On Tue, Sep 11, 2012 at 03:55:04PM -0700, Ian Lance Taylor wrote:
> +/* The DWARF abbreviations for a compilation unit. This structure
> + only exists while reading the compilation unit. Most DWARF readers
> + seem to a hash table to map abbrev ID's to abbrev entries.
> + However, we primar
Can you make definition of location_adhoc_data available in both
libcpp and gcc? In gcc side, the data structure will be annotated
with GTY(). The root global variables should also be visible in gcc
side.
David
On Tue, Sep 11, 2012 at 10:06 PM, Dehao Chen wrote:
> Now I think we are facing a mo
Now I think we are facing a more complex problem. The data structure
we use to store the location_adhoc_data are file-static in linemap.c
in libcpp. These data structures are not guarded by GTY(()).
Meanwhile, as we have removed the block data structure from
gimple.gsbase as well as tree.exp (encod
On Tue, Sep 11, 2012 at 11:49 PM, Ian Lance Taylor wrote:
> On Tue, Sep 11, 2012 at 9:36 PM, Chris Lattner wrote:
>>
>> On Sep 11, 2012, at 4:24 PM, Ian Lance Taylor wrote:
>>
>>> On Tue, Sep 11, 2012 at 4:07 PM, Chris Lattner wrote:
FYI, "backtrace" is a well-known function provide b
On Tue, Sep 11, 2012 at 9:36 PM, Chris Lattner wrote:
>
> On Sep 11, 2012, at 4:24 PM, Ian Lance Taylor wrote:
>
>> On Tue, Sep 11, 2012 at 4:07 PM, Chris Lattner wrote:
>>>
>>> FYI, "backtrace" is a well-known function provide by glibc (and other
>>> libc's). It might be best to pick another
On Sep 11, 2012, at 4:24 PM, Ian Lance Taylor wrote:
> On Tue, Sep 11, 2012 at 4:07 PM, Chris Lattner wrote:
>>
>> FYI, "backtrace" is a well-known function provide by glibc (and other
>> libc's). It might be best to pick another name.
>
> Good point. Thanks. I renamed it to backtrace_ful
On Fri, 31 Aug 2012, Hans-Peter Nilsson wrote:
> On Fri, 3 Aug 2012, Hans-Peter Nilsson wrote:
> Sorry, but I need more time to get
> test-results for a recent trunk revision to a state where
> testing patches makes sense. I'll test the patch and commit
> that or a variant. Thanks for your work.
On Tue, Sep 11, 2012 at 5:13 PM, Cary Coutant wrote:
>> The enclosed patch fixes yet another bug with the gdb-index v7 flag byte
>> by reporting certain structs and types as static only for non-C++ and
>> non-java.
>>
>> OK for Google 4.7?
>>
>> Sterling
>>
>> 2012-09-11 Sterling Augustine
>>
>
> The enclosed patch fixes yet another bug with the gdb-index v7 flag byte
> by reporting certain structs and types as static only for non-C++ and
> non-java.
>
> OK for Google 4.7?
>
> Sterling
>
> 2012-09-11 Sterling Augustine
>
> * dwarf2out.c (output_pubname):
OK for google/gcc-4_7.
On Tue, Sep 11, 2012 at 3:12 PM, Lawrence Crowl wrote:
> On 9/11/12, Andreas Schwab wrote:
>> Mark Kettenis writes:
>>> In file included from ../../../src/gcc/gcc/mcf.c:47:0:
>>> ../../../src/gcc/gcc/mcf.c: In function 'void dump_fixup_edge(FILE*,
>>> fixup_graph_type*, fixup_edge_p)':
>>> ../..
On Tue, Sep 11, 2012 at 4:36 PM, Lawrence Crowl wrote:
> On 9/11/12, Ian Lance Taylor wrote:
>> This patch is the interface to and configury of libbacktrace.
>> I've separated these out as the parts of libbacktrace that
>> require the most review. The interface to libbacktrace is in
>> the file
On 9/11/12, Ian Lance Taylor wrote:
> This patch is the interface to and configury of libbacktrace.
> I've separated these out as the parts of libbacktrace that
> require the most review. The interface to libbacktrace is in
> the file backtrace.h. This is what callers will use. The file
> backt
On Tue, Sep 11, 2012 at 4:07 PM, Chris Lattner wrote:
>
> FYI, "backtrace" is a well-known function provide by glibc (and other
> libc's). It might be best to pick another name.
Good point. Thanks. I renamed it to backtrace_full. I've included
the updated backtrace.h here. The other changes
On Tue, Sep 11, 2012 at 6:08 PM, Ian Lance Taylor wrote:
> On Tue, Sep 11, 2012 at 4:01 PM, Gabriel Dos Reis
> wrote:
>> On Tue, Sep 11, 2012 at 5:53 PM, Ian Lance Taylor wrote:
>>
>>> This patch is the interface to and configury of libbacktrace. I've
>>> separated these out as the parts of lib
The enclosed patch fixes yet another bug with the gdb-index v7 flag byte
by reporting certain structs and types as static only for non-C++ and
non-java.
OK for Google 4.7?
Sterling
2012-09-11 Sterling Augustine
* dwarf2out.c (output_pubname):
Index: dwarf2out.c
=
On Tue, Sep 11, 2012 at 4:01 PM, Gabriel Dos Reis
wrote:
> On Tue, Sep 11, 2012 at 5:53 PM, Ian Lance Taylor wrote:
>
>> This patch is the interface to and configury of libbacktrace. I've
>> separated these out as the parts of libbacktrace that require the most
>> review. The interface to libba
On Sep 11, 2012, at 3:53 PM, Ian Lance Taylor wrote:
> I have finished the initial implementation of the backtrace library I
> proposed at http://gcc.gnu.org/ml/gcc/2012-08/msg00317.html . I've
> separated the work into three patches. These patches only implement the
> backtrace library itself
On Tue, Sep 11, 2012 at 5:53 PM, Ian Lance Taylor wrote:
> This patch is the interface to and configury of libbacktrace. I've
> separated these out as the parts of libbacktrace that require the most
> review. The interface to libbacktrace is in the file backtrace.h. This
> is what callers will
I have finished the initial implementation of the backtrace library I
proposed at http://gcc.gnu.org/ml/gcc/2012-08/msg00317.html . I've
separated the work into three patches. These patches only implement the
backtrace library itself; actual use of the library will follow in
separate patches.
Th
I have finished the initial implementation of the backtrace library I
proposed at http://gcc.gnu.org/ml/gcc/2012-08/msg00317.html . I've
separated the work into three patches. These patches only implement the
backtrace library itself; actual use of the library will follow in
separate patches.
Th
On Tue, Sep 11, 2012 at 3:41 PM, Diego Novillo wrote:
>> @@ -778,6 +791,7 @@ type (options_p *optsp, bool nested)
>> return resolve_typedef (s, &lexer_line);
>>
>> case STRUCT:
>> +case CLASS:
>
>
> I think that as far as gengtype is concerned, 'struct' and 'class' should be
> e
On 9/11/12, Andreas Schwab wrote:
> Mark Kettenis writes:
>> In file included from ../../../src/gcc/gcc/mcf.c:47:0:
>> ../../../src/gcc/gcc/mcf.c: In function 'void dump_fixup_edge(FILE*,
>> fixup_graph_type*, fixup_edge_p)':
>> ../../../src/gcc/gcc/system.h:288:78: error: integer overflow in
>>
On 2012-08-29 20:31 , Aaron Gray wrote:
2012-08-30 Aaron Gray
* gengtype-lex.l: Support for FILE
Support for C++ single line Comments
Support for classes
Support for enums
ignore 'static'
ignore 'inline'
ignore 'public:'
i
Can you resend your patch in text form (also need to resolve the
latest conflicts) so that it can be commented inline?
Please also provide as summary a more up-to-date description of
1) Command line option syntax and semantics
2) New dumping APIs and semantics
3) Conversion changes
Looking at the
On 09/11/2012 10:14 AM, Iyer, Balaji V wrote:
> The function mangling handles several of the version inconsistencies
> you have mentioned. If the CPU revisions, vector lengths are not the
> same between the function declaration and the function, then the name
> of the function will be different and
On 09/11/2012 10:53 AM, Steven Bosscher wrote:
> + force_expand_binop (index_mode, code_to_optab (MINUS),
Use sub_optab directly, rather than code_to_optab.
Otherwise ok.
r~
Hello,
Thanks for the quick review!
On Tue, Sep 11, 2012 at 5:03 PM, Richard Henderson wrote:
> On 09/10/2012 04:26 PM, Steven Bosscher wrote:
>> + rtx index = force_reg (index_mode, dispatch_index);
>
> You can't modify the result of force_reg. Use copy_to_{mode_,}reg instead.
Done.
>> +
On Sep 11, 2012, at 8:06 AM, Christophe Lyon wrote:
> Ping?
> http://gcc.gnu.org/ml/gcc-patches/2012-09/msg00068.html
Since the arm people haven't rejected it… Ok.
On Sep 11, 2012, at 6:12 AM, Kyrylo Tkachov wrote:
> Fixed the format of the test options, as per Jakub's comment.
> Ok for trunk?
Ok.
Please see my answers below
>-Original Message-
>From: Richard Henderson [mailto:r...@redhat.com]
>Sent: Monday, September 10, 2012 12:38 PM
>To: Iyer, Balaji V
>Cc: Richard Guenther; gcc-patches@gcc.gnu.org; Gabriel Dos Reis; Aldy
>Hernandez (al...@redhat.com); Jeff Law
>Subject: Re: [PAT
On Tue, Sep 11, 2012 at 10:37 AM, Jakub Jelinek wrote:
> On Tue, Sep 11, 2012 at 05:29:12PM +0200, Paolo Carlini wrote:
>> PS: slightly interesting, in a couple of cases -
>> write_unnamed_type_name, wrap_cleanups_r - the parameters were
>> actually used.
>
> Just a general comment, often an argum
On 08/03/2012 08:05 AM, Bernd Schmidt wrote:
This patch allows us to change
rn++
rm=[rn]
into
rm=[rn + 4]
rn++
That is an interesting optimization. I think analogous optimization
could be done for INC/DEC addressing (probably it might be beneficial
for ppc which has such addressing and disp
On Tue, Sep 11, 2012 at 05:40:30PM +0200, Steven Bosscher wrote:
> On Tue, Sep 11, 2012 at 5:31 PM, Christian Bruel
> wrote:
> > Actually, the edge is fairly simple. I have
> >
> > BB5 (BB_COLD_PARTITION) -> BB10 (BB_HOT_PARTITION) -> EXIT
> >
> > and BB10 has no other incoming edges. and we are
when running a cfg dump, I get many messages like:
Invalid sum of incoming frequencies 1667, should be 3334
So it looks like a profile information was not correctly propagated
somewhere. which could lead to such partitioning incoherency. I have no
idea for the moment if this is local problem or n
On 09/11/2012 05:37 PM, Jakub Jelinek wrote:
On Tue, Sep 11, 2012 at 05:29:12PM +0200, Paolo Carlini wrote:
PS: slightly interesting, in a couple of cases -
write_unnamed_type_name, wrap_cleanups_r - the parameters were
actually used.
Just a general comment, often an argument is only conditiona
I saw comments in tree-streamer-out.c:
/* Do not stream BLOCK_SOURCE_LOCATION. We cannot handle debug information
for early inlining so drop it on the floor instead of ICEing in
dwarf2out.c. */
streamer_write_chain (ob, BLOCK_VARS (expr), ref_p);
However, what the code is doing se
On Sun, 9 Sep 2012, Marc Glisse wrote:
Hello,
this patch lets the compiler try to rewrite:
(vec_concat (vec_select x [a]) (vec_select x [b]))
as:
vec_select x [a b]
or even just "x" if appropriate.
In a first iteration I was restricting it to b-a==1, but it seemed better not
to: it helps
Hi!
As discussed on IRC, the picochip-* port doesn't have an active maintainer
anymore, this patch adds it to deprecated ports for 4.7.2+ so that it can be
removed in
GCC 4.8 unless somebody steps up to maintain it.
Ok for trunk/4.7?
2012-09-11 Jakub Jelinek
* config.gcc: Obsolete
On 09/11/2012 05:40 PM, Steven Bosscher wrote:
> On Tue, Sep 11, 2012 at 5:31 PM, Christian Bruel
> wrote:
>> Actually, the edge is fairly simple. I have
>>
>> BB5 (BB_COLD_PARTITION) -> BB10 (BB_HOT_PARTITION) -> EXIT
>>
>> and BB10 has no other incoming edges. and we are duplicating it.
>
>
Caught on a sparc build.
Testing on sparc. Will commit once it finishes.
Diego.
* var-tracking.c (vt_add_function_parameter): Adjust for VEC
changes.
diff --git a/gcc/var-tracking.c b/gcc/var-tracking.c
index 8c9ec48..9f5bc12 100644
--- a/gcc/var-tracking.c
+++ b/gcc/var-track
Mark Kettenis writes:
> In file included from ../../../src/gcc/gcc/mcf.c:47:0:
> ../../../src/gcc/gcc/mcf.c: In function 'void dump_fixup_edge(FILE*,
> fixup_graph_type*, fixup_edge_p)':
> ../../../src/gcc/gcc/system.h:288:78: error: integer overflow in expression
> [-Werror=overflow]
This is
On Tue, Sep 11, 2012 at 5:31 PM, Christian Bruel wrote:
> Actually, the edge is fairly simple. I have
>
> BB5 (BB_COLD_PARTITION) -> BB10 (BB_HOT_PARTITION) -> EXIT
>
> and BB10 has no other incoming edges. and we are duplicating it.
That is wrong, should never happen. Is there a test case to pla
On Tue, Sep 11, 2012 at 05:29:12PM +0200, Paolo Carlini wrote:
> PS: slightly interesting, in a couple of cases -
> write_unnamed_type_name, wrap_cleanups_r - the parameters were
> actually used.
Just a general comment, often an argument is only conditionally used,
e.g. depending on some preproces
On 2012-09-11 08:42 , Dominique Dhumieres wrote:
This is ok, of course.
Then could you please commit it (I don't have write access)?
Done. Rev 191192.
2012-09-11 Dominique Dhumieres
* config/darwin.c (darwin_asm_named_section): Adjust for
VEC changes.
(darwin_as
> Index: gcc/ChangeLog
>
> 2012-09-04 Lawrence Crowl
>
> * double-int.h (double_int::operator &=): New.
> (double_int::operator ^=): New.
> (double_int::operator |=): New.
> (double_int::mul_with_sign): Modify overflow parameter to bool*.
> (do
Hi,
On Tue, 11 Sep 2012, Dehao Chen wrote:
> Looks like we have two choices:
>
> 1. Stream out block info, and use LTO_SET_PREVAIL for TREE_CHAIN(t)
This will actually not work correctly in some cases. The problem is, if
the prevailing decl is already part of another chain (say in another
bl
Actually, the edge is fairly simple. I have
BB5 (BB_COLD_PARTITION) -> BB10 (BB_HOT_PARTITION) -> EXIT
and BB10 has no other incoming edges. and we are duplicating it.
My hypothesis, is that with a gcov based profile, we should never have
such partitioning on the edges, BB10 should be COLD as we
Hi,
since we are now using C++, I think we can remove the attributes and
just use unnamed parameters. For now I kept the names in comments for
documentation purposes, but would be glad to remove those too, if you like.
Booted and tested x86_64-linux.
Thanks,
Paolo.
PS: slightly interesting,
Hi!
As discussed on IRC, the picochip-* port doesn't have an active maintainer
anymore, this patch adds it to deprecated ports for 4.7.2+ so that it can be
removed in
GCC 4.8 unless somebody steps up to maintain it.
Ok for trunk/4.7?
2012-09-11 Jakub Jelinek
* config.gcc: Obsolete p
On Tue, 11 Sep 2012, Richard Guenther wrote:
On Tue, Sep 11, 2012 at 1:07 PM, Marc Glisse wrote:
Hello,
here is a patch that turns {v[1],v[0]} into vec_perm_expr(v,v,{1,0}) if the
target is ok with it.
I am attaching 2 versions of the patch. p-good is the one that passes
testing. p-bad, wher
On 09/11/2012 08:02 AM, Jakub Jelinek wrote:
> 2012-09-11 Jakub Jelinek
>
> PR libstdc++/54172
> * libsupc++/guard.cc (__cxa_guard_acquire): Fix up the last
> argument of the first __atomic_compare_exchange_n.
Looks good.
r~
Ping?
http://gcc.gnu.org/ml/gcc-patches/2012-09/msg00068.html
Thanks
Christophe.
On 3 September 2012 11:01, Christophe Lyon wrote:
> On 31 August 2012 18:14, Janis Johnson wrote:
>>
>> do something like
>>
>> /* { dg-final { scan-assembler-times "fmrrd\[\\t \]+r0,\[\\t \]*r1,\[\\t
>> \]*d0" 2
On Thu, Sep 06, 2012 at 11:10:37PM +0200, Jakub Jelinek wrote:
> > + int expected(0);
> > if (__atomic_compare_exchange_n(gi, &expected, pending_bit, false,
> > __ATOMIC_ACQ_REL,
> > __ATOMIC_RELAXED))
>
On 11 September 2012 12:52, Richard Earnshaw wrote:
> Try something like:
>
> short foo(int);
>
> short swaps (short x, int y)
> {
> int z = x;
> if (y)
> z = __builtin_bswap16(x);
> return foo (z);
> }
>
> If that's not enough, try adding 1 to z before calling foo.
>
Thanks, it wor
On 09/10/2012 04:26 PM, Steven Bosscher wrote:
> + rtx index = force_reg (index_mode, dispatch_index);
You can't modify the result of force_reg. Use copy_to_{mode_,}reg instead.
> + rtx tmp = expand_simple_binop (index_mode, MINUS,
> + index, CONST
The attached patch fixes the bootstrap on darwin to cope with the
VEC changes to remove unnecessary VEC function overloads. Tested on
x86_64-apple-darwin12. Okay for gcc trunk.
Jack
2012-09-11 Dominique d'Humieres
Jack Howarth
* config/darwin.c (darwin_asm_na
Hi,
the operands cache is ugly. This patch removes it at least for the def
operands, saving three pointers for roughly each normal statement (the
pointer in gsbase, and two pointers from def_optype_d). This is
relatively easy to do, because all statements except ASMs have at most one
def (an
On Tue, Sep 11, 2012 at 04:41:24PM +0200, Steven Bosscher wrote:
> > + if (args_to_skip)
> > +for (parm = DECL_ARGUMENTS (current_function_decl), num = 0;
> > +parm; parm = DECL_CHAIN (parm), num++)
> > + if (bitmap_bit_p (args_to_skip, num)
> > + && is_gimple_reg (parm))
> + if (args_to_skip)
> +for (parm = DECL_ARGUMENTS (current_function_decl), num = 0;
> +parm; parm = DECL_CHAIN (parm), num++)
> + if (bitmap_bit_p (args_to_skip, num)
> + && is_gimple_reg (parm))
> + {
> + tree ddecl;
> + gimple def_temp;
> +
> +
On Tue, 11 Sep 2012, Richard Guenther wrote:
On Tue, Sep 11, 2012 at 1:07 PM, Marc Glisse wrote:
Hello,
here is a patch that turns {v[1],v[0]} into vec_perm_expr(v,v,{1,0}) if the
target is ok with it.
I am attaching 2 versions of the patch. p-good is the one that passes
testing. p-bad, wher
On Tue, Sep 11, 2012 at 1:07 PM, Marc Glisse wrote:
> Hello,
>
> here is a patch that turns {v[1],v[0]} into vec_perm_expr(v,v,{1,0}) if the
> target is ok with it.
>
> I am attaching 2 versions of the patch. p-good is the one that passes
> testing. p-bad, where I rely on fold_stmt to detect ident
On Tue, Sep 11, 2012 at 3:30 PM, Michael Matz wrote:
> Hi,
>
> On Tue, 11 Sep 2012, Richard Guenther wrote:
>
>> >>> +++ gcc/lto/lto.c (working copy)
>> >>> @@ -1559,8 +1559,6 @@ lto_fixup_prevailing_decls (tree t)
>> >>> {
>> >>>enum tree_code code = TREE_CODE (t);
>> >>>LTO_NO_PRE
Hi!
As discussed in the PR, right now we do a very bad job for debug info
of partially inlined functions (both when they are kept only partially
inlined, or when partial inlining is performed, but doesn't seem to be
useful and foo.part.N is inlined back, either into the original function, or
into
This patch adds predefines for AArch64 code models. These code models are
added as an effective target for the AArch64 platform.
Tests for these predefines have been added to `gcc.target/aarch64/'.
Thanks,
Chris
ChangeLog:
[AArch64] Added predefines for AArch64 code models.
gc
On 09/11/12 14:17, Tobias Burnus wrote:
On 09/11/2012 03:08 PM, Tobias Burnus wrote:
your patch broke bootstrapping here:
/home/tob/projects/gcc-git/gcc/gcc/doc//sourcebuild.texi:1537: Node
`arm_neon_ok' previously defined at line 1532.
I fixed it (Rev. 191181) with the attached patch. arm_neo
This removes the unused gtc_mode param and moves lifetime management
of the various tables to a central place, avoiding repeated checks.
LTO bootstrapped on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2012-09-11 Richard Guenther
* lto.c (enum gtc_mode): Remove.
(str
Hi,
On Tue, 11 Sep 2012, Richard Guenther wrote:
> >>> +++ gcc/lto/lto.c (working copy)
> >>> @@ -1559,8 +1559,6 @@ lto_fixup_prevailing_decls (tree t)
> >>> {
> >>>enum tree_code code = TREE_CODE (t);
> >>>LTO_NO_PREVAIL (TREE_TYPE (t));
> >>> - if (CODE_CONTAINS_STRUCT (code, TS
On 09/11/2012 03:08 PM, Tobias Burnus wrote:
your patch broke bootstrapping here:
/home/tob/projects/gcc-git/gcc/gcc/doc//sourcebuild.texi:1537: Node
`arm_neon_ok' previously defined at line 1532.
I fixed it (Rev. 191181) with the attached patch. arm_neon_ok should
have been arm_neon2_ok. (I
Fixed the format of the test options, as per Jakub's comment.
Add -fno-short-enums flag to test c-c++-common/pr51712.c as discussed in
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51712.
This removes the excess warning that caused the test to fail.
Tested in arm-none-eabi configuration. The test no
> your patch broke bootstrapping here:
>
> /home/tob/projects/gcc-git/gcc/gcc/doc//sourcebuild.texi:1537: Node
> `arm_neon_ok' previously defined at line 1532.
>
> (Sorry for only complaining about those issues today.)
No need to feel sorry about that. It is Really Bad that people
apparently don't
Hi,
your patch broke bootstrapping here:
/home/tob/projects/gcc-git/gcc/gcc/doc//sourcebuild.texi:1537: Node
`arm_neon_ok' previously defined at line 1532.
(Sorry for only complaining about those issues today.)
Tobias
On 09/11/2012 02:54 PM, Ramana Radhakrishnan wrote:
Hi,
This allows the
Hi,
This allows the auto-vectorizer to use vfma under Ofast or ffast-math.
I have a follow-up patch which will add support for these from
arm_neon.h as well before someone asks. It's being regression tested as
we speak and that'll follow shortly.
Tested on A15 silicon native with no regressio
On Tue, Sep 11, 2012 at 01:46:37PM +0100, Kyrylo Tkachov wrote:
> Add -fno-short-enums flag to test c-c++-common/pr51712.c as discussed in
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51712.
> This removes the excess warning that caused the test to fail.
> Tested in arm-none-eabi configuration. Th
Add -fno-short-enums flag to test c-c++-common/pr51712.c as discussed in
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51712.
This removes the excess warning that caused the test to fail.
Tested in arm-none-eabi configuration. The test now passes.
Comment? Ok for trunk?
Thanks,
Kyrill
gcc/testsuite
> Does this restriction look right to you ? (regression tests are still
> running on x86 and sh)
Please generate your patches with diff -up (or svn diff -x -up).
> + && (BB_PARTITION (e->src) == BB_PARTITION (e->dest))
No need for parentheses around this check.
The shrink wrapping c
On Tue, Sep 11, 2012 at 5:03 AM, Diego Novillo wrote:
> On 2012-09-11 01:01 , Ian Lance Taylor wrote:
>>
>> On Mon, Sep 10, 2012 at 4:52 PM, Diego Novillo
>> wrote:
>>>
>>>
>>> Ian, could you commit the changes in go/gofrontend?
>>
>>
>> Done. Actually, it looks like you already committed them,
> This is ok, of course.
Then could you please commit it (I don't have write access)?
TIA
Dominique
Hello,
While testing the patch to enable shrink-wrapping on SH [PR54546], we
hit an the "error: EDGE_CROSSING missing across section boundary"
Indeed, shrink-wrap duplicates a bb with successors (containing the
return sequence) into an unlikely section. I first thought about setting
the EDGE_CROS
ok with
&& gimple_assign_lhs_ptr (loc->stmt) == loc->ref
instead. Let's hope we conservatively catch all writes to ref this way (which
is what we need, right)?
Yes.
Thanks. Committing the attached patch.
PR middle-end/54149
* tree-ssa-loop-im.c (execute_sm_if_changed_
On 2012-09-11 01:01 , Ian Lance Taylor wrote:
On Mon, Sep 10, 2012 at 4:52 PM, Diego Novillo wrote:
Ian, could you commit the changes in go/gofrontend?
Done. Actually, it looks like you already committed them, but I
brought the master repo up to date.
Yes, sorry. I'm not quite sure how t
On 09/11/2012 01:41 PM, Diego Novillo wrote:
On 2012-09-11 03:58 , Tobias Burnus wrote:
Did you test with or without Graphite?
I tested with and without release checking, all languages and all
targets that use VEC. So many combinations...
There is unfortunately always an N+1 configuration
On 2012-09-11 06:12 , Dominique Dhumieres wrote:
Fixed with the attached.
Followed by the same failure on darwin. Fixed with
--- ../_clean/gcc/config/darwin.c 2012-07-09 22:06:21.0 +0200
+++ ../p_work/gcc/config/darwin.c 2012-09-11 11:53:02.0 +0200
@@ -1878,7 +1878,
On Tue, Sep 11, 2012 at 1:41 PM, Diego Novillo wrote:
> On 2012-09-11 03:58 , Tobias Burnus wrote:
>
>> Did you test with or without Graphite?
>
>
> I tested with and without release checking, all languages and all targets
> that use VEC. So many combinations... how is graphite enabled?
By havin
On 2012-09-11 05:35 , Richard Guenther wrote:
On Tue, Sep 11, 2012 at 9:58 AM, Tobias Burnus wrote:
On 09/11/2012 01:52 AM, Diego Novillo wrote:
Remove unnecessary VEC function overloads.
Several VEC member functions that accept an element 'T' used to have
two overloads: one taking 'T', the
On 2012-09-11 03:58 , Tobias Burnus wrote:
Did you test with or without Graphite?
I tested with and without release checking, all languages and all
targets that use VEC. So many combinations... how is graphite enabled?
Diego.
Hello,
any advice?
http://gcc.gnu.org/ml/gcc-patches/2012-09/msg00044.html
On Sun, 2 Sep 2012, Marc Glisse wrote:
Hello,
this patch passes bootstrap+testsuite. It is probably wrong in many ways, but
I don't know enough to do more without some advice.
The goal is to recognize that v[0]+v[1
Any comment?
http://gcc.gnu.org/ml/gcc-patches/2012-08/msg02098.html
Maybe separately on the technical and political aspects?
On Sat, 1 Sep 2012, Marc Glisse wrote:
With the patch...
On Sat, 1 Sep 2012, Marc Glisse wrote:
Hello,
this patch copies some more vector extensions from the C fro
On Tue, 11 Sep 2012, Segher Boessenkool wrote:
> > > 2012-09-10 Maciej W. Rozycki
> > >
> > > gcc/
> > > * config/rs6000/rs6000.c (print_operand) <'c'>: Remove.
> > > * config/rs6000/spe.md: Remove a leftover comment.
> >
> > Okay.
>
> This patch wasn't sent to gcc-pa
Hello,
here is a patch that turns {v[1],v[0]} into vec_perm_expr(v,v,{1,0}) if
the target is ok with it.
I am attaching 2 versions of the patch. p-good is the one that passes
testing. p-bad, where I rely on fold_stmt to detect identity permutations,
ICEs towards the end of the pass while che
On Tue, 11 Sep 2012, Jakub Jelinek wrote:
On Tue, Sep 11, 2012 at 12:29:10PM +0200, Marc Glisse wrote:
Note that, apart from preventing external calls, it removes this use case:
__attribute__((vector(4))) double mysqrt(double x){return sqrt(x);}
__m256d var;
mysqrt(var);
I don't think those
On 11/09/12 11:25, Christophe Lyon wrote:
> On 10 September 2012 19:30, Richard Earnshaw wrote:
>> On 10/09/12 16:40, Christophe Lyon wrote:
>>> Why do we have to keep room for the predicate here? (%?) Doesn't this
>>> pattern match only in unconditional cases?
>>>
>>
>> Because the ARM back-end h
On Tue, Sep 11, 2012 at 12:29:10PM +0200, Marc Glisse wrote:
> >Btw, this then happily fits into my suggestion that the "elementalness"
> >can be autodetected by the compiler simply by means of a proper IPA
> >pass and thus be fully LTO / whole-program aware. No need for an
> >attribute (where you
The backport of the patch for PR53572 caused us to remove unused
decls at -O0, a regresion on the branch - fixed by the following.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2012-09-11 Richard Guenther
PR debug/54534
* cgraph.h (varpool_can_remove
On Tue, 11 Sep 2012, Richard Guenther wrote:
On Tue, Sep 11, 2012 at 10:41 AM, Richard Guenther
wrote:
On Mon, Sep 10, 2012 at 6:37 PM, Richard Henderson wrote:
Whether or not the compiler creates a clone COULD BE totally up to the
compiler, based on whether or not vectorization is enabled,
On 10 September 2012 19:30, Richard Earnshaw wrote:
> On 10/09/12 16:40, Christophe Lyon wrote:
>> Why do we have to keep room for the predicate here? (%?) Doesn't this
>> pattern match only in unconditional cases?
>>
>
> Because the ARM back-end has a very late conditionalizer pass that can
> als
> Fixed with the attached.
Followed by the same failure on darwin. Fixed with
--- ../_clean/gcc/config/darwin.c 2012-07-09 22:06:21.0 +0200
+++ ../p_work/gcc/config/darwin.c 2012-09-11 11:53:02.0 +0200
@@ -1878,7 +1878,7 @@ darwin_asm_named_section (const char *na
2012-09-10 Maciej W. Rozycki
gcc/
* config/rs6000/rs6000.c (print_operand) <'c'>: Remove.
* config/rs6000/spe.md: Remove a leftover comment.
Okay.
This patch wasn't sent to gcc-patches -- can we see it please?
Segher
On Tue, Sep 11, 2012 at 4:06 AM, Jakub Jelinek wrote:
> On Tue, Sep 11, 2012 at 03:57:44AM -0500, Gabriel Dos Reis wrote:
>> On Tue, Sep 11, 2012 at 3:42 AM, Richard Guenther
>> wrote:
>>
>> > Btw, this then happily fits into my suggestion that the "elementalness"
>> > can be autodetected by the
On Tue, Sep 11, 2012 at 11:06 AM, Jakub Jelinek wrote:
> On Tue, Sep 11, 2012 at 03:57:44AM -0500, Gabriel Dos Reis wrote:
>> On Tue, Sep 11, 2012 at 3:42 AM, Richard Guenther
>> wrote:
>>
>> > Btw, this then happily fits into my suggestion that the "elementalness"
>> > can be autodetected by the
On Tue, Sep 11, 2012 at 9:58 AM, Tobias Burnus wrote:
> On 09/11/2012 01:52 AM, Diego Novillo wrote:
>>
>> Remove unnecessary VEC function overloads.
>>
>> Several VEC member functions that accept an element 'T' used to have
>> two overloads: one taking 'T', the second taking 'T *'.
>
>
> They mig
1 - 100 of 115 matches
Mail list logo
