Hello,
In the test case for PR54146, build_insn_chain spends almost all its
time in this loop:
FOR_EACH_BB_REVERSE (bb)
{
bitmap_iterator bi;
rtx insn;
CLEAR_REG_SET (live_relevant_regs);
-- memset (live_subregs_used, 0, max_regno * sizeof (int));
The test case has
Hello,
In the test case for PR54146, compute_global_livein allocates/frees a
worklist for 400,000 basic blocks on each invocation. And it's called
a lot, for rewrite_into_loop_closed_ssa. But the maximum number of
basic blocks ever on the work list was only ~6500. So the work list
can be much
On Mon, Aug 6, 2012 at 10:44 PM, Dimitrios Apostolou ji...@gmx.net wrote:
What else is missing to make this patch appropriate for libiberty? Should I
change the prolog in strnlen.c, since I only copied it intact from gnulib?
We generally try to avoid straight GPL source code without runtime
On Tue, Aug 7, 2012 at 2:35 AM, Lawrence Crowl cr...@google.com wrote:
Convert double_int from a struct with function into a class with
operators and methods.
This patch adds the methods and operators. In general functions of
the form double_int_whatever become member functions whatever or,
On Tue, Aug 7, 2012 at 8:24 AM, Steven Bosscher stevenb@gmail.com wrote:
Hello,
In the test case for PR54146, compute_global_livein allocates/frees a
worklist for 400,000 basic blocks on each invocation. And it's called
a lot, for rewrite_into_loop_closed_ssa. But the maximum number of
On Tue, Aug 7, 2012 at 10:31 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Tue, Aug 7, 2012 at 8:24 AM, Steven Bosscher stevenb@gmail.com wrote:
Hello,
In the test case for PR54146, compute_global_livein allocates/frees a
worklist for 400,000 basic blocks on each invocation.
On Tue, Aug 7, 2012 at 11:03 AM, Steven Bosscher stevenb@gmail.com wrote:
On Tue, Aug 7, 2012 at 10:31 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Tue, Aug 7, 2012 at 8:24 AM, Steven Bosscher stevenb@gmail.com
wrote:
Hello,
In the test case for PR54146,
On Mon, 6 Aug 2012, Ian Lance Taylor wrote:
On Mon, Aug 6, 2012 at 10:44 PM, Dimitrios Apostolou ji...@gmx.net wrote:
What else is missing to make this patch appropriate for libiberty? Should I
change the prolog in strnlen.c, since I only copied it intact from gnulib?
We generally try to
On Tue, Aug 7, 2012 at 11:22 AM, Richard Guenther
richard.guent...@gmail.com wrote:
Another optimization would be to do
@@ -440,13 +442,13 @@ compute_global_livein (bitmap livein ATT
! bitmap_bit_p (def_blocks, pred_index)
bitmap_set_bit (livein, pred_index))
On Tue, Aug 7, 2012 at 11:45 AM, Steven Bosscher stevenb@gmail.com wrote:
On Tue, Aug 7, 2012 at 11:22 AM, Richard Guenther
richard.guent...@gmail.com wrote:
Another optimization would be to do
@@ -440,13 +442,13 @@ compute_global_livein (bitmap livein ATT
! bitmap_bit_p
On Tue, Aug 7, 2012 at 11:52 AM, Richard Guenther
richard.guent...@gmail.com wrote:
So I wonder why simply looping over all SSA defs in a loop body and checking
whether a use is outside of it is not enough to compute this information ...
(yes, we might end up creating too many loop closed PHIs,
ping?
re-attaching the updated patch with the fixed comment in genmultilib.
Matthias
On 08.07.2012 20:48, Matthias Klose wrote:
Please find attached v2 of the patch updated for trunk 20120706, x86 only,
tested on
x86-linux-gnu, KFreeBSD and the Hurd.
I left in the comment about the
It's the only LHS setter that does not so and that results in the
asymmetry of create_phi_node needing a SSA_NAME_DEF_STMT adjustment
if you fed it an SSA name result as opposed to a VAR_DECL result
in which case it will be already adjusted.
Bootstrapped and tested on x86_64-unknown-linux-gnu,
On Tue, Aug 7, 2012 at 11:58 AM, Steven Bosscher stevenb@gmail.com wrote:
On Tue, Aug 7, 2012 at 11:52 AM, Richard Guenther
richard.guent...@gmail.com wrote:
So I wonder why simply looping over all SSA defs in a loop body and checking
whether a use is outside of it is not enough to compute
This makes sure clobbers of SSA names are never retained in the IL.
First by verifying that in the gimple verifier, second by removing
them when we go into SSA form. Third by catching those we have to
get rid of when un-nesting - this avoids stuff like
finally
{
D.1234 = CLOBBER;
On 07/08/12 05:43, Ian Lance Taylor wrote:
This patch to libgo, from Shenghou Ma, adds support for NumCPU on
additional platforms: Solaris, Irix, *BSD, Darwin. Bootstrapped and ran
Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline.
Ian
Wouldn't it be more useful on Linux
On Tue, Aug 7, 2012 at 12:07 AM, Dehao Chen de...@google.com wrote:
Ping...
Richard, could you shed some lights on this?
Thanks,
Dehao
On Mon, Jul 30, 2012 at 8:29 PM, Dehao Chen de...@google.com wrote:
Hi,
This patch fixes the source location for automatically generated calls
to
On Thu, Aug 2, 2012 at 7:37 AM, Simon Baldwin sim...@google.com wrote:
Omit TARGET_LIB_PATH from RPATH_ENVVAR in HOST_EXPORTS on bootstrap builds.
Discussion and rationale at: http://gcc.gnu.org/ml/gcc/2012-06/msg00314.html
For google/main. Tested for bootstrap and regression.
2012-08-02
On Tue, Aug 7, 2012 at 2:30 AM, Hans-Peter Nilsson h...@bitrange.com wrote:
Just don't forget that libiberty isn't a target library anymore.
To wit, the (GCC) run-time exception is moot for that code, AFAIK.
Maybe enough reason to abandon that rule so its code can be
truly and freely shared
This adds copy_ssa_name similar to duplicate_ssa_name but not copying
any annotations (thus, just make a new SSA name that looks similar
to a template). This avoids a bunch of SSA_NAME_VAR uses which will
no longer work to create SSA names off.
The 2nd part is stuff I came along when working on
On Tue, Aug 7, 2012 at 3:31 AM, Matthias Klose d...@ubuntu.com wrote:
ping?
On 08.07.2012 20:48, Matthias Klose wrote:
Please find attached v2 of the patch updated for trunk 20120706, x86 only,
tested on
x86-linux-gnu, KFreeBSD and the Hurd.
+[case ${withval} in
+yes|no|auto-detect)
Richard Guenther wrote:
On Fri, Jul 27, 2012 at 5:24 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
ChangeLog:
* target.def (vector_alignment): New target hook.
* doc/tm.texi.in (TARGET_VECTOR_ALIGNMENT): Document new hook.
* doc/tm.texi: Regenerate.
*
On Tue, Aug 7, 2012 at 4:56 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
Richard Guenther wrote:
On Fri, Jul 27, 2012 at 5:24 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
ChangeLog:
* target.def (vector_alignment): New target hook.
* doc/tm.texi.in
On Tue, 7 Aug 2012, Ian Lance Taylor wrote:
On Tue, Aug 7, 2012 at 2:30 AM, Hans-Peter Nilsson h...@bitrange.com wrote:
Just don't forget that libiberty isn't a target library anymore.
To wit, the (GCC) run-time exception is moot for that code, AFAIK.
Maybe enough reason to abandon that
On 07/08/12 16:04, Richard Guenther wrote:
On Tue, Aug 7, 2012 at 4:56 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
Richard Guenther wrote:
On Fri, Jul 27, 2012 at 5:24 PM, Ulrich Weigand uweig...@de.ibm.com wrote:
ChangeLog:
* target.def (vector_alignment): New target hook.
(Note that while the patch contains changes to common code, those
should be no-ops for all targets that do not implement the new hook.)
I'll defer the decision to the target maintainers.
I'd rather have this consistent across all maintained release branches
today than to leave this for an
On 08/06/2012 05:45 PM, Sandra Loosemore wrote:
OK to commit the patch as originally posted?
Yes.
I mis-read skimming the part of that function about relocations.
r~
On Tue, Aug 7, 2012 at 5:21 AM, Richard Earnshaw rearn...@arm.com wrote:
Wouldn't it be more useful on Linux to check the task's affinity
settings? Then when a task is locked to a limited set of cores it won't
overload those cores with threads.
Good question. I'm not sure. This patch does
Richard Henderson wrote:
[ This ought to be exactly the patch you bootstrapped. It does
not include the SEQ follow-up. ]
Split out s390_two_part_insv from s390_expand_cs_hqi to try
harder to use bit insertion instructions in the CAS loop.
Reorg s390_expand_insv to aid that. Try RISBG
Richard Henderson wrote:
On 08/06/2012 11:34 AM, Ulrich Weigand wrote:
There is one particular inefficiency I have noticed. This code:
if (!__atomic_compare_exchange_n (v, expected, max, 0 , 0, 0))
abort ();
from atomic-compare-exchange-3.c gets compiled into:
l
On 08/07/2012 02:22 AM, Steven Bosscher wrote:
Hello,
In the test case for PR54146, build_insn_chain spends almost all its
time in this loop:
FOR_EACH_BB_REVERSE (bb)
{
bitmap_iterator bi;
rtx insn;
CLEAR_REG_SET (live_relevant_regs);
-- memset
On Aug 6, 2012, at 5:35 PM, Lawrence Crowl wrote:
Convert double_int from a struct with function into a class with
operators and methods.
We have a wide_int class that replaces this class. :-( It would have been
better to just convert it. Do you guys have a timeframe for the cxx-conversion
On 12-08-07 13:22 , Mike Stump wrote:
On Aug 6, 2012, at 5:35 PM, Lawrence Crowl wrote:
Convert double_int from a struct with function into a class with
operators and methods.
We have a wide_int class that replaces this class. :-( It would
have been better to just convert it. Do you guys
Hi,
this is an update on C++/54191, where in a number of cases, having to do
with inaccessible bases, we don't handle correctly access control under
SFINAE.
The attached draft patch p fixes a number of tests (and passes
regression testing), where we are currently emitting inaccessible base
On 8/7/12, Richard Guenther richard.guent...@gmail.com wrote:
On Tue, Aug 7, 2012 at 2:35 AM, Lawrence Crowl cr...@google.com wrote:
Convert double_int from a struct with function into a class with
operators and methods.
This patch adds the methods and operators. In general functions of
On 8/7/12, Mike Stump mikest...@comcast.net wrote:
On Aug 6, 2012, at 5:35 PM, Lawrence Crowl wrote:
Convert double_int from a struct with function into a class with
operators and methods.
We have a wide_int class that replaces this class. :-(
Really? Where? I don't see neither definition
Hello,
This patch fixes PR54150. The patch was approved by Richi in Bugzilla,
but I won't commit this for a few days so that others can speak up if
they don't agree.
Ciao!
Steven
PR54150.diff
Description: Binary data
On Tue, Aug 7, 2012 at 1:54 PM, Steven Bosscher stevenb@gmail.com wrote:
Hello,
This patch fixes PR54150. The patch was approved by Richi in Bugzilla,
but I won't commit this for a few days so that others can speak up if
they don't agree.
Ciao!
Steven
OK. Thanks for the cleanup!
--
On Sun, Aug 5, 2012 at 12:47 AM, Richard Sandiford
rdsandif...@googlemail.com wrote:
For the record, I can't approve this, but...
H.J. Lu hjl.to...@gmail.com writes:
i386,md has
(define_expand extzv
[(set (match_operand:SI 0 register_operand)
(zero_extract:SI (match_operand 1
On Aug 7, 2012, at 11:38 AM, Lawrence Crowl wrote:
Hm. There seems to be significant opinion that there should not be any
implicit conversions. I am okay with operations as above, but would like
to hear the opinions of others.
If there is an agreed upon and expected semantic, having them are
On Tue, Jun 12, 2012 at 10:00 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Thu, Nov 3, 2011 at 8:50 PM, Jason Merrill ja...@redhat.com wrote:
12.2 states that a temporary bound to a reference lives as long as the
reference itself. We have done that for reference variables, but not in
other
Looks good to me for google-4_7 branches.
-Rong
On Mon, Aug 6, 2012 at 5:03 PM, Cary Coutant ccout...@google.com wrote:
This patch is for the google/gcc-4_7 branch. It backports the following
patch from trunk at r190190:
http://gcc.gnu.org/ml/gcc-patches/2012-08/msg00321.html
GCC
Hello Everyone,
This patch is for the Cilk Plus branch affecting mainly the C++
compiler. This patch will store the initial value of a loop correctly and
remove the unnecessary static chain usage for some cases of Cilk_for.
Thanks,
Balaji V. Iyer.
Index: parser.c
Sorry, I was in the wrong directory when I was creating the patch. Here is the
fixed patch.
Thanks,
Balaji V. Iyer.
-Original Message-
From: Iyer, Balaji V
Sent: Tuesday, August 07, 2012 4:18 PM
To: 'gcc-patches@gcc.gnu.org'
Subject: [PATCH][Cilkplus] Remove unwanted static chain.
On Aug 7, 2012, at 11:42 AM, Lawrence Crowl wrote:
On 8/7/12, Mike Stump mikest...@comcast.net wrote:
On Aug 6, 2012, at 5:35 PM, Lawrence Crowl wrote:
Convert double_int from a struct with function into a class with
operators and methods.
We have a wide_int class that replaces this class.
This patch fixes a group of bugs that were causing link errors on
hard-float MIPS16 code built with a mips-linux-gnu toolchain. This is
Mark Mitchell's analysis of the original problem:
The MIPS16
instruction set cannot directly access hard-float registers, so helper
functions in libgcc are
Hello!
We can exit early from var_lowpart for matched modes. This avoids
situation, where var_lowpart tries to change
(gdb) p debug_rtx (incoming)
(concat:SC (reg:SF 48 $f16 [ x ])
(reg:SF 49 $f17 [ x+4 ]))
$1 = void
to SCmode, and returning NULL_RTX, since the RTX is neither REG, neither
I should mention that with my patch .ascii is used more aggresively than
before, so if a string is longer than ELF_STRING_LIMIT it will be written
as .ascii all of it, while in the past it would use .string for the
string's tail. Example diff to original behaviour:
.LASF15458:
-
On Sat, Aug 4, 2012 at 10:34 AM, Paolo Carlini paolo.carl...@oracle.com wrote:
First, I think we should add libstdc++ in CC.
Thus, I would recommend people working in this area to begin with
unordered_map, because in that case emplace is already available, assuming
that's really the point
Hi,
(adding in CC Francois too)
On 08/07/2012 11:43 PM, Ollie Wild wrote:
On Sat, Aug 4, 2012 at 10:34 AM, Paolo Carlini paolo.carl...@oracle.com wrote:
First, I think we should add libstdc++ in CC.
Thus, I would recommend people working in this area to begin with
unordered_map, because in
On 08/07/2012 10:02 AM, Ulrich Weigand wrote:
The following patch changes the builtin expander to pass a MEM oldval
as-is to the back-end expander, so that the back-end can move the
store to before the CC operation. With that patch I'm also seeing
all the IPMs disappear.
...
What do you
On Tue, Aug 7, 2012 at 2:24 PM, Dimitrios Apostolou ji...@gmx.net wrote:
BTW I can't find why ELF_STRING_LIMIT is only 256, it seems GAS supports
arbitrary lengths. I'd have to change my code if we ever set it too high (or
even unlimited) since I allocate the buffer on the stack.
See the
Removes references in libgcov.c to functions and structs removed from
pmu-profile.c
For google/main. Tested with crosstools and bootstrap.
2012-08-07 Chris Manghane cm...@google.com
* libgcc/pmu-profile.c
(enum pmu_tool_type): Remove pfmon-specific functions/structs.
On 08/06/2012 05:35 PM, Lawrence Crowl wrote:
+inline double_int
+double_int::operator ++ ()
+{
+ *this + double_int_one;
+ return *this;
+}
+
+inline double_int
+double_int::operator -- ()
+{
+ *this - double_int_one;
+ return *this;
+}
Surely unused results there?
r~
Pretty minor change, as per PR. This version seems more appropriate for
templatized types.
I'll wait a bit before putting this on 4.7_branch
-benjamin
tested x86/linux
2012-08-07 Benjamin Kosnik b...@redhat.com
PR libstdc++/54005
* include/std/atomic: Use
On Tue, 7 Aug 2012, Ian Lance Taylor wrote:
On Tue, Aug 7, 2012 at 2:24 PM, Dimitrios Apostolou ji...@gmx.net wrote:
BTW I can't find why ELF_STRING_LIMIT is only 256, it seems GAS supports
arbitrary lengths. I'd have to change my code if we ever set it too high (or
even unlimited) since I
On Tue, Aug 7, 2012 at 4:27 PM, Dimitrios Apostolou ji...@gmx.net wrote:
On Tue, 7 Aug 2012, Ian Lance Taylor wrote:
On Tue, Aug 7, 2012 at 2:24 PM, Dimitrios Apostolou ji...@gmx.net wrote:
BTW I can't find why ELF_STRING_LIMIT is only 256, it seems GAS supports
arbitrary lengths. I'd have
With --std=c++11, a template parameter can refer to a local type defined
within a function. Because that local type doesn't qualify for its own
type unit, we copy it as an unworthy type into the type unit that refers
to it, but we copy too much, leading to a comdat type unit that contains a
DIE
Hi,
Here is the patch with some obvious fixes. If there are no objections,
could anyone please check it in?
Changelog entry:
2012-08-08 Michael Zolotukhin michael.v.zolotuk...@intel.com
* common/config/i386/i386-common.c (OPTION_MASK_ISA_ADX_SET): New.
59 matches
Mail list logo
