Hi!
On Mon, Feb 27, 2012 at 03:28:32PM -0800, Benjamin Kosnik wrote:
On Tue, 28 Feb 2012 00:12:33 +0100
Jakub Jelinek ja...@redhat.com wrote:
and $(EGREP) -v '#(#| |$$)' just throws away the whole
};# Appended to version file.
line. I wonder if
sed -e 's/#[# $].*$//'
wouldn't be
Il 28/02/2012 09:50, Jakub Jelinek ha scritto:
- $(EGREP) -v '#(#| |$$)' $@.tmp | \
+ $(EGREP) -v '^[ ]*#(#| |$$)' $@.tmp | \
I don't know this part very well, so I wonder why you have to remove
comments at all... hence I wonder if sed 's/##.*//;s/# .*//;s/#$//'
(alternation is
On Mon, 27 Feb 2012, Mike Stump wrote:
On Feb 27, 2012, at 1:01 PM, Jack Howarth wrote:
Since this is just a testsuite issue in boehm-gc, wouldn't this be Hans'
call? I would definitely
like to get this fixed in gcc 4.7 so that the boehm-gc testsuite doesn't
suffer timeouts. Also
On Mon, 27 Feb 2012, Richard Guenther wrote:
This makes the patch for PR50444 less conservative by also looking
at TYPE_ALIGN of the base we offset. I guess we do not need to
worry about the '???' as IPA SRA uses sth different (see PR52402)
and the only case I definitely see only creates
On Tue, Feb 28, 2012 at 09:57:38AM +0100, Paolo Bonzini wrote:
Il 28/02/2012 09:50, Jakub Jelinek ha scritto:
- $(EGREP) -v '#(#| |$$)' $@.tmp | \
+ $(EGREP) -v '^[ ]*#(#| |$$)' $@.tmp | \
I don't know this part very well, so I wonder why you have to remove
comments at all...
Il 28/02/2012 10:19, Jakub Jelinek ha scritto:
- $(EGREP) -v '#(#| |$$)' $@.tmp | \
+ $(EGREP) -v '^[ ]*#(#| |$$)' $@.tmp | \
I don't know this part very well, so I wonder why you have to remove
comments at all... hence I wonder if sed 's/##.*//;s/# .*//;s/#$//'
On Mon, Feb 27, 2012 at 6:29 PM, Aldy Hernandez al...@redhat.com wrote:
On 02/27/12 11:02, Michael Matz wrote:
Hi,
On Mon, 27 Feb 2012, Aldy Hernandez wrote:
For that matter, didn't rth add a memory barrier at the beginning of
transactions last week? That would mean that we can't hoist
On Mon, Feb 27, 2012 at 5:44 PM, Andrew MacLeod amacl...@redhat.com wrote:
On 02/27/2012 11:22 AM, Aldy Hernandez wrote:
Ok. I see. So, I think what would be best is to have a way to check
whether
a store/load is part of a transaction - do we have a way to do that right
now?
(For example
On Tue, 28 Feb 2012, Martin Jambor wrote:
Hi,
the first patch in the series deals with plain MEM_REFs on LHS of
assignments to handle situations such as the testcase which currently
fails on strict alignment platforms (the array and the loop are there
to cross a cache line on ia64 so that
On Tue, 28 Feb 2012, Martin Jambor wrote:
Hi,
the second patch in the series handles MEM_REFs on LHS which are parts
of a handled_component, usually a COMPONENT_REF. The failing testcase
which requires it is not actually the one in the patch but it is
gcc.c-torture/execute/mayalias-3.c
On Tue, 28 Feb 2012, Martin Jambor wrote:
Hi,
this patch fixes misaligned reads through MEM_REFs such as the one in
the testcase which currently fails on both sparc64 and ia64 (again,
the array and the loop are there to cross ia64 cache line and fail
there too). The patch has to be
On Tue, Feb 28, 2012 at 9:25 AM, Thomas Koenig tkoe...@netcologne.de wrote:
Hello world,
the attached patch documents -floop-flatten as unsafe (see the PR).
Tested with 'make dvi', 'make info' and 'make pdf'. I'd like to
get this into the 4.7 release.
OK for trunk?
I have modified your
This patch is addendum to support skip of 2-word instructions as of PR49939 and
the extension made in
http://gcc.gnu.org/viewcvs?view=revisionrevision=177049
gas emits a warning if it sees a skip of 2-word instruction for cores that
might have the skip bug. However, only specific devices are
Thanks Jakub for taking care of this.
Paolo.
This fixes element ordering for V2SFmode, V2SImode and V2DImode
concats in ix86_expand_vector_set.
Boostrap / regtest pending on x86_64-unknown-linux-gnu.
Ok everywhere? (the testcase only fails on trunk)
Thanks,
Richard.
2012-02-28 Richard Guenther rguent...@suse.de
PR
Hello!
This fixes element ordering for V2SFmode, V2SImode and V2DImode
concats in ix86_expand_vector_set.
Boostrap / regtest pending on x86_64-unknown-linux-gnu.
Ok everywhere? (the testcase only fails on trunk)
2012-02-28 Richard Guenther rguent...@suse.de
PR target/52407
On Tue, 28 Feb 2012, Uros Bizjak wrote:
Hello!
This fixes element ordering for V2SFmode, V2SImode and V2DImode
concats in ix86_expand_vector_set.
Boostrap / regtest pending on x86_64-unknown-linux-gnu.
Ok everywhere? (the testcase only fails on trunk)
2012-02-28 Richard
On Tue, Feb 28, 2012 at 02:01:45PM +0100, Richard Guenther wrote:
I guess using __INT64_TYPE__ is more portable to weird targets
though. I integrated the other testcase changes though.
To which weird targets? We support long long on all targets,
and the C standard guarantees it is at least
On Tue, 28 Feb 2012, Jakub Jelinek wrote:
On Tue, Feb 28, 2012 at 02:01:45PM +0100, Richard Guenther wrote:
I guess using __INT64_TYPE__ is more portable to weird targets
though. I integrated the other testcase changes though.
To which weird targets? We support long long on all targets,
This patch fixes a regression that we've discovered in GCC's conformance
to the ARM ABI for the hard-float calling convention. The problem
occurred because of changes in the mid-end in the way we mark some
structures as having a machine mode now rather than BLKmode. We missed
this happening
2012/2/28 Georg-Johann Lay a...@gjlay.de:
This patch is addendum to support skip of 2-word instructions as of PR49939
and
the extension made in
http://gcc.gnu.org/viewcvs?view=revisionrevision=177049
gas emits a warning if it sees a skip of 2-word instruction for cores that
might have the
On 02/28/2012 04:39 AM, Richard Guenther wrote:
On Mon, Feb 27, 2012 at 5:44 PM, Andrew MacLeodamacl...@redhat.com wrote:
I t would seem appropriate to me that in the future, perhaps the CFG could
have a flag set for any basic block which is in a transaction... This would
make it pretty
Richard, are you going to check in these deprecations for 4.7 (with the
additional deprecation of old-ABI uClinux as I noted in
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01260.html)? Time is
rapidly running out for doing so.
Note that current glibc ports no longer supports ARM old-ABI; I
Richard Guenther wrote:
On Fri, Feb 24, 2012 at 2:16 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
this looks to me that, since the check is intended to verify that
misalignment remains the same throughout the execuction of the loop,
we actually want to check the inner-loop step here as
On Tue, 28 Feb 2012, Ulrich Weigand wrote:
Richard Guenther wrote:
On Fri, Feb 24, 2012 at 2:16 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
this looks to me that, since the check is intended to verify that
misalignment remains the same throughout the execuction of the loop,
we
I am testing the following patch to fix PR52406. We cannot simply
feed DR_BASE_OBJECT to the alias-oracle as it does not reflect
a real memory access. Only if we query two 'structurally compatible'
references we may do this. So the following patch reverts to
using DR_REF instead and improves
Richard Guenther wrote:
On Mon, Feb 20, 2012 at 11:19 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
we've noticed that the loop optimizer sometimes inserts weirdly
inefficient code to compute the value of an induction variable
at the end of the loop.
[snip]
The issue is that (start + 1) +
On Fri, 2012-01-27 at 09:29 +1100, Richard Henderson wrote:
Two of the patches have been posted here before; the libstdc++
patch was approved by Benjamin.
All of the patches tested on sparc64-linux, and sanity checked
on x86_64-linux. I've cross-compiled for m68k-linux, but I've
only been
On Tue, Feb 28, 2012 at 4:10 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
Richard Guenther wrote:
On Mon, Feb 20, 2012 at 11:19 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
we've noticed that the loop optimizer sometimes inserts weirdly
inefficient code to compute the value of an induction
The pattern to match a big-endian machine for linux is ambiguous as
reported by the PR and can cause some little-endian triplets to be
confused as big-endian.
This patch makes the string unambiguous.
R.
* config.gcc (arm*-*-linux*): Use an unambiguous pattern for
detecting
On 16/12/11 18:53, Joseph S. Myers wrote:
On Fri, 9 Dec 2011, Richard Earnshaw wrote:
I think we've reached the point where the following target
configurations should be End-of-Life'd. As such, I'd like to mark them
as deprecated in gcc-4.7, prior to removal after the branch.
I agree
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
So how about we just deprecate all targets that are still using FPE format?
This patch appears to add back the i[34567]86-*-interix3* deprecation -
removed deliberately when the target got a maintainer - and
arm-wrs-vxworks is *not* using FPE
On 28/02/12 15:38, Joseph S. Myers wrote:
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
So how about we just deprecate all targets that are still using FPE format?
This patch appears to add back the i[34567]86-*-interix3* deprecation -
Oops, botched merge conflict.
removed deliberately
On 21/02/12 16:17, Matthew Gretton-Dann wrote:
The attached patch fixes instruction generation for unsigned vector
comparisons against a known-zero vector.
ARM's Neon extensions only allow unsigned equality comparison against
unsigned zero, not less than or greater than comparisons.
This
Hi all,
This patch implements 64-bit immediate constant loads in NEON.
The current state is that you can load const_vector, but not const_int.
This is clearly not ideal. The result is a constant pool entry when it's
not necessary.
The patch disables the movdi_vfp patterns for loading DImode
On 28/02/12 15:51, Richard Earnshaw wrote:
On 28/02/12 15:38, Joseph S. Myers wrote:
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
So how about we just deprecate all targets that are still using FPE format?
This patch appears to add back the i[34567]86-*-interix3* deprecation -
Oops,
On 02/27/12 08:22, Aldy Hernandez wrote:
transform by making transaction load/store stmts behave the same as
potentially trapping stmts (thus, only optimize if the memory is accessed
unconditional somewhere else). That would work for PRE as well.
[easiest would be to make *_could_trap_p
This patch improves existing peephole optimizations that merge individual
LDRs into LDM, in the case that the order of registers in LDR instructions
is not ascending, but the loaded values can be reordered because their uses
commute.
There are two changes:
* use rtx__equal_p to compare operands
The following patch adds reload behaviour to LRA for some cases. Some
targets assumes such behaviour.
The patch was successfully bootstrapped on x86/x86-64.
Committed as rev. 184631.
2012-02-28 Vladimir Makarov vmaka...@redhat.com
* lra-constraints.c (check_and_process_move):
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
Here's an updated patch and a suggested web-page patch:
uClinux targets do not have -gnu in their target triplets.
You missed the WinCE port from the list in the webpage patch of ports
being deprecated for using FPA floating-point format and
On Mon, 27 Feb 2012, Bernhard Reutner-Fischer wrote:
2010-01-03 Bernhard Reutner-Fischer al...@gcc.gnu.org
* gcc.c (display_help): Document --help=common and sort entries
alphabetically.
I would liberaly label this a documentation change and commit it
so that it makes GCC
On 28/02/12 17:10, Joseph S. Myers wrote:
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
Here's an updated patch and a suggested web-page patch:
uClinux targets do not have -gnu in their target triplets.
You missed the WinCE port from the list in the webpage patch of ports
being
On 02/28/12 11:05, Richard Henderson wrote:
On 02/27/12 08:22, Aldy Hernandez wrote:
transform by making transaction load/store stmts behave the same as
potentially trapping stmts (thus, only optimize if the memory is accessed
unconditional somewhere else). That would work for PRE as well.
Hi all,
This patch adds a DImode negate pattern for NEON.
Unfortunately, the NEON vneg instruction only supports vectors, not
singletons, so there's no direct way to do it in DImode, and the
compiler ends up moving the value back to core registers, negating it,
and returning to NEON
After a few recent changes, I found that arm bootstrap broken again. I
found a latent bug whose occurrence is triggered by recent changes. The
following patch mostly fixes the bug.
The patch was successfully bootstrapped on x86/x86-64 (arm bootstrap
result will be available only this
On 02/28/12 09:44, Aldy Hernandez wrote:
PR middle-end/51752
* gimple.h (gimple_in_transaction): New.
(gimple_set_in_transaction): New.
(struct gimple_statement_base): Add in_transaction field.
* tree-ssa-loop-im.c: (movement_possibility): Restrict movement of
The following two mainline commits from Andreas fix the two ICEs mentioned
in PR50181 on the 4.5 and 4.6 release branches. Speaking with Uli offline,
he thought the two patches should be safe enough to backport.
This is a regression from GCC 4.4, so is this ok for 4.6 and 4.5 assuming my
On 02/28/12 13:12, Richard Henderson wrote:
On 02/28/12 09:44, Aldy Hernandez wrote:
PR middle-end/51752
* gimple.h (gimple_in_transaction): New.
(gimple_set_in_transaction): New.
(struct gimple_statement_base): Add in_transaction field.
*
Jack,
Can I ask you to close this PR?
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52042
Indeed, my bugzilla account is a simple one and I cannot to change the
bug status...
--
Patrick
On 02/07/2012 10:36 PM, Patrick Marlier wrote:
Hi,
The problem in this PR is that with PIE, getsectdata
On Feb 28, 2012, at 12:20 PM, Patrick Marlier wrote:
Can I ask you to close this PR?
Done.
as requested. all new symbols should be in new symbol versoning names.
With this in, check-abi will now fail if new symbols are inadvertently
added to previously-released versions.
I'm expecting this to make solaris symbol versioning, as reported by
check-abi, fail at first. Then we can
On Mon, Feb 27, 2012 at 04:38:01PM +0100, Richard Guenther wrote:
This fixes PR52402 in a similar way as PR50444. IPA SRA needs to be
careful about alignment when constructing accesses.
Bootstrap and regtest pending on x86_64, any comments?
The test fails on i386-linux, adding -w -Wno-psabi
Hi,
The problem in this PR is that the CLASSTYPE_SORTED_FIELDS is created
too early (in finish_struct_1) to handle those late unscoped enum
definitions.
Consequently, I propose to lately add those names in
CLASSTYPE_SORTED_FIELDS when they are encountered, in
build_enumerator.
Tested
On Tue, Feb 28, 2012 at 12:34:34PM -0800, Benjamin De Kosnik wrote:
as requested. all new symbols should be in new symbol versoning names.
With this in, check-abi will now fail if new symbols are inadvertently
added to previously-released versions.
I'm expecting this to make solaris symbol
Minor nit: the correct spelling is threshold, not threashold.
Thanks,
Paolo.
The libgo runtime was calling _exit when a Go program finished. There
was no particular reason for this, and it meant that --coverage did not
work becase the .gcda file was never generated. This patch fixes that.
Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu.
Committed to
On 02/28/2012 03:36 PM, Fabien Chêne wrote:
* decl.c (build_enumerator): Call
insert_into_classtype_sorted_fields if an enumerator-definition
referring
to a class scope opaque enum has been encountered.
This will insert and sort the vector again for each enumerator; we
2012/2/28 Jason Merrill ja...@redhat.com:
On 02/28/2012 03:36 PM, Fabien Chêne wrote:
* decl.c (build_enumerator): Call
insert_into_classtype_sorted_fields if an enumerator-definition
referring
to a class scope opaque enum has been encountered.
This will insert and
Hello,
The attached patch adds some SH target options that have been missing in
the documentation.
Tested with 'make info dvi pdf'.
OK to commit?
Cheers,
Oleg
2012-02-28 Oleg Endo olege...@gcc.gnu.org
* doc/invoke.texi (-msoft-atomic): Add more detailed
description.
In message 4f4d0e64.1020...@arm.com
Richard Earnshaw rearn...@arm.com wrote:
On 28/02/12 17:10, Joseph S. Myers wrote:
On Tue, 28 Feb 2012, Richard Earnshaw wrote:
Here's an updated patch and a suggested web-page patch:
uClinux targets do not have -gnu in their target
Oleg Endo oleg.e...@t-online.de wrote:
The attached patch adds some SH target options that have been missing in
the documentation.
Tested with 'make info dvi pdf'.
OK to commit?
Looks OK to me. Thanks for the patch.
Regards,
kaz
Hi!
Bootstrap fails on ia64-linux with Ada currently, because we get
a -fcompare-debug failure on one of the files, where a debug_insn
is the only insn that references in5 hard register (which is live on
entry of the first bb). hard_regs_live_count array is only used to compute
regs_ever_live,
On Tue, Feb 28, 2012 at 09:34:37PM +0100, Jakub Jelinek wrote:
On Mon, Feb 27, 2012 at 04:38:01PM +0100, Richard Guenther wrote:
This fixes PR52402 in a similar way as PR50444. IPA SRA needs to be
careful about alignment when constructing accesses.
Bootstrap and regtest pending on
Richard Guenther wrote:
On Tue, 28 Feb 2012, Ulrich Weigand wrote:
On powerpc64-linux, testing is still ongoing (I had to re-do the
baseline run ... b.t.w. this showed that the vect-outer- tests
also fail on this target currently).
OK once testing has completed successfully on powerpc?
On Tue, Feb 28, 2012 at 05:09:05PM -, Greta Yorsh wrote:
Is it OK for GCC 4.7 Stage 4 ?
Technically this is a regression in 4.7 compared to 4.6,
so I'd like to get this in.
However given the stage we are and given that it's not a
correctness issue, I would defer to the RMs.
In any case
On Tue, Feb 21, 2012 at 11:37:07AM -0800, Mike Stump wrote:
On Feb 21, 2012, at 10:26 AM, Ulrich Weigand wrote:
I just noticed that check_effective_target_vect_condition returns false
for ARM/NEON, even though the platforms in fact supports vectorized
conditional expressions. This causes a
On Fri, Feb 24, 2012 at 10:53:35PM -0500, Matt Turner wrote:
The x86/amd64 mmintrin.h provides the _mm_empty intrinsic for the 'emms'
MMX instruction. Although ARM does not need such an instruction, we
should provide an empty _mm_empty function nonetheless for source
compatibility.
OK for 4.8
On Wed, Feb 08, 2012 at 04:29:25PM +0800, Bin Cheng wrote:
Hi,
Julian Brown once posted a patch fixing ARM EABI violation, which I think
also essential to 4.6 branch.
I created a patch against 4.6 branch as attached. Is it ok to back port?
Ok after 4.6 reopens.
Ramana
[Sorry about the duplicate mail. My mailer seems to have eaten up the
original reply I sent. ]
On Tue, Feb 28, 2012 at 05:09:05PM -, Greta Yorsh wrote:
Is it OK for GCC 4.7 Stage 4 ?
This is stage4 - I'd like to hear what the RM's think. Technically
it's fixing a regression and is low risk
On Feb 28, 2012, at 3:39 PM, Jakub Jelinek wrote:
Here is a fix, bootstrapped/regtested on x86_64-linux and i686-linux,
ok for trunk?
Sure, but, I can't find -Wpsabi in the docs. :-( If -Wno-abi works, at least
it is documented. Ok either way.
On Feb 28, 2012, at 4:08 PM, Ramana Radhakrishnan wrote:
Certainly no objections to it - in fact I'd suggest it for 4.6 as well.
The backport is fine, if someone wants to do the work. Just let it bake a
little in the top of the tree as usual first.
This patch fixes PR52424 by pushing a missing marker onto
const_and_copies_stack before calling dom_thread_across_edge. This bug
has resulted in lost copy propagation opportunities in the presence of
edge threading.
Bootstrapped and tested with no regressions on powerpc64-linux-gnu. I'm
waiting
2012/2/28 Fabien Chêne fabien.ch...@gmail.com:
2012/2/28 Jason Merrill ja...@redhat.com:
On 02/28/2012 03:36 PM, Fabien Chêne wrote:
[...]
Will unqualified lookup work
because we're in the enumeration scope, or do we need to make lookup in the
class work?
Unqualified lookup works because
Regression fix okayed by Tobias Burnus on #gfortran and committed as
revision 184651.
Cheers
Paul
2012-02-29 Paul Thomas pa...@gcc.gnu.org
PR fortran/52386
* trans-expr.c (fcncall_realloc_result): Dereference the
descriptor if needed.
2012-02-29 Paul Thomas pa...@gcc.gnu.org
On Tue, Feb 28, 2012 at 05:58:14PM -0800, Mike Stump wrote:
On Feb 28, 2012, at 3:39 PM, Jakub Jelinek wrote:
Here is a fix, bootstrapped/regtested on x86_64-linux and i686-linux,
ok for trunk?
Sure, but, I can't find -Wpsabi in the docs. :-( If -Wno-abi works, at
least it is
Il 29/02/2012 00:37, Jakub Jelinek ha scritto:
Hi!
Bootstrap fails on ia64-linux with Ada currently, because we get
a -fcompare-debug failure on one of the files, where a debug_insn
is the only insn that references in5 hard register (which is live on
entry of the first bb).
76 matches
Mail list logo