Hi,
This patch fixes PR 49352 by ignoring debug uses in SLP reduction detection.
While fixing it Jakub also discovered that an incorrect statement may
be analyzed and operands of not commutative operation may be swapped.
The patch fixes those as well.
Bootstrapped and tested on powerpc64-suse-lin
On Jun 12, 2011, at 3:55 AM, Richard Guenther wrote:
> In almost all cases(*) the need for a lvalue VIEW_CONVERT_EXPR can be avoided
> by moving the VIEW_CONVERT_EXPR to the rvalue assigned too it. Remember that
> VIEW_CONVERT_EXPR always conver the full object and are not allowed to
> change size
On Jun 12, 2011, at 4:03 AM, Richard Guenther wrote:
> Btw, see tree.def which says
>
> /* Represents viewing something of one type as being of a second type.
> This corresponds to an "Unchecked Conversion" in Ada and roughly to
> the idiom *(type2 *)&X in C. The only operand is the value to
On Mon, Jun 13, 2011 at 12:24:02AM +0200, Thomas Koenig wrote:
> Hello world,
>
> this is the last round of TRIM optimizations. This patch extends the
> treatment of trailing TRIMs in concatenations to comparisions. It also
> does a bit of code cleanup by removing some duplication, and by not
>
Hello world,
this is the last round of TRIM optimizations. This patch extends the
treatment of trailing TRIMs in concatenations to comparisions. It also
does a bit of code cleanup by removing some duplication, and by not
changing the rhs in optimize_assignment.
OK for trunk?
Thomas
2
On 06/12/2011 06:59 AM, Richard Guenther wrote:
The please provide a specification on what a VIEW_CONVERT_EXPR does
to type-based alias analysis.
If the alias set of the VIEW_CONVERT_EXPR type the same as the set for
the operand, ignore it; if it's a subset, handle it like a
COMPONENT_REF; ot
Hi,
I had this patch sitting in my tree for a while, but somehow forgot to commit
it.
The issue here is that edge redirection affects vops and thus imply need for
SSA update in some cases.
Bootstrapped/regtested x86_64-linux, will commit it shortly.
PR middle-end/48836
* ipa-inlin
Hi,
this patch solves the bultin/strlen-3.c LTO linker plugin problem.
While removing alias code I was bit overactive and removed the check that makes
us to implicitly
do -fwhole-program when resolution info is around.
It is not quite clear to me why in LTO we need -fwhole-program to get the
tes
On Sun, Jun 12, 2011 at 9:40 AM, Jan Hubicka wrote:
> Hi,
> this is the fix (or rather a workaround) i comitted. Thanks!
>
> Index: ChangeLog
> ===
> --- ChangeLog (revision 174968)
> +++ ChangeLog (working copy)
> @@ -1,3 +1,9 @
On Sun, Jun 12, 2011 at 9:33 AM, Jan Hubicka wrote:
>> On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka wrote:
>> >>
>> >> This also pretty much destroyed C++ for ia32:
>> >>
>> >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
>> >> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
>> >
Hi,
this is the fix (or rather a workaround) i comitted. Thanks!
Index: ChangeLog
===
--- ChangeLog (revision 174968)
+++ ChangeLog (working copy)
@@ -1,3 +1,9 @@
+2011-06-11 Jan Hubicka
+
+ PR middle-end/49378
+ *
> On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka wrote:
> >>
> >> This also pretty much destroyed C++ for ia32:
> >>
> >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
> >> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
> >
> > Hi,
> > It seems somewhat amazing that we hit kernel sen
On Sun, Jun 12, 2011 at 7:33 AM, H.J. Lu wrote:
> On Sun, Jun 12, 2011 at 7:00 AM, H.J. Lu wrote:
>> On Sun, Jun 12, 2011 at 6:50 AM, Richard Guenther
>> wrote:
>>> On Sun, Jun 12, 2011 at 3:18 PM, H.J. Lu wrote:
On Sun, Jun 12, 2011 at 3:48 AM, Richard Guenther
wrote:
> On Sat,
On Sun, Jun 12, 2011 at 8:01 AM, H.J. Lu wrote:
> On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka wrote:
>>>
>>> This also pretty much destroyed C++ for ia32:
>>>
>>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
>>> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
>>
>> Hi,
>> It seem
On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka wrote:
>>
>> This also pretty much destroyed C++ for ia32:
>>
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
>> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
>
> Hi,
> It seems somewhat amazing that we hit kernel sensitive miscompilat
Hi,
Java HelloWorld compiled by cross gcj for RTEMS can run without
problem now, I think it's time for get the boehm-gc patch reviewed and
merged. :) Here is the patch:
Index: mach_dep.c
===
--- mach_dep.c (revision 172224)
+++ ma
> I am testing it now. Will know the results in 2 hours.
Thanks.
Could you also send me the preprocessed source for future.o? The object file I
am getting don't have the bug you report, that is most probably due to glibc
difference.
Honza
>
> Thanks.
>
> --
> H.J.
On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka wrote:
>>
>> This also pretty much destroyed C++ for ia32:
>>
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
>> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
>
> Hi,
> It seems somewhat amazing that we hit kernel sensitive miscompilat
>
> This also pretty much destroyed C++ for ia32:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
Hi,
It seems somewhat amazing that we hit kernel sensitive miscompilation here.
The problem most probably is the fact that thunks a
On Sun, Jun 12, 2011 at 7:00 AM, H.J. Lu wrote:
> On Sun, Jun 12, 2011 at 6:50 AM, Richard Guenther
> wrote:
>> On Sun, Jun 12, 2011 at 3:18 PM, H.J. Lu wrote:
>>> On Sun, Jun 12, 2011 at 3:48 AM, Richard Guenther
>>> wrote:
On Sat, Jun 11, 2011 at 5:09 PM, H.J. Lu wrote:
> Hi,
>
On Sun, Jun 12, 2011 at 6:50 AM, Richard Guenther
wrote:
> On Sun, Jun 12, 2011 at 3:18 PM, H.J. Lu wrote:
>> On Sun, Jun 12, 2011 at 3:48 AM, Richard Guenther
>> wrote:
>>> On Sat, Jun 11, 2011 at 5:09 PM, H.J. Lu wrote:
Hi,
expand_builtin_strlen has
src_reg = gen_reg_
On Sun, Jun 12, 2011 at 3:18 PM, H.J. Lu wrote:
> On Sun, Jun 12, 2011 at 3:48 AM, Richard Guenther
> wrote:
>> On Sat, Jun 11, 2011 at 5:09 PM, H.J. Lu wrote:
>>> Hi,
>>>
>>> expand_builtin_strlen has
>>>
>>> src_reg = gen_reg_rtx (Pmode);
>>> ...
>>> pat = expand_expr (src, src_reg, ptr_mode,
Hi,
Revision 174952:
http://gcc.gnu.org/ml/gcc-cvs/2011-06/msg00441.html
totally breaks C++ on x86:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
One symptom is we are using uninitialized registers, which leads
to writing random me
On Thu, 9 Jun 2011, Douglas B Rupp wrote:
Comments welcome.
--- gcc.orig/gcc/doc/install.texi 2011-06-08 14:05:22.0 -0700
+++ gcc/gcc/doc/install.texi2011-06-09 10:04:38.0 -0700
@@ -4499,7 +4499,7 @@
The Interix target is used by OpenNT, Interix, Services For UNIX (SFU
On Sun, Jun 12, 2011 at 3:48 AM, Richard Guenther
wrote:
> On Sat, Jun 11, 2011 at 5:09 PM, H.J. Lu wrote:
>> Hi,
>>
>> expand_builtin_strlen has
>>
>> src_reg = gen_reg_rtx (Pmode);
>> ...
>> pat = expand_expr (src, src_reg, ptr_mode, EXPAND_NORMAL);
>> if (pat != src_reg)
>> emit_move_insn (sr
On Sun, Jun 12, 2011 at 12:59 PM, Richard Guenther
wrote:
> On Sat, Jun 11, 2011 at 10:01 PM, Jason Merrill wrote:
>> On 06/10/2011 10:20 AM, Richard Guenther wrote:
>>>
>>> no, a VIEW_CONVERT_EXPR is generally not an lvalue (fold for example
>>> would turn the above to (volatile int) a[1]).
>>
>
On Sat, Jun 11, 2011 at 10:01 PM, Jason Merrill wrote:
> On 06/10/2011 10:20 AM, Richard Guenther wrote:
>>
>> no, a VIEW_CONVERT_EXPR is generally not an lvalue (fold for example
>> would turn the above to (volatile int) a[1]).
>
> The gimplifier seems to consider it an lvalue: gimplify_expr uses
On Sat, Jun 11, 2011 at 7:45 PM, Mike Stump wrote:
> On Jun 10, 2011, at 7:20 AM, Richard Guenther wrote:
>> On Fri, 10 Jun 2011, Jason Merrill wrote:
>>
>>> On 06/10/2011 10:03 AM, Richard Guenther wrote:
>> *((volatile int *)&a[0] + 1)
>
> It would be correct to fold it to
>
On Sat, Jun 11, 2011 at 5:09 PM, H.J. Lu wrote:
> Hi,
>
> expand_builtin_strlen has
>
> src_reg = gen_reg_rtx (Pmode);
> ...
> pat = expand_expr (src, src_reg, ptr_mode, EXPAND_NORMAL);
> if (pat != src_reg)
> emit_move_insn (src_reg, pat);
>
> But src_reg may be in ptr_mode, wich may not be the
On Fri, Jun 10, 2011 at 9:15 PM, Jan Hubicka wrote:
>> On Fri, Jun 10, 2011 at 8:42 PM, Jan Hubicka wrote:
>> > Hi,
>> > by some mistake we enable functions called once at -O0 and it actually
>> > happens from
>> > time to time.
>>
>> Why do it for -O1? It definitely makes debugging less reliab
I've just committed this patch. It makes match_operator modeless for
moxie comparisons. This fixes a regression introduced in March by this
patch: http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01344.html
2011-06-12 Anthony Green
* config/moxie/moxie.md (cbranchsi4): Remove mode from
2011/6/10 Georg-Johann Lay :
> Then I have a question on spill failures:
>
> There is PR46278, an optimization flaw that goes as follows:
>
> The avr BE defines fake addressing mode X+const that has to be written
> down in asm as
> X += const
> a = *X
> X -= const
>
> The comment says that this
Hi,
gcc.dg/vect/vect-72.c is not expected to use loop peeling for
alignment, but it does on ARM with double-word vectors. The loop
contains two data-refs of type char: one is aligned and the other is
misaligned by 1. When the cost model is disabled the peeling heuristic
chooses to peel a number of
33 matches
Mail list logo
