2013-04-11 Richard Biener <rguenther@suse.de>
PR tree-optimization/56878
* tree-flow.h (outermost_invariant_loop_for_expr): Declare.
* tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
New function.
* tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
Prefer to align the DR with the most invariant base address.
From-SVN: r197769
2013-04-11 Robert Dewar <dewar@adacore.com>
* sem_res.adb: Minor code reorganization and comment fixes.
* sem_type.adb: Minor reformatting.
From-SVN: r197768
2013-04-11 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch4.adb (Process_Transient_Object): Add new
local variable Fin_Call. Remove and explain ??? comment. Use the
Actions of logical operators "and then" and "or else" to insert
the generated finalization call.
2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
* gnat_rm.texi: Fix typo.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_res.adb: Minor reformatting.
From-SVN: r197767
2013-04-11 Robert Dewar <dewar@adacore.com>
* atree.h: Add declarations for Flag255-Flag289 Fix declaration
of Field30 (was wrong, but no effect, since not yet referenced by
back end) Add declarations for Field31-Field35 Add declarations
for Node31-Node35.
* einfo.ads, einfo.adb (Has_Invariants): No longer applies to
procedures.
(Has_Predicates): No longer applies to functions.
(Is_Predicate_Function): New flag.
(Is_Predicate_Function_M): New flag.
(Is_Invariant_Procedure): New flag.
(Predicate_Function_M): New function.
(Set_Predicate_Function_M): New procedure.
* exp_ch11.adb (Expand_N_Raise_Expression): Take care of special
case of appearing in predicate used for membership test.
* exp_ch3.adb (Insert_Component_Invariant_Checks): Set
Is_Invariant_Procedure flag.
* exp_ch4.adb (Expand_Op_In): Call special predicate function
that takes care of raise_expression nodes in the predicate.
* exp_util.ads, exp_util.adb (Make_Predicate_Call): Add argument Mem for
membership case.
* sem_ch13.adb (Build_Predicate_Functions): New name for
Build_Predicate_Function. Major rewrite to take care of raise
expression in predicate for membership tests.
* sem_res.adb (Resolve_Actuals): Include both predicate functions
in defense against infinite predicate function loops.
* sinfo.ads, sinfo.adb (Convert_To_Return_False): New flag.
2013-04-11 Robert Dewar <dewar@adacore.com>
* sem_prag.adb: Minor reformatting.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* lib-xref.adb: Generate reference for component of anonymous
access type.
From-SVN: r197766
2013-04-11 Vincent Celier <celier@adacore.com>
* gnat_ugn.texi: Remove section "The Development Environments"
now that all predefined attributes are documented, including
those in package IDE.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb: Preserve parent link in copy of expression.
From-SVN: r197763
2013-04-11 Vincent Celier <celier@adacore.com>
* projects.texi: Complete rewrite of the subsection Attributes
in section "Project file Reference".
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_ch4.adb: Minor reformatting.
From-SVN: r197762
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_ch4.adb (Expand_Concatenate): Remove wrapping in
expression-with-actions node. No longer needed given fix to
sem_prag and caused loss of some useful warnings.
* sem.ads: Minor reformatting.
* sem_prag.adb (Check_Disabled): Removed, to be replaced by not
Check_Enabled. These two routines were curiously incompatible
causing confusion.
(Analyze_Pragma, case Check): Make sure we do
not expand the string argument if the check is disabled. Avoid
use of Check_Disabled, which resulted in missing analysis in
some cases.
* sem_prag.ads (Check_Disabled): Removed, to be replaced by not
Check_Enabled. These two routines were curiously incompatible
causing confusion.
From-SVN: r197761
2013-04-11 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch4.adb (Process_Transient_Object): Use
an unchecked conversion when associating a transient controlled
object with its "hook".
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Analyze_Pragma, case
Preelaborable_Initialization): The pragma is legal if it comes
from an aspect on the private view of the type, even though its
analysis point takes place later at the freeze point.
From-SVN: r197760
2013-04-11 Robert Dewar <dewar@adacore.com>
* sem_ch6.adb: Minor reformatting.
2013-04-11 Yannick Moy <moy@adacore.com>
* ali-util.adb (Read_Withed_ALIs): Do not consider it an error to
read ALI files with No_Object=True in Alfa mode.
* gnat1drv.adb: Set appropriately Back_End_Mode in Alfa mode, whether
this is during frame condition generation of translation to Why.
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_ch4.adb: Minor code reorganization
* types.ads: Minor reformatting.
From-SVN: r197759
2013-04-11 Johannes Kanig <kanig@adacore.com>
* opt.ads New global boolean Frame_Condition_Mode to avoid
referring to command line switch.
* gnat1drv.adb (Gnat1drv) set frame condition mode when -gnatd.G
is present, and disable Code generation in that case. Disable
ALI file generation when switch is *not* present.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Analyze_Expression_Function): Perform the
pre-analysis on a copy of the expression, to prevent downstream
visbility issues involving operators and instantiations.
From-SVN: r197758
2013-04-11 Johannes Kanig <kanig@adacore.com>
* debug.adb: Reservation and documentation for -gnatd.G switch.
* gnat1drv.adb (Adjust_Global_Switches) Take into account -gnatd.G
switch, and set ALI file generation accordingly.
From-SVN: r197757
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_ch4.adb, exp_dist.adb: Minor reformatting.
* gnat_rm.texi, gnat_ugn.texi: -020 Add documentation clarifying that
check names introduced with pragma Check_Name are suppressed by -gnatp.
From-SVN: r197756
2013-04-11 Yannick Moy <moy@adacore.com>
* gnat1drv.adb (Adjust_Global_Switches): Allow missing body in Alfa
mode.
2013-04-11 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch4.adb (Expand_N_Allocator): Detect the
allocation of an anonymous controlled object where the type of
the context is named. Use the pool and finalization master of
the named access type to allocate the object.
2013-04-11 Vincent Celier <celier@adacore.com>
* gnat_ugn.texi: Remove most mentions of gprbuild.
* projects.texi: Remove all mentions of asociative array
attributes.
From-SVN: r197753
2013-04-11 Arnaud Charlet <charlet@adacore.com>
* debug.adb, sem_prag.adb, par-ch2.adb, sem_attr.adb, gnat1drv.adb,
exp_disp.adb, opt.ads, sem_ch13.adb (Relaxed_RM_Semantics): New flag.
Enable this flag in CodePeer mode, and also via -gnatd.M.
Replace some uses of CodePeer_Mode by Relaxed_RM_Semantics.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_ch8.adb (Check_Constrained_Object): If a subtype is created
from the renamed object in an object renaming declaration with
an unconstrained nominal subtype, freeze the created subtype at
once, to prevent order of elaboration issues in the backend.
2013-04-11 Arnaud Charlet <charlet@adacore.com>
* exp_aggr.adb (Aggr_Size_OK): Refine setting of Max_Aggr_Size
in particular in CodePeer mode.
2013-04-11 Vincent Celier <celier@adacore.com>
* gnat_ugn.texi: Add documentation for backup copies of project
files for gnatname.
From-SVN: r197751
2013-04-11 Tristan Gingold <gingold@adacore.com>
* gnat_rm.texi: Add Detect_BLocking in the ravenscar profile
pragma list.
2013-04-11 Vincent Celier <celier@adacore.com>
* gnatname.adb (Scan_Args): Recognize new switch --no-backup
(Usage): Add line for --no-backup.
* opt.ads (No_Nackup): New Boolean variable, initialized to False.
* prj-makr.adb (Initialize): Create a backup for an existing
project file if gnatname is not invoked with --no-backup.
From-SVN: r197750
2013-04-11 Thomas Quinot <quinot@adacore.com>
* exp_ch4.adb: Minor code improvement: replace various calls to
Make_If_Statement in expansion with Make_Implicit_If_Statement.
2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
* ali.adb: Fix minor typo.
From-SVN: r197749
2013-04-11 Thomas Quinot <quinot@adacore.com>
* exp_ch4.adb (Find_Enclosing_Context): Add missing case of
N_Procedure_Call_Statement.
2013-04-11 Robert Dewar <dewar@adacore.com>
* debug.adb: Minor comment fix.
From-SVN: r197748
2013-04-11 Johannes Kanig <kanig@adacore.com>
* debug.adb: Remove comment for -gnatd.G.
2013-04-11 Thomas Quinot <quinot@adacore.com>
* exp_ch4.adb (Expand_Record_Equality.Suitable_Element):
Remove recursive routine, replace with...
(Expand_Record_Equality.Element_To_Compare): New subroutine,
implement iterative search for next element to compare.
Add explanatory comment in the tagged case.
From-SVN: r197747
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_ch5.adb: remove spurious warning from non-empty loop.
* sem_ch8.adb (Enclosing_Instance): Make public to other routines
in the package, in order to suppress redundant semantic checks
on subprogram renamings in nested instantiations.
From-SVN: r197746
2013-04-11 Robert Dewar <dewar@adacore.com>
* errout.ads: Minor reformatting.
* sem_eval.adb (Why_Not_Static): Now issues continuation messages
(Why_Not_Static): Test for aggregates behind string literals.
* sem_eval.ads (Why_Not_Static): Now issues continuation messages.
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_ch4.adb (Expand_Concatenation): Wrap expansion in
Expressions_With_Actions.
2013-04-11 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Base_Types_Match): For an actual type in an
instance, the base type may itself be a subtype, so find true
base type to determine compatibility.
From-SVN: r197745
2013-04-11 Robert Dewar <dewar@adacore.com>
* s-osprim-mingw.adb, sem_ch3.adb, sem_prag.adb, sem_util.adb.
makeutl.adb, sem_ch8.adb: Minor reformatting.
2013-04-11 Vincent Celier <celier@adacore.com>
* gnat_ugn.texi: Minor fixes for VMS.
* ug_words: Minor addition: -gnato? => /OVERFLOW_CHECKS=?.
2013-04-11 Robert Dewar <dewar@adacore.com>
* usage.adb (Usage): Minor edit to -gnatW message
2013-04-11 Robert Dewar <dewar@adacore.com>
* exp_aggr.adb (Expand_N_Aggregate): Add circuit for handling
others for string literal case. Also add big ??? comment about
this new code, which should be redundant, but is not.
* sem_eval.adb (Eval_Concatenation): Handle non-static case
properly (Eval_String_Literal): Handle non-static literal properly
From-SVN: r197744
gcc/
* config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
floating-point vector comparisons against 0.
gcc/testsuite/
* gcc.target/aarch64/vect-fcm.x: Add check for zero forms of
inverse operands.
* gcc.target/aarch64/vect-fcm-eq-d.c: Check that new zero form
loop is vectorized.
* gcc.target/aarch64/vect-fcm-eq-f.c: Likewise.
* gcc.target/aarch64/vect-fcm-ge-d.c: Check that new zero form
loop is vectorized and that the correct instruction is generated.
* gcc.target/aarch64/vect-fcm-ge-f.c: Likewise.
* gcc.target/aarch64/vect-fcm-gt-d.c: Likewise.
* gcc.target/aarch64/vect-fcm-gt-f.c: Likewise.
From-SVN: r197741
PR tree-optimization/56899
* fold-const.c (extract_muldiv_1): Apply distributive law
only if TYPE_OVERFLOW_WRAPS (ctype).
* gcc.c-torture/execute/pr56899.c: New test.
From-SVN: r197692
* config/sparc/sparc.c: Include tree-pass.h.
(TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
(sparc_reorg): Rename to sparc_do_work_around_errata. Move to
head of file. Change return type. Split off gate function.
(sparc_gate_work_around_errata): New function.
(pass_work_around_errata): New pass definition.
(insert_pass_work_around_errata) New pass insert definition to
insert pass_work_around_errata just after delayed-branch scheduling.
(sparc_option_override): Insert the pass.
* config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
From-SVN: r197681
One major suboptimal area of the sparc back end is cstore generation
on 64-bit.
Due to the way arguments and return values of functions must be
promoted, the ideal mode for cstore's result would be DImode.
But this hasn't been done because of a fundamental limitation
of the cstore patterns. They require a fixed mode be used for
the boolean result value.
I've decided to work around this by building a target hook which
specifies the type to use for conditional store results, and then I
use a special predicate for operans 0 in the cstore expanders so
that they still match even when we use DImode.
The default version of the target hook just does what it does now,
so no other target should be impacted by this at all.
Regstrapped on 32-bit sparc-linux-gnu and I've run the testsuite
with "-m64" to validate the 64-bit side.
gcc/
* target.def (cstore_mode): New hook.
* target.h: Include insn-codes.h
* targhooks.c: Likewise.
(default_cstore_mode): New function.
* targhooks.h: Declare it.
* doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
* doc/tm.texi: Rebuild.
* expmed.c (emit_cstore): Obtain cstore boolean result mode using
target hook, rather than inspecting the insn_data.
* config/sparc/sparc.c (sparc_cstore_mode): New function.
(TARGET_CSTORE_MODE): Redefine.
(emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
result patterns.
* config/sparc/predicates.md (cstore_result_operand): New special
predicate.
* config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
Use it for operand 0.
(*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
(*snesi_special): Likewise.
(*snesi_zero): Likewise.
(*seqsi_zero): Likewise.
(*sltu_insn): Likewise.
(*sgeu_insn): Likewise.
(*seqdi_special): Make operand 0 and comparison operation be of
DImode.
(*snedi_special): Likewise.
(*snedi_special_vis3): Likewise.
(*neg_snesi_zero): Rename to *neg_snesisi_zero.
(*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
(*snesi_zero_extend): Delete, covered by 'P' mode iterator.
(*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
(*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
(*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
(*sltu_extend_sp64): Likewise.
(*neg_sltu_insn): Rename to *neg_sltusi_insn.
(*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
(*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
(*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
(*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
gcc/testsuite/
* gcc.target/sparc/setcc-4.c: New test.
* gcc.target/sparc/setcc-5.c: New test.
From-SVN: r197679
2013-04-10 Richard Biener <rguenther@suse.de>
* passes.c (execute_todo): Do not call ggc_collect conditional here.
(execute_one_ipa_transform_pass): But unconditionally here.
(execute_one_pass): And here.
(init_optimization_passes): Remove reload pass.
* tree-pass.h (TODO_ggc_collect): Remove.
(pass_reload): Likewise.
* ira.c (do_reload): Merge into ...
(ira): ... this.
(rest_of_handle_reload): Remove.
(pass_reload): Likewise.
* config/i386/i386.c (ix86_option_override): Refer to ira instead
of reload for vzeroupper pass placement.
* <everywhere>: Remove TODO_ggc_collect from todo_flags_start
and todo_flags_finish of all passes.
* g++.dg/pr55604.C: Use -fdump-rtl-ira.
From-SVN: r197671
gcc:
2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
PR tree-optimization/55524
* tree-ssa-math-opts.c
(convert_mult_to_fma): Don't use an fms construct
when we don't have an fms operation, but fnma, and it looks
likely that we'll be able to use the latter.
gcc/testsuite:
2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
PR tree-optimization/55524
* gcc.target/epiphany/fnma-1.c: New test.
From-SVN: r197668
2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
* cif-code.def (OVERWRITABLE): Correct the comment for overwritable
function.
* ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
inline fail caused by overwritable functions.
* gcc.dg/tree-ssa/inline-11.c: New test
From-SVN: r197667
2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
* combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
unnecessary bits in the constant power of two case.
From-SVN: r197666
PR c++/56895
* typeck.c (cp_build_binary_op): Call fold_non_dependent_expr_sfinae
first before calling maybe_constant_value for warn_for_div_by_zero
or invalid shift count warning purposes.
* g++.dg/template/arrow3.C: New test.
From-SVN: r197660
