Hello!
> This patch adds a test to the cmpstrnsi pattern in i386.md so that it
> will bail out (FAIL) if neither of the strings is a constant string. It
> can only work as a proper strncmp if the length is not longer than both
> of the strings. This change is required if expand_builtin_strncmp is
Most of the AVR's 8k memorydevices have only rjmp instruction, not jmp.
So, it
is important to wrap around jump destination to check if it can reach
backwards.
Currently link specs passes --pmem-wrap-around=xxK when mrelax and
mpmem-wrap-around options are enabled. Attached patch changes the sp
On Wed, Nov 2, 2016 at 4:27 PM, Segher Boessenkool
wrote:
> On Wed, Nov 02, 2016 at 02:51:41PM +0100, Richard Biener wrote:
>> >> I don't believe it needs a cleanup on every iteration. One cleanup at
>> >> the end should work fine.
>> >
>> > But as the comment there says:
>> >
>> > /* Merge
Hi,
The below patch requires int32plus for
gcc.c-torture/execute/pr78170.c, as it has int bitfields that are more
than 16 bits wide.
Committed to trunk as obvious.
Regards
Senthil
gcc/testsuite/ChangeLog
2016-11-03 Senthil Kumar Selvaraj
* gcc.c-torture/execute/pr78170.c: R
Hi,
The below patch adds a new effective target keyword (store_merge) for
use in the store_merging_xxx.c tests.
The tests currently require non_strict_align, but they still fail for the avr.
Eyeballing the dump, I found that the pass doesn't attempt merging as it is
unprofitable for a t
On Wed, 2 Nov 2016, Richard Biener wrote:
>
> The following fixes PR78185 by properly honoring possibly infinite child
> loops when computing what blocks are always executed during loop invariant
> motion. Such loops behave as if the loop would exit at this point.
>
> Both GIMPLE and RTL level
This adds an mcpu option for the upcoming Qualcomm Falkor core. This
is identical to the qdf24xx part that was added earlier and hence
retains the same tuning structure and continues to have the a57
pipeline for now. The part number has also been changed and this
patch fixes this for both qdf24xx
Hi Richard,
The attached patch tries to fix PR35691, by adding the following two
transforms to match.pd:
(x == 0 && y == 0) -> (x | typeof(x)(y)) == 0.
(x != 0 || y != 0) -> (x | typeof(x)(y)) != 0.
For GENERIC, the "and" operator is truth_andif_expr, and it seems for GIMPLE,
it gets transformed t
On Wed, Nov 02, 2016 at 01:39:22PM -0400, Jason Merrill wrote:
> On Wed, Nov 2, 2016 at 12:37 PM, Joseph Myers wrote:
> > On Wed, 2 Nov 2016, Jason Merrill wrote:
> >
> >> It seems to me that the general principle is that we should consider
> >> the location where the thing we're warning about is
Paul Hua writes:
> Loongson3a has 4 operand fused madd instrcution. This patch set
> loongson3a use fused madd.d.
Hi Paul,
Thanks for the fix. I was vaguely aware that this was wrong for
loongson-3a but never confirmed it.
I suspect this change is mechanical enough that it can bypass
copyright
The attached patch fixes the stack layout problems on AIX and
Power as described here:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77359
The patch has been bootstrapped on AIX (32 Bit) and bootstrappend
and regression tested on Power (biarch). It needs more testing
that I cannot do with the h
On Thu, 3 Nov 2016, Prathamesh Kulkarni wrote:
> Hi Richard,
> The attached patch tries to fix PR35691, by adding the following two
> transforms to match.pd:
> (x == 0 && y == 0) -> (x | typeof(x)(y)) == 0.
> (x != 0 || y != 0) -> (x | typeof(x)(y)) != 0.
>
> For GENERIC, the "and" operator is tr
Hi Vineet,
Thank you for your contribution.
> gcc/
> 2016-10-28 Vineet Gupta
>
> * config/arc/arc.h (SIZE_TYPE): define as unsigned int.
> * (PTRDIFF_TYPE): define as int.
>
Approved and committed. However, the entry changelog line is not as expected
and the patch didn't appl
Hi,
The gcc.target/mips/wrap-delay.c test was failing on mips-img-* toolchains
because it was using -mbranch-likely with an R6 target, and branch-likely
instructions were removed in R6.
This patch makes the testsuite downgrade to R5 if the -mbranch-likely option
is present and we're targeting R6.
On Tue, Nov 01, 2016 at 02:53:58PM +0100, Jakub Jelinek wrote:
> On Tue, Nov 01, 2016 at 09:41:20AM -0400, Jason Merrill wrote:
> > On Tue, Oct 25, 2016 at 9:59 AM, Marek Polacek wrote:
> > > On Mon, Oct 24, 2016 at 04:10:21PM +0200, Marek Polacek wrote:
> > >> On Thu, Oct 20, 2016 at 12:28:36PM +
> The same approach can also be used on other targets without hardware
> divmod insn.
This should be made generic instead: can't we automatically create a divmod
libfunc for double-word mode & define a default TARGET_EXPAND_DIVMOD_LIBFUNC?
This would save the code duplication in the ~50 back-ends
Ping.
Regards,
Andrew
> -Original Message-
> From: Andrew Bennett
> Sent: 28 August 2015 16:50
> To: Matthew Fortune; Moore, Catherine; gcc-patches@gcc.gnu.org
> Subject: RE: [PATCH] MIPS: If a test in the MIPS testsuite requires standard
> library support check the sysroot supports t
I'm concerned about the number of false positives for this warning, and
judging by previous discussions, I'm not alone in this. This patch
limits it to level 1 (any comment before the case label disables the
warning) for cases where the user specified no explicit level. It'll
still generate eno
On Thu, Nov 3, 2016 at 12:29 PM, Eric Botcazou wrote:
>> The same approach can also be used on other targets without hardware
>> divmod insn.
>
> This should be made generic instead: can't we automatically create a divmod
> libfunc for double-word mode & define a default TARGET_EXPAND_DIVMOD_LIBFU
On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schmidt wrote:
> I'm concerned about the number of false positives for this warning, and
> judging by previous discussions, I'm not alone in this. This patch limits it
> to level 1 (any comment before the case label disables the warning) for
> cases w
Hi Matthew,
Thanks for your comments, update the patch.
*** gcc/ChangeLog ***
2016-11-03 Chenghua Xu
* config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
TARGET_LOONGSON_3A.
(ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
Thanks,
Paul
On Thu, Nov 3, 2016 at 6:3
On 09/14/2016 03:00 PM, Andrew Burgess wrote:
In an attempt to get this patch merged (as I still think that its
correct) I've investigated, and documented a little more about how I
think things currently work. I'm sure most people reading this will
already know this, but hopefully, if my underst
On Thu, Nov 03, 2016 at 12:58:55PM +0100, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schmidt wrote:
> > I'm concerned about the number of false positives for this warning, and
> > judging by previous discussions, I'm not alone in this. This patch limits it
> > to level 1
On 02/11/16 11:36, Eric Botcazou wrote:
I think you're right. I suppose the new condition should be:
#ifdef LOAD_EXTEND_OP
/* If this is a typical RISC machine, we only have to worry
about the way loads are extended. */
if (!WORD_REGISTER_OPERATIONS
On 03/11/16 12:05, Kyrill Tkachov wrote:
On 02/11/16 11:36, Eric Botcazou wrote:
I think you're right. I suppose the new condition should be:
#ifdef LOAD_EXTEND_OP
/* If this is a typical RISC machine, we only have to worry
about the way loads are extended. */
if (
>What's your decision on this?
I think that we ought to standardize on a single order for note copying in the
RTL middle-end and the best way to enforce it is to have a single primitive in
rtlanal.c, with an optional filtering. Bernd's patch is a step in the right
direction, but doesn't en
Hello.
This is small follow-up of the patches I sent to string folding.
The patch transforms character array defined in an initializer to string
constant:
+const char global[] = {'a', 'b', 'c', 'd', '\0'};
Apart from that, it also enables string folding of local symbols like:
+ const char local
Andrew Pinski wrote:
> On Tue, Oct 25, 2016 at 10:08 AM, Wilco Dijkstra
> wrote:
> > SHA1H instructions may be scheduled after a SHA1C instruction
> > that uses the same input register. However SHA1C updates its input,
> > so if SHA1H is scheduled after it, it requires an extra move.
> > Increas
On 08/04/2016 02:52 PM, Nathan Sidwell wrote:
> On 08/04/16 08:27, Martin Liška wrote:
>> Hi.
>>
>> Following patch is grabbed from the PR, where I just applied the patch
>> and wrote a test-case which removes x.gcda file before running gcov tool.
>>
>> Ready to be installed?
>
> 2016-08-04 Marti
On 10/29/2016 07:47 AM, Bernd Edlinger wrote:
Find attached a new patch with test case.
Boot-strapped on x86_64-pc-linux-gnu.
Is it OK for trunk?
Ok.
Bernd
> Thanks, here is the patch doing this.
> Committing to trunk after bootstrap and testing on x86_64.
Thanks for the quick turn around!
--
Eric Botcazou
> libfunc is already implemented for all targets to use, there is also:
>
> OPTAB_NC(sdivmod_optab, "divmod$a4", UNKNOWN)
> OPTAB_NC(udivmod_optab, "udivmod$a4", UNKNOWN)
>
> in optabs.def that can probably be changed for generic expansion.
So what's the purpose of ix86_init_libfuncs if the lib
Fix ldrd offsets of Thumb-2 - for TARGET_LDRD the range is +-1020,
without -255..4091. This reduces the number of addressing instructions
when using DI mode operations (such as in PR77308).
Bootstrap & regress OK.
ChangeLog:
2015-11-03 Wilco Dijkstra
gcc/
* config/arm/arm.c (arm_
On 03/11/16 12:06, Eric Botcazou wrote:
What's your decision on this?
I think that we ought to standardize on a single order for note copying in the
RTL middle-end and the best way to enforce it is to have a single primitive in
rtlanal.c, with an optional filtering. Bernd's patch is a step
On 11/03/2016 12:58 PM, Jakub Jelinek wrote:
On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schmidt wrote:
I'm concerned about the number of false positives for this warning, and
judging by previous discussions, I'm not alone in this. This patch limits it
to level 1 (any comment before the case
During Richi's debugging of the section type conflict, he found a
problem with rs6000_xcoff_declare_object_name trying to access a decl
that did not exist. Fixed thusly.
* config/rs6000/rs6000.c (rs6000_xcoff_declare_object_name): Use
symtab_node::get_create.
Index: rs6000.c
On Thu, Nov 03, 2016 at 01:22:11PM +0100, Bernd Schmidt wrote:
> On 11/03/2016 12:58 PM, Jakub Jelinek wrote:
> >On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schmidt wrote:
> >>I'm concerned about the number of false positives for this warning, and
> >>judging by previous discussions, I'm not al
> Thanks for the feedback, I'll try to work through this.
Thanks, but since there doesn't seem to be a consensus on the goal, feel free
to disregard it and just implement what you need for your initial work.
--
Eric Botcazou
Hi Richard,
I did not understand your last remark:
> That is, here (and avoid the FOR_EACH_LOOP change):
>
> @@ -580,12 +586,21 @@ vectorize_loops (void)
> && dump_enabled_p ())
> dump_printf_loc (MSG_OPTIMIZED_LOCATIONS, vect_location,
>"loop vecto
On Thu, Nov 3, 2016 at 1:12 PM, Martin Liška wrote:
> Hello.
>
> This is small follow-up of the patches I sent to string folding.
> The patch transforms character array defined in an initializer to string
> constant:
>
> +const char global[] = {'a', 'b', 'c', 'd', '\0'};
>
> Apart from that, it al
On Thu, Nov 3, 2016 at 1:18 PM, Eric Botcazou wrote:
>> libfunc is already implemented for all targets to use, there is also:
>>
>> OPTAB_NC(sdivmod_optab, "divmod$a4", UNKNOWN)
>> OPTAB_NC(udivmod_optab, "udivmod$a4", UNKNOWN)
>>
>> in optabs.def that can probably be changed for generic expansio
On 11/03/2016 01:12 PM, Martin Liška wrote:
+ tree init = DECL_INITIAL (decl);
+ if (init
+ && TREE_READONLY (decl)
+ && can_convert_ctor_to_string_cst (init))
+DECL_INITIAL (decl) = build_string_cst_from_ctor (init);
I'd merge these two new functions since they're only ever cal
On 03.11.2016 08:58, Pitchumani Sivanupandi wrote:
Most of the AVR's 8k memorydevices have only rjmp instruction, not jmp. So, it
is important to wrap around jump destination to check if it can reach backwards.
Currently link specs passes --pmem-wrap-around=xxK when mrelax and
mpmem-wrap-around
> -Original Message-
> From: Toma Tabacu [mailto:toma.tab...@imgtec.com]
> Sent: Thursday, November 3, 2016 6:58 AM
> Subject: [PATCH,testsuite] MIPS: Downgrade R6 to R5 if tests need
> branch-likely instructions.
>
>
> gcc/testsuite/
> * gcc.target/mips/mips.exp: Add check for
On 2016.11.03 at 13:32 +0100, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 01:22:11PM +0100, Bernd Schmidt wrote:
> > On 11/03/2016 12:58 PM, Jakub Jelinek wrote:
> > >On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schmidt wrote:
> > >>I'm concerned about the number of false positives for this w
> libfunc, as in "__{,u}divmod{di,ti}4 library function" is already
> implemented in libgcc. But the enablement of this function inside the
> compiler has to be performed by each target.
So can we do it generically instead of duplicating it ~50 times?
--
Eric Botcazou
> On 11/03/2016 01:12 PM, Martin Liška wrote:
> >+ tree init = DECL_INITIAL (decl);
> >+ if (init
> >+ && TREE_READONLY (decl)
> >+ && can_convert_ctor_to_string_cst (init))
> >+DECL_INITIAL (decl) = build_string_cst_from_ctor (init);
>
> I'd merge these two new functions since the
On 11/03/2016 01:33 PM, Eric Botcazou wrote:
Thanks for the feedback, I'll try to work through this.
Thanks, but since there doesn't seem to be a consensus on the goal, feel free
to disregard it and just implement what you need for your initial work.
FWIW here's a more complete version of my
Toma Tabacu writes:
> The gcc.target/mips/wrap-delay.c test was failing on mips-img-*
> toolchains
> because it was using -mbranch-likely with an R6 target, and branch-
> likely
> instructions were removed in R6.
>
> This patch makes the testsuite downgrade to R5 if the -mbranch-likely
> option
>
On Thu, Nov 3, 2016 at 1:58 PM, Eric Botcazou wrote:
>> libfunc, as in "__{,u}divmod{di,ti}4 library function" is already
>> implemented in libgcc. But the enablement of this function inside the
>> compiler has to be performed by each target.
>
> So can we do it generically instead of duplicating
Hi all,
the attached patch restructures gfortran's way of initializing components of
derived types in ALLOCATE. The old way was to generate a new gfc_code-node and
add it after the ALLOCATE node to initialize the the derived type on certain
conditions (like initializer or allocatable components ex
> On 08/04/2016 02:52 PM, Nathan Sidwell wrote:
> > On 08/04/16 08:27, Martin Liška wrote:
> >> Hi.
> >>
> >> Following patch is grabbed from the PR, where I just applied the patch
> >> and wrote a test-case which removes x.gcda file before running gcov tool.
> >>
> >> Ready to be installed?
> >
>
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Andrew Bennett
> Sent: 03 November 2016 11:33
> To: Matthew Fortune; 'Moore, Catherine'; 'gcc-patches@gcc.gnu.org'
> Subject: RE: [PATCH] MIPS: If a test in the MIPS testsuite r
On Thu, Nov 03, 2016 at 01:55:33PM +0100, Markus Trippelsdorf wrote:
> On 2016.11.03 at 13:32 +0100, Jakub Jelinek wrote:
> > On Thu, Nov 03, 2016 at 01:22:11PM +0100, Bernd Schmidt wrote:
> > > On 11/03/2016 12:58 PM, Jakub Jelinek wrote:
> > > >On Thu, Nov 03, 2016 at 12:51:15PM +0100, Bernd Schm
On Tue, Oct 18, 2016 at 4:26 PM, Paolo Carlini wrote:
> Thus, I'm back to one of my first tries earlier today: a much more
> conservative change which uses fold_non_dependent_expr only for the purpose
> of suppressing the unwanted warnings, see the below.
This looks right; in general we use early
On Thu, Nov 3, 2016 at 7:24 AM, Marek Polacek wrote:
> On Tue, Nov 01, 2016 at 02:53:58PM +0100, Jakub Jelinek wrote:
>> On Tue, Nov 01, 2016 at 09:41:20AM -0400, Jason Merrill wrote:
>> > On Tue, Oct 25, 2016 at 9:59 AM, Marek Polacek wrote:
>> > > On Mon, Oct 24, 2016 at 04:10:21PM +0200, Marek
On 11/02/2016 03:51 PM, Jakub Jelinek wrote:
> On Wed, Nov 02, 2016 at 03:38:25PM +0100, Martin Liška wrote:
>> it converts:
>> foo ()
>> {
>> char a;
>> char * p;
>> char _1;
>> int _2;
>> int _8;
>> int _9;
>>
>> :
>> ASAN_MARK (2, &a, 1);
>> a = 0;
>> p_6 = &a;
>> ASAN_MARK
Hello,
I'm facing the following problem related to ivopts. The problem is that
GCC generates a lot of induction variables and as a result there is an
unnecessary increase of stack usage and register pressure.
For instance, for the attached testcase (tc_ivopts.c) GCC generates 26
induction va
On Thu, Nov 03, 2016 at 09:27:55AM -0400, Jason Merrill wrote:
> On Thu, Nov 3, 2016 at 7:24 AM, Marek Polacek wrote:
> > On Tue, Nov 01, 2016 at 02:53:58PM +0100, Jakub Jelinek wrote:
> >> On Tue, Nov 01, 2016 at 09:41:20AM -0400, Jason Merrill wrote:
> >> > On Tue, Oct 25, 2016 at 9:59 AM, Marek
On 2016.11.03 at 14:24 +0100, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 01:55:33PM +0100, Markus Trippelsdorf wrote:
> > On 2016.11.03 at 13:32 +0100, Jakub Jelinek wrote:
> > > On Thu, Nov 03, 2016 at 01:22:11PM +0100, Bernd Schmidt wrote:
> > > > On 11/03/2016 12:58 PM, Jakub Jelinek wrote:
Hi!
FYI, I think it is much more important to get the initial patch in, so
resolve the switch + declarations inside of its body stuff, add testcases
for that and post for re-review and check in.
These optimizations can be done even in stage3.
On Thu, Nov 03, 2016 at 02:34:47PM +0100, Martin Liška
Hi,
The patch looks correct, however I would suggest to rewrite this bit of the code
urgently in separate patch as it is way too complex to assert it is now bug
free -
there are too many possible failure scenarios to list... Also it generates quite
inefficient code - pushable_regs should include
On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
> I don't have gathered detailed statistics. But for example a simple
> /* drop through */ in a package header file will of course cause many
> bogus warnings during the build on level 2.
> For the Linux kernel false positives dec
On 11/03/2016 02:47 PM, Jakub Jelinek wrote:
On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
I don't have gathered detailed statistics. But for example a simple
/* drop through */ in a package header file will of course cause many
bogus warnings during the build on level 2.
On Thu, Nov 03, 2016 at 02:48:33PM +0100, Bernd Schmidt wrote:
>
> On 11/03/2016 02:47 PM, Jakub Jelinek wrote:
> >On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
> >>I don't have gathered detailed statistics. But for example a simple
> >>/* drop through */ in a package header
On 2016.11.03 at 14:47 +0100, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
> > I don't have gathered detailed statistics. But for example a simple
> > /* drop through */ in a package header file will of course cause many
> > bogus warnings during the b
On Thu, Nov 03, 2016 at 02:55:03PM +0100, Markus Trippelsdorf wrote:
> On 2016.11.03 at 14:47 +0100, Jakub Jelinek wrote:
> > On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
> > > I don't have gathered detailed statistics. But for example a simple
> > > /* drop through */ in a
On 11/03/2016 02:44 PM, Jakub Jelinek wrote:
> Hi!
>
> FYI, I think it is much more important to get the initial patch in, so
> resolve the switch + declarations inside of its body stuff, add testcases
> for that and post for re-review and check in.
> These optimizations can be done even in stage3
> I guess it can be done. Currently the expander goes:
>
> --cut here--
> /* Check if optab_handler exists for divmod_optab for given mode. */
> if (optab_handler (tab, mode) != CODE_FOR_nothing)
> {
> quotient = gen_reg_rtx (mode);
> remainder = gen_reg_rtx (mode);
> ex
On Thu, Nov 03, 2016 at 03:02:21PM +0100, Martin Liška wrote:
> > But how would you be able to find out if there isn't any return *ptr; after
> > the scope or similar (as MEM_REF)? With is_gimple_reg, they will be turned
> > into SSA form and you can easily verify (uses of ASAN_POISON are a proble
On 2016.11.03 at 14:57 +0100, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 02:55:03PM +0100, Markus Trippelsdorf wrote:
> > On 2016.11.03 at 14:47 +0100, Jakub Jelinek wrote:
> > > On Thu, Nov 03, 2016 at 02:35:57PM +0100, Markus Trippelsdorf wrote:
> > > > I don't have gathered detailed statisti
> FWIW here's a more complete version of my patch which I'm currently
> testing. Let me know if you think it's at least a good enough
> intermediate step to be installed.
It is, thanks.
> I think, the sel-sched example notwithstanding, this is something that
> normally should not be needed outsid
On 3 November 2016 at 18:36, Uros Bizjak wrote:
> On Thu, Nov 3, 2016 at 1:58 PM, Eric Botcazou wrote:
>>> libfunc, as in "__{,u}divmod{di,ti}4 library function" is already
>>> implemented in libgcc. But the enablement of this function inside the
>>> compiler has to be performed by each target.
>
Hi,
According to analysis given by
https://gcc.gnu.org/ml/gcc/2016-10/msg00228.html, calls to
pedantic_non_lvalue_loc and code handling lvalue in
fold_cond_expr_with_comparison are useless now. Given this is complicated
legacy code, it may be better to change code step by step, rather than doi
As I've noticed some time ago, recent versions of Solaris
include this little gem:
#if __cplusplus >= 201103L
#undef _GLIBCXX_USE_C99_MATH
#undef _GLIBCXX_USE_C99_MATH_TR1
#endif
This renders a couple of perfectly good libstdc++ tests as UNSUPPORTED
and is completely unsustainable. Agreement
On 11/03/2016 03:03 PM, Jakub Jelinek wrote:
> On Thu, Nov 03, 2016 at 03:02:21PM +0100, Martin Liška wrote:
>>> But how would you be able to find out if there isn't any return *ptr; after
>>> the scope or similar (as MEM_REF)? With is_gimple_reg, they will be turned
>>> into SSA form and you can
Ping this patch again.
On 2016/9/21 12:43 AM, Cesar Philippidis wrote:
>> +/* Returns the number of mappings associated with the pointer or pset. PSET
>> > + have three mappings, whereas pointer have two. */
>> > +
>> > static int
>> > -find_pset (int pos, size_t mapnum, unsigned short *kinds
On 3 November 2016 at 16:13, Richard Biener wrote:
> On Thu, 3 Nov 2016, Prathamesh Kulkarni wrote:
>
>> Hi Richard,
>> The attached patch tries to fix PR35691, by adding the following two
>> transforms to match.pd:
>> (x == 0 && y == 0) -> (x | typeof(x)(y)) == 0.
>> (x != 0 || y != 0) -> (x | ty
Hello.
As Honza noticed we spent quite some time in __gcov_time_profiler:
perf report for Postgres make check command:
4.10% postgres postgres[.]__gcov_time_profiler
Thus I rewrote the profiling code directly to GIMPLE statements:
_4 = __gcov7.main[0];
if (_4 == 0)
goto ;
else
I've been trying to fix some bad tree-ssa related optimisation for
s390x and come up with the attached experimental patch. The patch
is not really good - it breaks some situations in which the
optimisation was useful. With this code:
void bar(long);
void foo (char a)
{
long l;
char
On Thu, Nov 03, 2016 at 04:03:22PM +0100, Dominik Vogt wrote:
> I've been trying to fix some bad tree-ssa related optimisation for
> s390x and come up with the attached experimental patch. The patch
> is not really good - it breaks some situations in which the
> optimisation was useful. With this
>
> 2016-10-31 Martin Liska
>
> * libgcov-profiler.c (__gcov_time_profiler): Remove.
> (__gcov_time_profiler_atomic): Likewise.
>
> gcc/ChangeLog:
>
> 2016-10-31 Martin Liska
>
> * profile.c (instrument_values): Fix coding style.
> (branch_prob): Use renamed funct
On 28.10.2016 17:58, Mike Stump wrote:
On Oct 27, 2016, at 3:16 AM, Georg-Johann Lay wrote:
Now imagine some arithmetic like &&LAB2 - &&LAB1. This might result in
one or two stub addresses, and difference between such addresses is a
complete different thing than the difference between the ori
[Please cc me and Segher on PowerPC / AIX patches.]
I bootstrapped successfully with the patch and ran the testsuite.
Without patch:
https://gcc.gnu.org/ml/gcc-testresults/2016-11/msg00186.html
With patch:
https://gcc.gnu.org/ml/gcc-testresults/2016-11/msg00233.html
Not exactly the same revisio
> Is VRP the right pass to do this optimisation or should a later
> pass rather attempt to eliminate the new use of b_5 instead? Uli
> has brought up the idea a mini "sign extend elimination" pass that
> checks if the result of a sign extend could be replaced by the
> original quantity in all plac
On Thu, Nov 03, 2016 at 09:29:07AM +0100, Richard Biener wrote:
> On Wed, Nov 2, 2016 at 4:27 PM, Segher Boessenkool
> wrote:
> > On Wed, Nov 02, 2016 at 02:51:41PM +0100, Richard Biener wrote:
> >> >> I don't believe it needs a cleanup on every iteration. One cleanup at
> >> >> the end should wor
On Thu, Nov 03, 2016 at 04:29:05PM +0100, Eric Botcazou wrote:
> > Is VRP the right pass to do this optimisation or should a later
> > pass rather attempt to eliminate the new use of b_5 instead? Uli
> > has brought up the idea a mini "sign extend elimination" pass that
> > checks if the result of
On Thu, Nov 3, 2016 at 1:35 PM, Evgeny Kudryashov wrote:
> Hello,
>
> I'm facing the following problem related to ivopts. The problem is that GCC
> generates a lot of induction variables and as a result there is an
> unnecessary increase of stack usage and register pressure.
>
> For instance, for
I managed to forget to mask the thing inserted. Tested on powerpc64-linux
{-m32,-m64}, and Bin tested on arm. Applying to trunk.
Segher
2016-11-03 Segher Boessenkool
PR rtl-optimization/78186
* combine.c (change_zero_ext): Mask the RHS of a zero_extract as
well, wh
The vec_interleave_{low, high} standard patterns were removed 5 years ago.
Tested on SPARC/Solaris, applied on the mainline.
2016-11-03 Eric Botcazou
* config/sparc/sparc.md (vec_interleave_lowv8qi): Delete.
(vec_interleave_highv8qi): Likewise.
--
Eric BotcazouIndex: config
On Thu, Nov 3, 2016 at 3:03 AM, Siddhesh Poyarekar
wrote:
> This adds an mcpu option for the upcoming Qualcomm Falkor core. This
> is identical to the qdf24xx part that was added earlier and hence
> retains the same tuning structure and continues to have the a57
> pipeline for now. The part numb
OpenACC permits the user to request more gang, worker and vector level
parallelism than what the compiler can utilize. For instance, if the
user writes worker routine without including a worker-partitioned loop,
the compiler will not generate worker-partitioned code for that function.
The intent b
Hi,
On Fri, Oct 28, 2016 at 01:58:13PM +1100, kugan wrote:
> > Do I understand it correctly that extract_range_from_unary_expr deals
> > with any potential type conversions better (compared to what you did
> > before here)?
>
> Yes, this can be wrong at times too as reported in
> https://gcc.gnu.
Hi,
this patch comes from our GCN BE branch. Because vectors of that
architectures have many elements, RTL dumps can get quite unwieldy when
all elements are printed out all the time. However, pretty much all the
time it is the same value repeated over and over again, which lead us to
the follow
On Wed, Nov 02, 2016 at 01:19:09PM -0400, Jason Merrill wrote:
> On Wed, Nov 2, 2016 at 12:33 PM, Jakub Jelinek wrote:
> > On Wed, Nov 02, 2016 at 04:44:05PM +0100, Jakub Jelinek wrote:
> >> which means if gen_type_die or gen_type_die_with_usage doesn't
> >> use the langhook, then we'd emit a comp
On 11/03/2016 05:35 PM, Martin Jambor wrote:
* print-rtl.c (print_rtx_operand_codes_E_and_V): Print how many times
an element is repeated istead of printing each repeated element.
"instead"
---
gcc/print-rtl.c | 15 ++-
1 file changed, 14 insertions(+), 1 deletio
Hi!
My recent optimize_abbrev_table optimization apparently broke Solaris
bootstrap.
The bug is specific I think just to -gdwarf-{2,3} -gno-strict-dwarf, where
DW_FORM_exprloc can't be used for location expressions and some location
expression contains DW_OP_GNU_{{const,regval,deref}_type,convert,
Hi,
When using a callee-saved register to save the frame pointer the Thumb-1
prologue fails to save the callee-saved register before that. For ARM and
Thumb-2 targets the frame pointer is handled as a special case but nothing is
done for Thumb-1 targets. This patch adds the same logic for Thum
On Nov 3, 2016, at 2:29 AM, Senthil Kumar Selvaraj
wrote:
>
> The below patch adds a new effective target keyword (store_merge) for
> use in the store_merging_xxx.c tests.
>
> The tests currently require non_strict_align, but they still fail for the
> avr.
> Eyeballing the dump, I found th
On Thu, 3 Nov 2016, Prathamesh Kulkarni wrote:
On 3 November 2016 at 16:13, Richard Biener wrote:
On Thu, 3 Nov 2016, Prathamesh Kulkarni wrote:
Hi Richard,
The attached patch tries to fix PR35691, by adding the following two
transforms to match.pd:
(x == 0 && y == 0) -> (x | typeof(x)(y)) =
1 - 100 of 131 matches
Mail list logo
