On 04/27/2011 05:50 PM, Joseph S. Myers wrote:
Continuing the toplevel cleanups separating the cases disabling
different subdirectories - on the basis that eventually this
information should come from the subdirectories, and that as
previously discussed much of the default disabling of ${libgcj}
On 04/27/2011 06:33 PM, Joseph S. Myers wrote:
This patch, relative to a tree with
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg02123.html applied,
continues the cleanup of toplevel cases relating to disabling Java or
Java libraries by arranging for Java to be disabled (via
Hi
This patch is for google/main.
The bug http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43129 describes
how arm backend generates ineffecient code to load global variable's
address in terms of both code size and performance when option -fpic
is specified. It also describes how it can be improved
Gabriel Dos Reis g...@cs.tamu.edu writes:
Andrew Pinski pins...@gmail.com writes:
| On Wed, Apr 27, 2011 at 6:48 PM, Gabriel Dos Reis g...@cs.tamu.edu wrote:
| A local `extern' declaration does not give the entity an external linkage
| -- irrespective of the linkage of the function
This splits call handling from find_func_aliases into a separate
function. Following patches will then be easier to read.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2011-04-28 Richard Guenther rguent...@suse.de
* tree-ssa-structalias.c
This changes PTA graph dumps to DOT format and dumps before
and after solving. It also moves succ graph building after
freeing substitution info, reducing peak memory usage slightly.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2011-04-28 Richard Guenther
On 04/28/2011 03:37 AM, Doug Kwan (關振德) wrote:
Hi,
This patch causes cxxabi_tweaks.h to be installed in freestanding
mode. We need this because libsupc++ installs cxxabi.h, which
includes cxx_tweaks.h.
Makes sense. Did you install your other similar tweak to 4_6-branch too?
In that case,
On 04/28/2011 03:36 AM, Jason Merrill wrote:
In the discussion of 48760, Gaby suggested that we allow
list-initialization of built-in complex numbers. This made a lot of
sense to me, so I've gone ahead and implemented it.
Great. Thus, as regards the std::complex constructor application of
Ping for the ARM-specific parts of this patch:
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg01752.html
which improves the handling of arrays of vectors.
Richard
Committed.
Richard.
2011-04-28 Richard Guenther rguent...@suse.de
PR tree-optimization/40052
PR tree-optimization/15347
* gcc.dg/tree-ssa/vrp57.c: New testcase.
* gcc.dg/pr15347.c: Likewise.
Index: gcc/testsuite/gcc.dg/tree-ssa/vrp57.c
Hi,
tested x86_64-linux, committed to mainline.
Paolo.
/cp
2011-04-28 Paolo Carlini paolo.carl...@oracle.com
PR c++/48530
* tree.c (build_cplus_new): Check build_target_expr return
value for error_mark_node.
/testsuite
2011-04-28 Paolo Carlini
Hi,
now that Jason implemented list-initialization of complex in mainline,
we can fix this completely in the library. For 4_6-branch I mean to fix
it only in C++03 mode, where we can do work in the body (the constructor
is constexpr in C++0x mode).
Tested x86_64-linux, applied to mainline.
Hi,
tthis patch makes inliner to handle nested expressions when trying to figure
out what
will be constant after inlining. This is done by simple propagation across SSA
names.
We can't really handle PHIs well, since we know name will be constant, but we
don't know
what, so this is a lot
Hi please review the trivial patch below. It reduces race conditions
in value profiling. Another trivial change (to initialize
function_list struct) is also included.
Bootstrapped and regression tested on x86-64/linux.
OK,
thanks!
Honza
Honza,
I continue to receive an ICE:
/farm/dje/src/src/libstdc++-v3/include/precompiled/stdc++.h:94:0:
/tmp/20110427/powerpc-ibm-aix5.3.0.0/libstdc++-v3/include/valarray:1163:1:
internal compiler error: vector VEC(tree,base) index domain error, in
evaluate_conditions_for_edge at
Ok?
Ping?
PS: For the maintainer who will (eventually) review this patch,
the latest version, tested and with all the comments and contributions
from Joseph and Mike merged in, is --
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg01930.html
So, that's the one to review.
Hi,
Sometimes loop vectorization factor changes during the analysis, while
statement analysis depends on it. This patch moves the update of the
vectorization before statements, avoiding current difference between the
analysis and the transformations phases that caused the problem described
in
Looking closer reveals a use-after-free, so this reverts the
portion of the patch likely causing 48804.
Committed as obvious.
Richard.
2011-04-28 Richard Guenther rguent...@suse.de
Revert
2011-04-28 Richard Guenther rguent...@suse.de
* tree-ssa-structalias.c
On Thu, 28 Apr 2011, Paolo Bonzini wrote:
Do you plan to finish the switch to unsupported_languages instead of
noconfigdirs soon )for Java)?
I plan one further patch to simplify the libgcj-disabling logic (and in
particular to eliminate the disabling for *-*-*), but not to convert it
Richard Sandiford richard.sandif...@linaro.org writes:
| Gabriel Dos Reis g...@cs.tamu.edu writes:
| Andrew Pinski pins...@gmail.com writes:
|
| | On Wed, Apr 27, 2011 at 6:48 PM, Gabriel Dos Reis g...@cs.tamu.edu
wrote:
| | A local `extern' declaration does not give the entity an external
Gabriel Dos Reis g...@cs.tamu.edu writes:
| FWIW, I prefer Andrew's patch, but since yours has been applied,
| I suppose there's no point changing it.
when we come to agree on coding style guidelines for GCC in C++, I hope we
recommend against local extern declarations.
Oh, for pure C++, I
On 28 April 2011 01:40, Gabriel Dos Reis g...@integrable-solutions.net wrote:
My configuration command line is:
/home/gdr/src/gcc.svn/configure --enable-languages=c,c++
--enable-build-with-cxx --disable-multilib --disable-nls
and the builds is progressing quite very well with g++.
There is
On Thu, Apr 28, 2011 at 8:57 AM, Ville Voutilainen
ville.voutilai...@gmail.com wrote:
On 28 April 2011 01:40, Gabriel Dos Reis g...@integrable-solutions.net
wrote:
My configuration command line is:
/home/gdr/src/gcc.svn/configure --enable-languages=c,c++
--enable-build-with-cxx
I only have some stylistic comments for this patch. The new pass looks
OK to me, but I do not know this area well enough to do a good review.
In your ChangeLog entries, please remove the directory prefix from the
file names.
http://codereview.appspot.com/4433079/diff/1/gcc/hooks.c
File
Hi Richard,
Ping for the ARM-specific parts of this patch:
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg01752.html
ARM part approved.
Cheers
Nick
Richard Sandiford richard.sandif...@linaro.org writes:
| Gabriel Dos Reis g...@cs.tamu.edu writes:
| | FWIW, I prefer Andrew's patch, but since yours has been applied,
| | I suppose there's no point changing it.
|
| when we come to agree on coding style guidelines for GCC in C++, I hope we
|
Will you be proposing this patch for trunk as well?
Diego.
On Wed, Apr 27, 2011 at 22:09, Sharad Singhai sing...@google.com wrote:
Hi Diego,
Thanks for the quick feedback. Here is a an updated version of the patch.
2011-04-27 Sharad Singhai sing...@google.com
ChangeLog.google-main
* params.def: Add new parameters to control
On Thu, Apr 28, 2011 at 4:47 PM, Diego Novillo dnovi...@google.com wrote:
On Wed, Apr 27, 2011 at 22:09, Sharad Singhai sing...@google.com wrote:
Hi Diego,
Thanks for the quick feedback. Here is a an updated version of the patch.
2011-04-27 Sharad Singhai sing...@google.com
Hello,
At the moment G++ considers a call expression involving a member
function which implicit 'this' pointer is type dependant, as being a
type dependent expression (PR c++/47172). The problem in this PR is
that it fails to recognize a BASELINK node as being a member function.
As a result,
Hi,
On Thu, Apr 28, 2011 at 03:39:11PM +0200, Richard Guenther wrote:
On Thu, Apr 28, 2011 at 3:31 PM, Martin Jambor mjam...@suse.cz wrote:
Hi,
when I was removing the cgraph_node function I noticed that
cgraph_preserve_function_body_p takes a tree decl parameter which it
immediately
This patch fixes the following warnings seen during a powerpc64
bootstrap.
libgomp/config/linux/sem.c: In function ‘gomp_sem_wait_slow’:
libgomp/config/linux/sem.c:33:1: note: non-delegitimized UNSPEC UNSPEC_TOCREL
(44) found in variable location
libgomp/config/linux/bar.c: In function
Hi!
Both final and mergeable are passed to GOMP_task in flags, but mergeable
is currently ignored, we might want to clone it on the compiler side in
that case. taskyield calls new GOMP_taskyield function, which is currently
empty though.
2011-04-28 Jakub Jelinek ja...@redhat.com
*
OK.
Jason
Hi,
since the split of warn_uninitialized to warn_maybe_uninitialized the
fortran and Ada frontends have changed behaviours (causing uninit_func.adb
to fail). I've left out the Java frontend because it also didn't set
warn_uninitialized with -Wall before, and go because it doesn't do
Hi,
minor nit:
--- libgomp/testsuite/libgomp.c++/task-8.C (revision 0)
+++ libgomp/testsuite/libgomp.c++/task-8.C (revision 0)
@@ -0,0 +1,44 @@
+// { dg-do run }
+
+#includeomp.h
+#includecstdlib
+ #pragma omp atomic read
+e = err;
+ if (e)
+abort ();
if you include cstdlib
Shouldn't there be a warning from newer gcc, i.e. from trunk?
The code in GCC 4.3 reads:
/* Inline declaration after use or definition.
??? Should we still warn about this now we have unit-at-a-time
mode and can get it right?
Definitely don't complain if the
On 04/27/2011 11:13 AM, Anatoly Sokolov wrote:
* config/sparc/sparc.h (PRINT_OPERAND, PRINT_OPERAND_ADDRESS,
PRINT_OPERAND_PUNCT_VALID_P): Remove.
* config/sparc/sparc-protos.h (print_operand): Remove declaration.
* config/sparc/sparc.c
On 04/27/2011 12:16 PM, Joseph S. Myers wrote:
Ping^2. This patch
http://gcc.gnu.org/ml/gcc-patches/2011-04/msg00919.html is still pending
review (except insofar as it relates to GNU Hurd).
Ok.
r~
The patch below is a backport of those fixes to the 4.5 branch, which
fixes the profiled-bootstrap failure for me. (Note that on current
mainling, the ifcvt.c dead_or_predictable routine has been significantly
rewritten beyond what was done by those patches. These additional
changes do not
I think this could use a comment along the lines of make sure we flush
any queued register saves in case this clobbers affected registers. OK
with that change.
Jason
On 04/28/2011 06:17 PM, Joseph S. Myers wrote:
This patch cleans up the libgcj-disabling case statement at toplevel,
removing the default disabling for *-*-* (which allows other empty
cases to be removed) and disabling based on architecture, or most
cases of architecture-OS combination where the
This patch moves the libgloss_dir settings out of the general toplevel
case over targets, simplifying various target patterns in the process.
Resulting empty cases will be cleaned up when the disabling of newlib
and libgloss for various target OSes are moved to a separate case
statement.
OK to
Hi,
Thanks. I will wait for 2 days before checking this into trunk.
The previous fix was in the 4.6 release. Can I also check in this fix
into 4.6 branch as well?
-Doug
On Thu, Apr 28, 2011 at 2:34 AM, Paolo Carlini paolo.carl...@oracle.com wrote:
On 04/28/2011 03:37 AM, Doug Kwan (關振德)
gcc-patches-ow...@gcc.gnu.org wrote on 28/04/2011 05:30:35 PM:
When I started looking at PR 48765, I noticed that vectorizable_reduction
set STMT_VINFO_TYPE before checking the reduction cost. This probably
doesn't matter in practice, and certainly has nothing to do with fixing
the PR
While investigating the remaining testsuite failures on Tru64 UNIX, I
found that both two libffi failures and several libjava failures had a
common cause: libffi/src/alpha/osf.S lacks unwind info that is usable on
Tru64 UNIX.
The following patch fixes this, along the same lines of what was done
On Thu, Apr 28, 2011 at 11:22 AM, Alan Modra amo...@gmail.com wrote:
This patch fixes the following warnings seen during a powerpc64
bootstrap.
libgomp/config/linux/sem.c: In function ‘gomp_sem_wait_slow’:
libgomp/config/linux/sem.c:33:1: note: non-delegitimized UNSPEC UNSPEC_TOCREL
(44)
On 04/28/2011 07:16 PM, Doug Kwan (關振德) wrote:
Hi,
Thanks. I will wait for 2 days before checking this into trunk.
The previous fix was in the 4.6 release. Can I also check in this fix
into 4.6 branch as well?
Yes.
Paolo.
commit 76c5256b18ce0357dc0fc38ce143b8979318b823
Author: Cary Coutant ccout...@google.com
Date: Thu Apr 28 10:48:56 2011 -0700
Track discriminators by instruction instead of by basic block
Rework discriminator assignment so that it attaches the discriminator
to the source
On Thu, Apr 28, 2011 at 14:04, Cary Coutant ccout...@google.com wrote:
Rework discriminator assignment so that it attaches the discriminator
to the source location of each instruction instead of to the basic
block itself.
Any idea on how this affects memory consumption?
This
Hi,
I committed the below patchlet, per the exchange with Jason on the audit
trail, after having regtested it on x86_64-linux.
Thanks,
Paolo.
/cp
2011-04-28 Paolo Carlini paolo.carl...@oracle.com
PR c++/48798
* semantics.c (finish_base_specifier):
Rework discriminator assignment so that it attaches the discriminator
to the source location of each instruction instead of to the basic
block itself.
Any idea on how this affects memory consumption?
Richard had the same question when I first proposed this patch. This
was my answer
Like HP-UX, Tru64 UNIX needs -fno-common to make gcc.dg/pr48616.c pass.
Tested with the appropriate runtest invocation, installed on mainline.
Rainer
2011-04-28 Rainer Orth r...@cebitec.uni-bielefeld.de
PR tree-optimization/48775
* gcc.dg/pr48616.c: Also add
2011-04-27 Doug Kwan dougk...@google.com
* include/Makefile.am (install-freestanding-headers): Also
install cxxabi_tweaks.h.
* include/Makefile.in: Regnerate.
OK for trunk and 4.6
-benjamin
This patch disables -ftracer for profile use. Okay for google/main?
Thanks,
Sharad
2011-04-28 Sharad Singhai sing...@google.com
Google Ref 40087
* opts.c (common_handle_option): Disable -ftracer for profile use.
* doc/invoke.texi: Update doc that -ftracer is no longer
On Thu, Apr 28, 2011 at 14:51, Sharad Singhai sing...@google.com wrote:
This patch disables -ftracer for profile use. Okay for google/main?
Could you elaborate on why doing this is beneficial? Are you
proposing this for trunk as well?
2011-04-28 Sharad Singhai sing...@google.com
Just FYI, as I was going through some of my recent changes to backport them to
the IBM 4.6 branch, I noticed I forgot to add the ChangeLog entries to the
changes I made on March 21st. I updated the ChangeLog files. Sorry about
that.
Index: gcc/ChangeLog
This patch from Silvius Rus adds a new flag (-fstrict-enum-precision).
While porting the patch to 4.6, I noticed that the C++ FE now has a
similar flag that seems to have similar semantics (-fstrict-enums).
Silvius's patch is used to disable some switch() optimizations that
assume enum types can
gcc-patches-ow...@gcc.gnu.org wrote on 28/04/2011 03:42:01 PM:
Hi,
Sometimes loop vectorization factor changes during the analysis, while
statement analysis depends on it. This patch moves the update of the
vectorization before statements, avoiding current difference between the
On Thu, Apr 28, 2011 at 03:50:45PM -0400, Diego Novillo wrote:
Committed to google/main. Jason, Silvius, what do you think would be
the best approach to merge this into trunk?
When this code does get merged to trunk, can the testcases abort() on
failure rather than returning 1? This is
... are the testcases formatted according to the GNU guidelines, in terms, for
example, of open and closed curly braces? I don't think so, I see some weird
(sorry, after all those years unavoidably look to me *really* weird) open
braces ending lines?
Paolo
On Thu, Apr 28, 2011 at 15:57, Paolo Carlini pcarl...@gmail.com wrote:
... are the testcases formatted according to the GNU guidelines
They weren't. I've run indent -gnu on all of them. Thanks for noticing.
Diego.
On 04/28/2011 10:55 AM, Rainer Orth wrote:
I cannot test the Alpha/Linux side, though.
Neither can I, at the moment.
2011-04-28 Rainer Orth r...@cebitec.uni-bielefeld.de
* src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
Define.
Use them to handle ELF vs.
Introducing the parameters for FDO allows FDO specific tunings. In
general, these parameters are kludges lacking a better way of doing
it. In the long run, we are working on smarter mechanism to make
decisions based on hot program traces and locality regions as well as
information such as uArch
I just backported the following patches from the trunk to the GCC 4.6 branch
and committed them after doing a bootstrap and make check with no regressions.
[gcc]
2011-04-28 Michael Meissner meiss...@linux.vnet.ibm.com
Backport from mainline
2011-04-01 Andrew Pinski
Hi,
a small patch for a [4.7 Regression] ICE on invalid: apparently, as
happens in other places in the same file, we want to check here too the
return value of build_value_init for error_mark_node and bail out.
Tested x86_64-linux. Ok?
Thanks,
Paolo.
//
/cp
2011-04-29
The attached patch fixes two regressions:
a) PR 48810: For type-bound procedures, the access flags should be
checked only for the generic function, not for the specific function the
generic resolves to. (4.6/4.7 rejects-valid regression.)
b) PR 48800: Function-results shall not be
This patch added regression tests for LIPO in google/main and fixes a couple of
bugs found
in app testing:
1) duplicate assembler labels in .s file
2) missing icall profling for static functions indirect called
3) assertion in type 'merging'
4) gcov-dump bug.
Bootstrap compiler, regression
On Thu, Apr 28, 2011 at 8:53 PM, Diego Novillo dnovi...@google.com wrote:
On Thu, Apr 28, 2011 at 14:51, Sharad Singhai sing...@google.com wrote:
This patch disables -ftracer for profile use. Okay for google/main?
Could you elaborate on why doing this is beneficial? Are you
proposing this
On Thu, Apr 28, 2011 at 9:50 PM, Diego Novillo dnovi...@google.com wrote:
This patch from Silvius Rus adds a new flag (-fstrict-enum-precision).
While porting the patch to 4.6, I noticed that the C++ FE now has a
similar flag that seems to have similar semantics (-fstrict-enums).
Silvius's
Sharad can provide some some performance data -- we have seen up to 2%
degradation to with tracer turned on for one of google's most
important program. Perhaps Sharad can collect some SPEC numbers.
I agree a better approach should be to fix the problem in tracer
instead of turning it off in
On Fri, Apr 29, 2011 at 1:34 AM, Xinliang David Li davi...@google.com wrote:
Sharad can provide some some performance data -- we have seen up to 2%
degradation to with tracer turned on for one of google's most
important program. Perhaps Sharad can collect some SPEC numbers.
I agree a better
This patch from Silvius Rus adds support for sampled edge profile collection
to reduce instrumentation run overhead. Bootstraps and no test regressions. Ok
for google/main?
2011-04-28 Silvius Rus silvius@gmail.com
* doc/invoke.texi: Document -fprofile-generate-sampling option.
Tracer is not turned on by default for non-FDO case -- Sharad's change
simply turns it off by default for all cases.
David
On Thu, Apr 28, 2011 at 4:38 PM, Richard Guenther
richard.guent...@gmail.com wrote:
On Fri, Apr 29, 2011 at 1:34 AM, Xinliang David Li davi...@google.com wrote:
Sharad can
This patch stops m68k_handle_option from using global state. The
general approach is similar to that used for ARM: new .def files are
added for the -march= and -mtune= arguments, to go along with the
existing one for -mcpu= arguments, and as well as initializing arrays
in m68k.c these are used to
+ Honza
This patch may be a candidate for trunk as well. This feature not only
allows profile collection with much less overhead (for multi-thread
programs with hot regions, the slow down can be significant due to
cache ping-pong effect of counter update) without sacrificing too much
the
On Thu, Apr 28, 2011 at 01:58:24PM -0400, David Edelsohn wrote:
On Thu, Apr 28, 2011 at 11:22 AM, Alan Modra amo...@gmail.com wrote:
* config/rs6000/rs6000.c (rs6000_delegitimize_address): Handle
unspec plus offset. Tidy macho code.
Looks good.
Committed mainline revision 173141. I
On Thu, Apr 28, 2011 at 4:58 PM, Xinliang David Li davi...@google.com wrote:
+ Honza
This patch may be a candidate for trunk as well. This feature not only
allows profile collection with much less overhead (for multi-thread
programs with hot regions, the slow down can be significant due to
This patch saves and restores the PARM_DECL DECL_ARG_TYPE in the PPH file.
Index: gcc/cp/ChangeLog.pph
2011-04-28 Lawrence Crowl cr...@google.com
* pph-streamer-out.c (pph_stream_write_tree): Write PARM_DECL
DECL_ARG_TYPE.
* pph-streamer-in.c (pph_stream_read_tree):
On Apr 28, 2011, at 4:42 PM, Easwaran Raman wrote:
This patch from Silvius Rus adds support for sampled edge profile collection
to reduce instrumentation run overhead.
Sounds interesting for trunk...
On Thu, Apr 28, 2011 at 10:08 PM, dnovi...@google.com wrote:
I only have some stylistic comments for this patch. The new pass looks
OK to me, but I do not know this area well enough to do a good review.
In your ChangeLog entries, please remove the directory prefix from the
file names.
Yes, after porting it to google/main.
Carrot
On Thu, Apr 28, 2011 at 10:26 PM, Diego Novillo dnovi...@google.com wrote:
Will you be proposing this patch for trunk as well?
Diego.
Ok for google/main.
David
On Thu, Apr 28, 2011 at 8:18 PM, Carrot Wei car...@google.com wrote:
Yes, after porting it to google/main.
Carrot
On Thu, Apr 28, 2011 at 10:26 PM, Diego Novillo dnovi...@google.com wrote:
Will you be proposing this patch for trunk as well?
Diego.
Ok to install?
for gcc/ChangeLog
from Alexandre Oliva aol...@redhat.com
PR other/48093
* doc/invoke.texi: Document -mtls-dialect and GCC_COMPARE_DEBUG.
Index: gcc/doc/invoke.texi
===
--- gcc/doc/invoke.texi.orig 2011-04-27
The testcase in PR 47994, as well as one of the tests in 47919, exposed
our failure to skip debug insns when testing whether it is safe to split
the unrecognizable insn resulting from combining 3 insns.
Unfortunately, both testcases are far too sensitive to other changes in
the compiler, to the
On Apr 23, 2011, Nicola Pero nicola.p...@meta-innovation.com wrote:
Ok to commit ?
Yeah, thanks.
Index: ChangeLog
+2011-04-22 Nicola Pero nicola.p...@meta-innovation.com
+
+ * Makefile.in (ENABLE_MAINTAINER_RULES): New.
+
Index: cp/ChangeLog
+2011-04-23 Nicola Pero
86 matches
Mail list logo