On Wed, Jan 4, 2012 at 5:24 PM, Eric Botcazou wrote:
>> OK. But passing small structures by value doesn't seem that rare --
>> especially in C++ -- and it doesn't feel right to disable SRA just because
>> the backend likes to increase the alignment of stack vars.
>
> Agreed.
>
>> ...something lik
On Wed, Jan 4, 2012 at 7:20 PM, Aldy Hernandez wrote:
> I fixed this PR, and then it got reopened because the testcase triggered a
> different problem on Alpha, Mips, and other architectures. The problem is
> actually totally different than the previous fix for 51472, and has nothing
> to do with
On Thu, Jan 5, 2012 at 1:13 AM, Jakub Jelinek wrote:
> Hi!
>
> profiledbootstrap with --disable-poststage-build-with-cxx and -fexceptions
> currently fails on x86_64-linux. The problem is that no EDGE_FAKE edge is
> added from noreturn fatal_error call to the exit block, eventhough
> fatal_error
On Thu, Jan 5, 2012 at 3:25 AM, Richard Henderson wrote:
> On 01/05/2012 12:53 PM, Jakub Jelinek wrote:
>> On Thu, Jan 05, 2012 at 12:38:17PM +1100, Richard Henderson wrote:
>>> +# Disable libitm if we're not building C++
>>> +case ,${enable_languages}, in
>>> + *,c++) ;;
>>
>> Shouldn't that be
On Thu, Jan 5, 2012 at 5:28 AM, Richard Henderson wrote:
> On 01/04/2012 11:27 PM, Rainer Orth wrote:
>> [libgfortran, libitm] Link with -shared-libgcc
>> http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01382.html
>>
>> This will need a fortran resp. libitm maintainer.
>
> Does the follow
This code change intends to improve scev for array element and reduce the
common sub-expressions in loop, which may be introduced by multiple
reference of expression &a[i]. In the end the register pressure may be
reduced in the loop. The test case is simplified from a real benchmark, and
only want
Approved after you fix some minor fomatting nits. Your code has:
Thank you.
I committed the attached nit adjusted patch.
/Marcusdiff --git a/gcc/reload.c b/gcc/reload.c
index 53dcd2d..206fb36 100644
--- a/gcc/reload.c
+++ b/gcc/reload.c
@@ -4212,7 +4212,12 @@ find_reloads (rtx insn, int repl
Ping for backporting this expand patch, which fixes an ice-on-valid
regression from 4.4 while compiling certain C++ packages on ARM:
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01550.html
As I understand it, this bug is the only reason Ubuntu is keeping
a GCC 4.4 package: there's no known wor
Ping!
2012/1/2 Paolo Carlini :
> Hi,
>
>> Hello,
>>
>> additionally to the suggested patch by Pawel Sikora, I added the
>> adjustments for mt-allocator to it too.
>>
>> ChangeLog
>>
>> 2012-01-01 Kai Tietz
>>
>> PR libstc++/51673
>> * config/abi/pre/gnu-versioned-namespace.ver: Adju
Ping for this patch.
Tristan.
On Dec 21, 2011, at 12:04 PM, Tristan Gingold wrote:
> Hi,
>
> this patch fixes two build libgcc issues on VMS/ia64.
>
> * Because VMS doesn't define LIBGCC2_HAS_TF_MODE, compatibility thunks
> shouldn't be compiled. Otherwise, the shared libgcc library fails to
Which was fixed in 4.6 already. We only warn for type mismatches
for symbols in different TUs if they are not structurally compatible.
Richard.
2012-01-05 Richard Guenther
PR lto/41576
* gfortran.dg/lto/pr41576_0.f90: New testcase.
* gfortran.dg/lto/pr41576_1.f90: Li
We can't directly drop to VARYING from UNDEFINED in likely_value
as an UNDEFINED value can later become CONSTANT which would make
us go up the lattice.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2012-01-05 Richard Guenther
PR tree-optimization/51760
On 01/04/12 17:16, Patrick Marlier wrote:
On 01/04/2012 01:20 PM, Aldy Hernandez wrote:
I fixed this PR, and then it got reopened because the testcase triggered
a different problem on Alpha, Mips, and other architectures. The problem
is actually totally different than the previous fix for 51472,
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01088.html
On 01/04/12 18:54, Patrick Marlier wrote:
In the reentrant.c testcase, the first transaction has no transactional
access and thus removed.
In order to keep this transaction, I added a shared access inside.
Tested on i686.
OK to commit?
(I don't have an account, so thanks in advance to committer)
On 01/04/12 20:03, Patrick Marlier wrote:
Hi,
(Initially from http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00225.html)
PR51516 shows up a problem due to the TM IPA rework.
Indeed, an alias can get called but tm cg data is in the parent of the
alias.
The fix consists to set and use information fr
This should improve the situation a bit, avoiding ICEs and
irritating "Please submit a full bug report" messages.
It restricts the command-line options to the set of supported
languages and rejects compiling with -flto -fmudflap
(we then silently drop -fmudflap if you choose to only apply
it at l
You want is_gimple_reg () instead of is_gimple_non_addressable () as you
can't simply make something addressable at this point.
is_gimple_non_addressable
looks like a weird redundant predicate to me - it's only used once and its use
should be replaced (and the predicate removed).
Ok, thanks.
As you suggested here:
http://gcc.gnu.org/ml/gcc-patches/2012-01/msg00232.html
Bootregged on x86-64 Linux.
OK?
* gimple.c (is_gimple_non_addressable): Remove.
* gimple.h: Remove is_gimple_non_addressable.
* gimplify.c (gimplify_modify_expr_rhs): Use is_gimple_reg instead
Easy, just mimics what we do for the target attribute.
Bootstrap / regtest pending on x86_64-unknown-linux-gnu.
Richard.
2012-01-05 Richard Guenther
PR lto/50490
* tree-streamer-out.c (write_ts_optimization): New function.
(streamer_write_tree_body): Call it.
Hi!
Jason has added var ={v} {CLOBBER}; cleanup generation in the gimplifier,
unfortunately it works only for C++/Obj-C++, because other FEs don't add
the needed CLEANUP_POINT_EXPRs to the IL (because they don't use cleanups or
emit them differently). Fixed by only adding the clobber cleanups
if
Hi!
The attached testcase on i?86 (32-bit) ICEs, because the csa pass
cross-jumps the two noret calls, eventhough they have different args size
depths.
The following patch attempts to fix that up by forcibly adding REG_ARGS_SIZE
notes to all noreturn calls when not accumulating outgoing args, whi
Committed.
Richard.
2012-01-05 Richard Guenther
* g++.dg/torture/pr49309.C: Skip for -flto.
Index: gcc/testsuite/g++.dg/torture/pr49309.C
===
--- gcc/testsuite/g++.dg/torture/pr49309.C (revision 182912)
+++ gcc/tes
On Thu, Jan 5, 2012 at 3:32 PM, Aldy Hernandez wrote:
>
>> You want is_gimple_reg () instead of is_gimple_non_addressable () as you
>> can't simply make something addressable at this point.
>> is_gimple_non_addressable
>> looks like a weird redundant predicate to me - it's only used once and its
>
On Thu, Jan 5, 2012 at 3:48 PM, Aldy Hernandez wrote:
> As you suggested here:
>
> http://gcc.gnu.org/ml/gcc-patches/2012-01/msg00232.html
>
> Bootregged on x86-64 Linux.
>
> OK?
No, this pessimizes things too much at gimplification time, please simply
inline the predicate instead.
Richard.
Hello,
We crash upon these alias-declarations in which the identifier part is
invalid. This is because make_id_declarator is called with an
error_mark_node for the identifier.
The patch below just avoids calling make_id_declarator in this case.
It try to get out as early as possible to avoid the
Is there a better way to detect early inline phase and ipa_inline
pass? Use always_inline_functions_inlined flag seems hacky.
David
On Wed, Jan 4, 2012 at 1:12 AM, Dehao Chen wrote:
> Hi,
>
> This patch:
>
> * dump inline decisions with profile info whenever available.
> * disable dump of einlin
On Tue, Jan 3, 2012 at 9:36 PM, Richard Henderson wrote:
> On 01/03/2012 09:50 PM, Uros Bizjak wrote:
>> I tried to investigate -mbig-endian a bit, but unfortunately almost
>> all gcc.dg/torture/vshuf-*.c tests FAIL with -O2 on unpatched gcc.
>> Tests pass with -O0, though.
>
> Tests without -O do
On Thu, Jan 5, 2012 at 4:01 PM, Richard Guenther wrote:
> Index: gcc/tree-streamer-out.c
> ===
> --- gcc/tree-streamer-out.c (revision 182908)
> +++ gcc/tree-streamer-out.c (working copy)
> @@ -767,6 +767,27 @@ write_ts_target
OK.
Jason
Hi.
This patch removes obsolete PREFERRED_RELOAD_CLASS macro from the SCORE
back end in the GCC and introduces equivalent TARGET_PREFERRED_RELOAD_CLASS
target hook.
Untested.
OK to install?
* config/score/score.h (PREFERRED_RELOAD_CLASS): Remove.
* config/score/score
Hi.
This patch removes obsolete PREFERRED_RELOAD_CLASS macro from the BFIN back
end in the GCC and introduces equivalent TARGET_PREFERRED_RELOAD_CLASS target
hook.
Compiled. Untested.
OK to install?
* config/bfin/bfin.h (PREFERRED_RELOAD_CLASS): Remove.
* config/bfin
Hi.
This patch removes obsolete PREFERRED_RELOAD_CLASS macro from the ARM back
end in the GCC and introduces equivalent TARGET_PREFERRED_RELOAD_CLASS target
hook.
Bootstrapped and regression tested on arm-unknown-linux-gnueabi.
OK to install?
* config/arm/arm.h (PREFERRED_RE
Hi,
the testcase ICE on negative function body size estimates. This is result of
code in estimate_edge_devirt_benefit that attempts to benefit inlining of
functions with callback whose callee can be devirtualized and become inlinable
and the inlining is profitable.
The logic here is wrong: the si
Hi!
On the following testcase we ICE, because one of the asm gotos degenerates
(has the same fallthru block as jump label destination) and the target
basic block doesn't immediately follow the block ending in the asm goto.
During outof_cfglayout force_nonfallthru_and_redirect incorrectly
adds an u
Hi!
We use the same goto label for duplicate names of outputs/inputs/labels,
but unlike outputs/inputs names where the STRING_CSTs are at
TREE_PURPOSE (TREE_PURPOSE (i)), the STRING_CSTs for labels are
at TREE_PURPOSE (i), thus if we need to report duplicate labels
(either duplicate to some input/
On 01/06/2012 07:05 AM, Jakub Jelinek wrote:
> PR middle-end/51768
> * stmt.c (check_unique_operand_names): Don't ICE during error
> reporting if i is from labels chain.
>
> * c-c++-common/pr51768.c: New test.
Ok.
r~
On 01/06/2012 07:00 AM, Jakub Jelinek wrote:
> PR rtl-optimization/51767
> * cfgrtl.c (force_nonfallthru_and_redirect): Force addition
> of jump_block and add an extra edge for degenerated asm gotos.
>
> * gcc.c-torture/compile/pr51767.c: New test.
Ok.
r~
On 01/06/2012 04:46 AM, Uros Bizjak wrote:
> 2012-01-05 Uros Bizjak
>
> PR target/51681
> * config/ia64/ia64.c (expand_vec_perm_shrp): Use correct operands
> for shrp pattern. Correctly handle and fixup shift variable.
> Return false when shift > nelt for BYTES_BIG_ENDI
On 01/06/2012 02:17 AM, Jakub Jelinek wrote:
> PR debug/51762
> * calls.c (emit_call_1): For noreturn calls force a REG_ARGS_SIZE
> note when !ACCUMULATE_OUTGOING_ARGS.
>
> * gcc.dg/pr51762.c: New test.
Ok.
r~
This removes a duplicated
if (code == RSHIFT_EXPR)
condition in extract_range_from_binary_expr_1 and avoids useless computations
in extract_range_from_unary_expr_1. No functional changes.
Tested on i586-suse-linux, applied on the mainline as obvious.
2012-01-05 Eric Botcazou
*
*ping*
On 30 December 2011, Tobias Burnus wrote:
Dear all,
first, I want to wish all of you a happy New Year.
Attached you find a patch which calls _gfortran_caf_deregister for
allocatable coarrays - for -fcoarray=lib. In the caf libraray version,
coarrays are allocated/deallocated in the li
On 01/05/2012 08:30 PM, Richard Guenther wrote:
> Do all targets support sth like weak relocs?
Nearly. Those that don't probably don't pass the pthreads requirement.
r~
On 01/06/2012 12:53 AM, Aldy Hernandez wrote:
> http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01088.html
It looks like it would be cleaner to put the .alias dereferencing
directly into get_cg_data.
r~
This patch fixes a testcase that fails whenever "120" appears in the
date. I first noticed the problem at the start of December, and now I
expect it to return and continue until the end of September, if it's not
fixed.
OK?
Andrew
2012-01-05 Andrew Stubbs
gcc/testsuite/
* gcc.target/arm/
Hello,
In the example of the patch below, the DIE for the definition of the
class 'Executor', when put in the type unit section is not put under
the DIE for the 'thread' namespace. Rather, it's put under the DIE
for the global namespace.
One thing to note is that the DIE for the declaration of '
On Thu, Jan 05, 2012 at 11:47:49PM +0100, Dodji Seketeli wrote:
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/debug/dwarf2/nested-3.C
> @@ -0,0 +1,34 @@
> +// Origin: PR debug/45682
> +// { dg-options "-g -gdwarf-4 -dA -fdebug-types-section" }
> +
> +namespace thread {
> +class Executor {};
> +}
On 12/21/2011 10:04 PM, Tristan Gingold wrote:
> * config/ia64/t-ia64 (LIB1ASMFUNCS): Move backward
> compatibility thunks...
> (CUSTOM_CRTSTUFF, crtbegin.o, crtend.o)
> (crtbeginS.o, crtendS.o): ... and these to ...
> * config/ia64/t-ia64-elf: ... this new file.
>
On Darwin, the memcpy is a macro to provide secure string operations
(FORTIFY_LEVEL > 0). I propose to remove the string.h include and add
manually the declaration.
Tested on x86_64-apple-darwin and on x86_64-unknown-linux-gnu.
This solves PR51655.
Patrick Marlier.
testsuite/ChangeLog
2012-01
On Wed, 2012-01-04 at 09:59 +1100, Richard Henderson wrote:
> On 01/03/2012 09:42 PM, Torvald Riegel wrote:
> >> Why does the explicit CFG approach not work exactly? cfun->calls_setjmp
> >> is
> >> thought to be quite a big hammer.
> >
> > I don't know, actually. When I looked at the miscompil
> In the example of the patch below, the DIE for the definition of the
> class 'Executor', when put in the type unit section is not put under
> the DIE for the 'thread' namespace. Rather, it's put under the DIE
> for the global namespace.
I don't think this is the problem that PR 45682 is complai
This fixes a known, obvious performance problem in libitm's undo log:
Before, each write was put into a separately allocated buffer, which
kills performance. With the patch, everything goes into a single
buffer: Undo data first, then the length of the data, then the address.
We only need this data
Ping.
I think this is a useful improvement to the docs and could prevent the
most commonly-encountered bootstrap failure for inexpert users
building GCC.
OK for trunk?
On 30 December 2011 13:29, Jonathan Wakely wrote:
On 22 December 2011 00:23, Jonathan Wakely wrote:
> The most frequently asked
2012-01-06 Jonathan Wakely
PR libstdc++/51504
* doc/xml/manual/debug.xml: Suggest using symbol interposition
to override symbols using annotation macros.
committed to trunk
Index: doc/xml/manual/debug.xml
=
Hello Iain,
On 11/21/2011 10:17 AM, Iain Sandoe wrote:
This checked to DTRT on i686-darwin9 (weak ref doesn't behave like
ELF's) and x86-64-darwin10/XCode 3.2.5 (weak ref behaves as per ELF).
It is my understanding from some off-list discussion reported with the
ld64 engineer -
- that Darwin >=
Or use a new global variable to denote whether it's in early-inline or
ipa-inline?
Dehao
On Fri, Jan 6, 2012 at 1:46 AM, Xinliang David Li wrote:
>
> Is there a better way to detect early inline phase and ipa_inline
> pass? Use always_inline_functions_inlined flag seems hacky.
>
> David
>
> On W
Hi Anatoly,
The patch looks OK.
But I cannot apply your patch by saving your email as a patch file. If
you take a look at this:
http://gcc.gnu.org/cgi-bin/get-raw-msg?listname=gcc-patches&date=2012-01&msgid=4F05F12F.607%40post.ru
you will find that there is a extra white space before each c
Ping. Reposting the patch as it required updates to HEAD.
r~
commit 04fbf0c29c0836c6ac8a1342a9eb664a88b696ee
Author: Richard Henderson
Date: Thu Dec 8 12:11:08 2011 -0800
arm: Implement vec_perm and vec_perm_const for NEON.
diff --git a/gcc/config/arm/arm-protos.h b/gcc/config/arm/arm-p
Not ideal but better. Ok with this change.
David
On Thu, Jan 5, 2012 at 5:47 PM, Dehao Chen wrote:
> Or use a new global variable to denote whether it's in early-inline or
> ipa-inline?
>
> Dehao
>
> On Fri, Jan 6, 2012 at 1:46 AM, Xinliang David Li wrote:
>>
>> Is there a better way to detect
59 matches
Mail list logo
