On Mon, 28 Nov 2016, Richard Biener wrote:
>
> The following fixes an optimization regression noted in PR78546 where
> moving patterns from forwprop to match.pd missed one variant.
>
> Bootstrap & regtest running on x86_64-unknown-linux-gnu.
The following is what I have applied (note the fix, t
On Mon, 28 Nov 2016, Jeff Law wrote:
> On 11/15/2016 07:58 AM, Richard Biener wrote:
> >
> > Appearantly for some unknown reason we refuse to inline anything into
> > functions calling cilk_spawn. That breaks fortified headers and
> > all other always-inline function calls (intrinsics come to my
While rewriting PHI-OPT to use match and simplify infrastructure, I
ran into a problem where VRP pass would create a EQ_EXPR which has a
non boolean type inside the VRP pass. This currently works on
accident as it seems we don't check the type of the argument of
COND_EXPR gimple to be boolean type
Hi,
add common defines _REENTRANT and _POSIX_SOURCE for bfin.
Patch is used in Buildroot for a while to fix issues compiling
some software.
See here, why this should be always enabled:
https://lists.gnu.org/archive/html/autoconf-archive-maintainers/2016-06/msg1.html
2016-11-29 Waldemar B
On Tue, Nov 15, 2016 at 09:06:23AM -0500, David Edelsohn wrote:
> > * config/rs6000/rs6000.c (rs6000_emit_stack_reset): Emit the
> > stack_restore_tie insn instead of stack_tie, for the SVR4 and
> > SPE ABIs.
> > * config/rs6000/rs6000.md (stack_restore_tie): New def
On 11/28/2016 06:35 PM, David Edelsohn wrote:
Martin,
I am seeing a number of new failures with the testcases on AIX.
FAIL: gcc.dg/tree-ssa/builtin-sprintf-warn-1.c (test for excess errors)
Excess errors:
/nasfarm/edelsohn/src/src/gcc/testsuite/gcc.dg/tree-ssa/builtin-sprintf-warn-1.c:1485:3:
PR 78512 - r242674 miscompiles Linux kernel observes that the Linux
kernel fails to boot as a result of enabling the -fprintf-return-value
optimization in GCC. This is likely because the kernel has its own
sprintf with a large set of extensions to the %p directive that
conflict with the optimizat
On Mon, Nov 28, 2016 at 02:44:44PM -0700, Jeff Law wrote:
> On 11/23/2016 04:55 PM, Segher Boessenkool wrote:
> >Combine can turn a conditional trap into an unconditional trap. If it
> >does that it should make the code after it unreachable (an unconditional
> >trap should be the last insn in its
On Mon, Nov 28, 2016 at 01:56:10PM -0700, Jeff Law wrote:
> On 11/26/2016 04:19 PM, Segher Boessenkool wrote:
> >simplify_truncation changes the truncation of many operations into
> >the operation on the truncation. This patch makes this code also
> >handle extracts.
> I'm OK with this for the tr
Martin,
I am seeing a number of new failures with the testcases on AIX.
FAIL: gcc.dg/tree-ssa/builtin-sprintf-warn-1.c (test for excess errors)
Excess errors:
/nasfarm/edelsohn/src/src/gcc/testsuite/gcc.dg/tree-ssa/builtin-sprintf-warn-1.c:1485:3:
warning: specified destination size 2147483647 i
On Sun, 27 Nov 2016, Martin Sebor wrote:
> Finally, the patch also tightens up the constraint on the upper bound
> of bounded functions like snprintf to be INT_MAX. The functions cannot
> produce output in excess of INT_MAX + 1 bytes and some implementations
> (e.g., Solaris) fail with EINVAL whe
On 11/15/2016 07:58 AM, Richard Biener wrote:
Appearantly for some unknown reason we refuse to inline anything into
functions calling cilk_spawn. That breaks fortified headers and
all other always-inline function calls (intrinsics come to my mind as
well).
Bootstrapped and tested on x86_64-unk
On 11/15/2016 02:52 PM, David Malcolm wrote:
This patch addresses various bugs in the spellcheck code in which
the goal string somehow makes it into the candidate list.
The goal string will always have an edit distance of 0 to itself, and
thus is the "closest" string to the goal, but offering it
On 11/10/2016 03:43 AM, Richard Biener wrote:
On Thu, Nov 10, 2016 at 12:05 AM, Segher Boessenkool
wrote:
On Wed, Nov 09, 2016 at 11:29:45PM +0100, Marc Glisse wrote:
This patch makes RTL simplify transform (xor (and (xor A B) C) B) back
to (ior (and A C) (and B ~C)) for constant C (and simila
No functional changes.
2016-11-28 Uros Bizjak
* config/i386/i386.md (*and_1): Merge insn pattern from
*andsi_1 and *andhi_1 using SWI24 mode iterator. Use multi-line
output template string.
(*anddi_1): Use multi-line output template string.
(*andqi_1): Ditto.
Bootstrapped
On 11/09/2016 02:46 PM, Segher Boessenkool wrote:
This patch changes spread_components to use a simpler algorithm that
puts prologue components as early as possible, and epilogue components
as late as possible. This allows better scheduling, and also saves a
bit of code size. The blocks that ru
On 11/28/2016 04:51 PM, Jeff Law wrote:
On 11/25/2016 07:02 AM, Andreas Schwab wrote:
On Nov 23 2016, Christophe Lyon wrote:
Since you committed this patch (r242733), I've noticed a regression:
gcc.dg/uninit-pred-6_c.c warning (test for warnings, line 43)
now fails on some arm targets, for
On 11/06/2016 10:17 AM, Iain Sandoe wrote:
Hi Folks,
This might make some configuration modes easier on modern Darwin (although
hitting the problem used to be an unusual circumstance, it will likely become
more common if we need to use build-sysroots - e.g. targeting a minimum OS X
version of
On 11/05/2016 02:46 PM, cheng...@emindsoft.com.cn wrote:
From: Chen Gang
r10 may also be as parameter stack pointer for the nested function, so
need save it before call mcount.
Also clean up code: use '!' instead of "== 0" for checking
static_chain_decl and compute_total_frame_size.
2016-11-0
On 11/28/2016 12:35 PM, Jakub Jelinek wrote:
Hi!
On the following case we ICE in DF checking, because
reload_combine_recognize_pattern starts with apply_change_group
that in the end on success calls df_insn_rescan, but if it is successful,
we then remove_reg_equal_equiv_notes and nothing rescans
On 11/24/2016 03:58 AM, Pitchumani Sivanupandi wrote:
GCC inlines small functions if the code size after expansion is not
excedded.
For test case (inline.c, avr-gcc -Os -S inline.c) code size become
higher if
'func2' is inlined. It happens because the CONVERT_EXPR/ NOP_EXPR are
considered
as zero
On Mon, Nov 28, 2016 at 03:00:16PM -0700, Jeff Law wrote:
> On 11/28/2016 02:55 PM, Segher Boessenkool wrote:
> >On Mon, Nov 28, 2016 at 01:56:10PM -0700, Jeff Law wrote:
> >>On 11/26/2016 04:19 PM, Segher Boessenkool wrote:
> >>>simplify_truncation changes the truncation of many operations into
>
On 11/25/2016 01:29 AM, Dominik Vogt wrote:
On Thu, Nov 24, 2016 at 08:33:32AM -0700, Jeff Law wrote:
On 11/24/2016 02:59 AM, Dominik Vogt wrote:
On Wed, Nov 23, 2016 at 01:22:31PM -0700, Jeff Law wrote:
PR target/77822
* system.h (SIZE_POS_IN_RANGE): New.
OK. Though system.h
On 11/24/2016 05:11 AM, Richard Biener wrote:
where CST is unsigned implying that the lower bound of the offset
is the greater of CST and MIN. For instance, in the following it
determines that bos(p, 0) is 4 (and if the 3 were greater than 7
and overflowed the addition the result would be zer
On Mon, Nov 28, 2016 at 05:09:50PM -0500, David Malcolm wrote:
> Segher: it looks like the "PR78432" in the Subject line is a typo; that
> bug number appears to relate to go-dump.c and is marked as RESOLVED
> FIXED.
It's PR78342, oops. Thanks,
Segher
On 11/24/2016 07:11 AM, Jiong Wang wrote:
gcc/
2016-11-11 Jiong Wang
* function.c (expand_function_end): Guard stack_protect_epilogue
with
ENABLE_DEFAULT_SSP_RUNTIME.
* cfgexpand.c (pass_expand::execute): Likewise guard for
stack_protect_prologue.
* defa
On Mon, 2016-11-28 at 15:02 -0700, Jeff Law wrote:
> On 11/28/2016 03:00 PM, David Malcolm wrote:
> > On Mon, 2016-11-28 at 14:44 -0700, Jeff Law wrote:
> > > On 11/23/2016 04:55 PM, Segher Boessenkool wrote:
> > > > Combine can turn a conditional trap into an unconditional trap.
> > > > If
> > >
On 11/24/2016 02:40 PM, Andrew Burgess wrote:
* Christophe Lyon [2016-11-21 13:47:09 +0100]:
On 20 November 2016 at 18:27, Mike Stump wrote:
On Nov 19, 2016, at 1:59 PM, Andrew Burgess wrote:
So, your new test fails on arm* targets:
After a little digging I think the problem might be tha
On 11/28/2016 03:00 PM, David Malcolm wrote:
On Mon, 2016-11-28 at 14:44 -0700, Jeff Law wrote:
On 11/23/2016 04:55 PM, Segher Boessenkool wrote:
Combine can turn a conditional trap into an unconditional trap. If
it
does that it should make the code after it unreachable (an
unconditional
trap
On Mon, 2016-11-28 at 14:44 -0700, Jeff Law wrote:
> On 11/23/2016 04:55 PM, Segher Boessenkool wrote:
> > Combine can turn a conditional trap into an unconditional trap. If
> > it
> > does that it should make the code after it unreachable (an
> > unconditional
> > trap should be the last insn in
On 11/28/2016 02:55 PM, Segher Boessenkool wrote:
On Mon, Nov 28, 2016 at 01:56:10PM -0700, Jeff Law wrote:
On 11/26/2016 04:19 PM, Segher Boessenkool wrote:
simplify_truncation changes the truncation of many operations into
the operation on the truncation. This patch makes this code also
hand
On 11/25/2016 01:04 AM, Richard Biener wrote:
On Thu, 24 Nov 2016, Senthil Kumar Selvaraj wrote:
Richard Biener writes:
On Thu, 24 Nov 2016, Richard Biener wrote:
On Thu, 24 Nov 2016, Senthil Kumar Selvaraj wrote:
Hi,
I've been analyzing a failing regtest (gcc.dg/strlenopt-8.c) for th
On Mon, Nov 28, 2016 at 01:56:10PM -0700, Jeff Law wrote:
> On 11/26/2016 04:19 PM, Segher Boessenkool wrote:
> >simplify_truncation changes the truncation of many operations into
> >the operation on the truncation. This patch makes this code also
> >handle extracts.
> >
> >Tested on powerpc64-lin
On 11/25/2016 07:02 AM, Andreas Schwab wrote:
On Nov 23 2016, Christophe Lyon wrote:
Since you committed this patch (r242733), I've noticed a regression:
gcc.dg/uninit-pred-6_c.c warning (test for warnings, line 43)
now fails on some arm targets, for instance arm-none-linux-gnueabihf
--with
On 11/24/2016 12:24 AM, Senthil Kumar Selvaraj wrote:
Jeff Law writes:
On 11/23/2016 02:54 AM, Senthil Kumar Selvaraj wrote:
Hi,
The below patch fixes uninit-19.c for avr by adding
-finline-small-functions for avr.
The test fails for avr because fn1 does not get inlined into fn2. Inli
On 11/23/2016 04:55 PM, Segher Boessenkool wrote:
Combine can turn a conditional trap into an unconditional trap. If it
does that it should make the code after it unreachable (an unconditional
trap should be the last insn in its bb, and that bb has no successors).
This patch seems to work. It
Ping?
On 16 November 2016 at 22:18, Christophe Lyon
wrote:
> On 15 November 2016 at 12:50, Jonathan Wakely wrote:
>> On 14/11/16 14:32 +0100, Christophe Lyon wrote:
>>>
>>> On 20 October 2016 at 19:40, Jonathan Wakely wrote:
On 20/10/16 10:33 -0700, Mike Stump wrote:
>
>
>
On 11/06/2016 09:39 AM, Waldemar Brodkorb wrote:
Hi,
This commit makes the Blackfin platform use the generic linker version
information, rather than a completely duplicated file, specific for the
Blackfin architecture.
This is made possible using the newly introduced skip_underscore
variable of
On 11/26/2016 02:42 AM, Waldemar Brodkorb wrote:
Hi,
Some platforms, such as Blackfin, have a special prefix for assembly
symbols as opposed to C symbols. For this reason, a function named
"foo()" in C will in fact be visible as a symbol called "_foo" in the
ELF binary.
The current linker versi
I was digging into issues around the patches for 78120 when I stumbled
upon undesirable bb copying in bb-reorder.c on the m68k.
The core issue is that the m68k does not define a length attribute and
therefore generic code assumes that the length of all insns is 0 bytes.
That in turn makes
Hi
Here is a patch to fix pretty printers when versioned namespace is
activated.
You will see that I have hesitated in making the fix independant of
the version being used. In source files you will find (__7::)? patterns
while in xmethods.py I chose (__\d+::)? making it ready for __8
On 11/26/2016 04:19 PM, Segher Boessenkool wrote:
simplify_truncation changes the truncation of many operations into
the operation on the truncation. This patch makes this code also
handle extracts.
Tested on powerpc64-linux. With this patch the rlwimi testcases work.
Is this okay for trunk?
On 11/27/2016 08:34 PM, Martin Sebor wrote:
PR 78521 notes that the gimple-ssa-sprintf pass doesn't do the right
thing (i.e., the -Wformat-length and -fprintf-return-value options
behave incorrectly) when a conversion specification includes a width
or precision with a non-constant value. The cod
On 11/25/16 12:30, Ramana Radhakrishnan wrote:
> On Sun, Nov 6, 2016 at 2:18 PM, Bernd Edlinger
> wrote:
>> Hi!
>>
>> This improves the stack usage on the sha512 test case for the case
>> without hardware fpu and without iwmmxt by splitting all di-mode
>> patterns right while expanding which is si
Hi!
On the following case we ICE in DF checking, because
reload_combine_recognize_pattern starts with apply_change_group
that in the end on success calls df_insn_rescan, but if it is successful,
we then remove_reg_equal_equiv_notes and nothing rescans the notes.
Fixed by rescanning just the notes
On Nov 28, 2016, at 8:52 AM, Thomas Preudhomme
wrote:
>
> Ping?
Ok.
> On 17/11/16 20:42, Thomas Preudhomme wrote:
>> Ping?
>>
>> Best regards,
>>
>> Thomas
>>
>> On 08/11/16 13:35, Thomas Preudhomme wrote:
>>> Ping,
>>>
>>> Best regards,
>>>
>>> Thomas
>>>
>>> On 02/11/16 10:04, Thomas P
On 11/28/2016 07:50 PM, Jeff Law wrote:
So how would you suggest this be fixed right now? I'd really like to
get the regression addressed.
The regression is still fixed. That wasn't the case at all stages while
I was working on it, but the i386 patch seems to suffice now.
I would claim tha
On 11/24/2016 09:14 AM, Segher Boessenkool wrote:
On Thu, Nov 24, 2016 at 08:48:04AM -0700, Jeff Law wrote:
On 11/24/2016 07:53 AM, Segher Boessenkool wrote:
That we compare different kinds of costs (which really has no meaning at
all, it's a heuristic at best) in various places is a known pro
On 11/28/2016 10:30 AM, Jakub Jelinek wrote:
Hi!
For TImode and wider, neg_const_int returns wrong value for
(const_int 0x8000) - the argument unmodified.
The function comment above the function kind of suggests that it
truncates, but I fail to see which caller might benefit from suc
On 11/28/2016 06:10 AM, Paolo Bonzini wrote:
On 27/11/2016 00:28, Marc Glisse wrote:
On Sat, 26 Nov 2016, Paolo Bonzini wrote:
--- match.pd(revision 242742)
+++ match.pd(working copy)
@@ -2554,6 +2554,19 @@
(cmp (bit_and@2 @0 integer_pow2p@1) @1)
(icmp @2 { build_zero_cst (TREE_T
PING!
I know it's a lengthy patch, but comments would be nice anyway.
- Andre
On Tue, 22 Nov 2016 20:46:50 +0100
Andre Vehreschild wrote:
> Hi all,
>
> attached patch addresses the need of extending the API of the caf-libs to
> enable allocatable components asynchronous allocation. Allocatabl
On Mon, Nov 28, 2016 at 07:12:40PM +0100, Markus Trippelsdorf wrote:
> > Can't you just make small_val an unsigned WIDE_INT, instead?
>
> No, it doesn't work. First of all you'll get -Wsign-compare warnings.
> And gcc.target/powerpc/altivec-11.c also starts failing.
Yuck. Let's go with your orig
On 2016.11.28 at 12:02 -0600, Segher Boessenkool wrote:
> Hi Markus,
>
> On Mon, Nov 28, 2016 at 02:58:19PM +0100, Markus Trippelsdorf wrote:
> > Running bootstrap-ubsan on ppc64le shows many instances of e.g.:
> > config/rs6000/rs6000.c:6217:36: runtime error: left shift of negative
> > value -
Hi Markus,
On Mon, Nov 28, 2016 at 02:58:19PM +0100, Markus Trippelsdorf wrote:
> Running bootstrap-ubsan on ppc64le shows many instances of e.g.:
> config/rs6000/rs6000.c:6217:36: runtime error: left shift of negative value
> -12301
>
> The attached patch fixes the issue and was tested on ppc6
Hi!
If we add shared (FRAME.NN) clause to some parallel/task OpenMP construct
or map (tofrom: FRAME.NN) clause to some target construct, we need to
add one of those to all outer parallel/task/target constructs too.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
committed to
Hi!
For TImode and wider, neg_const_int returns wrong value for
(const_int 0x8000) - the argument unmodified.
The function comment above the function kind of suggests that it
truncates, but I fail to see which caller might benefit from such
behavior, it would mean having to guard again
Hi Jeff,
On Mon, Nov 28, 2016 at 08:46:05AM -0700, Jeff Law wrote:
> On 11/28/2016 07:27 AM, Martin Jambor wrote:
> > Hi,
> >
> > one of a number of symptoms of an otherwise unrelated HSA bug I've
> > been debugging today is gcc crashing or hanging in the C++ pretty
> > printer when attempting to
OK.
On Mon, Nov 28, 2016 at 10:38 AM, Jakub Jelinek wrote:
> On Mon, Nov 28, 2016 at 09:44:30AM -0500, Jason Merrill wrote:
>> On Fri, Nov 25, 2016 at 12:31 PM, Jakub Jelinek wrote:
>> > * decl.c (build_enumerator): Call fixup_type_variants on
>> > current_class_type after finish
On Thu, Nov 24, 2016 at 8:50 AM, Richard Biener wrote:
>> > + /* ??? We can't annotate types late, but for LTO we may not
>> > +generate a location early either (gfortran.dg/save_5.f90).
>> > +The proper way is to handle it like VLAs though it is told
>> > +that DW_AT_string_leng
Richard!
I attached vect dump for hte part of attached test-case which
illustrated how vectorization of epilogues works through masking:
#define SIZE 1023
#define ALIGN 64
extern int posix_memalign(void **memptr, __SIZE_TYPE__ alignment,
__SIZE_TYPE__ size) __attribute__((weak));
extern void free
On Fri, Nov 25, 2016 at 10:51 AM, Uros Bizjak wrote:
> Attached patch tries to fix the mess with operations where register
> allocator is free to use either mask or general registers. There are
> plenty of problems with this approach:
> a) missing operations wth general registers
> - kxnor operat
Hi Richard,
Ping?
Best regards,
Thomas
On 17/11/16 20:42, Thomas Preudhomme wrote:
Ping?
Best regards,
Thomas
On 08/11/16 13:35, Thomas Preudhomme wrote:
Ping,
Best regards,
Thomas
On 02/11/16 10:04, Thomas Preudhomme wrote:
Ping?
Best regards,
Thomas
On 28/10/16 10:49, Thomas Preu
OK.
I was initially a little surprised you didn't just delete the call from
the IL and emit the trivial constant assignment. Then I realized you
probably would need to update the vuse/vdefs that were on the original
call. Using update_call_from_tree seems to take care of that for you.
I didn't
Hi Richard,
I am just wondering if this might be a mistake,
because this looks somehow odd:
> + /* CST - (CST - A) -> CST - A */
> + (simplify
> + (minus CONSTANT_CLASS_P@1 (minus CONSTANT_CLASS_P@2 @0))
> + (with { tree cst = const_binop (MINUS_EXPR, type, @1, @2); }
> +(if (cst
Robert Suchanek writes:
> The patch primarily fixes an ICE with out-of-range values to the
> __builtin_msa_insve* intrinsics.
>
> The compiler segfaults in mips_legitimize_const_move () as it tries to
> split symbol that has NULL_RTX value and gets here because the patterns
> reject the operand a
On 11/28/2016 07:27 AM, Martin Jambor wrote:
Hi,
one of a number of symptoms of an otherwise unrelated HSA bug I've
been debugging today is gcc crashing or hanging in the C++ pretty
printer when attempting to emit a warning because dump_decl() ended up
in an infinite recursion calling itself on
On Mon, Nov 28, 2016 at 09:44:30AM -0500, Jason Merrill wrote:
> On Fri, Nov 25, 2016 at 12:31 PM, Jakub Jelinek wrote:
> > * decl.c (build_enumerator): Call fixup_type_variants on
> > current_class_type after finish_member_declaration.
>
> Let's do this in finish_enum_value_list,
OK.
On Thu, Nov 24, 2016 at 4:47 PM, Jakub Jelinek wrote:
> Hi!
>
> Delayed folding caused following regression where we no longer warn about
> returning reference to local variable.
>
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
> trunk?
>
> 2016-11-24 Jakub Jel
On Fri, Nov 25, 2016 at 12:31 PM, Jakub Jelinek wrote:
> * decl.c (build_enumerator): Call fixup_type_variants on
> current_class_type after finish_member_declaration.
Let's do this in finish_enum_value_list, in the block controlled by
if (at_class_scope_p ()
&& COMPLETE_
On Thu, 24 Nov 2016, Yuri Rumyantsev wrote:
> Hi All,
>
> Here is the second patch which supports epilogue vectorization using
> masking without cost model. Currently it is possible
> only with passing parameter "--param vect-epilogues-mask=1".
>
> Bootstrapping and regression testing did not sh
Hi,
one of a number of symptoms of an otherwise unrelated HSA bug I've
been debugging today is gcc crashing or hanging in the C++ pretty
printer when attempting to emit a warning because dump_decl() ended up
in an infinite recursion calling itself on the DECL_ABSTRACT_ORIGIN of
the decl it was loo
> > Thanks. This is in the #else branch of '#ifdef HAVE_PTHREAD_AFFINITY_NP'
> > (so the ChangeLog entry should have [!HAVE_PTHREAD_AFFINITY_NP]) --
> > I'm curious why you don't have this defined. One reason I can think of is,
> > if you're building libgomp against musl libc
(I was mistaken he
Running bootstrap-ubsan on ppc64le shows many instances of e.g.:
config/rs6000/rs6000.c:6217:36: runtime error: left shift of negative value
-12301
The attached patch fixes the issue and was tested on ppc64le.
OK for trunk?
Thanks.
PR target/78556
* config/rs6000/rs6000.h (vsp
Been looking at this off and on, and I'm still not sure I entirely get
it - sorry.
On 11/11/2016 10:15 PM, David Malcolm wrote:
Implementing an RTL frontend by using the RTL reader from read
-rtl.c
means that we now need a diagnostics subsystem on the *host* for
handling errors in RTL files, ra
> On Mon, Nov 28, 2016 at 11:52:53AM +0100, Jan Hubicka wrote:
> > > > IPA-ICF performs some code-generation visible changes from hash table
> > > > traversal, where the hash values can be different between -g and -g0
> > > > (I bet it hashes DECL_UID in somewhere, perhaps other things).
> > >
> >
On Mon, 28 Nov 2016, Pitchumani Sivanupandi wrote:
> On Thursday 24 November 2016 04:54 PM, Richard Biener wrote:
> > On Thu, 24 Nov 2016, Pitchumani Sivanupandi wrote:
> >
> > > GCC inlines small functions if the code size after expansion is not
> > > excedded.
> > > For test case (inline.c, avr
Hi all,
the attached patch was posted on bugzilla by Tobias three years ago,
but left unattended since then. It is simple, works well (fixing a
bogus warning) and regtests cleanly on x86_64-linux-gnu.
If no one objects, I will commit this to trunk by tomorrow.
Cheers,
Janus
2016-11-28 Tobias
>> + /* Sign-extend @1 to TYPE. */
>> + w1 = w1.from (w1, TYPE_PRECISION (type), SIGNED);
>>
>> not sure why you do always sign-extend. If the inner op is unsigned
>> and we widen then that's certainly bogus considering your UINT_MAX
>> example above. Does
>>
>>
On Mon, Nov 28, 2016 at 03:50:54PM +0300, Alexander Monakov wrote:
> > * config/nvptx/affinity.c: Delete to use fallback implementation.
>
> Right. This file was (correctly) deleted on the branch, but I missed the
> deletion when writing the Changelog and then subsequently missed it again
> d
On Thursday 24 November 2016 04:54 PM, Richard Biener wrote:
On Thu, 24 Nov 2016, Pitchumani Sivanupandi wrote:
GCC inlines small functions if the code size after expansion is not excedded.
For test case (inline.c, avr-gcc -Os -S inline.c) code size become higher if
'func2' is inlined. It happe
On 27/11/2016 00:28, Marc Glisse wrote:
> On Sat, 26 Nov 2016, Paolo Bonzini wrote:
>
>> --- match.pd(revision 242742)
>> +++ match.pd(working copy)
>> @@ -2554,6 +2554,19 @@
>> (cmp (bit_and@2 @0 integer_pow2p@1) @1)
>> (icmp @2 { build_zero_cst (TREE_TYPE (@0)); })))
>>
>> +/* If w
> -Original Message-
> From: Alexander Monakov [mailto:amona...@ispras.ru]
> Sent: Monday, November 28, 2016 1:51 PM
> To: Hahnfeld, Jonas
> Cc: gcc-patches@gcc.gnu.org; Jakub Jelinek (ja...@redhat.com)
> Subject: Re: [PATCH] Fix OpenMP offloading to NVPTX and correct location of
>
> On Mon
On Mon, 28 Nov 2016, Hahnfeld, Jonas wrote:
> Hi,
>
> find attached a patch that fixes OpenMP offloading to NVPTX for me. Without
> the changes, I got errors from the linker for affinity symbols.
> 2016-11-28 Jonas Hahnfeld
>
> * config/linux/affinity.c: Include ../../affinity.c as fall
The following fixes another issue with marrying the SSA propagator
and match-and-simplify during propagation (which is tricky at best).
Bootstrap and regtest running on x86_64-unknown-linux-gnu, I've
applied the VRP patch below as well to see if the theory works
out (VRP misses to handle simplifi
The following fixes an optimization regression noted in PR78546 where
moving patterns from forwprop to match.pd missed one variant.
Bootstrap & regtest running on x86_64-unknown-linux-gnu.
Richard.
2016-11-28 Richard Biener
PR middle-end/78546
* match.pd: Add CST - (CST - A)
On Fri, Nov 25, 2016 at 10:46:49AM +, Tamar Christina wrote:
> Hi all,
>
> This adds the following NEON intrinsics
> to the Aarch64 back-end of GCC:
>
> * vsriq_n_p64
> * vsri_n_p64
> * vextq_p64
> * vext_p64
> * vceq_p64
> * vbslq_p64
> * vbsl_p64
>
> Added new tests for these and ran regre
On Sat, 26 Nov 2016, Prathamesh Kulkarni wrote:
> On 25 November 2016 at 14:48, Richard Biener wrote:
> > On Fri, 25 Nov 2016, Prathamesh Kulkarni wrote:
> >
> >> On 25 November 2016 at 13:43, Richard Biener wrote:
> >> > On Fri, 25 Nov 2016, Jakub Jelinek wrote:
> >> >
> >> >> On Fri, Nov 25, 2
Hi,
find attached a patch that fixes OpenMP offloading to NVPTX for me. Without
the changes, I got errors from the linker for affinity symbols.
Kind regards,
Jonas
--
Jonas Hahnfeld, MATSE-Auszubildender
IT Center
Group: High Performance Computing
Division: Computational Science and Engineering
On Fri, Nov 25, 2016 at 2:46 PM, Richard Biener
wrote> On Wed, Nov 16, 2016 at 4:54 PM,
Robin Dapp wrote:
>> Found some time to look into this again.
>>
>>> Index: tree-ssa-propagate.c
>>> ===
>>> --- tree-ssa-propagate.c(rev
On Mon, Nov 28, 2016 at 11:52:53AM +0100, Jan Hubicka wrote:
> > > IPA-ICF performs some code-generation visible changes from hash table
> > > traversal, where the hash values can be different between -g and -g0
> > > (I bet it hashes DECL_UID in somewhere, perhaps other things).
> >
> > It should
> > Hi!
> >
> > IPA-ICF performs some code-generation visible changes from hash table
> > traversal, where the hash values can be different between -g and -g0
> > (I bet it hashes DECL_UID in somewhere, perhaps other things).
>
> It should not - the hash is stable for LTO streaming.
> I think ICF
On Mon, 28 Nov 2016, Christophe Lyon wrote:
> Hi Richard,
>
>
> On 25 November 2016 at 11:20, Richard Biener wrote:
> > On Thu, 24 Nov 2016, Richard Biener wrote:
> >
> >>
> >> I am testing the following patch for an optimization regression where
> >> a loop made dead by final value replacement
> Hi!
>
> IPA-ICF performs some code-generation visible changes from hash table
> traversal, where the hash values can be different between -g and -g0
> (I bet it hashes DECL_UID in somewhere, perhaps other things).
It should not - the hash is stable for LTO streaming.
I think ICF merging -g and
> On 28 Nov 2016, at 10:34, Iain Sandoe wrote:
>
> Hi,
>
> I used a non-portable shell idiom in the second patch for pr71767. This
> caused configuration issues resulting in bootstrap fails for FreeBSD and
> hppa64-*-*.
>
> Fix tested by Andreas and John (and re-checked on Darwin and Linux
Hi,
I used a non-portable shell idiom in the second patch for pr71767. This caused
configuration issues resulting in bootstrap fails for FreeBSD and hppa64-*-*.
Fix tested by Andreas and John (and re-checked on Darwin and Linux by me).
Applied as 242912.
Iain
gcc/
2016-11-28 Iain Sandoe
Hi Richard,
On 25 November 2016 at 11:20, Richard Biener wrote:
> On Thu, 24 Nov 2016, Richard Biener wrote:
>
>>
>> I am testing the following patch for an optimization regression where
>> a loop made dead by final value replacement was made used again by
>> DOM 20 passes later. The real issue
On Fri, 25 Nov 2016, Jakub Jelinek wrote:
> Hi!
>
> This patch optimizes a = {}; b = a; into a = {}; b = {};
> and similarly for memset instead of the first stmt and/or memcpy
> instead of the second one.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> 2016-11-25 J
http://gcc.gnu.org/r242909
This patch adds treatment of reg_unused_after for *(X+const) that only
restores X if X is known to be used.
Applied as obvious.
Johann
* config/avr/avr.c (out_movhi_r_mr) [REG_X + PLUS]: Only SBIW if
X is not unused after.
Index: config/avr/avr.c
Hi Andrew,
Approved, but ...
> +/* Return true if a load instruction (CONSUMER) uses the same address
> + as a store instruction (PRODUCER). This function is used to avoid
> + st/ld address hazard in ARC700 cores. */
Missing space after dot.
> + /* Peel the producer and the consumer for
On Fri, 25 Nov 2016, Jakub Jelinek wrote:
> Hi!
>
> IPA-ICF performs some code-generation visible changes from hash table
> traversal, where the hash values can be different between -g and -g0
> (I bet it hashes DECL_UID in somewhere, perhaps other things).
> The following patch fixes it by addin
1 - 100 of 104 matches
Mail list logo
