Statement): Account for statements
that appear in assertion expressions.
(Is_Subject_To_Ghost):
Moved from spec.
* ghost.ads (Is_Subject_To_Ghost): Moved to body.
* rtsfind.ads (Load_RTU): Redo the capture and restore of the
Ghost mode.
* sem.adb Add with and use clauses for Ghost.
(Analyze): Redo
the capture and restore of the Ghost mode. Set the Ghost mode
when analyzing a declaration.
(Do_Analyze): Redo the capture
and restore of the Ghost mode.
* sem_ch3.adb (Analyze_Full_Type_Declaration): Do not capture, set
and restore the Ghost mode.
(Analyze_Incomplete_Type_Decl):
Do not capture, set and restore the Ghost mode.
(Analyze_Number_Declaration): Do not capture, set and restore the
Ghost mode.
(Analyze_Object_Declaration): Do not capture, set and
restore the Ghost mode.
(Analyze_Private_Extension_Declaration):
Do not capture, set and restore the Ghost mode.
(Analyze_Subtype_Declaration): Do not capture, set and restore
the Ghost mode.
(Restore_Globals): Removed.
* sem_ch5.adb (Analyze_Assignment): Redo the capture and restore
of the Ghost mode.
(Restore_Globals): Removed.
* sem_ch6.adb (Analyze_Abstract_Subprogram_Declaration):
Do not capture, set and restore the Ghost mode.
(Analyze_Procedure_Call): Redo the capture and restore of the
Ghost mode.
(Analyze_Subprogram_Body_Helper): Redo the capture
and restore of the Ghost mode. (Analyze_Subprogram_Declaration):
Do not capture, set and restore the Ghost mode.
(Restore_Globals): Removed.
* sem_ch7.adb (Analyze_Package_Body_Helper): Redo the capture and
restore of the Ghost mode.
(Analyze_Package_Declaration):
Do not capture, set and restore the Ghost mode.
(Analyze_Private_Type_Declaration): Do not capture, set and
restore the Ghost mode.
(Restore_Globals): Removed.
* sem_ch8.adb (Analyze_Exception_Renaming): Do not capture,
set and restore the Ghost mode.
(Analyze_Generic_Renaming): Do not capture, set and restore the Ghost
mode.
(Analyze_Object_Renaming): Do not capture, set and restore the
Ghost mode.
(Analyze_Package_Renaming): Do not capture, set and restore the Ghost
mode.
(Analyze_Subprogram_Renaming): Do not capture, set and restore the
Ghost mode.
(Restore_Globals): Removed.
* sem_ch11.adb (Analyze_Exception_Declaration): Do not capture,
set and restore the Ghost mode.
* sem_ch12.adb (Analyze_Generic_Package_Declaration):
Do not capture, set and restore the Ghost mode.
(Analyze_Generic_Subprogram_Declaration): Do not capture, set
and restore the Ghost mode.
* sem_ch13.adb (Build_Invariant_Procedure_Declaration): Redo
the capture and restore of the Ghost mode.
* sem_prag.adb (Analyze_Contract_Cases_In_Decl_Part):
Redo the capture and restore of the Ghost mode.
(Analyze_External_Property_In_Decl_Part):
Redo the capture and restore of the Ghost mode.
(Analyze_Initial_Condition_In_Decl_Part): Redo the
capture and restore of the Ghost mode. (Analyze_Pragma):
Do not capture, set and restore the Ghost mode for Assert.
Redo the capture and restore of the Ghost mode for Check. Do
not capture and restore the Ghost mode for Invariant.
(Analyze_Pre_Post_Condition_In_Decl_Part): Redo the capture and
restore of the Ghost mode.
* sem_res.adb (Resolve): Capture, set and restore the Ghost mode
when resolving a declaration.
* sem_util.adb (Build_Default_Init_Cond_Procedure_Body):
Redo the capture and restore of the Ghost mode.
(Build_Default_Init_Cond_Procedure_Declaration): Redo the capture
and restore of the Ghost mode.
Arnaud Charlet [Fri, 16 Oct 2015 10:44:09 +0000 (12:44 +0200)]
[multiple changes]
2015-10-16 Javier Miranda <miranda@adacore.com>
* inline.adb (Add_Inlined_Body): Ensure that
Analyze_Inlined_Bodies will be invoked after completing the
analysis of the current unit.
2015-10-16 Arnaud Charlet <charlet@adacore.com>
* sem_ch13.adb (Adjust_Record_For_Reverse_Bit_Order): Fix error
message for bad last bit position.
* sem_ch3.adb, sem_util.adb, sem_util.ads: Minor reformatting.
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* exp_ch5.adb (Expand_N_Case_Statement): If expression is
compile-time known but does not obey a static predicate on
its type, replace the case statement with a raise statement,
as with other statically detected constraint violations.
Ed Schonberg [Fri, 16 Oct 2015 10:28:37 +0000 (10:28 +0000)]
sem_util.adb (Gather_Components): When gathering components of a nested variant...
2015-10-16 Ed Schonberg <schonberg@adacore.com>
* sem_util.adb (Gather_Components): When gathering components
of a nested variant, the record type used in legality checks is
the enclosing record type.
Richard Biener [Fri, 16 Oct 2015 07:52:32 +0000 (07:52 +0000)]
gimple-fold.c (gimple_fold_builtin_memory_op): Use gimple_build and get rid of force_gimple_operand_gsi.
2015-10-16 Richard Biener <rguenther@suse.de>
* gimple-fold.c (gimple_fold_builtin_memory_op): Use gimple_build
and get rid of force_gimple_operand_gsi.
(gimple_fold_builtin_memory_chk): Likewise.
(gimple_fold_builtin_stxcpy_chk): Likewise.
(rewrite_to_defined_overflow): Likewise.
(gimple_convert_to_ptrofftype): New function.
* gimple-fold.h (gimple_convert_to_ptrofftype): New overload,
declare.
* tree-vectorizer.h (vect_get_new_ssa_name): Declare.
* tree-vect-data-refs.c (vect_get_new_ssa_name): New helper.
* tree-vect-loop.c (get_initial_def_for_induction): Drop
use of force_gimple_operand in favor of gimple_build.
Use vect_get_new_ssa_name.
* tree-vect-stmts.c (vect_init_vector): Use vect_get_new_ssa_name.
(vectorizable_mask_load_store): Likewise.
(vectorizable_call): Likewise.
(vectorizable_store): Likewise.
(vectorizable_load): Likewise.
(vect_get_vec_def_for_stmt_copy): Remove redundant stmt.
PR67945: Fix oscillation between pow representations
This patch fixes some fallout from my patch to move the sqrt and cbrt
folding rules to match.pd. The rules included canonicalisations like:
sqrt(sqrt(x))->pow(x,1/4)
which in the original code was only ever done at the generic level.
My patch meant that we'd do it whenever we tried to fold a gimple
statement, and eventually it would win over the sincos optimisation
that replaces pow(x,1/4) with sqrt(sqrt(x)).
Following a suggestion from Richard B, the patch adds a new
PROP_gimple_* flag to say whether fp routines have been optimised
for the target. If so, match.pd should only transform calls to math
functions if the result is actually an optimisation, not just an
IL simplification or canonicalisation. The question then of course
is: which rules are which? I've added block comments that describe
the criteria I was using.
A slight wart is that we need to use the cfun global to access
the PROP_gimple_* flag; there's no local function pointer available.
Bootstrapped & regression-tested on x86_64-linux-gnu. Also tested
on powerc64-linux-gnu.
gcc/
PR tree-optimization/67945
* tree-pass.h (PROP_gimple_opt_math): New property flag.
* generic-match-head.c (canonicalize_math_p): New function.
* gimple-match-head.c: Include tree-pass.h.
(canonicalize_math_p): New function.
* match.pd: Group math built-in rules into simplifications
and canonicalizations. Guard the latter with canonicalize_math_p.
* tree-ssa-math-opts.c (pass_data_cse_sincos): Provide the
PROP_gimple_opt_math property.
Uros Bizjak [Wed, 14 Oct 2015 21:52:37 +0000 (23:52 +0200)]
mips.h (MIPS_STACK_ALIGN): Implement using ROUND_UP macro.
* config/mips/mips.h (MIPS_STACK_ALIGN): Implement using
ROUND_UP macro.
* config/mips/mips.c (mips_setup_incoming_varargs): Use
ROUND_DOWN to calculate off.
(mips_gimplify_va_arg_expr): Use ROUND_UP to calculate rsize.
(mips_emit_probe_stack_range): Use ROUND_DOWN to calculate
rounded_size.
Jeff Law [Wed, 14 Oct 2015 17:12:40 +0000 (11:12 -0600)]
[PATCH] Split ssa-dom-thread-2.c into separate files/tests
* gcc.dg/tree-ssa/ssa-dom-thread-2.c: Deleted. The six functions
contained within have their own file/test now.
* gcc.dg/tree-ssa/ssa-dom-thread-2a.c: New test extracted from
ssa-dom-thread-2.c. Tighten expected output slightly and comment
expectations a bit more.
* gcc.dg/tree-ssa/ssa-dom-thread-2b.c: Likewise.
* gcc.dg/tree-ssa/ssa-dom-thread-2c.c: Likewise.
* gcc.dg/tree-ssa/ssa-dom-thread-2d.c: Likewise.
* gcc.dg/tree-ssa/ssa-dom-thread-2e.c: Likewise.
* gcc.dg/tree-ssa/ssa-dom-thread-2f.c: Likewise.
projects / thirdparty / gcc.git / log
