RE: [PATCH][ARM][thumb1] Reduce lr save for leaf function with non-far jump

Ping^2 > -Original Message- > From: Joey Ye > Sent: Saturday, January 05, 2013 15:41 > To: Ramana Radhakrishnan > Cc: Joey Ye; gcc-patches@gcc.gnu.org > Subject: RE: [PATCH][ARM][thumb1] Reduce lr save for leaf function with > non-far jump > > Ping > > > -Original Message- > > Fr

[PATCH, RFA]: Properly mask memory model argument with MEMMODEL_MASK

Hello! Attached fairly mechanical patch adds missing memory model argument masks when comparing comparing with various MEMMODEL_* types. Please note that we can modify the memory model with target flags, IX86_HLE_* in case of x86. Also, please note that this patch doesn't touch tsan.c. 2013-01-1

[PATCH] Remove clang-ism -Wno-c99-extensions from libsanitizer AM_CXXFLAGS

Hi! The -Wno-c99-extensions option isn't supported by GCC (and never was), guess Wei copied it over from llvm makefiles. GCC normally doesn't error out on it, just warns. Fixed thusly, ok for trunk? 2013-01-15 Jakub Jelinek * sanitizer_common/Makefile.am (AM_CXXFLAGS): Remove

Re: [PATCH] Allow new ISL/CLooG versions

On Mon, 14 Jan 2013, Jack Howarth wrote: > On Mon, Jan 14, 2013 at 08:27:12PM +0100, Dominique Dhumieres wrote: > > In order to bootstrap r195167 with the new ISL/CLooG versions, > > I had to apply the following patch: > > > > --- ../work/configure 2013-01-14 19:32:00.0 +0100 > > ++

RE: [RFA] New port: CR16: BFD Changes required by the gdb port

Hi Pedro, >> - cr16-*-*) >> -noconfigdirs="$noconfigdirs gdb" >> -;; >> Dunno why these were removed with that merge. The >> ChangeLog entries brought in don't seen to mention these changes. Thanks for looking this up. The below patch includes these above entries into configure.ac and

Re: [PATCH] Fix up ICE with reduction % (PR tree-optimization/55955)

On Mon, 14 Jan 2013, Jakub Jelinek wrote: > Hi! > > For *SHIFT_EXPR/*ROTATE_EXPR with VECTOR_TYPE types, optab_for_tree_code > requires us to decide if we want optab_vector or optab_scalar. It is likely > not worth it to handle these four, they would fail soon anyway. > > Bootstrapped/regtested

Re: [PATCH] Allow new ISL/CLooG versions

On Tue, 15 Jan 2013, Richard Biener wrote: > On Mon, 14 Jan 2013, Jack Howarth wrote: > > > On Mon, Jan 14, 2013 at 08:27:12PM +0100, Dominique Dhumieres wrote: > > > In order to bootstrap r195167 with the new ISL/CLooG versions, > > > I had to apply the following patch: > > > > > > --- ../work/

Re: [RFC, x86] Changes for AVX and AVX2 processors

Hi, this was checked in: http://gcc.gnu.org/ml/gcc-cvs/2013-01/msg00442.html Thanks, K On Fri, Jan 11, 2013 at 4:38 PM, Vladimir Yakovlev wrote: > Kirill, > > Could you commit patch? > > 2013-01-11 Vladimir Yakovlev > > * config/i386/i386-c.c (ix86_target_macros_internal): New case. >

[SH] re-fix the sp_switch attribute

Hello. It seems that the .md part from the patch posted in http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01797.html is missing after rev 150306 Thus compiling the sp_switch attribute documentation example still fail with an "unrecognizable insn" error. However, after re-applying the missing part

Re: [PATCH][RFC] Fix PR55964

On Mon, Jan 14, 2013 at 04:58:09PM +0100, Richard Biener wrote: > I happen to have a patch for PR55964 around - preparatory work > to make loop distribution (and vectorization) handle nested loops. > It mostly kills the broken way loop distribution copies loops > (which is where we ICE in this PR).

Re: Fix Thumb-1 jump table alignment length calculation (PR target/43961)

On 13/01/13 22:37, Joseph S. Myers wrote: Bug 43961 (a regression in 4.5 and later) is the ARM back end generating an out-of-range Thumb-1 branch because the calculation (in architecture-independent code) of the size of a jump table fails to allow for alignment. Jump table alignment should be al

[PATCH] Fix PR54767

This fixes PR54767 - in VRP we may not propagate symbolical range information (such as equivalences - see the original fix for PR53465) across backedges. The following patch makes sure we don't by cleaning all such information when visiting PHI nodes for incoming edges marked with EDGE_DFS_BACK.

Re: [PATCH] Fix PR55882

On Fri, 11 Jan 2013, Richard Biener wrote: > On Thu, 10 Jan 2013, Eric Botcazou wrote: > > > > The problem with this intermediate inconsistency is that ... > > > > > > > That's why only the final result should matter and need be correct. > > > > > > ... this can't be ensured anymore. In some c

Re: [PATCH] Allow new ISL/CLooG versions

On Tue, Jan 15, 2013 at 11:05:51AM +0100, Richard Biener wrote: > 2013-01-15 Richard Biener > > PR other/55973 > * configure: Re-generate. > > config/ > * isl.m4 (ISL_INIT_FLAGS): Warn about disabled version check > for in-tree build. > (ISL_CHECK_VERSION):

[PATCH] Fix PR55961, do not include gmp.h for building host tools

This makes sure to not include gmp.h for building generator programs on the host (and thus require gmp.h on the host). Configury might need updating for the use of GENERATOR_FILE from system.h as well, but the patch below is a strict improvement over the current situation, restoring previous beha

Re: [PATCH] Allow new ISL/CLooG versions

On Tue, Jan 15, 2013 at 11:05:51AM +0100, Richard Biener wrote: > On Tue, 15 Jan 2013, Richard Biener wrote: > > > On Mon, 14 Jan 2013, Jack Howarth wrote: > > > > > On Mon, Jan 14, 2013 at 08:27:12PM +0100, Dominique Dhumieres wrote: > > > > In order to bootstrap r195167 with the new ISL/CLooG v

Re: [PATCH] Fix PR55882

> I have applied the minimal fix to trunk now. We retain the clearly > incorrect fact that the !align_computed path only ever increases > alignment (it's only ever set previously to either BITS_PER_UNIT > or via the MEM_REF block). I will momentarily test a followup > removing that block for trun

Re: [PATCH] Fix PR55882

On Tue, 15 Jan 2013, Eric Botcazou wrote: > > I have applied the minimal fix to trunk now. We retain the clearly > > incorrect fact that the !align_computed path only ever increases > > alignment (it's only ever set previously to either BITS_PER_UNIT > > or via the MEM_REF block). I will momenta

[patch,avr] Tidy builtins.def

This patch factors out stuff from builtins.def that is used again and again. Strings like the "__builtin_avr_" common to all builtins is moved to the implementation of DEF_BUILTIN. Similar for ICODE and CODE_FOR_. ID is no more needed, the needed enums are derived from NAME. There is no function

Re: [PATCH 1/2] [pr53679] libgo: add a --enable-werror configure flag

On Sun, Dec 23, 2012 at 3:30 PM, Mike Frysinger wrote: > diff --git a/libgo/configure.ac b/libgo/configure.ac > index 8cde50b..63d8cbc 100644 > --- a/libgo/configure.ac > +++ b/libgo/configure.ac > @@ -50,8 +50,11 @@ AC_PROG_AWK > WARN_FLAGS='-Wall -Wextra -Wwrite-strings -Wcast-qual' > AC_SUBS

Re: [PATCH] Fix PR55961, do not include gmp.h for building host tools

On Tue, Jan 15, 2013 at 5:16 AM, Richard Biener wrote: > 2013-01-15 Richard Biener > > PR bootstrap/55961 > * system.h: Do not include gmp.h for building host tools. Looks OK to me. Ian

[PATCH] Ensure matching types in debug stmts created by tree-sra.c (PR tree-optimization/55920)

Hi! This patch fixes ICE on the attached testcase on ARM, where we ended up creating invalid debug stmt, which had SImode var on the lhs, but BLKmode MEM_REF on the rhs. Fixed by matching (almost, with the build_ref_for_model -> build_debug_ref_for_model change) what we do for normal stmt modific

Re: PING: replace gcc.target/arm/ftest-*.c with compile-only tests

Janis Johnson writes: > Richard, > > In December I submitted a patch to replace ftest*.c in gcc.target/arm with > compile-only tests: . > There has been no response to the patch. > > One reason for the replacement is that the arch_v*_multil

[PATCH] No SRA debug replacements for variables that cannot be removed

Hi, as discussed in PR 55920, this patch prevents SRA from creating debug replacements and debug statements for aggregate variables which cannot be scalarized away. Jakub bootstrapped and tested the patch and I suppose he is OK with it too so I will commit is shortly. Thanks, Martin 2013-01-1

Re: [PATCH] No SRA debug replacements for variables that cannot be removed

On Tue, Jan 15, 2013 at 04:52:16PM +0100, Martin Jambor wrote: > 2013-01-11 Martin Jambor > > PR tree-optimization/55920 > * tree-sra.c (analyze_access_subtree): Do not mark non-removable > accesses as grp_to_be_debug_replaced. Ok, thanks. > --- src.orig/gcc/tree-sra.c > +++

Re: [PATCH] Ensure matching types in debug stmts created by tree-sra.c (PR tree-optimization/55920)

On Tue, 15 Jan 2013, Jakub Jelinek wrote: > Hi! > > This patch fixes ICE on the attached testcase on ARM, where we ended up > creating invalid debug stmt, which had SImode var on the lhs, but BLKmode > MEM_REF on the rhs. Fixed by matching (almost, with the build_ref_for_model > -> build_debug_r

[PATCH, AArch64] Make zero_extends explicit for some SImode patterns

Greetings! I've made zero_extend versions of SI mode patterns that write to W registers in order to make the implicit zero_extend that they do explicit, so GCC can be smarter about when it actually needs to plant a zero_extend (uxtw). If that sounds familiar, it's because this patch continues th

Re: [testsuite] fix ARM test gcc.target/arm/neon-vld1_dupQ.c

On 01/14/2013 03:04 PM, Janis Johnson wrote: > Test gcc.target/arm/neon-vld1_dupQ.c started failing with r194594, a C > front end change that causes the test to get warnings. The test passes > local variables of type int64x1_t to functions declared with arguments > of type int64_t. This patch cas

dg-extract-results.sh truncates logs containing "Running "

The regex to detect the beginning of a test file execution was too broad, matching any line beginning with "Running ". This caused dg-extract-results.sh to truncate some test logs, including gdb.base/help.exp in the GDB test suite (which tests "help running", printing "Running the program.") The a

Re: [PATCH, AArch64] Make zero_extends explicit for some SImode patterns

On 15/01/13 16:13, Ian Bolton wrote: Greetings! I've made zero_extend versions of SI mode patterns that write to W registers in order to make the implicit zero_extend that they do explicit, so GCC can be smarter about when it actually needs to plant a zero_extend (uxtw). If that sounds familiar

Re: [testsuite] skip ARM tests fma.c and fma-sp.c for conflicting options

On 01/14/13 23:03, Janis Johnson wrote: The options specified for gcc.target/arm/fma.c and fma-sp.c can conflict with several multilib options. This patch skips the tests for multilibs with conflicting options, and it adds option "-mfloat-abi=hard" which is needed for the test. OK for trunk? J

patch to fix PR55153

The following patch fixes http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55153 The reason for the crash was in moving prefetch after return insn. That is because there was no dependency from prefetch to the return as the prefetch did not generate any pending memory or register and MOVE_BARRIER

Re: [SH] re-fix the sp_switch attribute

Hi, On Tue, 2013-01-15 at 11:16 +0100, Christian Bruel wrote: > Hello. > > It seems that the .md part from the patch posted in > http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01797.html > is missing after rev 150306 > > Thus compiling the sp_switch attribute documentation example still fail > wit

Re: [PATCH] Fix PR55876 - Make generation of paradoxical subreg in widen_operand more robust

On 01/14/2013 12:27 AM, Tom de Vries wrote: 2013-01-14 Tom de Vries PR target/55876 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG. Update comment. Ok. r~

Re: dg-extract-results.sh truncates logs containing "Running "

& now with an actual GCC patch, instead of the GDB one - sorry about that. On Tue, Jan 15, 2013 at 8:31 AM, David Blaikie wrote: > The regex to detect the beginning of a test file execution was too > broad, matching any line beginning with "Running ". This caused > dg-extract-results.sh to trunca

Fwd: summarize kfails in dg-extract-results.sh

For some reason KFAILs weren't being summarized by dg-extract-results.sh. I assume this was just an oversight, so here's a patch to add it alongside all the other result types. I encountered this in the GDB test suite but was asked to submit the change upstream to GCC as GDB would prefer to avoid

Re: [PATCH 1/2] [pr53679] libgo: add a --enable-werror configure flag

On Tuesday 15 January 2013 09:56:06 Ian Lance Taylor wrote: > On Sun, Dec 23, 2012 at 3:30 PM, Mike Frysinger wrote: > > diff --git a/libgo/configure.ac b/libgo/configure.ac > > index 8cde50b..63d8cbc 100644 > > --- a/libgo/configure.ac > > +++ b/libgo/configure.ac > > @@ -50,8 +50,11 @@ AC_PROG_A

[google gcc-4_7] Backport r195207 (fix for PR55982) into google/gcc-4_7

Ok for google/gcc-4_7 ? Ref b/8003094 Thanks, -- Paul Pluzhnikov 2013-01-15 Paul Pluzhnikov PR 55982 * strncat-chk.c (__strncat_chk): Fix loop unroll. Index: libssp/strncat-chk.c === --- libssp/strncat-chk.c

Re: [google gcc-4_7] Backport r195207 (fix for PR55982) into google/gcc-4_7

On Tue, Jan 15, 2013 at 1:42 PM, Paul Pluzhnikov wrote: > Ok for google/gcc-4_7 ? > > Ref b/8003094 > > Thanks, > -- > Paul Pluzhnikov > > > 2013-01-15 Paul Pluzhnikov > > PR 55982 > * strncat-chk.c (__strncat_chk): Fix loop unroll. > > > Index: libssp/strncat-chk.c > ==

[testsuite] add option to LTO flags for c-torture/execute/builtins tests

Most of the tests in gcc.c-torture/execute/builtins that use LTO torture options "-O2 -flto -fuse-linker-plugin -fno-fat-lto-objects" fail to link on EABI and ELF targets with multiple definitions for either memset or strlen. I filed PR testsuite/55994 and was told that the fix is to run the tests

Re: Fwd: summarize kfails in dg-extract-results.sh

On 01/15/2013 09:24 AM, David Blaikie wrote: > For some reason KFAILs weren't being summarized by > dg-extract-results.sh. I assume this was just an oversight, so here's > a patch to add it alongside all the other result types. > > I encountered this in the GDB test suite but was asked to submit t

Re: dg-extract-results.sh truncates logs containing "Running "

On 01/15/2013 09:16 AM, David Blaikie wrote: > & now with an actual GCC patch, instead of the GDB one - sorry about that. > > On Tue, Jan 15, 2013 at 8:31 AM, David Blaikie wrote: >> The regex to detect the beginning of a test file execution was too >> broad, matching any line beginning with "Run

Re: [PATCH, RFA]: Properly mask memory model argument with MEMMODEL_MASK

Uros Bizjak writes: > > 2013-01-15 Uros Bizjak > > * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument > with MEMMODEL_MASK before comparing with MEMMODEL_* memory types. > * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto. > (expand_mem_thread_f

Re: [PATCH, RFA]: Properly mask memory model argument with MEMMODEL_MASK

On 01/15/2013 12:09 AM, Uros Bizjak wrote: 2013-01-15 Uros Bizjak * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument with MEMMODEL_MASK before comparing with MEMMODEL_* memory types. * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto. (expand

Re: testsuite: missing or wrong dg-* directives?

Le 14/01/2013 23:16, Harald Anlauf a écrit : Attached there is a partial patch for the obvious parts, plus other missed cases (missing options). No failures, just a few more passes from the fixed dg-do run's. Thanks, applied as revision 195217. Mikael

Re: testsuite: missing or wrong dg-* directives?

Attached there is a partial patch for the obvious parts, plus other missed cases (missing options). No failures, just a few more passes from the fixed dg-do run's. 2013-01-14 Manfred Schwarb Harald Anlauf * gfortran.dg/bounds_check_4.f90: Add dg-options "-fbounds-check".

[PATCH] Fix indention in vax.md

Hi! While searching for a VAX-related ICE, I saw that some indention could be fixed a bit. Committed as obvious. 2013-01-15 Jan-Benedict Glaw * config/vax/vax.md (add3, sub3, mul3, div3, and3, *and_const_int, ior3, xor3, ashrsi3, *call): Fix indention. diff --git a/

Re: testsuite: missing or wrong dg-* directives?

Am 14.01.2013 20:49, schrieb Mike Stump: On Jan 14, 2013, at 6:23 AM, Mikael Morin wrote: Le 14/01/2013 00:37, Manfred Schwarb a écrit : Am 14.01.2013 00:10, schrieb Manfred Schwarb: There are several other oddities: d_g-final, braces have to be separated by spaces. Want to send a patch?

[testsuite] fix loop index for gcc.dg/webizer.c

Execution of test gcc.dg/webizer.c fails with a segfault for powerpc-eabi with the GNU simulator. The test has an array of size 2 and accesses that array with indices of 1 and 2. This patch fixes the bounds of the loop index. Tested on powerpc-none-eabi; OK for trunk? Janis 2013-01-15 Janis Jo

Re: [testsuite] fix loop index for gcc.dg/webizer.c

On 01/15/2013 02:57 PM, Janis Johnson wrote: Execution of test gcc.dg/webizer.c fails with a segfault for powerpc-eabi with the GNU simulator. The test has an array of size 2 and accesses that array with indices of 1 and 2. This patch fixes the bounds of the loop index. Tested on powerpc-none-

Re: [testsuite] fix loop index for gcc.dg/webizer.c

On 01/15/2013 02:01 PM, Jeff Law wrote: > On 01/15/2013 02:57 PM, Janis Johnson wrote: >> Execution of test gcc.dg/webizer.c fails with a segfault for >> powerpc-eabi with the GNU simulator. The test has an array of size 2 >> and accesses that array with indices of 1 and 2. This patch fixes the >

Re: [testsuite] fix loop index for gcc.dg/webizer.c

On 01/15/2013 03:08 PM, Janis Johnson wrote: On 01/15/2013 02:01 PM, Jeff Law wrote: On 01/15/2013 02:57 PM, Janis Johnson wrote: Execution of test gcc.dg/webizer.c fails with a segfault for powerpc-eabi with the GNU simulator. The test has an array of size 2 and accesses that array with indic

Re: [PATCH] Fix PR55833 + cheaper checking

Hi, > > Yes, you should check whether you are in an irreducible loop. This is done > > by > > testing EDGE_IRREDUCIBLE_LOOP flag, > > Alright, I was wondering whether there's any other way. Unfortunately, > here I couldn't do something like > > if (loop_preheader_edge (loop)->flags & EDGE_IRR

Re: [SH] re-fix the sp_switch attribute

Christian Bruel wrote: > This patch fixes the sp_switch unspec pattern to be recognized by > broken_move, where it is now handled to force a dump_table when necessary. > > The example from the documentation now compiles fine. Added it to the > regression tests. OK for trunk ? OK with the PR numb

[PATCH] Fix shrink-wrapping with vDRAP (PR target/55940)

Hi! As the following testcase shows, even when stack_realign_drap we might need to prevent crtl->drap_reg accesses in the bbs considered for shrink-wrapping, even if reload decides stack realignment isn't needed, the vDRAP (in the testcase %edi) can be used by the first bbs and initialized only la

Re: Fwd: summarize kfails in dg-extract-results.sh

On Tue, Jan 15, 2013 at 12:22 PM, Janis Johnson wrote: > On 01/15/2013 09:24 AM, David Blaikie wrote: >> For some reason KFAILs weren't being summarized by >> dg-extract-results.sh. I assume this was just an oversight, so here's >> a patch to add it alongside all the other result types. >> >> I en

Re: [PATCH] Fix shrink-wrapping with vDRAP (PR target/55940)

On 01/15/2013 02:38 PM, Jakub Jelinek wrote: 2013-01-15 Jakub Jelinek PR target/55940 * function.c (thread_prologue_and_epilogue_insns): Always add crtl->drap_reg to set_up_by_prologue.set, even if stack_realign_drap is false. * gcc.dg/pr55940.c: New tes

Re: [PATCH] Remove clang-ism -Wno-c99-extensions from libsanitizer AM_CXXFLAGS

On 01/15/2013 01:10 AM, Jakub Jelinek wrote: 2013-01-15 Jakub Jelinek * sanitizer_common/Makefile.am (AM_CXXFLAGS): Remove -Wno-c99-extensions. * interception/Makefile.am (AM_CXXFLAGS): Likewise. * asan/Makefile.am (AM_CXXFLAGS): Likewise. * sanitizer_com

Re: [PATCH] Fix PR55882

> Not necessarily. The following is a 4.7 variant of the patch > (on trunk get_object_alignment_1 got one extra output which > moved the align return value to by-reference). OK, I obviously didn't try very hard here... > Can you give it testing on the branch and a strict-align target? > > Thank

Re: [testsuite] add option to LTO flags for c-torture/execute/builtins tests

On Tue, 15 Jan 2013, Janis Johnson wrote: > Most of the tests in gcc.c-torture/execute/builtins that use LTO torture > options "-O2 -flto -fuse-linker-plugin -fno-fat-lto-objects" fail to > link on EABI and ELF targets with multiple definitions for either memset > or strlen. I filed PR testsuite/5

[PR55547] fix alias regression on alpha on misaligned symbols (was: Re: do you have time to review this alpha P1 patch?)

On Jan 15, 2013, Richard Henderson wrote: > On 01/15/2013 08:24 AM, Aldy Hernandez wrote: >> Ok, it's really an alias.c bug, but it is Alpha, and aoliva has already >> provided an unreviewed patch... >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55547 > The patch in #C4 is ok. Thanks, I'm che

Allow widening multiplication in tree-ssa/slsr-*.c

Hi, The slsr-*.c testsuite failures are due to not matching widening add (or multiply-add) in AARCH64. The LP64 targets generates WIDEN_MULT_PLUS_EXPR which is correct and even better code generation will be produced. However, the testcase expects +OP2 in these cases and hence FAIL's. Please find

Re: Allow widening multiplication in tree-ssa/slsr-*.c

On Tue, Jan 15, 2013 at 9:21 PM, Hurugalawadi, Naveen wrote: > Hi, > > The slsr-*.c testsuite failures are due to not matching widening add > (or multiply-add) in AARCH64. > The LP64 targets generates WIDEN_MULT_PLUS_EXPR which is correct and > even better code generation will be produced. However

[google gcc-4_7] Inlining and devirtualization tests

David, This patch adds tests for inlining and devirtualization optimizations, some of which are already in google's gcc-4_7 branch. It is only right to add tests for these optimizations. Main gcc-4_7 branch and trunk pass tests 1 and 5, while Google's gcc-4_7 branch also passes tests 2 and 3.

Re: [google gcc-4_7] Inlining and devirtualization tests

Looks fine. Why adding tests that are expected to fail? Are these tests passing with trunk? David On Tue, Jan 15, 2013 at 9:33 PM, Maxim Kuvyrkov wrote: > David, > > This patch adds tests for inlining and devirtualization optimizations, some > of which are already in google's gcc-4_7 branch.

Re: [PR55547] fix alias regression on alpha on misaligned symbols (was: Re: do you have time to review this alpha P1 patch?)

On Wed, Jan 16, 2013 at 5:29 AM, Alexandre Oliva wrote: >> On 01/15/2013 08:24 AM, Aldy Hernandez wrote: >>> Ok, it's really an alias.c bug, but it is Alpha, and aoliva has already >>> provided an unreviewed patch... > >>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55547 > >> The patch in #C4 is