Hello,
+ if (reg_mentioned_p (reg, insn) !DEBUG_INSN_P (insn))
It probably makes sense to test for !DEBUG_INSN_P first, since it's much
cheaper.
Thanks, will commit the following to fix that:
* modulo-sched.c (doloop_register_get): Check !DEBUG_INSN_P
first.
Index:
On Wed, May 11, 2011 at 5:46 PM, Kai Tietz ktiet...@googlemail.com wrote:
2011/5/11 Richard Guenther richard.guent...@gmail.com:
The most important thing is to get predicate types sane - that affects
tcc_comparison codes and the TRUTH_* codes. After that, the TRUTH_*
codes are redundant with
On Wed, May 11, 2011 at 9:22 PM, Nathan Froyd froy...@codesourcery.com wrote:
The comment for pointer_map_traverse says:
/* Pass each pointer in PMAP to the function in FN, together with the pointer
to the value and the fixed parameter DATA. If FN returns false, the
iteration stops.
On Thu, May 12, 2011 at 12:14 AM, pier...@pvittet.com wrote:
After using function flow_loops_find in cfgloop.c, it is needed to use
free_dominance_info to clear structure which have beend added by the use of
calculate_dominance_info.
I have added a comment precising this.
My contributor
On Thu, May 12, 2011 at 10:09:42AM +0100, Hari Sandanagobalane wrote:
The REGNO_REG_CLASS is generally an array of size
FIRST_PSEUDO_REGISTER. So, the indexes go from 0 to
FIRST_PSEUDO_REGISTER-1.
That is true.
I think the = condition is fine in that
case. Do you agree?
That is wrong. It
This changes all uses of AC_ARG_(WITH|ENABLE) to use AS_HELP_STRING,
fixing a few quoting bugs on the way.
OK for trunk?
Andreas.
2011-05-12 Andreas Schwab sch...@redhat.com
* configure.ac: Use AS_HELP_STRING throughout.
* configure: Regenerate.
diff --git a/gcc/configure.ac
Hello, gentle maintainer.
This is a message from the Translation Project robot.
A revised PO file for textual domain 'cpplib' has been submitted
by the Chinese (simplified) team of translators. The file is available at:
http://translationproject.org/latest/cpplib/zh_CN.po
(This file,
cpplib-4.6.0.zh_CN.po.gz
Description: Binary data
The Translation Project robot, in the
name of your translation coordinator.
coordina...@translationproject.org
Hi Guys,
I am applying the patch below to the RX backend on the 4.6 branch. It
fixes several small, obvious bugs, as described in the changelog.
Cheers
Nick
gcc/ChangeLog
2011-05-12 Nick Clifton ni...@redhat.com
* config/rx/rx.h (HAVE_PRE_DECREMENT): Fix typo.
*
Hi Guys,
I am applying the patch below to fix some small, obvious problems with
the RX backend in the mainline sources.
Cheers
Nick
gcc/ChangeLog
2011-05-12 DJ Delorie d...@redhat.com
Nick Clifton ni...@redhat.com
* config/rx/rx.h (HAVE_PRE_DECREMENT): Fix typo.
Hi Guys,
I am checking in this patch, created by DJ Delorie, to add
TARGET_BUILTIN_DECL to the RX backend.
Cheers
Nick
gcc/ChangeLog
2011-05-12 DJ Delorie d...@redhat.com
(rx_builtins): New arrays - holds builtin functions.
(ADD_RX_BUILTIN1, ADD_RX_BUILTIN2,
On 12/05/11 10:18, Jakub Jelinek wrote:
On Thu, May 12, 2011 at 10:09:42AM +0100, Hari Sandanagobalane wrote:
The REGNO_REG_CLASS is generally an array of size
FIRST_PSEUDO_REGISTER. So, the indexes go from 0 to
FIRST_PSEUDO_REGISTER-1.
That is true.
I think the = condition is fine in
Hi,
the problem is that Andi's test build is with -O0 and we should not allocate
ipa-prop datastructures then. We do so via duplication hook by accident since
do_whole_program_analysis for some illogical reason calls
ipa_register_cgraph_hooks.
Comitted as obvoius.
Honza
PR lto/48952
This removes the mode param from the generic type merging machinery
and simplifies code accordingly.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Hi,
looks the cleanup series makes a progress, now I get 459s for Mozilla WPA (it
was 540 last time I tried).
This fixes PR48172 by properly doing the runtime alias check for
vectorization.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk
sofar.
Richard.
2011-05-12 Richard Guenther rguent...@suse.de
PR tree-optimization/48172
* tree-vect-loop-manip.c
Hello,
The attached patch fixes SMS testsuite failures seen on PowerPC and SPU.
Tested ppc64-redhat-linux on both with -m32,-m64 and SPU.
OK for mainline?
Thanks,
Revital
testsuite/Changelog
PR rtl-optimization/47013
* gcc.dg/sms-2.c: Change scan-tree-dump-times and the code
While testing some other issues with C++'s __java_boolean type
occurred. So I adjusted check in test-cfg.c as you suggested.
Additionally due the fact that we are now boolifying conditions for
even BOOLEAN_TYPE'ed cases (for making sure inner arms are boolified,
too), we possibly would alter here
Hello Ramana,
Following our conversation; here is the patch again without
the arm specific flags. Tested on ARM machine configured with
[--with-arch=armv7-a] --with-mode=thumb.
OK for mainline?
Thanks,
Revital
testsuite/Changlog:
* gcc.target/arm/sms-9.c: New file.
*
This improves hashing and comparing both regular and canonical type
merging.
Bootstrapped and tested on x86_64-unknonw-linux-gnu and SPEC2k6 tested,
applied to trunk.
Richard.
2011-05-12 Richard Guenther rguent...@suse.de
* gimple.c (gtc_visit): Compare TREE_ADDRESSABLE, handle
On Thu, May 12, 2011 at 3:29 PM, Kai Tietz ktiet...@googlemail.com wrote:
While testing some other issues with C++'s __java_boolean type
occurred. So I adjusted check in test-cfg.c as you suggested.
Additionally due the fact that we are now boolifying conditions for
even BOOLEAN_TYPE'ed cases
If an intrinsic module proc [such as ISO_Fortran_env's
compiler_options()] is resolved multiple times, with -Wall a warning is
printed, claiming that one has manually typed an intrinsic.
The fix is to avoid resolving the proc multiple times. Note: Typing of
use-associated symbols will always
Hi!
The read from a 1-bit signed bitfield initialized by a comparison
is optimized into the comparison, which has that 1-bit signed bitfield
comparison. Unfortunately that is still expanded as setting the result
to 0 resp. 1 instead of this case 0 resp. -1 QImode pseudo, which is then
sign
Hi!
combine_blocks at the end removes most of the bbs, keeps around just
loop header, maybe latch and maybe exit_bb. We need to free bb-aux
through free_bb_predicate, but that is done in the caller, using array of
(former) loop bbs, with the new loop-num_nodes count (at most 3).
While that
On Thu, May 12, 2011 at 4:42 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
combine_blocks at the end removes most of the bbs, keeps around just
loop header, maybe latch and maybe exit_bb. We need to free bb-aux
through free_bb_predicate, but that is done in the caller, using array of
Hi!
- http://gcc.gnu.org/ml/gcc-patches/2011-05/msg00403.html
debug info improvement for unused parameters passed in memory
- http://gcc.gnu.org/ml/gcc-patches/2011-04/msg01669.html
PR48688 optimization, I know Richard asked for trying it during
combine, but that attempt failed due to
You can invoke a procedure on a cast (type conversion) in Ada, even if the
formal parameter is Out or In Out; if this case, the conversion is deemed a
view conversion and is applied in both directions.
Geert just found out that we don't properly handle these view conversions for
unconstrained
On Wed, May 11, 2011 at 12:28:18PM -0700, H.J. Lu wrote:
This patch restores the old behavior for Pmode. OK for trunk if there
are no regressions?
That is IMHO wrong, ignoring subregs is a very bad idea.
While you can workaround generation of the DW_OP_GNU_convert that way
(why do you want
On Thu, May 12, 2011 at 7:56 AM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, May 11, 2011 at 12:28:18PM -0700, H.J. Lu wrote:
This patch restores the old behavior for Pmode. OK for trunk if there
are no regressions?
That is IMHO wrong, ignoring subregs is a very bad idea.
While you can
Hi,
the original intention of this patch was not only to make the
interface a bit nicer and internals of the lattice isolated but also
so that we do not use struct ipcp_lattice in ipa-inline-analysis.c and
thus my large modifications of theat structure do not have to be
reflected in the file.
On 05/10/2011 06:57 PM, Joseph S. Myers wrote:
On Tue, 10 May 2011, Bernd Schmidt wrote:
On C6X, we have PARM_BOUNDARY == 8 (one byte), but some function
argument slots still must be rounded to a larger value. As far as I
could tell there's currently no way of telling gcc about this, hence a
On 05/11/2011 12:45 PM, Alexander Monakov wrote:
Could the above be conditional on whether the target CPU is exposed-pipeline?
Can do. New patch below.
I'm concerned this may degrade scheduling for other targets in some cases.
On the other hand, it may also improve scheduling. Hard to say
By the way, in the future we may want ipa-cp to merge clones that are
equivalent (let's say one of the predicate says that a parameter is
greater than 10 and we have a bunch of callers supplying values 1, 2
and 8, we will want to create just one clone for all three values
because the effect
This patch fixes the problem in two steps:
(i) It reverts r162289; and
(ii) It adds the correct initialization for loop.reverse[] in
gfc_trans_assignment_1. This was implemented incorrectly in the fix
for PR24524 (in spite of the correct comment in dependency.c!) and
removed at sometime, I do not
H.J. == H J Lu hjl.to...@gmail.com writes:
H.J. On Thu, May 12, 2011 at 7:56 AM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, May 11, 2011 at 12:28:18PM -0700, H.J. Lu wrote:
This patch restores the old behavior for Pmode. OK for trunk if there
are no regressions?
That is IMHO wrong,
Hi,
in my current efforts to enhance IPA-CP I propagate values in a single
pass over the call graph nodes in the topological order and only
resort to propagation until no change in SCCs.
In order to do that I need to quickly figure out whether two nodes on
both sides of a call graph edge are in
On 05/10/2011 06:55 PM, Joseph S. Myers wrote:
Unless the documentation is based on pre-existing GFDL-only documentation
in tm.texi.in, it's preferable for the documentation of a new hook to go
in the doc string in target.def and get to tm.texi that way, rather than
putting it directly in
On 05/12/2011 05:46 PM, Paul Richard Thomas wrote:
This patch fixes the problem in two steps:
(i) It reverts r162289; and
(ii) It adds the correct initialization for loop.reverse[] in
gfc_trans_assignment_1. This was implemented incorrectly in the fix
for PR24524 (in spite of the correct
While thinking about this issue some more it occurred to me that this
deferral can be problematic for constructors, since we use whether or
not a constructor is really constexpr to decide whether or not a class
is literal. The rule that seems to me to make the most sense is to say
that if we
Just for the record...
Tom What my patch does now is that it generally uses a signed integer type
Tom of the appropriate width for legacy DWARF values. Then, for mod, if
Tom the value's type is this special type, it converts it to an
Tom identically-sized unsigned type, and converts back after
2011/5/12 Richard Guenther richard.guent...@gmail.com:
On Thu, May 12, 2011 at 3:29 PM, Kai Tietz ktiet...@googlemail.com wrote:
While testing some other issues with C++'s __java_boolean type
occurred. So I adjusted check in test-cfg.c as you suggested.
Additionally due the fact that we are
Tom I have since changed this. For explicitly-typed values, I tried
Tom DW_OP_shr and DW_OP_shra identically. The type encodes the desired
Tom operation.
Jakub pinged me on irc to say that this is not what GCC actually emits.
So, I am going to change GDB to follow. That is, GDB will pick a
This patch to the Go frontend fixes a bug when taking the address of a
variable when the address of the variable does not escape the function.
When the address does escape the variable is moved onto the heap, to
avoid danling pointers. When the address does not escape this is not
necessary.
* Andreas Schwab wrote on Thu, May 12, 2011 at 11:22:12AM CEST:
This changes all uses of AC_ARG_(WITH|ENABLE) to use AS_HELP_STRING,
fixing a few quoting bugs on the way.
OK for trunk?
Yes, if you visually compared the diff of configure and of './configure
--help' before and after the
Hi,
This patch changes maybe_eh_return to bool. OK for trunk?
Thanks.
H.J.
---
2011-05-12 H.J. Lu hongjiu...@intel.com
* config/i386/i386.c (ix86_save_reg): Change maybe_eh_return
to bool.
(ix86_emit_restore_regs_using_mov): Likewise.
On Thu, May 12, 2011 at 11:19 AM, Kai Tietz ktiet...@googlemail.com wrote:
2011/5/12 Richard Guenther richard.guent...@gmail.com:
On Thu, May 12, 2011 at 3:29 PM, Kai Tietz ktiet...@googlemail.com wrote:
While testing some other issues with C++'s __java_boolean type
occurred. So I adjusted
[ Strong Typing Is For People With Weak Memories ]
The attached patch fixes the C++ (--disable-werror) bootstrap:
2011-05-12 Toon Moene t...@moene.org
* objc-next-runtime-abi-02.c (objc_build_internal_classname):
Add const qualifier to constant variable pointer declaration.
On May 12, 2011, at 1:33 PM, Toon Moene wrote:
[ Strong Typing Is For People With Weak Memories ]
:-)
2011-05-12 Toon Moene t...@moene.org
* objc-next-runtime-abi-02.c (objc_build_internal_classname):
Add const qualifier to constant variable pointer declaration.
Apply as
On Thu, May 12, 2011 at 9:55 PM, H.J. Lu hongjiu...@intel.com wrote:
Hi,
This patch changes maybe_eh_return to bool. OK for trunk?
Thanks.
H.J.
---
2011-05-12 H.J. Lu hongjiu...@intel.com
* config/i386/i386.c (ix86_save_reg): Change maybe_eh_return
to bool.
Hello, Rainer.
this is the patch I've checked in after it passed sparc-sun-solaris2.11
andi i386-pc-solaris2.11 bootstrap.
Rainer
2011-05-05 Rainer Orth r...@cebitec.uni-bielefeld.de
* config/sparc/sol2.h (ASM_OUTPUT_CALL): Use
targetm.asm_out.print_operand.
This patch cleans up x86 option handling to use .opt features where
possible for options with string arguments, instead of decoding those
strings in ix86_option_override_internal. (The only options with
string arguments that may be generated through target attributes are
-march=, -mtune= and
On Thu, May 12, 2011 at 2:09 PM, Uros Bizjak ubiz...@gmail.com wrote:
On Thu, May 12, 2011 at 9:55 PM, H.J. Lu hongjiu...@intel.com wrote:
Hi,
This patch changes maybe_eh_return to bool. OK for trunk?
Thanks.
H.J.
---
2011-05-12 H.J. Lu hongjiu...@intel.com
*
The patch is a correctness issue as the Fortran standard has those as
constraints - and the results of passing kind=4 strings is also odd ...
I used also the chance to resolve INQUIRE's decimal ...
Build and partially regtested on x86-64-linux.
OK for the trunk?
Tobias
PS: I will regtest
On Thu, May 12, 2011 at 2:09 PM, Uros Bizjak ubiz...@gmail.com wrote:
FWIW, the return type of ix86_save_reg (and ix86_hard_regno_mode_ok
and perhaps many others...) can also be changed to bool.
This patch changes the return type of ix86_save_reg to bool and return
value of
On Wed, May 11, 2011 at 1:19 AM, Diego Novillo dnovi...@google.com wrote:
Please add a ChangeLog entry. OK with that change.
Committed after adding the ChangeLog entry.
--
Cheers
Sandy
Is this patch ok for trunk?
Hi,
the patch is OK.
Sorry for delayed reply, I was very busy by the inliner/thunk bits and this code
is not really in my area of expertise, so it needed some consideration to make
sense
of it.
Honza
Allowing relative path in -fprofile-generate= is very useful
This patch add support to build Linux kernel with FDO.
Building Linux kernel with FDO needs both compiler change and
kernel changes. This part only contains the gcc changes. I'll
attach kernel changes to the reveiew thread for reference.
Source files gcov-io.c gcov-io.h and libgcov.c will be
56 matches
Mail list logo