On Fri, 8 Nov 2013, Cong Hou wrote:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59050
This is my bad. I forget to check the test result for gfortran. With
this patch the bug should be fixed (tested on x86-64).
Ok.
Btw, requirements are to bootstrap and test with all default
languages
On Sat, 9 Nov 2013, Tom de Vries wrote:
Richard,
This patch simplifies code in gimple_equal_p.
Bootstrapped and regtested on x86_64.
OK for trunk?
Ok.
Thanks,
Richard.
Thanks,
- Tom
2013-11-06 Tom de Vries t...@codesourcery.com
* tree-ssa-tail-merge.c
However, that brings up an couple interesting questions.
Let's say we find a NULL pointer which reaches a return statement in a
function which is marked as returns_nonnull. In that case there is no
dereference. Presumably for that kind of scenario we'll just keep the
builtin trap.
On Sat, 9 Nov 2013, Tom de Vries wrote:
Richard,
This patch factors out gimple_dont_merge_p from gimple_equal_p and
find_duplicate.
Bootstrapped and regtested on x86_64.
OK for trunk?
+static bool
+gimple_dont_merge_p (gimple stmt)
+{
+ switch (gimple_code (stmt))
+{
+case
On Sat, 9 Nov 2013, Tom de Vries wrote:
Richard,
This patch factors out gimple_operand_equal_value_p from gimple_equal_p.
Bootstrapped and regtested on x86_64.
OK for trunk?
Ok.
Thanks,
Richard.
Thanks,
- Tom
2013-11-06 Tom de Vries t...@codesourcery.com
*
On Sat, 9 Nov 2013, Tom de Vries wrote:
Richard,
Consider the test-case test.c:
...
int z;
int x;
void
f (int c, int d)
{
if (c)
z = 5;
else
{
if (d)
x = 4;
z = 5;
}
}
...
Atm, we don't tail-merge the 'z = 5' blocks, because
On Sun, Nov 10, 2013 at 04:27:00PM +0100, Marc Glisse wrote:
Hello,
I am posting this patch to get some feedback on the approach. The
goal is to replace malloc+free with a stack allocation (a decl
actually) when the size is a small constant.
Why constraint yourself to small sizes. Stack
Hi,
this is an old bug report from Jan, which was closed, then reopened by Tom at
some point, but the patch never got reviewed. The original submission is at:
http://gcc.gnu.org/ml/gcc-patches/2008-05/msg01857.html
Tested on x86_64-suse-linux, OK for the mainline?
2013-11-11 Jan
On Mon, Nov 11, 2013 at 11:08:14AM +0100, Ondřej Bílka wrote:
On Sun, Nov 10, 2013 at 04:27:00PM +0100, Marc Glisse wrote:
I am posting this patch to get some feedback on the approach. The
goal is to replace malloc+free with a stack allocation (a decl
actually) when the size is a small
Hello,
As it appeared that concerns about the speed of
location_get_source_line were as present as the need of just fixing
this bug, I have conflated the two concerns in a new attempt below,
trying to address the points you guys have raised during the previous
reviews.
The patch below introduces
2013-11-11 Joern Rennecke joern.renne...@embecosm.com
* config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
Index: config/arc/arc.h
===
--- config/arc/arc.h(revision 204665)
+++ config/arc/arc.h(working
On Fri, Nov 8, 2013 at 5:02 PM, Jeff Law l...@redhat.com wrote:
On 11/08/13 07:45, Steven Bosscher wrote:
On Fri, Nov 8, 2013 at 3:40 PM, Joern Rennecke
joern.renne...@embecosm.com wrote:
bootstrapped / regtested on i686-pc-linux-gnu.
Not a very elaborate description of the patch, eh? :-)
On Mon, Nov 11, 2013 at 11:19:05AM +0100, Jakub Jelinek wrote:
On Mon, Nov 11, 2013 at 11:08:14AM +0100, Ondřej Bílka wrote:
On Sun, Nov 10, 2013 at 04:27:00PM +0100, Marc Glisse wrote:
I am posting this patch to get some feedback on the approach. The
goal is to replace malloc+free with a
Hi,
since the switch to SSA form at -O0, the compiler generates wrong debug info
for something like:
void
foo (int i)
{
int j = 0;
i = 1;
j = j + 1; /* BREAK */
}
If you try to display the value of i after breaking in GDB, you don't get 1.
The reason is that there is no default def
Hi,
in Ada 2012 it is allowed to omit components of aggregates (the equivalent of
initializers/constructors); in this case, the contents of the corresponding
fields in the record become undefined. Now the gimplifier implements the C
semantics of clearing the missing components, so this patch
Hi,
this is something I forgot to submit right after submitting
http://gcc.gnu.org/ml/gcc-patches/2013-05/msg01906.html
We want to use the standard t-elf fragment on VxWorks as well.
Tested on ARM/VxWorks, OK for the mainline?
2013-11-11 Eric Botcazou ebotca...@adacore.com
*
In this case it's fold-all-builtins folding a strlen call with a
PHI foo, bar argument. IMHO not presenting RTL with such
non-sense is best achieved by not letting TER do constant propagation
(because it doesn't fold the result). We can never rule out such
stray non-propagated constants, so
On Mon, Nov 11, 2013 at 12:21 PM, Eric Botcazou ebotca...@adacore.com wrote:
In this case it's fold-all-builtins folding a strlen call with a
PHI foo, bar argument. IMHO not presenting RTL with such
non-sense is best achieved by not letting TER do constant propagation
(because it doesn't fold
On 8 November 2013 10:29, Bernhard Reutner-Fischer wrote:
On 04/11/2013 02:04 PM, Bernhard Reutner-Fischer wrote:
I would have expected that somebody would tell me that omitting ::tmpnam
violates 27.9.2 cstdio from the spec but noone yelled at me yet?
std::tmpnam, like std::gets, should be
On 11/11/13 11:11, Eric Botcazou wrote:
Hi,
this is something I forgot to submit right after submitting
http://gcc.gnu.org/ml/gcc-patches/2013-05/msg01906.html
We want to use the standard t-elf fragment on VxWorks as well.
Tested on ARM/VxWorks, OK for the mainline?
Ok, please apply.
On Mon, Nov 11, 2013 at 12:26:09PM +0100, Richard Biener wrote:
The question is whether you for example want to handle
a_2 = 1 + 0;
at RTL expansion time? I'd say it's better to have
I think we already handle that just fine, there are tons of various
simplify_gen_* calls during
On Fri, Nov 8, 2013 at 6:41 PM, Steven Bosscher stevenb@gmail.com wrote:
On Fri, Nov 8, 2013 at 6:20 PM, Steven Bosscher wrote:
On Wed, Oct 30, 2013 at 5:03 AM, Andrew Pinski wrote:
Here is what I applied in the end; Jeff told me just to remove the
testcase. I added the comment trying to
On Sat, Nov 9, 2013 at 12:18 AM, Marc Glisse marc.gli...@inria.fr wrote:
On Wed, 6 Nov 2013, Richard Biener wrote:
So the only thing that remains is the mem_ref_offset thing and yes, I
guess
I'd prefer to use double-ints because we deal with bit offsets in the end.
Here it is
On Sat, Nov 9, 2013 at 3:23 PM, Kenneth Zadeck zad...@naturalbridge.com wrote:
On 11/08/2013 05:30 AM, Richard Sandiford wrote:
From tree-vrp.c:
@@ -1893,6 +1884,10 @@ vrp_int_const_binop (enum tree_code code
/* If the singed operation wraps then int_const_binop has done
On Sun, Nov 10, 2013 at 6:11 PM, Eric Botcazou ebotca...@adacore.com wrote:
This is the first patch of what I think will be four to fix those few
places.
Bootstrappedtested on powerpc64-unknown-linux-gnu. Also built SH to be
sure.
OK for trunk?
The generic part is OK (modulo the
On Mon, Nov 11, 2013 at 1:39 AM, Trevor Saunders tsaund...@mozilla.com wrote:
On Fri, Nov 08, 2013 at 10:37:00AM +0100, Richard Biener wrote:
On Thu, Nov 7, 2013 at 5:00 PM, tsaund...@mozilla.com wrote:
From: Trevor Saunders tsaund...@mozilla.com
Hi,
This is the result of seeing what
On Thu, Oct 31, 2013 at 1:46 AM, Sandra Loosemore
san...@codesourcery.com wrote:
On 10/29/2013 02:51 AM, Bernd Edlinger wrote:
On Mon, 28 Oct 2013 21:29:24, Sandra Loosemore wrote:
On 10/28/2013 03:20 AM, Bernd Edlinger wrote:
I have attached an update to your patch, that should
a) fix
On Mon, Nov 11, 2013 at 8:29 AM, Bin.Cheng amker.ch...@gmail.com wrote:
On Fri, Nov 8, 2013 at 10:06 PM, Richard Biener
richard.guent...@gmail.com wrote:
On Fri, Nov 8, 2013 at 2:41 PM, bin.cheng bin.ch...@arm.com wrote:
Hi,
This patch refactors force_expr_to_var_cost and handles type
On Mon, Nov 11, 2013 at 4:11 AM, Jeff Law l...@redhat.com wrote:
On 11/10/13 05:34, Eric Botcazou wrote:
But I think that you cannot transform
foo ()
{
*0 = 1;
}
to __builtin_trap as you can catch the trap via an exception handler
in a caller of foo, no?
That is true. OK, I can
On Mon, Nov 11, 2013 at 11:10 AM, Eric Botcazou ebotca...@adacore.com wrote:
Hi,
this is an old bug report from Jan, which was closed, then reopened by Tom at
some point, but the patch never got reviewed. The original submission is at:
On 11/11/13 10:50, Richard Biener wrote:
On Sat, 9 Nov 2013, Tom de Vries wrote:
Bootstrapped and regtested on x86_64.
OK for trunk?
Comments inline
diff --git a/gcc/tree-ssa-tail-merge.c b/gcc/tree-ssa-tail-merge.c
index 98b5882..43516a7 100644
--- a/gcc/tree-ssa-tail-merge.c
+++
On Sun, Nov 10, 2013 at 4:27 PM, Marc Glisse marc.gli...@inria.fr wrote:
Hello,
I am posting this patch to get some feedback on the approach. The goal is to
replace malloc+free with a stack allocation (a decl actually) when the size
is a small constant.
For testing, I highjacked the leaf
On Mon, Nov 11, 2013 at 11:56 AM, Eric Botcazou ebotca...@adacore.com wrote:
Hi,
since the switch to SSA form at -O0, the compiler generates wrong debug info
for something like:
void
foo (int i)
{
int j = 0;
i = 1;
j = j + 1; /* BREAK */
}
If you try to display the value of i
On Mon, Nov 11, 2013 at 11:55 AM, Eric Botcazou ebotca...@adacore.com wrote:
Hi,
in Ada 2012 it is allowed to omit components of aggregates (the equivalent of
initializers/constructors); in this case, the contents of the corresponding
fields in the record become undefined. Now the gimplifier
On Mon, 11 Nov 2013, Tom de Vries wrote:
On 11/11/13 10:50, Richard Biener wrote:
On Sat, 9 Nov 2013, Tom de Vries wrote:
Bootstrapped and regtested on x86_64.
OK for trunk?
Comments inline
diff --git a/gcc/tree-ssa-tail-merge.c b/gcc/tree-ssa-tail-merge.c
index
This ensures that the standard timed mutexes use steady_clock for
relative timeouts and the same clock as pthread_mutex_timedlock for
absolute timeouts.
I'm not adding a new test as the only ones I could come up with
involve waiting for longer than is suitable in a testcase, and don't
reliably
Hello all,
I'm pinging the patch
http://gcc.gnu.org/ml/gcc-patches/2013-10/msg00056.html
##
Index: gcc/toplev.c
===
--- gcc/toplev.c(revision 204671)
+++ gcc/toplev.c(working copy)
@@ -1968,11 +1968,13
On Mon, Nov 11, 2013 at 8:36 AM, Basile Starynkevitch
bas...@starynkevitch.net wrote:
2013-11-11 Basile Starynkevitch bas...@starynkevitch.net
* toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
diagnostic_finish.
OK.
Diego.
On Thu, Nov 7, 2013 at 10:07 PM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/7 Jeff Law l...@redhat.com:
On 11/04/13 05:40, Richard Biener wrote:
Effectively the bounds are passed on the side.
Well, not exactly. I can see __bound_tmp.0_4 passed to access_and_store.
I'm referring
On Fri, Nov 8, 2013 at 11:03 AM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/8 Richard Biener richard.guent...@gmail.com:
On Thu, Nov 7, 2013 at 7:55 PM, Jeff Law l...@redhat.com wrote:
On 11/07/13 04:50, Ilya Enkovich wrote:
Hi,
Here is an updated patch version.
I think this
Hello,
On 05 Nov 16:05, Kirill Yukhin wrote:
Is it ok with that change?
Ping?
--
Thanks, K
Hi all,
My patch last Friday introduced a warning about reaching the end of a non-void
function which breaks bootstrap. On second thought, instead of breaking at the
end of the comparisons handling, we should just return instead.
Tested arm-none-eabi on qemu and checked the build log to make
On 11/11/13 13:48, Kyrill Tkachov wrote:
Hi all,
My patch last Friday introduced a warning about reaching the end of a non-void
function which breaks bootstrap. On second thought, instead of breaking at the
end of the comparisons handling, we should just return instead.
Tested arm-none-eabi on
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Thu, Nov 7, 2013 at 10:07 PM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/7 Jeff Law l...@redhat.com:
On 11/04/13 05:40, Richard Biener wrote:
Effectively the bounds are passed on the side.
Well, not exactly. I can see
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Fri, Nov 8, 2013 at 11:03 AM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/8 Richard Biener richard.guent...@gmail.com:
On Thu, Nov 7, 2013 at 7:55 PM, Jeff Law l...@redhat.com wrote:
On 11/07/13 04:50, Ilya Enkovich wrote:
Hi,
Hello,
Since a couple of days I am seeing failure on the tests above on my
Fedora system. The errors look like:
FAIL: c-c++-common/asan/memcmp-1.c -O0 output pattern test, is
=
==21832==ERROR: AddressSanitizer:
On Mon, Nov 11, 2013 at 03:01:53PM +0100, Dodji Seketeli wrote:
Since a couple of days I am seeing failure on the tests above on my
Fedora system. The errors look like:
FAIL: c-c++-common/asan/memcmp-1.c -O0 output pattern test, is
On Mon, Nov 11, 2013 at 3:00 PM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Fri, Nov 8, 2013 at 11:03 AM, Ilya Enkovich enkovich@gmail.com
wrote:
2013/11/8 Richard Biener richard.guent...@gmail.com:
On Thu, Nov 7, 2013 at 7:55 PM,
On Mon, Nov 11, 2013 at 11:19:21AM +0100, Dodji Seketeli wrote:
.../c-c++-common/cpp/warning-zero-in-literals-1.c | Bin 0 - 240 bytes
libcpp/include/line-map.h | 8 +
libcpp/line-map.c | 40 ++
8 files changed, 585 insertions(+),
On 11/11/2013 06:49 AM, Richard Biener wrote:
On Sat, Nov 9, 2013 at 3:23 PM, Kenneth Zadeck zad...@naturalbridge.com wrote:
On 11/08/2013 05:30 AM, Richard Sandiford wrote:
From tree-vrp.c:
@@ -1893,6 +1884,10 @@ vrp_int_const_binop (enum tree_code code
/* If the singed operation
Hello,
when implementing the new ABI for powerpc64le-linux, I ran into an assertion
failure in store_unaligned_arguments_into_pseudos:
gcc_assert (args[i].partial % UNITS_PER_WORD == 0);
This can happen in the new ABI since we pass homogeneous structures
consisting of soleley
Working to address a user question, I noticed that many of our pages use
the spelling of backend when http://gcc.gnu.org/codingconventions.html
suggest back end (noun) and back-end (adjective).
Joseph, if you confirm that back end it is (as a noun), I'll apply
the patch below.
Gerald
Index:
Hello,
this is another tweak to the middle-end to help support the new
powerpc64le-linux ABI.
In the new ABI, we make a distinction between functions that pass
all arguments solely in registers, and those that do not. Only when
calling one the latter type (or a varags routine) does the caller
Hello,
this fixes another issue related to CR unwinding, this time only on
64-bit little-endian systems.
The problem is linux-unwind.h:ppc_fallback_frame_state, which notes
that the kernel places the CR value in the low bits of a 64-bit
field in the signal handler frame; but the offset of that
Hello,
this patch fixes a corner case bug in unwinding CR values. The
underlying problem is as follows:
In order to save the CR in the prologue, two insns are necessary.
The first is a mfcr that loads the CR value into a GPR, and the
second is a store of the GPR to the stack. The back-end
The following patch removes now dead code from input_gimple_stmt
(which also could be quite slow). Type mismatches can only
occur at the decl level after the new tree merging code went
in and those are handed by wrapping all decls in MEM_REFs to
transparently have them view-converted.
Hello,
this patch refactors code to expand calls in preparation for adding support
for the new little-endian ABI.
Currently, the call expanders always generate the same RTX pattern, which
is matched by different insns patterns depending on circumstances (ABI,
local vs. global calls, ...). This
Hello,
another patch in preparation for the new ABI.
When calling a function in the absence of a prototype, we need to pass
arguments that are normally passed in floating-point or vector registers
also on the stack and/or in GPRs.
The code currently handling this in rs6000_function_arg is
The following patch is necessary to make DECL_BIT_FIELD_REPRESENTATIVE
useable in COMPONENT_REFs (which is the easiest way to get
some bitfield lowering).
An earlier patch passed bootstrapped and testing on
x86_64-unknown-linux-gnu (with bitfield lowering), re-testing
after minor changes.
Hello,
this is another patch to prepare for the new ABI. Since this will introduce
a new setting of DEFAULT_ABI, which will have to be treated like ABI_AIX in
many cases, it would be better if there were fewer explicit references to
ABI_AIX in the back-end. This patch eliminates a number of
Hello,
another patch in preparation for the new ABI.
USE_FP_FOR_ARG_P and USE_ALTIVEC_FOR_ARG_P take a TYPE argument
which they never use. Since passing a correct TYPE for the
homogeneous struct case would be a bit problematic, it seems
cleaner to just remove the unused argument.
No change in
Hello,
this is the final patch preparing for the new ABI.
The logic in rs6000_arg_partial_bytes is a bit complex, since it apparently
still contains remnants from the time where this routine was used even for
arguments that are returned both in GPRs and FPRs/VRs (*and* memory).
These days, all
Hi,
I apologize for long time for the review.
diff --git a/gcc/c-family/c-common.c b/gcc/c-family/c-common.c
index 8f7f5e5..f3ad003 100644
--- a/gcc/c-family/c-common.c
+++ b/gcc/c-family/c-common.c
@@ -343,6 +343,8 @@ static tree handle_tls_model_attribute (tree *, tree,
tree, int,
On Fri, Nov 08, 2013 at 05:11:39PM +, Joern Rennecke wrote:
On 8 November 2013 15:50, Steven Bosscher stevenb@gmail.com wrote:
Even with the gimple opts disabled, a const-const comparison would
normally be folded by the RTL expanders.
Well, in this spirit, attached is another way
Hello,
this patch adds another ELFv2 ABI feature: explicit tracking of CR fields
in DWARF CFI.
In the current ABI, DWARF CFI contains only a single record describing
the save location of the whole CR field. It is implicit that all (or
at least all call-clobbered) fields are present at that
Hello,
this patch adds the first major feature of the ELFv2 ABI: removal
of function descriptors.
In the current ABI, it is the responsibility of a caller to set
up the TOC pointer needed by the callee by loading the correct
value into r2. This typically happens in one of these ways:
- For a
On Mon, Nov 11, 2013 at 3:26 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 06:49 AM, Richard Biener wrote:
On Sat, Nov 9, 2013 at 3:23 PM, Kenneth Zadeck zad...@naturalbridge.com
wrote:
On 11/08/2013 05:30 AM, Richard Sandiford wrote:
From tree-vrp.c:
@@ -1893,6
Andi Kleen a...@firstfloor.org writes:
PING^3
Since it doesn't look like a generic solution for the
LTO options problem will appear this development cycle,
I would still like to pursue this option for 4.9.
This would help fixing parts of the linux kernel LTO
build.
Can someone please review
Hello,
this is the first patch in the series to add support for the ELFv2 ABI.
The ELFv2 ABI is the intended ABI for the new powerpc64le-linux port.
However, it is not inherently tied to the byte order; it it possible
in principle to use the ELFv2 ABI in big-endian mode too.
Therefore, it is
Hello,
this patch implements the new struct calling convention for ELFv2.
With the new ABI, so-called homogeneous aggregates, i.e. struct, arrays,
or unions that (recursively) contain only elements of the same floating-
point or vector type are passed as if those elements were passed as
separate
Hello,
this is the second part of reducing stack space consumption for the
ELFv2 ABI: the old ABI reserved six words in every stack frame for
various purposes, and two of those were basically unused: one word
for compiler use and one word for linker use.
Since neither the compiler nor the linker
Hello,
analogously to the previous patch handling homogeneous aggregates
as arguments, this patch handles such aggregates as return values.
In addition, the ELFv2 ABI returns any aggregate of up to 16 bytes
in size (that is not otherwise handled) in up to two GPRs.
In either case, the return
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Mon, Nov 11, 2013 at 3:00 PM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Fri, Nov 8, 2013 at 11:03 AM, Ilya Enkovich enkovich@gmail.com
wrote:
2013/11/8 Richard Biener
Hello,
the final area of changes implemented by the ELFv2 ABI is stack space
consumption. Reducing the miminum stack size requirements is important
for environments where stack space is restricted, e.g. Linux kernel code,
or where we have a large number of stacks (e.g. heavily multi-threaded
Hello,
this patch finally throws the switch and enables the ELFv2 ABI
by default on powerpc64le-linux.
Bye,
Ulrich
ChangeLog:
2013-11-11 Ulrich Weigand ulrich.weig...@de.ibm.com
* config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
Index: gcc/gcc/config/rs6000/sysv4le.h
+2013-10-29 Martin Liska marxin.li...@gmail.com
+ Jan Hubicka j...@suse.cz
+
+ * cgraph.c (dump_cgraph_node): Profile dump added.
+ * cgraph.h (struct cgraph_node): New time profile variable added.
+ * cgraphclones.c
2013-11-04 Rong Xu x...@google.com
* libgcc/libgcov.c: Delete as part of re-factoring.
* libgcc/libgcov-profiler.c (__gcov_interval_profiler): Moved from
libgcov.c
(__gcov_pow2_profiler): Ditto.
(__gcov_one_value_profiler_body): Ditto.
On 11/11/2013 09:42 AM, Richard Biener wrote:
On Mon, Nov 11, 2013 at 3:26 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 06:49 AM, Richard Biener wrote:
On Sat, Nov 9, 2013 at 3:23 PM, Kenneth Zadeck zad...@naturalbridge.com
wrote:
On 11/08/2013 05:30 AM, Richard Sandiford
On Mon, Nov 11, 2013 at 3:45 PM, Ilya Enkovich enkovich@gmail.com wrote:
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Mon, Nov 11, 2013 at 3:00 PM, Ilya Enkovich enkovich@gmail.com
wrote:
2013/11/11 Richard Biener richard.guent...@gmail.com:
On Fri, Nov 8, 2013 at 11:03
Hello!
Building latest gcc+go on CentOS 5.10 breaks with following build failure:
../../../gcc-svn/trunk/libgo/go/net/fd_unix.go:414:72: error:
reference to undefined identifier ‘syscall.F_DUPFD_CLOEXEC’
r0, _, e1 := syscall.Syscall(syscall.SYS_FCNTL, uintptr(fd),
syscall.F_DUPFD_CLOEXEC, 0)
Here is the patch that includes profile-tool.
Profile-tool now has two functions: merge and rewrite. I'll add diff later.
Compiler is tested with spec2006 and profiledbootstrap.
profile-tool is tested with spec2006 profiles.
Hi,
it would be nice if you could elaborate bit more on the tool
2013-11-08 Teresa Johnson tejohn...@google.com
Jan Hubicka j...@suse.cz
* predict.c (drop_profile): New function.
(handle_missing_profiles): Ditto.
(counts_to_freqs): Don't overwrite estimated frequencies
when function has no profile counts.
On Mon, Nov 11, 2013 at 4:04 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 09:42 AM, Richard Biener wrote:
On Mon, Nov 11, 2013 at 3:26 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 06:49 AM, Richard Biener wrote:
On Sat, Nov 9, 2013 at 3:23 PM,
On 11/11/2013 10:04 AM, Kenneth Zadeck wrote:
On 11/11/2013 09:42 AM, Richard Biener wrote:
On Mon, Nov 11, 2013 at 3:26 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 06:49 AM, Richard Biener wrote:
On Sat, Nov 9, 2013 at 3:23 PM, Kenneth Zadeck
zad...@naturalbridge.com
On 11/11/2013 10:29 AM, Richard Biener wrote:
On Mon, Nov 11, 2013 at 4:04 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 09:42 AM, Richard Biener wrote:
On Mon, Nov 11, 2013 at 3:26 PM, Kenneth Zadeck
zad...@naturalbridge.com wrote:
On 11/11/2013 06:49 AM, Richard Biener
Hi!
Dunno how I've managed to break the branch on Friday, anyway, here is an
obvious fix, committed to the branch.
2013-11-11 Jakub Jelinek ja...@redhat.com
* tree-vect-data-refs.c (vect_analyze_data_refs): Check loop-safelen
rather than loop-simdlen.
*
On Mon, Nov 11, 2013 at 7:23 AM, Jan Hubicka hubi...@ucw.cz wrote:
2013-11-08 Teresa Johnson tejohn...@google.com
Jan Hubicka j...@suse.cz
* predict.c (drop_profile): New function.
(handle_missing_profiles): Ditto.
(counts_to_freqs): Don't overwrite
On 11/11/13 05:16, Richard Biener wrote:
On Mon, Nov 11, 2013 at 4:11 AM, Jeff Law l...@redhat.com wrote:
On 11/10/13 05:34, Eric Botcazou wrote:
But I think that you cannot transform
foo ()
{
*0 = 1;
}
to __builtin_trap as you can catch the trap via an exception handler
in a caller
On Mon, Nov 11, 2013 at 12:58:36PM +0100, Richard Biener wrote:
On Mon, Nov 11, 2013 at 1:39 AM, Trevor Saunders tsaund...@mozilla.com
wrote:
On Fri, Nov 08, 2013 at 10:37:00AM +0100, Richard Biener wrote:
On Thu, Nov 7, 2013 at 5:00 PM, tsaund...@mozilla.com wrote:
From: Trevor
Jakub Jelinek ja...@redhat.com writes:
-OBJS-libcommon = diagnostic.o diagnostic-color.o pretty-print.o intl.o
input.o version.o
+OBJS-libcommon = diagnostic.o diagnostic-color.o pretty-print.o intl.o
vec.o input.o version.o
Too long line?
Fixed in my local copy of the patch, thanks.
Jakub Jelinek ja...@redhat.com writes:
On Mon, Nov 11, 2013 at 03:01:53PM +0100, Dodji Seketeli wrote:
Since a couple of days I am seeing failure on the tests above on my
Fedora system. The errors look like:
FAIL: c-c++-common/asan/memcmp-1.c -O0 output pattern test, is
I have a warning like that already in drop_profile(). Is that
I think it should be warning (or silent) for COMDAT and error/note for
other functions (depending on flag_profile_correction).
I guess drop_profile is better place for it indeed.
sufficient? Also, Steven Bosscher suggested putting
On 11/11/13 02:33, Eric Botcazou wrote:
However, that brings up an couple interesting questions.
Let's say we find a NULL pointer which reaches a return statement in a
function which is marked as returns_nonnull. In that case there is no
dereference. Presumably for that kind of scenario we'll
On Mon, Nov 11, 2013 at 09:24:27AM -0700, Jeff Law wrote:
On 11/11/13 02:33, Eric Botcazou wrote:
However, that brings up an couple interesting questions.
Let's say we find a NULL pointer which reaches a return statement in a
function which is marked as returns_nonnull. In that case there
On Mon, Nov 11, 2013 at 8:27 AM, Jakub Jelinek ja...@redhat.com wrote:
On Mon, Nov 11, 2013 at 09:24:27AM -0700, Jeff Law wrote:
On 11/11/13 02:33, Eric Botcazou wrote:
However, that brings up an couple interesting questions.
Let's say we find a NULL pointer which reaches a return statement
On Mon, Nov 4, 2013 at 11:06 AM, David Edelsohn dje@gmail.com wrote:
Balaji,
I am seeing a large number of libcilkrts failures on AIX. These all
are of the form:
Executing on host: /tmp/20131103/gcc/xgcc -B/tmp/20131103/gcc/
/nasfarm/edelsohn
Ok. Can you update doc/generic.texi?
Thanks, done (it was still talking about TREE_LIST).
--
Eric Botcazou
On Mon, Nov 11, 2013 at 8:22 AM, Jan Hubicka hubi...@ucw.cz wrote:
I have a warning like that already in drop_profile(). Is that
I think it should be warning (or silent) for COMDAT and error/note for
other functions (depending on flag_profile_correction).
I guess drop_profile is better place
Due to the different interfaces of int_size_in_bytes and
simple_type_size_in_bits (and 'size' in add_byte_size_attribute being
unsigned and not [unsigned] HWI) it would be cleaner to
add an early return after the call to int_size_in_bytes if its
return value is -1 (and make sure the return
Hi Yufeng,
The idea is a good one but I don't like your implementation of adding an
extra expression parameter to look at on the find_basis_for_candidate
lookup. This goes against the design of the pass and may not be
sufficiently general (might there be situations where a third possible
basis
1 - 100 of 163 matches
Mail list logo