On 8 April 2013 22:08:54 Jan Hubicka wrote:
Hi,
this patch makes change_decl_assembler_name to do the right thing with inline
clones. My original plan was to remove inline clones from assembler_name_hash,
but it hits the problem that we currently need to make them unique for purposes
of LTO sre
On Tue, Apr 9, 2013 at 9:42 AM, Igor Zamyatin wrote:
> Richard,
>
> Any plans to commit your new fix?
I did already.
Richard.
>
> Thanks,
> Igor
>
>> On Thu, Apr 4, 2013 at 3:33 PM, Yuri Rumyantsev
>> wrote:
>> > Hi Richard,
>> >
>> > You slightly change behavior of find_base_term, namely befo
This patch makes type_hash_lookup and type_hash_add in tree.c static. These
functions are used only within tree.c, and I don't think there is any reason to
export low-level interface to type hashing anyway.
Will commit in 2 days unless someone complains. Bootstrapped on
x86_64-linux-gnu.
Tha
In propagate_tree_value and replace_exp_1 operating on SSA form
are the only users of unsave_expr_now where really unsharing
matters. Thus, use unshare_expr and remove unsave_expr_now.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2013-04-09 Richard Biener
On Mon, 8 Apr 2013, Jakub Jelinek wrote:
> Hi!
>
> lhs ={v} {CLOBBER};
> stmts right now allow only VAR_DECL or MEM_REF lhs, but the forwprop code
> below on the attached testcase attempts to propagate an ARRAY_REF (of
> MEM_REF) into it. Fixed by not propagating in that case, allowing arbitrary
On Mon, 8 Apr 2013, Jan Hubicka wrote:
> Hi,
> currently lto-symtab is trying to resolve all duplicated declarations,
> including static variables where such duplicates should not happen.
> This conflicts with the plan to solve PR54095 by postponning renaming to
> the partitioning. This patch add
> The attached patch has passed bootstrap/regression test
> on x86_64-unknown-linux-gnu with current main trunk.
> A plaintext gcc/ChangeLog is as below:
>
>
> 2013-04-06 Chung-Ju Wu
>
> * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter
> out unnecessary bits in the co
On Mon, Apr 8, 2013 at 5:10 PM, Joern Rennecke
wrote:
> This is basically the same patch as attached to the PR, except that I
> have changed the goto-loop into a do-while loop with a new comment;
> this caused the need for a lot of reformatting.
Can you please include a testcase that shows the ef
On Mon, Apr 8, 2013 at 9:13 PM, Laurent Alfonsi wrote:
> Hello,
>
> I have identified a big performance regression between 4.6 and 4.7. (I have
> enclosed a pathological test).
>
> After investigation, it is because of the += statement applied on 2 signed
> chars.
> - It is now type-promoted
On Mon, Apr 8, 2013 at 10:47 PM, Marc Glisse wrote:
> On Sun, 7 Apr 2013, Marc Glisse wrote:
>
>> extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__,
>> __artificial__))
>> _mm_slli_epi16 (__m128i __A, int __B)
>> {
>> - return (__m128i)__builtin_ia32_psllwi128 ((__v8hi)_
> * config/sparc/sparc.md: Use define_c_enum for "unspec" and
> "unspecv".
OK, thanks.
--
Eric Botcazou
On Mon, Apr 8, 2013 at 10:39 PM, Lawrence Crowl wrote:
> On 4/8/13, Kenneth Zadeck wrote:
>> The other problem, which i invite you to use the full power of
>> your c++ sorcery on, is the one where defining an operator so
>> that wide-int + unsigned hwi is either rejected or properly
>> zero exten
On Tue, Apr 9, 2013 at 3:54 AM, Jeff Law wrote:
>
> This incorporates the concrete suggestions from Steven & Richi -- it doesn't
> do any refactoring of the VRP code. There's still stuff I'm looking at that
> might directly lead to some refactoring. In the mean time I'm submitting
> the obvious
On Tue, Apr 9, 2013 at 5:40 AM, Zhouyi Zhou wrote:
> On Mon, Apr 8, 2013 at 5:48 PM, Richard Biener
> wrote:
>>Can you trigger this message to show up with -Winline before/after the patch?
>>Can you please add a testcase then?
> Thanks Richard for reviewing, from my point of view about gcc and m
> Bootstrapped and tested on sparc64-unknown-linux-gnu.
> OK for trunk?
OK modulo a couple of nits:
@@ -804,6 +801,136 @@ char sparc_hard_reg_printed[8];
struct gcc_target targetm = TARGET_INITIALIZER;
+/* We use the machine specific reorg pass to enable workarounds for errata.
+ We need t
On Tue, Apr 9, 2013 at 10:20 AM, Maxim Kuvyrkov wrote:
> This patch makes type_hash_lookup and type_hash_add in tree.c static. These
> functions are used only within tree.c, and I don't think there is any reason
> to export low-level interface to type hashing anyway.
>
> Will commit in 2 days u
On Tue, 9 Apr 2013, Richard Biener wrote:
On Mon, Apr 8, 2013 at 10:47 PM, Marc Glisse wrote:
On Sun, 7 Apr 2013, Marc Glisse wrote:
extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__,
__artificial__))
_mm_slli_epi16 (__m128i __A, int __B)
{
- return (__m128i)__buil
On Mon, Apr 8, 2013 at 11:45 PM, Lawrence Crowl wrote:
> Ping?
You didn't commit the ones I already approved? I don't want to go over
them again ...
Richard.
> On 3/31/13, Lawrence Crowl wrote:
>> On 3/28/13, Richard Biener wrote:
>>> On Mar 27, 2013 Lawrence Crowl wrote:
>>> > On 3/27/13,
On Tue, Apr 09, 2013 at 11:08:38AM +0200, Marc Glisse wrote:
> The *intrin.h files already use __extension__ to create vectors, like:
> return __extension__ (__m128d){ __F, 0.0 };
> but even when I remove it it does not warn with -std=c89 -pedantic.
Even with -Wsystem-headers ?
Jakub
On Tue, 9 Apr 2013, Jakub Jelinek wrote:
On Tue, Apr 09, 2013 at 11:08:38AM +0200, Marc Glisse wrote:
The *intrin.h files already use __extension__ to create vectors, like:
return __extension__ (__m128d){ __F, 0.0 };
but even when I remove it it does not warn with -std=c89 -pedantic.
Even w
On 04/09/2013 01:47 AM, Robert Dewar wrote:
Well the back end has all the information to figure this out I think!
But anyway, for Ada, the current situation is just fine, and has
the advantage that the -gnatG expanded code listing clearly shows in
Ada source form, what is going on.
Isn't this a
> Alternatively, we can bump the minimum of that param, as usual ;)
Let's do that and bump it to 1, my understanding is that 0 and 1 are
equivalent for this param.
--
Eric Botcazou
On Mon, Apr 8, 2013 at 5:38 PM, Tobias Burnus wrote:
> Dear all,
>
> attached is an updated version of the patch, which address the raised issues
> and some minor problems and omissions I found.
>
> OK for the trunk?
+For Boolean (@code{LOGICAL}) arguments, please note that GCC expects
+only the
Janne Blomqvist wrote:
+For Boolean (@code{LOGICAL}) arguments, please note that GCC expects
+only the integer value 0 and 1. If a GNU Fortran @code{LOGICAL}
+variable contains another integer value, the result is undefined.
+As some other Fortran compilers use @math{-1} for @code{.TRUE.},
+extr
2013/4/9 Dave Korn :
> On 22/03/2013 08:44, Kai Tietz wrote:
>
>> 2013-03-22 Kai Tietz
>>
>> * config/i386/cygming-crtbegin.c (__register_frame_info): Make weak.
>> (__deregister_frame_info): Likewise.
>
> Hi Kai,
>
> I read your explanation of the problem relating to x86-64 me
from my local tree.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2013-04-09 Richard Biener
java/
* expr.c (build_java_binop): Pass a type to build_int_cst.
* tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
work that is n
Hello,
sched-vis.c would print insns in a SEQUENCE as a pattern, in effect
printing only GET_CODE(insn). Fixed with this patch.
Bootstrapped&tested on powerpc64-unknown-linux-gnu and
sparc64-unknown-linux-gnu with some dump inspections to make sure
everything looks as expected now.
OK for trunk?
> Bootstrapped&tested on powerpc64-unknown-linux-gnu and
> sparc64-unknown-linux-gnu with some dump inspections to make sure
> everything looks as expected now.
> OK for trunk?
Yes, thanks.
--
Eric Botcazou
On Tue, Apr 09, 2013 at 11:45:04AM +0200, Eric Botcazou wrote:
> > Alternatively, we can bump the minimum of that param, as usual ;)
>
> Let's do that and bump it to 1, my understanding is that 0 and 1 are
> equivalent for this param.
Alright. So ok to apply this one (trunk/4.8)?
2013-04-09 M
Quoting Richard Biener :
On Mon, Apr 8, 2013 at 5:10 PM, Joern Rennecke
wrote:
This is basically the same patch as attached to the PR, except that I
have changed the goto-loop into a do-while loop with a new comment;
this caused the need for a lot of reformatting.
Can you please include a te
This gets rid of slp_void_p and moves vect_get_place_in_interleaving_chain
next to its only user.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2013-04-09 Richard Biener
* tree-vectorizer.h (slp_void_p): Remove.
(slp_tree): Typedef before _slp_tree d
On 4/9/2013 5:39 AM, Florian Weimer wrote:
On 04/09/2013 01:47 AM, Robert Dewar wrote:
Well the back end has all the information to figure this out I think!
But anyway, for Ada, the current situation is just fine, and has
the advantage that the -gnatG expanded code listing clearly shows in
Ada s
On 04/09/2013 05:13 AM, Marek Polacek wrote:
On Tue, Apr 09, 2013 at 11:45:04AM +0200, Eric Botcazou wrote:
Alternatively, we can bump the minimum of that param, as usual ;)
Let's do that and bump it to 1, my understanding is that 0 and 1 are
equivalent for this param.
Alright. So ok to app
On 05/04/13 16:35, Kyrylo Tkachov wrote:
Hi all,
This patch is a backport of the fix for PR 56720 where we would ICE on
arm-*-* when trying to
expand vcond with a floating point unorderd comparison cases. The patch is
almost identical
to the trunk patch at:
http://gcc.gnu.org/ml/gcc-patches/2013
On Mon, Apr 08, 2013 at 03:23:01PM +0200, Jakub Jelinek wrote:
> Anyway, I've kept the default as never for now.
Here is an updated patch with %r/%R, fully bootstrapped/regtested now
on x86_64-linux and i686-linux, tested on a couple of small testcases
with -fdiagnostics-color{,=never,=auto,=alway
> -Original Message-
> From: Richard Earnshaw
> Sent: 09 April 2013 14:12
> To: Kyrylo Tkachov
> Cc: gcc-patches@gcc.gnu.org; Ramana Radhakrishnan; 'Jakub Jelinek'
> Subject: Re: [PATCH][Backport 4.7][ARM] Fix PR 56720
>
> On 05/04/13 16:35, Kyrylo Tkachov wrote:
> > Hi all,
> >
> > This p
Hi!
The omp for loop iterator can be addressable e.g. because of shared use of
it in a 2x nested parallel (in which case we can't use copy-in/out).
The following patch tweaks expand_omp_for* to allow the iterators to be
addressable.
Bootstrapped/regtested on x86_64-linux and i686-linux, committe
On Tue, Apr 9, 2013 at 1:25 PM, Joern Rennecke
wrote:
> Quoting Richard Biener :
>
>> On Mon, Apr 8, 2013 at 5:10 PM, Joern Rennecke
>> wrote:
>>>
>>> This is basically the same patch as attached to the PR, except that I
>>> have changed the goto-loop into a do-while loop with a new comment;
>>>
On 04/09/2013 02:41 PM, Robert Dewar wrote:
On 4/9/2013 5:39 AM, Florian Weimer wrote:
On 04/09/2013 01:47 AM, Robert Dewar wrote:
Well the back end has all the information to figure this out I think!
But anyway, for Ada, the current situation is just fine, and has
the advantage that the -gnatG
This splits discovering of a SLP opportunity (vect_build_slp_tree)
and calculating the cost of a SLP instance (now vect_analyze_slp_cost).
The immediate need for me is to support handling of mismatches
in the SLP tree due to commutative operand order mismatch, one more
cleanup patch for this reaso
Hello,
This patch fixes label alignments after a ADDR_DIFF_VEC with byte
offsets. The bug occurs with building the libgcc for a sh-elf target.
See a reduced case here. The funny thing is that the assembly error
given in Subject appears only on a debug section when compiled with -g,
thus the emitt
Quoting Richard Biener :
Oops, I missed your final comment when I wrote my first reply.
I fail to see why you need two passes for this rather than considering the
case that the immediate use stmt of the multiplication we start from
combines another multiplication with a MINUS_EXPR. That is ...
Looks good to me if nobody else has comments.
Jason
On Tue, Apr 9, 2013 at 4:53 PM, Joern Rennecke
wrote:
> Quoting Richard Biener :
>
> Oops, I missed your final comment when I wrote my first reply.
>
>
>> I fail to see why you need two passes for this rather than considering the
>> case that the immediate use stmt of the multiplication we start f
Hi Richard,
I do not have write access to GCC SVN repository, can you commit it for me?
Thanks alot
Cheers
On Tue, Apr 9, 2013 at 5:04 PM, Richard Biener
wrote:
>
> On Tue, Apr 9, 2013 at 5:40 AM, Zhouyi Zhou wrote:
> > On Mon, Apr 8, 2013 at 5:48 PM, Richard Biener
> > wrote:
> >>Can you t
I understand, Thanks for your answer.
Looking at the standard, I was thinking the example you pointed was
undefined.
I have created a bugzilla
(http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56894) to track this big
performance regression.
My investigations pointed on the scev_const_prop optimiz
On 02/04/13 17:06, Kyrylo Tkachov wrote:
From: Ramana Radhakrishnan [mailto:ramana@googlemail.com]
Sent: 02 April 2013 11:10
To: Kyrylo Tkachov
Cc: gcc-patches@gcc.gnu.org; Richard Earnshaw; Ramana Radhakrishnan
Subject: Re: [PATCH][ARM] minmax_arithsi for non-canonical operand
order with MIN
On 18/02/13 18:44, Greta Yorsh wrote:
Convert define_insn into define_insn_and_split for various min and max
patterns that output multiple assembly instructions. Use movsicc to emit
RTL. A separate patch will split movsicc.
gcc/
2013-02-14 Greta Yorsh
* config/arm/arm.md (arm_smax_
On 18/02/13 18:38, Greta Yorsh wrote:
Convert define_insn into define_insn_and_split for various patterns that
output multiple assembly instructions.
It appears that preparation statements in define_insn_and_split sometimes
are called with which_alternative set to -1 even after reload. Therefore
> Ah, OK, yea, it makes perfect sense now. Approved.
Thanks.
> If you wanted to get ambitious, rewriting reorg.c to compute and use
> dependency links to drive its candidate selection instead of the insane
> tracking code it uses would be a huge step forward, both in terms of
> cleanliness. It'
Quoting Richard Biener :
I don't see that. It's merely a complication of optimal handling of
a * b +- c * d vs. just a * b +- c. The pass does simple pattern matching
only, not doing a global optimal transform, so adding another special-case
is reasonable. Special-casing just for single-use 2
On Tue, Apr 9, 2013 at 6:15 PM, Eric Botcazou wrote:
>> If you wanted to get ambitious, rewriting reorg.c to compute and use
>> dependency links to drive its candidate selection instead of the insane
>> tracking code it uses would be a huge step forward, both in terms of
>> cleanliness. It'd also
Ok for google branches.
thanks,
David
On Mon, Apr 8, 2013 at 3:44 PM, Rong Xu wrote:
> Revised copyright info.
>
> -Rong
>
> 2013-04-08 Rong Xu
>
> * contrib/profile_merge.py: An offline profile merge tool.
>
> Index: contrib/profile_merge.py
> ===
Premature ping?
Also bootstrapped&tested on ia64-unknown-linux-gnu now.
On Sun, Apr 7, 2013 at 12:40 AM, Steven Bosscher wrote:
> Hello,
>
> In this PR43631, var-tracking notes are inserted on basic block
> boundaries and BB_HEAD/BB_END end up pointing to these notes. This
> breaks verify_flow_inf
On 09/04/2013 11:37, Kai Tietz wrote:
> Hmm, well in standard-case you are right. But well there is still a
> chance that GetProcAddress returns NULL-pointer ...
How would that actually happen? Removing any of those functions from libgcc
or libjava would be a very serious ABI breakage; we'r
Have not had a chance to look at the whole patch yet, but here are some
initial comments.
https://codereview.appspot.com/8508048/diff/4001/contrib/profile_merge.py
File contrib/profile_merge.py (right):
https://codereview.appspot.com/8508048/diff/4001/contrib/profile_merge.py#newcode53
contrib/
> "Jonathan" == Jonathan Wakely writes:
Marc> I thought you were going to suggest enhancing the configure test
Marc> so it fails on old systemtap (detects it as absent).
Jonathan> Ah yes, that's a much better idea!
Here's a patch to do that.
I tested it on x86-64 Fedora 18. I tested the "
On 04/09/2013 10:15 AM, Eric Botcazou wrote:
Yes, I agree that it's quite convoluted but, as Steven already said, rewriting
it to use a more modern framework is hard because of SEQUENCEs and, frankly, I
have neither the real incentive nor the time to start such an endeavor.
I understand complet
This patch is pending on trunk, but I would like to get this into google
branches now as it is causing spurious warnings.
Google ref b/8496800.
This patch allows the unused attribute to be used without warning
on C++ class members, which are of type FIELD_DECL. This is for
compatibility with clan
The standard is somewhat unclear here, but I think what makes sense is
to always check whether the address is null rather than confine the
check to when the immediate operand of typeid is an INDIRECT_REF.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 43d5a64843e3e18a74e59b1b368cd129d5f1
Hi,
the attached patch reduces the size of module files on disk by
compressing them with zlib and storing them in the gzip format (RFC
1952). I chose zlib because it's a) ubiquitous and b) there's already
a copy of zlib in the GCC source tree, so this doesn't introduce any
further build
dependenci
On 9 April 2013 18:47, Tom Tromey wrote:
>> "Jonathan" == Jonathan Wakely writes:
>
> Marc> I thought you were going to suggest enhancing the configure test
> Marc> so it fails on old systemtap (detects it as absent).
>
> Jonathan> Ah yes, that's a much better idea!
>
> Here's a patch to do th
On Tue, 9 Apr 2013, Marc Glisse wrote:
On Tue, 9 Apr 2013, Richard Biener wrote:
I seem to remember discussion in the PR(s) that the intrinsics should
(and do for other compilers) expand to the desired instructions even when
the corresponding instruction set is disabled.
emmintrin.h starts w
Ok -- there are legitimate use cases for this.
David
On Tue, Apr 9, 2013 at 11:01 AM, Teresa Johnson wrote:
> This patch is pending on trunk, but I would like to get this into google
> branches now as it is causing spurious warnings.
>
> Google ref b/8496800.
>
> This patch allows the unused att
Hi,
Patch set 3 for offline profile merge tool.
(1) Rename to profile_tool for future function expansion.
(2) Install to bin/ directory for easy user access.
Thanks,
-Rong
2013-04-09 Rong Xu
* gcc/Makefile.in: Install profile_tool to bin directory.
* contrib/profile_tool:
https://codereview.appspot.com/8508048/diff/4001/contrib/profile_merge.py
File contrib/profile_merge.py (right):
https://codereview.appspot.com/8508048/diff/4001/contrib/profile_merge.py#newcode53
contrib/profile_merge.py:53: data_file = open(path, 'rb')
On 2013/04/09 17:32:11, martint wrote:
H
Dear all,
attached is a first attempt to implement DW_TAG_namelist support in GCC.
(That's a DWARF2 features.)
It nicely works for:
module m
real :: cc
end module m
subroutine sub()
use m
implicit none
integer :: aa, bb
n
Hi,
this patch removes handling of alias pairs in ipa.c. This code is no longer
needed
now when we represent all aliases explicitly in the symbol table.
Bootstrapped/regtested x86_64-linux.
Honza
* ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
Drop
aliased
Christian Bruel wrote:
> This patch fixes label alignments after a ADDR_DIFF_VEC with byte
> offsets. The bug occurs with building the libgcc for a sh-elf target.
>
> See a reduced case here. The funny thing is that the assembly error
> given in Subject appears only on a debug section when compil
Hi!
As discussed in the PR, these 4 places call maybe_constant_value even
when processing_template_decl, just to try harder to see if the value isn't
constant, and that may crash if fold_non_dependent_expr hasn't been called
first. This is just optimization to get better warnings, so we don't wan
70 matches
Mail list logo
