I stumbled over this while looking at regressions triggered when moving
certain branch-cost driven transformations from fold-const.c to a later
point in the pipeline.
The coalescing we do as part of the out-of-ssa process restricts itself
to only coalescing when the types of the underlying
On 05/17/2013 01:27 AM, David Edelsohn wrote:
On Fri, May 10, 2013 at 11:43 PM, Chung-Ju Wu jasonw...@gmail.com wrote:
2013/5/10 Sebastian Huber sebastian.hu...@embedded-brains.de:
v2: Format changes
gcc/testsuite/ChangeLog
2013-05-10 Sebastian Huber sebastian.hu...@embedded-brains.de
On Fri, Jun 7, 2013 at 8:07 AM, Jeff Law wrote:
Rather than using strict pointer equality, we can do better by looking at
TYPE_CANONICAL when it's available. Thus objects of the following two types
(T1 T2) become candidates for coalescing if they are tied together by a
copy or PHI node.
On 06/06/13 17:26, Richard Henderson wrote:
On 06/06/2013 08:02 AM, Richard Earnshaw wrote:
(define_insn addmode3
- [(set (match_operand:FIXED 0 s_register_operand =r)
-(plus:FIXED (match_operand:FIXED 1 s_register_operand
r)
-(match_operand:FIXED 2
On Fri, Jun 7, 2013 at 4:00 AM, Alan Modra amo...@gmail.com wrote:
force_const_mem() isn't supposed to handle VOIDmode or BLKmode, so the
check for VOIDmode when aligning is needless. If we ever did get one
of these modes in a constant pool, this
pool-offset += GET_MODE_SIZE (mode);
On Thu, Jun 6, 2013 at 5:10 PM, Dehao Chen de...@google.com wrote:
Hi, Martin,
Yes, your patch can fix my case. Thanks a lot for the fix.
With the fix, value profiling will still promote the wrong indirect
call target. Though it will not be inlining, but it results in an
additional check.
On Fri, Jun 7, 2013 at 8:07 AM, Jeff Law l...@redhat.com wrote:
I stumbled over this while looking at regressions triggered when moving
certain branch-cost driven transformations from fold-const.c to a later
point in the pipeline.
The coalescing we do as part of the out-of-ssa process
On 06/06/2013 06:34 PM, Uros Bizjak wrote:
Hello!
This patch avoids ABI check failure on alpha.
2013-06-06 Uros Bizjak ubiz...@gmail.com
* config/abi/post/alpha-linux-gnu/baseline_symbols.txt: Update.
Tested on alphaev68-pc-linux-gnu.
OK for mainline?
Ok, thanks. Don't you need to
On Fri, Jun 7, 2013 at 11:40 AM, Paolo Carlini paolo.carl...@oracle.com wrote:
This patch avoids ABI check failure on alpha.
2013-06-06 Uros Bizjak ubiz...@gmail.com
* config/abi/post/alpha-linux-gnu/baseline_symbols.txt: Update.
Tested on alphaev68-pc-linux-gnu.
OK for mainline?
2013/6/7 Sebastian Huber sebastian.hu...@embedded-brains.de:
On 05/17/2013 01:27 AM, David Edelsohn wrote:
On Fri, May 10, 2013 at 11:43 PM, Chung-Ju Wu jasonw...@gmail.com wrote:
2013/5/10 Sebastian Huber sebastian.hu...@embedded-brains.de:
v2: Format changes
gcc/testsuite/ChangeLog
tree-stdarg.c:check_all_va_list_escapes looks at va_arg accesses
to determine cfun-va_list_gpr_size. The value set is the offset of
the highest access in the va_list area, plus the size of that access.
For big-endian rs6000 this always results in a multiple of the gpr
register size, but that
On Fri, Jun 7, 2013 at 6:04 AM, Alan Modra amo...@gmail.com wrote:
tree-stdarg.c:check_all_va_list_escapes looks at va_arg accesses
to determine cfun-va_list_gpr_size. The value set is the offset of
the highest access in the va_list area, plus the size of that access.
For big-endian rs6000
On 06/06/13 17:08, Kyrylo Tkachov wrote:
Hi all,
This patch cleans up the anddi3_insn pattern by removing duplicate
alternatives and redundant attribute setters.
It also restricts the splitting to after reload and when we know that
we're not using the NEON versions.
Tested arm-none-eabi on
On Wed, Jun 05, 2013 at 07:50:52PM +, Joseph S. Myers wrote:
On Wed, 5 Jun 2013, Marek Polacek wrote:
It works by creating a COMPOUND_EXPR around original expression, so e.g.
it creates:
if (b 0 || (b 31 || a 0))
{
__builtin___ubsan_handle_shift_out_of_bounds ();
}
Andreas Schwab sch...@linux-m68k.org writes:
Rainer Orth r...@cebitec.uni-bielefeld.de writes:
Indeed, and I see -ansi -pedantic-errors -gdwarf-2 -O0 used for the test
on both x86_64-unknown-linux-gnu and i386-pc-solaris2.10 (if I augment
the target clause).
DEFAULT_CFLAGS is set by many
Jason Merrill ja...@redhat.com writes:
On 05/31/2013 07:13 AM, Rainer Orth wrote:
Also, both testcases are missing from gcc/testsuite/ChangeLog.
I prefer not to mess with logging testcase changes.
http://gcc.gnu.org/codingconventions.html#ChangeLogs
I hadn't seen this before. I think this
On 06/07/13 02:30, Steven Bosscher wrote:
On Fri, Jun 7, 2013 at 8:07 AM, Jeff Law wrote:
Rather than using strict pointer equality, we can do better by looking at
TYPE_CANONICAL when it's available. Thus objects of the following two types
(T1 T2) become candidates for coalescing if they are
Hi all,
In a similar vein to
http://gcc.gnu.org/ml/gcc-patches/2013-05/msg01838.html this patch lets
us use the immediate forms of the eor instruction when dealing with
64-bit constants.
The insn is changed into an insn_and_split and merged with the NEON
version. A new constraint is introduced to
Oops, once again with patch attached...
gcc/
2013-06-07 Kyrylo Tkachov kyrylo.tkac...@arm.com
PR target/56315
* config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
(xordi3): Change operand 2 constraint to arm_xordi_operand.
* config/arm/constraints.md
Vladimir,
If I introduce an unbalanced parentheses error in a reservation string, f.i. in
athlon.md using the following patch:
...
diff --git a/gcc/config/i386/athlon.md b/gcc/config/i386/athlon.md
index d872b8f..b1ed5cd 100644
--- a/gcc/config/i386/athlon.md
+++ b/gcc/config/i386/athlon.md
@@
On 07/06/13 14:04, Kyrylo Tkachov wrote:
Oops, once again with patch attached...
gcc/
2013-06-07 Kyrylo Tkachov kyrylo.tkac...@arm.com
PR target/56315
* config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
(xordi3): Change operand 2 constraint to
On Fri, Jun 7, 2013 at 2:05 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Thu, Jun 6, 2013 at 5:10 PM, Dehao Chen de...@google.com wrote:
Hi, Martin,
Yes, your patch can fix my case. Thanks a lot for the fix.
With the fix, value profiling will still promote the wrong indirect
call
It helps to initialize a block before calling gfc_add_block_to_block
(block ...
Committed as Rev. 199812 after building and regtesting on x86-64-gnu-linux.
Tobias
Index: gcc/fortran/ChangeLog
===
--- gcc/fortran/ChangeLog
On Fri, Jun 7, 2013 at 3:30 PM, Xinliang David Li davi...@google.com wrote:
On Fri, Jun 7, 2013 at 2:05 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Thu, Jun 6, 2013 at 5:10 PM, Dehao Chen de...@google.com wrote:
Hi, Martin,
Yes, your patch can fix my case. Thanks a lot for the
On 06/06/2013 09:50 PM, Alan Modra wrote:
* varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
reporting.
(get_named_section): Don't NULL !DECL_P decl.
Ok.
r~
Unreviewed patches:
* http://gcc.gnu.org/ml/fortran/2013-06/msg00027.html
* http://gcc.gnu.org/ml/fortran/2013-06/msg00048.html
* * *
As with ALLOCATE (type-spec :: ...), also array constructors take as
type-spec a type name (integer, my_dt, etc.) without an extra TYPE().
Thus, instead of
On 29 June 2012 04:59, Mike Stump mikest...@comcast.net wrote:
On Jun 28, 2012, at 5:15 PM, Bernhard Reutner-Fischer wrote:
On Thu, Jun 28, 2012 at 04:43:05PM -0700, Mike Stump wrote:
On Jun 28, 2012, at 3:27 PM, Bernhard Reutner-Fischer wrote:
Perhaps you want to pursue this? We'd need to
On 13-06-06 9:22 PM, David Edelsohn wrote:
The patch actually restore the LRA behaviour for x86/x86-64 before rev. 199298.
The revision was added for PPC SDmode value correct generation. So it is really
needed for PPC64 and badly hurts x86/x86-64 performance (by doing secondary
memory reloads
Greetings,
Jonathan had asked if I could make some simplifications to the
organization of C++ versioning macros and namespaces in my standard
literals patch. I plum forgot to do it before I put in the patch.
Also, I noticed I had picked up some garbage comments in the test cases.
This
On Fri, Jun 07, 2013 at 04:03:52PM +0200, Tobias Burnus wrote:
Unreviewed patches:
* http://gcc.gnu.org/ml/fortran/2013-06/msg00027.html
* http://gcc.gnu.org/ml/fortran/2013-06/msg00048.html
* * *
As with ALLOCATE (type-spec :: ...), also array constructors take as
type-spec a type
On Fri, May 31, 2013 at 02:11:38PM -0400, Vladimir Makarov wrote:
The following patch switches LRA on for s390. The patch introduces
a new option -mlra to use LRA instead of reload. I did not document
the option as I'd like to delete it for gcc4.9. By default, LRA will
be used for s390
On Fri, Jun 7, 2013 at 6:47 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Fri, Jun 7, 2013 at 3:30 PM, Xinliang David Li davi...@google.com wrote:
On Fri, Jun 7, 2013 at 2:05 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Thu, Jun 6, 2013 at 5:10 PM, Dehao Chen
On 13-06-07 10:57 AM, Andreas Krebbel wrote:
On Fri, May 31, 2013 at 02:11:38PM -0400, Vladimir Makarov wrote:
The following patch switches LRA on for s390. The patch introduces
a new option -mlra to use LRA instead of reload. I did not document
the option as I'd like to delete it for
Hello,
After seeing Steve Ellcey's patch at:
http://gcc.gnu.org/ml/gcc-patches/2013-05/msg00667.html
I've been playing with pushing the same logic closer to the
current jump-threading framework.
I've split off Steve's patch into path search code (added to
tree-ssa-threadedge.c) and path copy
This is a resubmission of a patch originally posted here
http://gcc.gnu.org/ml/gcc-patches/2013-05/msg01488.html
It was only intended as a stop-gap measure to avoid a couple of
libstdc++ testsuite failures on Solaris 9 and 10 where nanosleep only
lives in librt, which is now required
I haven't looked at the patch yet and my knowledge of spec files is
pretty weak, so sorry if these are stupid questions ...
On 7 June 2013 16:45, Rainer Orth wrote:
Follow libgfortran's lead and introduce libstdc++.spec. Here,
target-specific linker switches can be determined by the
Jonathan Wakely jwakely@gmail.com writes:
On 7 June 2013 16:45, Rainer Orth wrote:
Follow libgfortran's lead and introduce libstdc++.spec. Here,
target-specific linker switches can be determined by the libstdc++
configury, with configuration well localized, and one can easily handle
Hello Richard et al.,
I looked at my C++ implementation for Array Notations and a bunch of
helper functions (and one structure) can be shared between C and C++. Attached
is a patch that will move these functions from c/c-array-notation.c to
c-family/array-notation-common.c.
Is this OK
On 1 June 2013 14:06, Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch cleanups way we handle aliases. The main point is to merge code
that was previously done separately for variables and functions.
Hello, This patch appears to break both arm and aarch64. I don't
fully understand the
This patch fixes two issues:
* storage_size('aa') was rejected as constant expression - as
ts.u.cl-length == 0.
* In trans*.c, there was a fold_convert missing (- ICE). Additionally,
I have replaced the detour to generate a tree containing the value 8
via a fortran expression.
Build and
On 06/07/2013 09:08 AM, Iyer, Balaji V wrote:
Hello Richard et al.,
I looked at my C++ implementation for Array Notations and a bunch of
helper functions (and one structure) can be shared between C and C++.
Attached is a patch that will move these functions from c/c-array-notation.c
On 1 June 2013 14:06, Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch cleanups way we handle aliases. The main point is to merge code
that was previously done separately for variables and functions.
Hello, This patch appears to break both arm and aarch64. I don't
fully understand
On 6/6/13 9:00 AM, Richard Earnshaw wrote:
The pipeline offset is 4 for Thumb2 as well. So at the very least you
need to explain why your change doesn't apply then as well.
Yes some context is lost in that comment. Thunks are usually emitted in
ARM mode, except for Thumb-only targets. Is the
Hi,
on some testers attr-alias.c fails because ( in the template is interpreted by
dejagnu rather than part of the pattern. This patch fixes it by simply testing
for indentation of the call (to not match the declaration nor function body
dump header).
Comitted as obvious.
Honza
Index:
The attached patch cleans up the C++ pragma simd implementation to share
more things with the C front-end, particularly the type checking. For
instance, I got rid of all the specialized parsing routines to just use
generic cp_parser_*expression* and then perform the actual typechecking
in the
On Thu, May 30, 2013 at 6:10 PM, Iyer, Balaji V balaji.v.i...@intel.com wrote:
I think David is getting the correct output but just that dg-error is not
catching it correctly.
What version of expect are you using?
Fedora 18 apparently has 5.45, others are using 5.44, and AIX 7.1
provides
interesting_blocks is only used inside tree-into-ssa.c, so it can be
made static.
Successfully bootstrapped on x86_64-unknown-linux-gnu (using
gcc-4.7.2-2.fc17.x86_64).
OK for trunk?
2013-06-07 David Malcolm dmalc...@redhat.com
* tree-into-ssa.c (interesting_blocks): Make static.
The array unknown within tree-object-size.c is only ever read from,
not written to, so it can be marked as const.
Successfully bootstrapped on x86_64-unknown-linux-gnu (using
gcc-4.7.2-2.fc17.x86_64).
OK for trunk?
2013-06-07 David Malcolm dmalc...@redhat.com
* tree-object-size.c
Hi,
See http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57548
The ICE here is because of a multi-versioned function called from
global namespace and has no caller. This ICEs in target hook
ix86_can_inline_p as caller is 0x0. The following simple patch
attached fixes this problem.
*
On Fri, Jun 07, 2013 at 01:40:00PM -0400, David Malcolm wrote:
The array unknown within tree-object-size.c is only ever read from,
not written to, so it can be marked as const.
Successfully bootstrapped on x86_64-unknown-linux-gnu (using
gcc-4.7.2-2.fc17.x86_64).
OK for trunk?
On Fri, 2013-06-07 at 20:49 +0200, Jakub Jelinek wrote:
On Fri, Jun 07, 2013 at 01:40:00PM -0400, David Malcolm wrote:
The array unknown within tree-object-size.c is only ever read from,
not written to, so it can be marked as const.
Successfully bootstrapped on x86_64-unknown-linux-gnu
This patch adds instruction scheduling support for the Power8 processor.
Bootstrap/regression test with no new failures. Ok for trunk?
2013-06-07 Michael Meissner meiss...@linux.vnet.ibm.com
Pat Haugen pthau...@us.ibm.com
Peter Bergner berg...@vnet.ibm.com
*
Hi!
This PR is about DATA_ALIGNMENT macro increasing alignment of some decls
for optimization purposes beyond ABI mandated levels. It is fine to emit
the vars aligned as much as we want for optimization purposes, but if we
can't be sure that references to that decl bind to the definition we
OK.
Jason
On 1 June 2013 14:06, Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch cleanups way we handle aliases. The main point is to merge
code
that was previously done separately for variables and functions.
Hello, This patch appears to break both arm and aarch64. I don't
fully
Hi,
please see the PR log for details. My symtab patch made testcase anon6.C to
fail because it tests that unused static var is output into the assembly and we
now optimize it out.
Adding attribute used however shows problem that the var suddenly becomes WEAK
and COMDAT. It seems to me that this
On 06/07/2013 12:25 PM, Jakub Jelinek wrote:
This PR is about DATA_ALIGNMENT macro increasing alignment of some decls
for optimization purposes beyond ABI mandated levels. It is fine to emit
the vars aligned as much as we want for optimization purposes, but if we
can't be sure that references
On Fri, Jun 07, 2013 at 01:43:27PM -0700, Richard Henderson wrote:
On 06/07/2013 12:25 PM, Jakub Jelinek wrote:
This PR is about DATA_ALIGNMENT macro increasing alignment of some decls
for optimization purposes beyond ABI mandated levels. It is fine to emit
the vars aligned as much as we
Hello Everyone,
This patch below should fix the bug reported in PR 57541. The following
statements were not caught by the array notation expander, and they should be
caught and replaced with zero nodes:
A[:];
A[x:y];
A[x:y:z];
Here are the Changelogs
gcc/c/ChangeLog
2013-06-07
On Fri, 7 Jun 2013, Richard Henderson wrote:
I've had a brief look over the instances of D_A within the tree atm. Most of
them carry the cut-n-paste comment for the same reasons. These I believe
never intended an ABI change, and were really only interested in optimization.
But these I think
On Fri, Jun 7, 2013 at 6:47 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Fri, Jun 7, 2013 at 3:30 PM, Xinliang David Li davi...@google.com wrote:
On Fri, Jun 7, 2013 at 2:05 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Thu, Jun 6, 2013 at 5:10 PM, Dehao Chen
OK.
Jason
On 06/06/2013 10:55 AM, Ed Smith-Rowland wrote:
On 06/05/2013 04:01 PM, Jonathan Wakely wrote:
On 5 June 2013 20:18, Ed Smith-Rowland wrote:
Greetings,
This patch implements quoted string manipulators for C++14.
27.7.6 - Quoted manipulators[quoted.manip].
The idea is to allow round
63 matches
Mail list logo