On Tue, 7 May 2013, Bill Schmidt wrote:
> The intent of conditional candidate processing in SLSR was always to
> apply it only to CAND_MULT candidates [(base + index) * stride].
> However, I neglected to actually enforce this, leading to wrong code
> generated for CAND_ADD candidates. This patch
Hi,
I personally consider the fix for this this issue obvious: barring
objections I mean to commit it mainline and branch later today.
Thanks,
Paolo.
///
2013-05-08 Paolo Carlini
PR tree-optimization/57200
* tree-ssa-loop-niter.c (do_warn_aggressive_loo
DJ Delorie writes:
>> > - if (GET_MODE_BITSIZE (GET_MODE (op0)) < GET_MODE_BITSIZE (GET_MODE
>> > (op1)))
>> > + /* Don't try to sign-extend SImode to PSImode, for example. */
>> > + if (GET_MODE_BITSIZE (GET_MODE (op0)) <= GET_MODE_BITSIZE
>> > (GET_MODE (op1)))
>>
>> Can't you do just
As described in the PR, amd64-pc-solaris2.1[01] Ada bootstrap was failing
for some time. It has turned out that this patch is the culprit:
2013-04-23 Eric Botcazou
Pascal Obry
* gcc-interface/Makefile.in (targ): Fix target name check.
diff --git a/gcc/ada/gcc-interfac
On 30/04/13 16:27, Sofiane Naci wrote:
Hi,
The attached patch is part of the ongoing work to enhance instruction
scheduling opportunities in AArch64.
This patch splits a few alternatives of the movtf_aarch64 pattern, which
used to issue 2 consecutive instructions.
Tested successfully with a ful
Hi,
This patch implements the BFI variant of BFM. In doing so, it also
implements the insv standard pattern.
I've regression tested on bare-metal and linux.
It comes complete with its own compilation and execution testcase.
OK for trunk?
Cheers,
Ian
2013-05-08 Ian Bolton
gcc/
*
Hi,
The attached patch is part of the ongoing work to enhance instruction
scheduling opportunities in AArch64.
This patch splits the aarch64_combine pattern, which used to issue 2
consecutive instructions.
Tested successfully with a full aarch64-elf regression run.
OK for trunk?
Thanks
Sofiane
This patch should be applied to all open GCC branches.
This patch removes the mthumb/march=armv7 since it is virtually equal to
the mthumb/march=armv7-m multilib. Add variants for ARMv7-A and
ARMv7-R.
I tried to use the MULTILIB_REQUIRED option, but this didn't work for
patterns with more than t
Fixes an LTO documentation typo in gcc internals.
08-05-2013 Paulo Matos
* gcc/doc/lto.texi: Fix typo.
Paulo Matos
lto_docfix_pr50345.patch
Description: lto_docfix_pr50345.patch
On Wed, May 08, 2013 at 10:56:35AM +0200, Paolo Carlini wrote:
> I personally consider the fix for this this issue obvious: barring
> objections I mean to commit it mainline and branch later today.
>
> Thanks,
> Paolo.
>
> ///
> 2013-05-08 Paolo Carlini
>
> PR tree-
On Wed, May 8, 2013 at 12:10 PM, Paulo Matos wrote:
> Fixes an LTO documentation typo in gcc internals.
>
> 08-05-2013 Paulo Matos
> * gcc/doc/lto.texi: Fix typo.
Ok.
Thanks,
Richard.
> Paulo Matos
>
>
Hi DJ,
* Uses register bank 3 instead of register bank 0 inside the
handler.
I wonder if nested interrupts will cause problems, since the second
interrupt will occur with RB3 already selected.
Yes - in fact this appears to be a general problem with RL78 interrupt
handlers. Ther
This patch fixes a problem with section anchors. Found on powerpc, but also
appears on MIPS and ARM targets.
Section anchors can only be used for definitions known to bind in the current
object file. The default predicate uses the bind_local_p hook to determine
this. Unfortunately that hook
> 2013-05-08 Nathan Sidwell
>
> gcc/
> * varasm.c (default_use_anchors_for_symbol_p): Reject WEAK.
>
> gcc/testsuite/
> * gcc.dg/visibility-21.c: New.
>
> Index: gcc/varasm.c
> ===
> --- gcc/varasm.c
I would like to hear from an ARM maintainer before I commit this.
--joel
On 5/8/2013 5:06 AM, Sebastian Huber wrote:
This patch should be applied to all open GCC branches.
This patch removes the mthumb/march=armv7 since it is virtually equal to
the mthumb/march=armv7-m multilib. Add variants
On 05/08/13 14:11, Jan Hubicka wrote:
With LTO we play more of similar tricks, by making use of the resolution file.
I.e. for COMMON and EXTERNAL. Does it matter here?
probably. It'd be a missed optimization though, rather than wrong code
emission.
I would preffer the renaming excercise,
On 05/07/2013 03:46 AM, Jakub Jelinek wrote:
Here in a form of patch, seems to work and the patch is 12KB shorter
that way.
This patch is OK.
Jason
OK.
Jason
Hello!
2013-05-08 Uros Bizjak
* config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
(*vec_extract): Use VI12_128 mode iterator.
(*vec_extract_mem): Ditto.
(*vec_extract*_mem splitters): Merge splitters using VI_128 mode
attribute.
Tested on x86_64-pc-linux-gnu {,-m32}, c
I should also add generic new-port comments:
See the checklist under "Back End" in sourcebuild.texi. In particular,
add the target to contrib/config-list.mk so all-targets builds cover it.
Make sure that the port builds cleanly when configuring with
--enable-werror-always and building with a b
Thanks for fixing this.
Nathan Sidwell writes:
>> I would preffer the renaming excercise, since the name confused me few times,
>> too and the other predicate would be useful for IPA code :)
>
> before invoking sed, can we agree on the right names? I propose
>
> * renaming 'binds_local_p' to 'bi
Hello,
I'd like to make libbacktrace easier to import for use in other software.
Right now it's mostly standalone, but depends on rest of GCC in the following:
1. Build system.
2. Trivially depends on "filenames.h".
3. Depends on "dwarf2.{def,h}" to obtain DW_* enumeration values.
My autotools-f
> Yes - in fact this appears to be a general problem with RL78 interrupt
> handlers. There are no instructions to push/pop the current bank
> selection, so in order for nested interrupt handlers to work they would
> have to do something like:
The bank select bits are in PSW, which is automati
This corrects a signed/unsigned comparison. Regstrapped on
powerpc64-unknown-linux-gnu, committed as obvious.
Thanks,
Bill
2013-05-08 Bill Schmidt
* gimple-ssa-strength-reduction.c (count_candidates): Change
return value to int.
(analyze_candidates_and_replace): Chan
> On 05/08/13 14:11, Jan Hubicka wrote:
>
> >With LTO we play more of similar tricks, by making use of the resolution
> >file.
> >I.e. for COMMON and EXTERNAL. Does it matter here?
>
> probably. It'd be a missed optimization though, rather than wrong code
> emission.
Well, the resolution inf
Hello!
We don't have to use emit_move_insn here, the same can be achieved in
the (post-reload) splitter body:
- [(const_int 0)]
-{
- operands[1] = adjust_address (operands[1], SImode, 4);
- emit_move_insn (operands[0], operands[1]);
- DONE;
-})
+ [(set (match_dup 0) (match_dup 1))]
+ "opera
Hi,
initial patch for supporting x64 add some checks for MS_ABI. That
caused for 32-bit cygwin to some regression due cygwin-32 uses
SYSV_ABI as default ABI. So I introduce in this patch the define
TARGET_PECOFF to indicate if targets use pe-coff or ELF-object/image
sematics.
This approach has a
On Wed, May 8, 2013 at 8:56 AM, Alexander Monakov wrote:
>
> I'd like to make libbacktrace easier to import for use in other software.
> Comments?
> +#ifdef IN_GCC
Where is IN_GCC defined?
> +# Test if DWARF2 extensions enumeration values are declared
> +if test -n "${with_target_subdir}"; th
Hi,
see for details http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57212
ChangeLog libstdc++
2013-05-08 Andoni Morales Alastruey
PR libstdc++/57212
* include/Makefile.am (_GLIBCXX___MINGW32_GLIBCXX___): Modify
to __MINGW32__.
* include/Makefile: Regenerated.
I will apply this pa
On 8 May 2013 15:11:18 Jan Hubicka wrote:
> 2013-05-08 Nathan Sidwell
>gcc/
>* varasm.c (default_use_anchors_for_symbol_p): Reject WEAK.
>gcc/testsuite/
>* gcc.dg/visibility-21.c: New.
> Index: gcc/varasm.c
> ===
ok.
David
On Tue, May 7, 2013 at 7:27 PM, Dehao Chen wrote:
> This patch updated the unittest and doc for the new
> -frecord-compilation-info-in-elf flag.
>
> Bootstrapped and passed regression test.
>
> OK for google-4_7 branch?
>
> Thanks,
> Dehao
>
> Index: gcc/testsuite/gcc.dg/record-gcc-swi
On Wed, May 8, 2013 at 8:04 PM, Uros Bizjak wrote:
> We don't have to use emit_move_insn here, the same can be achieved in
> the (post-reload) splitter body:
>
> - [(const_int 0)]
> -{
> - operands[1] = adjust_address (operands[1], SImode, 4);
> - emit_move_insn (operands[0], operands[1]);
> -
VLA types in diagnostics have tended to look like
int[(((sizetype)(((ssizetype)i) + -1)) + 1)]
...which is horrible. This patch fixes that to
int[i]
and also specifically diagnoses trying to use a variably modified type
in throw or catch.
Tested x86_64-pc-linux-gnu, applying to trunk.
comm
On 05/07/2013 04:41 AM, Richard Biener wrote:
Eh, two NULL_TREE terminators are ugly ... callers know the number of elements,
so maybe instead pass that number as argument?
OK, sure.
Can we overload build_constructor with a variadic variant? Thus,
tree build_constructor (tree type, vec *val
> It looks like the patch also treats PSImode->SImode as a truncation though.
I'd have to check specifically for partial_int_mode, since there's
nothing in PARTIAL_INT_MODE() to set the precision, unless there's
some other way to set PSImode's bits *and* reliably use that info.
For some reason,
On Wed, May 08, 2013 at 07:56:18PM +0400, Alexander Monakov wrote:
> Hello,
>
> I'd like to make libbacktrace easier to import for use in other software.
> Right now it's mostly standalone, but depends on rest of GCC in the following:
>
> 1. Build system.
> 2. Trivially depends on "filenames.h".
On Wed, May 8, 2013 at 2:35 PM, Ondřej Bílka wrote:
>
> Is it possible to link staticaly without external dependencies? This
> simplifies generating backtrace when inside linker.
I'm not sure I understand the question. Is it possible to link
libbacktrace statically? Sure. That is what happens
Hello,
here are a few more changes to fold-const.c so vectors can use the
existing optimizations. Note that I made fold_truth_not_expr safe for use
with vector BIT_NOT_EXPR.
Passes bootstrap+testsuite on x86_64-linux-gnu.
2013-05-09 Marc Glisse
gcc/
* fold-const.c (fold_negate_ex
Hi, all,
There was a patch to remove libgcc_tm.stamp when developers
manually issue "make clean-target-libgcc" (under top-level build folder)
or "make clean" (under BUILD-FOLDER/TARGET/libgcc).
Without that patch, we cannot successfully do
"make all-target-libgcc" again due to that we
remove libg
A case of giving an error during tentative parsing; we need to wait
until we know for sure that we're doing something dodgy.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.8.
commit 4d01649f80e76af9be37b35d166a00990defa978
Author: Jason Merrill
Date: Wed May 8 12:33:26 2013 -0400
P
On 05/01/2013 07:40 PM, Mike Stump wrote:
$ bash -c "trap 'echo remove lock' 0; true"; echo $?
Thanks for the suggestion. Here's a revised patch:
commit 79b52becc6a8649933f7442d46e8054f8b6abd4b
Author: Jason Merrill
Date: Thu Apr 25 17:54:17 2013 -0400
* Makefile.in (link-mutex, re
On 2013/5/9 04:11 AM, Bernhard Reutner-Fischer wrote:
> On 8 May 2013 15:11:18 Jan Hubicka wrote:
>> > 2013-05-08 Nathan Sidwell
>> > gcc/
>> > * varasm.c (default_use_anchors_for_symbol_p): Reject WEAK.
>> > gcc/testsuite/
>> > * gcc.dg/visibility-21.c: New.
>> > Index: gcc/var
On 05/08/13 18:47, Jan Hubicka wrote:
On 05/08/13 14:11, Jan Hubicka wrote:
* a new function or hook 'binds_here_p' (and any default
implementations needed). Returns true if the symbol is known to
bind in the object file being emitted by the current compilation.
Thinking about it again, isn
Hi Maintainers,
The attach patch contains the following for aarch64 backend to enable
gprof support.
1. Changes to "aarch64_return_addr" to get return address from a stack frame.
2. Defines macros associated with generating code for profiling.
gcc/ChangeLog
-
2013-05-09 Ven
44 matches
Mail list logo
