On Thu, Feb 26, 2015 at 11:20 PM, Jeff Law wrote:
> On 02/26/15 10:30, Wilco Dijkstra wrote:
>>
>> Several GCC versions ago a conditional negate optimization was introduced
>> as a workaround for
>> PR45685. However the branchless expansion for conditional negate is
>> extremely inefficient on mos
On Fri, Feb 27, 2015 at 12:34 AM, Pat Haugen
wrote:
> The subject testcase is failing on older powerpc64 hardware that doesn't
> support vector instructions because the prolog code is saving callee save
> vector registers used in the loop before the check_vect() call has even been
> performed. Fol
On Fri, Feb 27, 2015 at 09:08:30AM +0100, Richard Biener wrote:
> On Fri, Feb 27, 2015 at 12:34 AM, Pat Haugen
> wrote:
> > The subject testcase is failing on older powerpc64 hardware that doesn't
> > support vector instructions because the prolog code is saving callee save
> > vector registers us
On 26 February 2015 at 06:22, Xingxing Pan wrote:
> Hi,
>
> This patch fix the type of mov_aarch64 in aarch64.md.
> Is it OK for trunk?
OK, thank you /Marcus
On 25 February 2015 at 09:53, Kyrill Tkachov wrote:
>
> On 13/02/15 10:14, Richard Earnshaw wrote:
>>
>> On 13/02/15 09:52, Kyrill Tkachov wrote:
>>>
>>> Hi all,
>>>
>>> This patch to changes.html mentions the xgene1 support in GCC 5 for arm
>>> and aarch64 and also the FreeBSD support for ARM.
>>
Hello!
> This is actually Jakub's patch from the PR, with a few minor tweaks that were
> needed to bootstrap
> and pass the regression suite. The splitter was using operand 0 without
> setting it first. It should've
> been operand 2. Also, there was a division by zero that was causing an
> inva
This PR points out that we can ICE in case we have __auto_type with undeclared
variable: grokdeclarator returns error_mark_node, so check for that before
accessing decl's TREE_CODE (can't use error_operand_p here as that would cause
bogus diagnostics be emitted).
Bootstrapped/regtested on x86_64-l
On 19/02/15 12:19, Matthew Wahab wrote:
> The LEGITIMIZE_RELOAD_ADDRESS macro is only needed for reload. Since the
> ARM backend no longer supports reload, this macro is not needed and this
> patch removes it.
>
> Tested arm-none-linux-gnueabihf with gcc-check.
>
> Ok for trunk? now or in stage 1
On 20/02/15 04:14, Hurugalawadi, Naveen wrote:
> Hi Marcus,
>
>>> The handling of SYMBOL_SMALL_TPREL is present in 4.9 and very clearly
>>> has exactly the same issue.
>
> Please find attached the patch ported for gcc-4.9.
>
> Please review the patch and let us know if its okay?
> Regression tes
Alexandre Oliva writes:
> Ok, I looked into it further, after patching dwlocstat to dump
> per-variable per-range coverage/length info, so as to be able to compare
> object files more easily.
If you send me those patches, I can finish them, bind the functionality
to a command line option, and me
Applied.
Richard.
2015-02-27 Richard Biener
PR lto/65193
* g++.dg/lto/pr65193_0.C: New testcase.
Index: gcc/testsuite/g++.dg/lto/pr65193_0.C
===
*** gcc/testsuite/g++.dg/lto/pr65193_0.C(revision 0)
--- g
Hi,
the attached patch gets rid of -m64/-m31 uses in our target specific
testcases in order to make
"make check RUNTESTFLAGS='--target_board=unix\{-m31,-m64\}'"
runs work fine again.
Committed to mainline.
Bye,
-Andreas-
2015-02-27 Andreas Krebbel
* gcc.target/s390/20140327-1.c: R
On Thu, Feb 26, 2015 at 09:25:57PM +0100, Kai Tietz wrote:
> Well, testcase for the pragma ...
>
> ChangeLog testsuite/
>
> 2015-02-26 Kai Tietz
>
> * gcc.dg/weak/weak-17.c: New file
Missing full stop.
> Updated patch (regression-tested):
> Index: c-pragma.c
> =
Hi,
In df_finish_pass, optional problems are removed manually making non null
entries in df->problems_in_order non contiguous. This may lead to null pointer
dereference when accessing all problems from df->problems_in_order[0] to
df->problems_in_order[df->num_problems_defined - 1] and miss some ot
Applied at rev.revision 221055.
--
Kai
On Fri, 27 Feb 2015, Marek Polacek wrote:
> This PR points out that we can ICE in case we have __auto_type with undeclared
> variable: grokdeclarator returns error_mark_node, so check for that before
> accessing decl's TREE_CODE (can't use error_operand_p here as that would cause
> bogus diagnosti
On Wed, Feb 25, 2015 at 8:18 PM, Richard Henderson wrote:
> On 02/25/2015 09:02 AM, Uros Bizjak wrote:
>> The patch was tested on alpha-linux-gnu and alphaev68-linux-gnu for
>> all default languages plus obj-c++ and go.
>>
>> OK for mainline?
>
> Ok. Thanks.
Unfortunately, while "normal" bootstr
Sebastian Pop writes:
> +void
> +baz (void)
> +{
> + while (1)
> +{
> + a = foo ();
> + b = foo ();
FAIL: gcc.dg/tree-ssa/ssa-dom-thread-9.c (test for excess errors)
Excess errors:
/daten/aranym/gcc/gcc-20150227/gcc/testsuite/gcc.dg/tree-ssa/ssa-dom-thread-9
On Thu, Feb 26, 2015 at 6:10 PM, Jan Hubicka wrote:
> Hi,
> this is the final version of patch I comitted. It has new fix to
> make_decl_local
> to set TREE_ADDRESSABLE becuase we leave the flag undefined for non-local
> decls.
> I also dropped Optimization from fmerge-all-constants, fmerge-con
On Thu, 2015-02-26 at 20:04 -0600, Peter Bergner wrote:
> On Thu, 2015-02-26 at 22:56 +0100, Jakub Jelinek wrote:
> > How do make check results (asan.exp/ubsan.exp) look like on ppc64le?
> > If it works as good as or better as ppc64be, then I'm fine with adding it
> > even in stage4.
>
> They have
On Thu, Feb 26, 2015 at 6:49 AM, Kai Tietz wrote:
> Hi,
>
> This is the remaining fix for re-enabling native boostrap for
> Windows-variant of gcc without disabling -Werror for libgcc.
>
> ChangeLog
>
> 2015-02-26 Kai Tietz
>
> PR target/65038
> * config.in: Regenerated.
> * configu
Many folks complained that -Wformat-signedness is overly pedantic, and
especially
with -Werror it can cause a lot of needless pain. This patch moves it ouf of
-Wformat=2.
Bootstrapped/regtested on x86_64-linux, applying to trunk.
2015-02-27 Marek Polacek
PR c/65040
* doc/inv
The following patch fixes PR65302
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65032
LRA rematerialization sub-pass did not update info about scratch pseudos
in rematerialized insns and this resulted in skipping reverse
transformation of scratch pseudos into scratches in such insns at the
LRA
On 03/02/15 15:18, Kyrill Tkachov wrote:
> Hi all,
>
> The ICE in this PR occurs when -mtune=xscale triggers a particular path
> through arm_gen_constant during expand
> that creates a 0xf00f mask but for a 64-bit HOST_WIDE_INT doesn't
> sign extend it into
> 0xf00f that signifies
On Tue, Feb 24, 2015 at 2:30 PM, Jeff Law wrote:
> On 02/23/15 20:38, Martin Sebor wrote:
>>
>> On 02/22/2015 11:45 AM, David Edelsohn wrote:
>>>
>>> Does this patch really fix the problem? The PR notes that the
>>> testcase fails and code quality has regressed. Has the code
>>> generation been
2015-02-27 14:48 GMT+01:00 H.J. Lu :
> On Thu, Feb 26, 2015 at 6:49 AM, Kai Tietz wrote:
>> Hi,
>>
>> This is the remaining fix for re-enabling native boostrap for
>> Windows-variant of gcc without disabling -Werror for libgcc.
>>
>> ChangeLog
>>
>> 2015-02-26 Kai Tietz
>>
>> PR target/6503
On 2015.02.27 at 07:47 -0600, Peter Bergner wrote:
>
> Ok, since the results met your criteria for inclusion, I committed the
> change as revision 221060. Thanks.
Are there any plans to fix:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63927 ?
--
Markus
9.c (test for excess errors)
> Excess errors:
> /daten/aranym/gcc/gcc-20150227/gcc/testsuite/gcc.dg/tree-ssa/ssa-dom-thread-9.c:45:11:
> error: too few arguments to function 'foo'
> /daten/aranym/gcc/gcc-20150227/gcc/testsuite/gcc.dg/tree-ssa/ssa-dom-thread-9.c:46:11:
>
On 02/27/2015 07:27 AM, David Edelsohn wrote:
On Tue, Feb 24, 2015 at 2:30 PM, Jeff Law wrote:
On 02/23/15 20:38, Martin Sebor wrote:
On 02/22/2015 11:45 AM, David Edelsohn wrote:
Does this patch really fix the problem? The PR notes that the
testcase fails and code quality has regressed.
On 02/26/2015 07:21 PM, Jan Hubicka wrote:
2015-02-25 Martin Liska
Jan Hubicka
PR ipa/64693
* ipa-icf.c (symbol_compare_collection::symbol_compare_collection): New.
(sem_item_optimizer::subdivide_classes_by_sensitive_refs): New function.
(sem_item
On Fri, 2015-02-27 at 15:30 +0100, Markus Trippelsdorf wrote:
> On 2015.02.27 at 07:47 -0600, Peter Bergner wrote:
> >
> > Ok, since the results met your criteria for inclusion, I committed the
> > change as revision 221060. Thanks.
>
> Are there any plans to fix:
> https://gcc.gnu.org/bugzilla/
On Fri, Feb 27, 2015 at 7:21 AM, wrote:
>
> As discussed in this bugzilla, the debug info from libgo should not be
> stripped or some things won't work as documented, like runtime.Callers. Is
> that information documented anywhere so that anyone who builds gccgo and
> libgo and provides it to oth
On 02/09/2015 11:20 PM, Pedro Alves wrote:
> Converting GDB to be a C++ program, I stumbled on 'basename' issues,
> like:
>
...
> So I thought of adding a m4 file that projects that use libiberty can
> source to pull in the autoconf checks that libiberty needs done in
> order to use its public
At a minimum I think it should be mentioned in libgo/README.
I'll have to ask around to find out where would be best so that anyone
who builds libgo for distribution knows that the debug info should not
be stripped.
I'm not sure about other places
On 02/27/2015 09:59 AM, Ian Lance Taylor wrot
> Hi.
>
> There's one missing vector comparison. Fix is obvious, ready for trunk?
>
> Thanks,
> Martin
> >From 3d03fb28ec21b6ed30d5179bd70aba79d246cd26 Mon Sep 17 00:00:00 2001
> From: mliska
> Date: Fri, 27 Feb 2015 16:35:31 +0100
> Subject: [PATCH] Fix missing condition in symbol_compare_hash
>
> This caused:
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65237
Hi,
this is patch I commited. gcc.dg/attr-noinline.c has template that counts
number of calls
in optimized assembler. Those do not match if one function is turned into
another's wrapper.
gcc.dg/noreturn-7.c misses one wa
On Fri, 2015-02-27 at 03:10 +0100, Jan Hubicka wrote:
> Bootstrapped/regtested x86_64-linux, comitted.
>
> Honza
> * ipa-icf.c (symbol_compare_collection::symbol_compare_colleciton):
> Use address_matters_p.
I think this patch is causing an ICE while building glibc on MIPS. I am
bui
On Fri, 2015-02-27 at 09:33 -0800, Steve Ellcey wrote:
> On Fri, 2015-02-27 at 03:10 +0100, Jan Hubicka wrote:
>
> > Bootstrapped/regtested x86_64-linux, comitted.
> >
> > Honza
> > * ipa-icf.c (symbol_compare_collection::symbol_compare_colleciton):
> > Use address_matters_p.
>
> I think
> Richard Biener wrote:
> On Thu, Feb 26, 2015 at 11:20 PM, Jeff Law wrote:
> > On 02/26/15 10:30, Wilco Dijkstra wrote:
> >>
> >> Several GCC versions ago a conditional negate optimization was introduced
> >> as a workaround for
> >> PR45685. However the branchless expansion for conditional nega
On 02/27/2015 07:04 PM, Steve Ellcey wrote:
> Following up to my own email. I can reproduce this with the following
> cut down test case if I compile with '-O2 -fmerge-all-constants' on
> MIPS.
>
> extern const char *const _sys_siglist[128];
> const char *const __new_sys_siglist[128] = { };
> ext
I tested Marek's proposed change and it works correctly,
i.e. arrays which are not part of a struct are now
instrumented when accessed through a pointer. This also
means that the following case is diagnosed (correctly)
as undefined behaviour as pointed out by Richard:
int
main (void)
{
int *t =
On February 27, 2015 12:42:43 PM GMT+01:00, Thomas Preud'homme
wrote:
>Hi,
>
>In df_finish_pass, optional problems are removed manually making non
>null
>entries in df->problems_in_order non contiguous. This may lead to null
>pointer
>dereference when accessing all problems from df->problems_in_o
On 02/27/2015 04:59 PM, Ian Lance Taylor wrote:
> On Fri, Feb 27, 2015 at 7:21 AM, wrote:
>>
>> As discussed in this bugzilla, the debug info from libgo should not be
>> stripped or some things won't work as documented, like runtime.Callers. Is
>> that information documented anywhere so that anyo
On Fri, Feb 27, 2015 at 12:07 PM, Matthias Klose wrote:
>
> is there anything which could be stripped without scarifying functionality?
> Linux distributions usually strip things by default, so a hint what exactly is
> needed to keep the functionality would be appreciated.
What is needed is file/
>
> ../sysdeps/gnu/siglist.c:72:1: internal compiler error: in address_matters_p,
> at symtab.c:1908
> versioned_symbol (libc, __new_sys_sigabbrev, sys_sigabbrev, GLIBC_2_3_3);
> ^
> 0x66a080 symtab_node::address_matters_p()
> /scratch/sellcey/repos/bootstrap/src/gcc/gcc/symtab.c:1908
>
On Thu, 2015-02-26 at 11:17 -0500, David Malcolm wrote:
> On Fri, 2015-02-20 at 10:29 -0700, Jeff Law wrote:
> > On 02/19/15 14:56, Chris Johns wrote:
> > > On 20/02/2015 8:23 am, Joel Sherrill wrote:
> > >>
> > >> On 2/19/2015 2:56 PM, Sandra Loosemore wrote:
> > >>> Jakub Jelinek wrote:
> >
On 25 February 2015 at 20:22, Matthias Klose wrote:
> When gdb is linked/used with Python 3, import of the pretty printers fails:
>
> Traceback (most recent call last):
> File
> "/usr/share/gdb/auto-load/usr/lib/i386-linux-gnu/libstdc++.so.6.0.21-gdb.py",
> line 58, in
>import libstdcxx.v6
>
On Feb 27, 2015, Petr Machata wrote:
> Alexandre Oliva writes:
>> Ok, I looked into it further, after patching dwlocstat to dump
>> per-variable per-range coverage/length info, so as to be able to compare
>> object files more easily.
> If you send me those patches, I can finish them, bind the f
Looks like a typo to me. Bootstrapped and regression tested on
x86_64-unknown-linux-gnu with no new regressions.
Is this OK for stage1?
Thanks,
Kugan
gcc/ChangeLog:
2015-02-28 Kugan Vivekanandarajah
* ira-costs.c (record_operand_costs): Fix typo (remove redundant code).
diff --git a
The following patch improves inheritance for PR64317 testcase
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64317
I ran a lot SPEC2000 benchmarks to get better default parameter value
for EBB_PROBABILITY_CUTOFF in LRA inheritance. The new default parameter
value improves SPECInt2000 by 0.4%
On 02/19/2015 07:03 PM, Jakub Jelinek wrote:
+ /* Avoid warning for !!b == y where b is boolean. */
+ && (!DECL_P (current.lhs)
+ || TREE_TYPE (current.lhs) == NULL_TREE
+ || TREE_CODE (TREE_TYPE (current.lhs)) != BOOLEAN_TYPE))
There's something wrong h
On February 27, 2015 11:03:14 PM GMT+01:00, Vladimir Makarov
wrote:
> The following patch improves inheritance for PR64317 testcase
>
>https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64317
>
> I ran a lot SPEC2000 benchmarks to get better default parameter value
>for EBB_PROBABILITY_CUTOFF in LRA
On Thu, Dec 4, 2014 at 8:46 AM, H.J. Lu wrote:
> On Thu, Dec 4, 2014 at 4:44 AM, Uros Bizjak wrote:
>> On Wed, Dec 3, 2014 at 10:35 PM, H.J. Lu wrote:
>>
It would probably help reviewers if you pointed to actual path
submission [1], which unfortunately contains the explanation
On Fri, Feb 27, 2015 at 3:23 PM, H.J. Lu wrote:
> On Thu, Dec 4, 2014 at 8:46 AM, H.J. Lu wrote:
>> On Thu, Dec 4, 2014 at 4:44 AM, Uros Bizjak wrote:
>>> On Wed, Dec 3, 2014 at 10:35 PM, H.J. Lu wrote:
>>>
> It would probably help reviewers if you pointed to actual path
> submi
On 2015-02-27 5:30 PM, Bernhard Reutner-Fischer wrote:
On February 27, 2015 11:03:14 PM GMT+01:00, Vladimir Makarov
wrote:
Committed as rev.221070.
+DEFPARAM (PARAM_LRA_INHERITANCE_EBB_PROBABILITY_CUTOFF,
+ "lra-inheritance-ebb-probability-cutoff",
+ "Minimal fall-through e
On 2015-02-27 4:52 PM, Kugan wrote:
Looks like a typo to me. Bootstrapped and regression tested on
x86_64-unknown-linux-gnu with no new regressions.
Is this OK for stage1?
Yes, Kugan. Thanks.
Given that Martin's fix to the testcase allowed it to succeed without
Richi's fix for the underlying problem, is there a modification to the
testcase or a new testcase that would really test the optimization?
Let me work on it.
Below is a patch with a couple of minor tweaks to the existing
tes
On 02/27/2015 03:03 PM, Vladimir Makarov wrote:
The following patch improves inheritance for PR64317 testcase
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64317
[snip]
Index: params.def
===
--- params.def (revision 220916)
+++
On 02/27/2015 04:30 PM, Marcus Shawcroft wrote:
On 26 February 2015 at 06:22, Xingxing Pan wrote:
Hi,
This patch fix the type of mov_aarch64 in aarch64.md.
Is it OK for trunk?
OK, thank you /Marcus
Hi,
Could someone help to apply the patch? Until now I don't have SVN write
access.
--
On 2015-02-27 8:06 PM, Sandra Loosemore wrote:
On 02/27/2015 03:03 PM, Vladimir Makarov wrote:
The following patch improves inheritance for PR64317 testcase
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64317
[snip]
Index: params.def
60 matches
Mail list logo
