On Mon, Oct 29, 2012 at 10:53:42PM -0700, H.J. Lu wrote:
> ---2012-10-30 H.J. Lu
>
> * gimple-pretty-print.c (dump_gimple_bb_header): Avoid alloca.
Ok, thanks.
Jakub
Hello,
one more optimization that needed help for vectors, it crashed on (xBecause of PR 55001, testcases are awkward to add (I could do a x86-only
one if needed).
bootstrap+testsuite.
2012-10-30 Marc Glisse
* fold-const.c (fold_binary_op_with_conditional_arg): Handle vectors.
Hi,
I would like to apply the patch below to trunk and gcc-4.7-branch.
This patch was originalyl submitted by Joel Sherrill back in May
(http://gcc.gnu.org/ml/gcc-patches/2012-05/msg01180.html),
but had never received any feedback.
It has been part of the rtems-gcc patches, since then.
Ralf
On 30 October 2012 05:20, Teresa Johnson wrote:
> Index: cfgrtl.c
> ===
> --- cfgrtl.c(revision 192692)
> +++ cfgrtl.c(working copy)
> @@ -912,7 +912,8 @@ rtl_can_merge_blocks (basic_block a, basic_block b
> partition bo
On Mon, Oct 29, 2012 at 02:07:55PM -0400, David Miller wrote:
> > I'd like to close the stage 1 phase of GCC 4.8 development
> > on Monday, November 5th. If you have still patches for new features you'd
> > like to see in GCC 4.8, please post them for review soon. Patches
> > posted before the fr
Hello,
This PR is due to the selective scheduling missing the dependencies with
implicit_sets. From the sched-deps.c code it looks like implicit sets
generate anti dependencies with either sets, uses or clobbers, so that's
that I've done with the below patch. Vlad, as it looks you've added
Adding myself to the list of members in "write after approval".
Index: ChangeLog
===
--- ChangeLog (revision 192977)
+++ ChangeLog (working copy)
@@ -1,3 +1,7 @@
+2012-10-30 Ganesh Gopalasubramanian
+
+ * MAINTAINERS (Writ
On 08/30/2012 11:45 AM, Jonathan Wakely wrote:
> On 29 August 2012 13:25, Michael Haubenwallner wrote:
>> On 08/28/2012 08:12 PM, Jonathan Wakely wrote:
>>> On 28 August 2012 18:27, Michael Haubenwallner wrote:
>
> Does it actually produce a segfault? I suppose it might on some
> platf
On Tue, Oct 30, 2012 at 08:05:13AM +0100, Marc Glisse wrote:
> Hello,
>
> one more optimization that needed help for vectors, it crashed on
> (x do a x86-only one if needed).
>
> bootstrap+testsuite.
>
> 2012-10-30 Marc Glisse
>
> * fold-const.c (fold_binary_op_with_conditional_arg): H
Hello,
This is the latest proposed patch from the PR.
Tested on rev 192482 with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
and no new failures.
Pre-approved by Kaz in the PR.
Committed as rev 192983.
Cheers,
Oleg
gcc/ChangeLo
On 30 October 2012 09:05, Michael Haubenwallner wrote:
> Any chance to get this in for 4.8?
I'm looking into it today.
On Mon, Oct 29, 2012 at 05:10:10PM +0100, Richard Biener wrote:
> On Mon, Oct 29, 2012 at 4:25 PM, Dehao Chen wrote:
> > On Mon, Oct 29, 2012 at 7:17 AM, Michael Matz wrote:
> >> Hi,
> >>
> >> On Mon, 29 Oct 2012, Richard Biener wrote:
> >>
> >>> > Well, you merely moved the bogus code to gimple-
On 10/29/2012 21:05, JonY wrote:
>> ChangeLog
>> 2012-10-29 Jonathan Yong
>>
>> * config/os/mingw32-w64/os_defines.h: Remove
>> _GLIBCXX_HAVE_BROKEN_VSWPRINTF
>> as no longer required.
>>
>>
>>
>> Index: libstdc++-v3/config/os/mingw32-w64/os_defines.h
>> ==
On 30 October 2012 09:28, Jonathan Wakely wrote:
> On 30 October 2012 09:05, Michael Haubenwallner wrote:
>> Any chance to get this in for 4.8?
>
> I'm looking into it today.
Consider the case where one object file containing
std::string().erase() is built with an older GCC without the fix for
PR
"H.J. Lu" writes:
>>> LRA has
>>>
>>> if (REG_P (reg) && (ep = get_elimination (reg)) != NULL)
>>> {
>>> rtx to_rtx = replace_p ? ep->to_rtx : ep->from_rtx;
>>>
>>> if (! replace_p)
>>> {
>>> offset += (ep->offset - ep->previous_offset);
On 30 October 2012 10:11, Jonathan Wakely wrote:
> On 30 October 2012 09:28, Jonathan Wakely wrote:
>> On 30 October 2012 09:05, Michael Haubenwallner wrote:
>>> Any chance to get this in for 4.8?
>>
>> I'm looking into it today.
>
> Consider the case where one object file containing
> std::string(
"H.J. Lu" writes:
> On Mon, Oct 29, 2012 at 5:11 PM, H.J. Lu wrote:
>> On Mon, Oct 29, 2012 at 4:41 PM, H.J. Lu wrote:
>>> On Mon, Oct 29, 2012 at 9:38 AM, Vladimir Makarov
>>> wrote:
On 12-10-29 12:21 PM, Richard Sandiford wrote:
>
> Vladimir Makarov writes:
>>
>> H.J
Hi,
for past week or two I was playing with ways to throttle down the complette
unrolling heuristics. I made complette unroller to use the tree-ssa-loop-niter
upper bound and unroll even in non-trivial cases and this has turned out to
increase number of complettely unrolled loops by great amount,
On Tue, 30 Oct 2012, Jan Hubicka wrote:
> Hi,
> for past week or two I was playing with ways to throttle down the complette
> unrolling heuristics. I made complette unroller to use the
> tree-ssa-loop-niter
> upper bound and unroll even in non-trivial cases and this has turned out to
> increase
On Tue, Oct 30, 2012 at 4:09 AM, Richard Sandiford
wrote:
> "H.J. Lu" writes:
>> On Mon, Oct 29, 2012 at 5:11 PM, H.J. Lu wrote:
>>> On Mon, Oct 29, 2012 at 4:41 PM, H.J. Lu wrote:
On Mon, Oct 29, 2012 at 9:38 AM, Vladimir Makarov
wrote:
> On 12-10-29 12:21 PM, Richard Sandiford
On Mon, 29 Oct 2012, Jan Hubicka wrote:
> > > ICK ...
> > >
> > > Why not sth as simple as
> > >
> > > return num_ssa_operands (stmt, SSA_OP_USE);
> > >
> > > ? a[1][2] and b[2] really have the same cost, variable length
> > > objects have extra SSA operands in ARRAY_REF/COMPONENT_REF for
On Tue, Oct 30, 2012 at 2:17 AM, Bin.Cheng wrote:
> On Tue, Oct 30, 2012 at 8:23 AM, Lawrence Crowl wrote:
>> On 10/29/12, Diego Novillo wrote:
>>> On Oct 29, 2012 Diego Novillo wrote:
>>> > Just to make sure. Testing on ppc should be fast, for example.
>>>
>>> And useless. Your patch does no
On Mon, Oct 29, 2012 at 4:25 PM, Steven Bosscher wrote:
> On Mon, Oct 29, 2012 at 7:54 PM, Diego Novillo wrote:
>
>> Sure. But the point is not to add more. We should mechanically strip
>> all the #if 0 code from the tree, btw. No point keeping all that
>> garbage around.
>
> Please no. A lot (
This fixes PR55111.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2012-10-30 Richard Biener
PR tree-optimization/55111
* tree-ssa-pre.c (eliminate_insert): Properly fold the built
stmt.
* gcc.dg/torture/pr55111.c: New testcase.
Inde
On Tue, 30 Oct 2012, Marek Polacek wrote:
On Tue, Oct 30, 2012 at 08:05:13AM +0100, Marc Glisse wrote:
Hello,
one more optimization that needed help for vectors, it crashed on
(x
* fold-const.c (fold_binary_op_with_conditional_arg): Handle vectors.
(fold_binary_loc): call it f
On 2012-10-29 15:01 , Lawrence Crowl wrote:
On 10/27/12, Marc Glisse wrote:
On Fri, 26 Oct 2012, Lawrence Crowl wrote:
2012-10-26 Lawrence Crowl
missing '>'
Fixed.
* is-a.h: New.
(is_a (U*)): New. Test for is-a relationship.
(as_a (U*)): New. Treat as a deri
On Tue, Oct 30, 2012 at 12:53 PM, Diego Novillo wrote:
> On Mon, Oct 29, 2012 at 4:25 PM, Steven Bosscher
> wrote:
>> On Mon, Oct 29, 2012 at 7:54 PM, Diego Novillo wrote:
>>
>>> Sure. But the point is not to add more. We should mechanically strip
>>> all the #if 0 code from the tree, btw. No
On Tue, Oct 30, 2012 at 1:14 PM, Marc Glisse wrote:
> On Tue, 30 Oct 2012, Marek Polacek wrote:
>
>> On Tue, Oct 30, 2012 at 08:05:13AM +0100, Marc Glisse wrote:
>>>
>>> Hello,
>>>
>>> one more optimization that needed help for vectors, it crashed on
>>> (x>> do a x86-only one if needed).
>>>
>>>
On Tue, Oct 30, 2012 at 1:20 PM, Diego Novillo wrote:
> On 2012-10-29 15:01 , Lawrence Crowl wrote:
>>
>> On 10/27/12, Marc Glisse wrote:
>>>
>>> On Fri, 26 Oct 2012, Lawrence Crowl wrote:
2012-10-26 Lawrence Crowl >>
>>>
>>> missing '>'
>>
>>
>> Fixed.
>>
* is-a.h: New.
As requested this adds predicates to check whether the lhs of
a assign or call is a store and whether rhs1 of an assignment
is a load. It uses this in place of the existing, slightly
bogus, check in the stmt estimate code.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Richard.
201
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Marcus Shawcroft
> Sent: 15 October 2012 12:37
> To: gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH] [AArch64] Add vcond, vcondu support.
>
> On 09/10/12 12:08, James Greenhalgh
> 2012-10-29 Lawrence Crowl
>
> * is-a.h: New.
> (is_a (U*)): New. Test for is-a relationship.
> (as_a (U*)): New. Treat as a derived type.
> (dyn_cast (U*)): New. Conditionally cast based on is_a.
> * cgraph.h (varpool_node): Rename to varpool_node_for_decl.
On Tue, Oct 30, 2012 at 4:38 AM, H.J. Lu wrote:
> On Tue, Oct 30, 2012 at 4:09 AM, Richard Sandiford
> wrote:
>> "H.J. Lu" writes:
>>> On Mon, Oct 29, 2012 at 5:11 PM, H.J. Lu wrote:
On Mon, Oct 29, 2012 at 4:41 PM, H.J. Lu wrote:
> On Mon, Oct 29, 2012 at 9:38 AM, Vladimir Makarov
>>
"H.J. Lu" writes:
> On Tue, Oct 30, 2012 at 4:38 AM, H.J. Lu wrote:
>> On Tue, Oct 30, 2012 at 4:09 AM, Richard Sandiford
>> wrote:
>>> The address in this case is:
>>>
>>> (plus:SI (mult:SI (reg/v:SI 223 [orig:154 j ] [154])
>>> (const_int 8 [0x8]))
>>> (subreg:SI (plus:DI (reg/f:DI
On Mon, Oct 29, 2012 at 1:56 PM, Jakub Jelinek wrote:
> Status
> ==
>
> I'd like to close the stage 1 phase of GCC 4.8 development
> on Monday, November 5th. If you have still patches for new features you'd
> like to see in GCC 4.8, please post them for review soon. Patches
> posted before t
On 10/26/2012 02:22 PM, Sebastian Huber wrote:
Hello,
here is a test case for PR55033.
Is there something wrong with this test case? It compiles well with Alan's
patch.
--
Sebastian Huber, embedded brains GmbH
Address : Obere Lagerstr. 30, D-82178 Puchheim, Germany
Phone : +49 89 18 90
Hello,
what needs to be done to get this committed?
--
Sebastian Huber, embedded brains GmbH
Address : Obere Lagerstr. 30, D-82178 Puchheim, Germany
Phone : +49 89 18 90 80 79-6
Fax : +49 89 18 90 80 79-9
E-Mail : sebastian.hu...@embedded-brains.de
PGP : Public key available on reque
On Tue, 30 Oct 2012, Richard Biener wrote:
>
> As requested this adds predicates to check whether the lhs of
> a assign or call is a store and whether rhs1 of an assignment
> is a load. It uses this in place of the existing, slightly
> bogus, check in the stmt estimate code.
>
> Bootstrap and r
> And tree expressions don't have TREE_BLOCK before gimple-low either.
> So IMNSHO it is gimple-low.c that should set TREE_BLOCK of all the gimple
> stmts as well as all expression in the operands. It is not overwriting
> anything, no frontend sets TREE_BLOCK for any expression, the way frontends
On 10/30/2012 09:39 AM, Richard Sandiford wrote:
"H.J. Lu" writes:
On Tue, Oct 30, 2012 at 4:38 AM, H.J. Lu wrote:
On Tue, Oct 30, 2012 at 4:09 AM, Richard Sandiford
wrote:
The address in this case is:
(plus:SI (mult:SI (reg/v:SI 223 [orig:154 j ] [154])
(const_int 8 [0x8]))
On Tue, Oct 30, 2012 at 3:17 PM, Dehao Chen wrote:
>> And tree expressions don't have TREE_BLOCK before gimple-low either.
>> So IMNSHO it is gimple-low.c that should set TREE_BLOCK of all the gimple
>> stmts as well as all expression in the operands. It is not overwriting
>> anything, no fronten
I changed the patch according Uros' remarks. Please, have a look.
Changelog:
2012-10-30 Andrey Turetskiy
* config/i386/i386.c (bdesc_args): Rename CODE_FOR_avx2_umulhrswv16hi3 to
CODE_FOR_avx2_pmulhrswv16hi3.
* config/i386/predicates.md (const1_operand): Extend for vector
Hi,
On Tue, 30 Oct 2012, Richard Biener wrote:
> On Tue, Oct 30, 2012 at 3:17 PM, Dehao Chen wrote:
> >> And tree expressions don't have TREE_BLOCK before gimple-low either.
> >> So IMNSHO it is gimple-low.c that should set TREE_BLOCK of all the gimple
> >> stmts as well as all expression in the
On Fri, Oct 26, 2012 at 8:30 AM, Alexandre Oliva wrote:
> Both jump threading and loop induction variable optimizations were
> dropping useful debug information, and it took improvements in both for
> debug info about relevant variables in the enclosed testcase to survive
> all the way to the end.
Hi,
this is first patch of change of tree-ssa-loop-niter to consider bounds that are
not in block dominating latch. This patch makes them to be recorded and they
are not used. I plan to followup with:
1) patch to add simple shortest path walk at the end of
estimate_numbers_of_iterations_loop
On Tue, Oct 30, 2012 at 3:49 PM, Michael Matz wrote:
> Hi,
>
> On Tue, 30 Oct 2012, Richard Biener wrote:
>
>> On Tue, Oct 30, 2012 at 3:17 PM, Dehao Chen wrote:
>> >> And tree expressions don't have TREE_BLOCK before gimple-low either.
>> >> So IMNSHO it is gimple-low.c that should set TREE_BLOC
On Fri, Oct 26, 2012 at 7:27 PM, Jakub Jelinek wrote:
> Hi!
>
> This patch extends optimize_range_tests optimization, so that it
> handles also the cases where the truth && or || has been gimplifed
> as a series of GIMPLE_CONDs or mixture thereof and BIT_{AND,IOR}_EXPR
> stmts.
>
> Example of code
On Sun, Oct 28, 2012 at 6:13 PM, Joern Rennecke
wrote:
> Quoting Richard Biener :
>
>>> Thus, you can allow the length to vary downwards as well as upwards
>>> across iterations with suitable definitions of the @code{length}
>>> attribute
>>> and/or @code{ADJUST_INSN_LENGTH}. Care has to be taken
On Tue, 30 Oct 2012, Jan Hubicka wrote:
> Hi,
> this is first patch of change of tree-ssa-loop-niter to consider bounds that
> are
> not in block dominating latch. This patch makes them to be recorded and they
> are not used. I plan to followup with:
>
> 1) patch to add simple shortest path wa
On Tue, Oct 30, 2012 at 03:38:11PM +0100, Richard Biener wrote:
> I question the need of BLOCK info on expression trees. If BLOCKs are
> relevant then the tree ends up referencing a declaration with a BLOCK as
> context, no? Thus, the case
>
> int tem, a;
> {
> int a;
> ...
> tem
OK.
Jason
On Tue, Oct 30, 2012 at 4:03 PM, Jakub Jelinek wrote:
> On Tue, Oct 30, 2012 at 03:38:11PM +0100, Richard Biener wrote:
>> I question the need of BLOCK info on expression trees. If BLOCKs are
>> relevant then the tree ends up referencing a declaration with a BLOCK as
>> context, no? Thus, the ca
On Tue, Oct 30, 2012 at 03:51:19PM +0100, Richard Biener wrote:
> + FOR_EACH_IMM_USE_STMT (stmt, imm_iter, def)
> + {
> + if (!gimple_debug_bind_p (stmt))
> + continue;
> +
> + FOR_EACH_IMM_USE_ON_STMT (use_p, imm_iter)
> +
On Tue, Oct 30, 2012 at 04:15:38PM +0100, Richard Biener wrote:
> So maybe TER (well, those looking up the stmt) should pick the location
> from the TERed statement properly then?
Perhaps, but Micha's patch doesn't do that.
But in that case IMHO it still would help to set all expr locations to
UNK
On Tue, Oct 30, 2012 at 9:21 AM, Sharad Singhai wrote:
> As per discussion in http://gcc.gnu.org/ml/gcc/2012-10/msg00225.html,
> I have added the -fopt-info pass filtering in the attached patch.
>
> The basic idea is that there are optimization pass groups and a user
> can selectively enable dumps
On Tue, Oct 30, 2012 at 4:21 PM, Jakub Jelinek wrote:
> On Tue, Oct 30, 2012 at 04:15:38PM +0100, Richard Biener wrote:
>> So maybe TER (well, those looking up the stmt) should pick the location
>> from the TERed statement properly then?
>
> Perhaps, but Micha's patch doesn't do that.
> But in tha
On 10/30/2012 06:34 AM, Richard Sandiford wrote:
"H.J. Lu" writes:
LRA has
if (REG_P (reg) && (ep = get_elimination (reg)) != NULL)
{
rtx to_rtx = replace_p ? ep->to_rtx : ep->from_rtx;
if (! replace_p)
{
offset += (ep->offset
On Tue, Oct 30, 2012 at 3:47 PM, Andrey Turetskiy
wrote:
> I changed the patch according Uros' remarks. Please, have a look.
>
> Changelog:
>
> 2012-10-30 Andrey Turetskiy
>
>* config/i386/i386.c (bdesc_args): Rename CODE_FOR_avx2_umulhrswv16hi3
> to
>CODE_FOR_avx2_pmulhrswv16h
On Tue, Oct 30, 2012 at 1:21 AM, Sharad Singhai wrote:
> As per discussion in http://gcc.gnu.org/ml/gcc/2012-10/msg00225.html,
> I have added the -fopt-info pass filtering in the attached patch.
>
> The basic idea is that there are optimization pass groups and a user
> can selectively enable dumps
> I'd say either we should do the TREE_BLOCK setting on all non-shareable
> trees during gimple-low and clear the block (but then likely whole
> location?; it doesn't make sense to say in the debugger that something
> has certain source location when you can't print variables declared in that
> loc
On 09/17/2012 12:54 PM, Florian Weimer wrote:
On 09/17/2012 12:15 PM, Paolo Carlini wrote:
Hi,
On 09/17/2012 11:51 AM, Florian Weimer wrote:
On 08/21/2012 12:37 PM, Florian Weimer wrote:
I don't think there are any callers out there, but let's fix this for
completeness.
A compiler emitting c
On Tue, Oct 30, 2012 at 8:28 AM, Richard Biener
wrote:
> On Tue, Oct 30, 2012 at 9:21 AM, Sharad Singhai wrote:
>> As per discussion in http://gcc.gnu.org/ml/gcc/2012-10/msg00225.html,
>> I have added the -fopt-info pass filtering in the attached patch.
>>
>> The basic idea is that there are opti
Quoting Richard Biener :
Apart from the iteration_threshold the hookization would be straight-forward.
Now I cannot decipher from the patch what functional change it introduces ;)
The only change occurs if we reach an iteration count of MAX_INT iterations -
which should already be indicative o
Hi,
Florian Weimer ha scritto:
>Ping?
>
>Patch is at: http://gcc.gnu.org/ml/gcc-patches/2012-08/msg01416.html
Sorry, I don't know the code well enough to review your patch, but since I'm in
CC, I still don't understand why, instead of adding a full libstdc++ testcase
you are extending a C++ t
Hello,
Hot/cold partitioning is apparently a hot topic all of a sudden, which
is a good thing of course, because it's in need of some TLC.
The attached patch adds another check the RTL cfg checking
(verify_flow_info) for the partitioning: A hot block can never be
dominated by a cold block (becaus
On Tue, Oct 30, 2012 at 8:29 AM, Richard Biener
wrote:
> On Tue, Oct 30, 2012 at 4:21 PM, Jakub Jelinek wrote:
>> On Tue, Oct 30, 2012 at 04:15:38PM +0100, Richard Biener wrote:
>>> So maybe TER (well, those looking up the stmt) should pick the location
>>> from the TERed statement properly then?
On Tue, Oct 30, 2012 at 8:49 AM, Matthew Gretton-Dann wrote:
> On 30 October 2012 05:20, Teresa Johnson wrote:
>> Index: cfgrtl.c
>> ===
>> --- cfgrtl.c(revision 192692)
>> +++ cfgrtl.c(working copy)
>> @@ -912,7 +912,8 @@ rtl_
BTW, one thing I found confusing is that in expr.c, some code is for
frontend, while some are for rtl. Shall we separate them into two
files? And we don't expect to see EXPR_LOCATION in the rtl side.
Thanks,
Dehao
On 10/30/2012 05:17 PM, Paolo Carlini wrote:
Sorry, I don't know the code well enough to review your patch, but since I'm in
CC, I still don't understand why, instead of adding a full libstdc++ testcase
you are extending a C++ testcase, in old-deja even, normally considered legacy.
AFAIK, th
> The debugger isn't the only consumer of debug info, and other tools might need
> a finer granularity than a GIMPLE location, so clearing EXPR_LOCATION to work
> around a debug info size issue seems very short-sighted (to say the least).
Hi, Eric,
There might be some misunderstanding here. Clear
I'm hitting the same bug as in PR53708 when compiling GLIBC's dlfcn.c when
vectorization is enabled on powerpc64-linux. A reduced test case is:
bergner@bns:~/gcc/BUGS> cat foo.i
static void (*const init_array []) (void)
__attribute__ ((section (".init_array"), aligned (sizeof (void *)), used))
On Tue, Oct 30, 2012 at 9:26 AM, Steven Bosscher wrote:
> Hello,
>
> Hot/cold partitioning is apparently a hot topic all of a sudden, which
> is a good thing of course, because it's in need of some TLC.
>
> The attached patch adds another check the RTL cfg checking
> (verify_flow_info) for the par
> gcc/ChangeLog:
> 2012-10-25 Dehao Chen
>
> * tree-eh.c (do_return_redirection): Set location for jump statement.
> (do_goto_redirection): Likewise.
> (frob_into_branch_around): Likewise.
> (lower_try_finally_nofallthru): Likewise.
> (lower_try_finally_co
On Tue, Oct 30, 2012 at 6:20 AM, Teresa Johnson wrote:
> Index: bb-reorder.c
> ===
> --- bb-reorder.c(revision 192692)
> +++ bb-reorder.c(working copy)
> @@ -2188,6 +2188,8 @@ insert_section_boundary_note (void)
>
Thanks for explanation, I understand it.
I fixed issue which you marked. Changelog is unchanged.
---
Best regards,
Andrey Turetskiy
On Tue, Oct 30, 2012 at 7:40 PM, Uros Bizjak wrote:
> On Tue, Oct 30, 2012 at 3:47 PM, Andrey Turetskiy
> wrote:
>> I changed the patch according Uros' remarks. Pl
On Tue, Oct 30, 2012 at 12:10 AM, Ralf Corsepius
wrote:
>
> I would like to apply the patch below to trunk and gcc-4.7-branch.
>
> This patch was originalyl submitted by Joel Sherrill back in May
> (http://gcc.gnu.org/ml/gcc-patches/2012-05/msg01180.html),
> but had never received any feedback.
>
On Tue, Oct 30, 2012 at 6:53 PM, Andrey Turetskiy
wrote:
> Thanks for explanation, I understand it.
> I fixed issue which you marked. Changelog is unchanged.
>
>>> I changed the patch according Uros' remarks. Please, have a look.
>>>
>>> Changelog:
>>>
>>> 2012-10-30 Andrey Turetskiy
>>>
>>>
On 10/30/12, Diego Novillo wrote:
> On Oct 30, 2012 Bin.Cheng wrote:
> > Just one question: Should we change the name of functions
> > "sbitmap_intersection_of_succs/sbitmap_intersection_of_preds/
> > sbitmap_union_of_succs/sbitmap_union_of_preds" too? It might
> > be a little confusing that sbit
Hi,
this patch implements the second part of planned change - to determine loop
bounds
based by shortest path discovery. This allows to bound number of iterations
on loops with bounds in statements that do not dominate the latch.
I originally planned to implement this as part of maybe_lower_iter
On Tue, 2012-10-30 at 11:58 -0500, Peter Bergner wrote:
> I'm hitting the same bug as in PR53708 when compiling GLIBC's dlfcn.c when
> vectorization is enabled on powerpc64-linux. A reduced test case is:
>
> bergner@bns:~/gcc/BUGS> cat foo.i
> static void (*const init_array []) (void)
> __attr
On Tue, Oct 30, 2012 at 6:48 PM, Steven Bosscher wrote:
> On Tue, Oct 30, 2012 at 6:20 AM, Teresa Johnson wrote:
>> Index: bb-reorder.c
>> ===
>> --- bb-reorder.c(revision 192692)
>> +++ bb-reorder.c(working copy)
>> @
On Tue, Oct 30, 2012 at 01:43:33PM -0500, Peter Bergner wrote:
> Commenting on Richard's question from the bugzilla:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53708#c10
>
> I suppose if attribute((__aligned__)) truly does just set a minimum alignment
> value (and the documentation seems t
I can't approve the whole thing of course, but I like the idea.
However...
Joern Rennecke writes:
> +@deftypevr {Target Hook} bool TARGET_HAVE_CC0
> +@deftypevrx {Target Hook} {bool} TARGET_AUTO_INC_DEC
> +@deftypevrx {Target Hook} {bool} TARGET_STACK_REGS
> +@deftypevrx {Target Hook} {bool} TARG
On Tue, 2012-10-30 at 19:55 +0100, Jakub Jelinek wrote:
> On Tue, Oct 30, 2012 at 01:43:33PM -0500, Peter Bergner wrote:
> > Commenting on Richard's question from the bugzilla:
> >
> > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53708#c10
> >
> > I suppose if attribute((__aligned__)) truly does
This patch replaces three separate default definitions of
SLOW_UNALIGNED_ACCESS with a single global one. Note that tm.texi
requires SLOW_UNALIGNED_ACCESS to be true if STRICT_ALIGNMENT.
Tested on x86_64-linux-gnu, powerpc64-linux-gnu and mipsisa64-elf.
Applied as obvious.
Richard
gcc/
On 10/27/2012 09:16 PM, Sriraman Tallam wrote:
+ /* See if there's a match. For functions that are multi-versioned,
+all the versions match. */
if (same_type_p (target_fn_type, static_fn_type (fn)))
- matches = tree_cons (fn, NULL_TREE, matches);
+
I'm finishing off some patches to allow insv, extv and extzv to be
defined as normal direct optabs (such as "insvsi" and "insvdi" rather
than just "insv"). This series of patches does some groundwork to make
that possible.
The patches are not supposed to change the generated code. I checked
that
This patch removes what I believe is a redundant check in store_bit_field_1
for whether the value to insert (i.e. the rhs) has BLKmode. We shouldn't
see BLKmode values here, and even if we did, the only effect of the test
is to fall through to store_fixed_bit_field, which can't handle BLKmode
eith
extract_bit_field_1 has a block beginning:
/* If OP0 is a memory, try copying it to a register and seeing if a
cheap register alternative is available. */
if (ext_mode != MAX_MACHINE_MODE && MEM_P (op0))
{
and within it there are tests for whether ext_mode != MAX_MACHINE_MODE.
This
This patch splits out the code to handle insv and ext(z)v from
store_bit_field_1 and extract_bit_field_1 respectively. I removed
"x" prefixes from some of the variables and tried to make the placement
of the REG and SUBREG handling more consistent, but there are no
behavioural changes.
Tested as
This patch simply separates out the MEM and non-MEM insv and ext(z)v cases.
On it's own, it's probably a wash whether this is an improvement or not,
but it makes the optabs patches much easier.
Tested as described in the covering note. OK to install?
Richard
gcc/
* expmed.c (store_bit_
On Tue, Oct 30, 2012 at 5:59 PM, Teresa Johnson wrote:
> I will try testing your patch on top of mine with our fdo benchmarks.
Thanks. But you should expect a lot of errors, hopefully you can make
something out of it for Bugzilla.
> For the others on the cc list, you may need to include my patch
This patch splits out a fairly common operation: that of narrowing a MEM
to a particular mode and adjusting the bit number accordingly.
I've kept with "bit_field" rather than "bitfield" for consistency with
the callers, although we do have "bitfield" in "adjust_bitfield_address".
Tested as descri
On a change of tack, this tackles some redundant code in combine.
It has code to convert a variable bit position to the mode required
by the bit position operand to insv, extv or extzv:
[A]
else if (pos_rtx != 0
&& GET_MODE_SIZE (pos_mode) < GET_MODE_SIZE (GET_MODE (pos_rtx)))
pos
On Tue, Oct 30, 2012 at 12:25 PM, Steven Bosscher wrote:
> On Tue, Oct 30, 2012 at 5:59 PM, Teresa Johnson wrote:
>> I will try testing your patch on top of mine with our fdo benchmarks.
>
> Thanks. But you should expect a lot of errors, hopefully you can make
> something out of it for Bugzilla.
>
Combine tries to optimise comparisons involving:
(zero_extract (const_int X)
(const_int 1)
(var Y))
and so on BITS_BIG_ENDIAN targets it tries gamely to work out what mode
X actually has. At the moment it tries reading the mode from operand 1
of extzv, but
On Tue, Oct 30, 2012 at 02:03:44PM -0500, Peter Bergner wrote:
> Ok, then I'll bootstrap and regtest your suggested change while we
> wait for richi to comment. I'm fine with whatever you and richi
> decide is best. The ObjC guys should probably test it though too.
>
> I assume you think we shou
The optabs patches that I'm working on treat patterns with names like
"insvsi" and "insvdi" as part of the public target interface.
Those names clash with some existing rs6000.md patterns, so this patch adds
some _internal suffixes (a bit like fix_truncsi2_internal, for example).
AFAICT there are
This patch adds a @direntry for gcov.
I noticed that it was missing today, when I tried to find the gcov
manual from the info "dir" node. Then I found out that I had filed PR
50899 for this ages ago.
Ok?
Tom
2012-10-30 Tom Tromey
PR other/50899
* doc/gcc.texi: Add @direntry
On Tue, 2012-10-30 at 20:37 +0100, Jakub Jelinek wrote:
> On Tue, Oct 30, 2012 at 02:03:44PM -0500, Peter Bergner wrote:
> > Ok, then I'll bootstrap and regtest your suggested change while we
> > wait for richi to comment. I'm fine with whatever you and richi
> > decide is best. The ObjC guys sho
1 - 100 of 153 matches
Mail list logo
