On Fri, Nov 3, 2017 at 12:11 AM, Wilco Dijkstra wrote:
> The Arm backend sets the default sched-pressure algorithm to
> SCHED_PRESSURE_MODEL. Benchmarking on AArch64 shows this
> speeds up floating point performance on SPEC - eg. CactusBSSN improves
> by ~16%. The gains are mostly due to less sp
The x86 backend defines a PROBE_INTERVAL which is supposed to be used by
the -fstack-check= mechanisms.
Some stack-clash code was using PROBE_INTERVAL rather than querying the
PARAM system for the right value. If the former is larger than the
latter and we allocate a large stack, then the
The x86 port relies on implicit probes that occur when a call
instruction pushes the return address onto the stack. Those implicit
calls allow the target to avoid emitting explicit stack probes.
Of course we have to account for tail call optimizations which turn the
call into a jump -- in p
Internal testing within Red Hat tripped an assert in the x86 target.
The assert wants to verify that the integer registers were saved prior
to probing to deal with a special case where the probe turns into a call
(which occurs on Windows IIUC).
In the case of stack-clash we aren't bound by
Hi Martin,
As mentioned in PR, the issue here for propagating value of 'm' from
f_c1 to foo() is that the jump function operation is FLOAT_EXPR, and
the type of input param 'm' is int, so fold_unary() doesn't do the
conversion to real_type. The attached patch fixes that by calling
fold_convert if o
The testcase gcc.target/mips/msa.c gives dozens of FAILs if it's tested
with a GCC configured to default to -fno-common, because of patterns like
/* { dg-final { scan-assembler-times "\t.comm\tv16i8_\\d+,16,16" 3 } } */
Seems like the simplest solution is to force -fcommon for this test. OK?
Several passes which perform dominator walks want to identify when block
has a single incoming edge, ignoring loop backedges.
I'm aware of 4 implementations of this code. 3 of the 4 are identical
in function. The 4th (tree-ssa-dom.c) has an additional twist that it
also ignores edges tha
On 11/02/2017 06:30 AM, Claudiu Zissulescu wrote:
From: claziss
The _Uncached type qualifier can be used to bypass the cache without
resorting to declaring variables as volatile.
gcc/
2017-07-12 Claudiu Zissulescu
* config/arc/arc-protos.h (arc_is_uncached_mem_p): Function proto.
On Thu, 2 Nov 2017, Richard Biener wrote:
You can handle floats as followup but some testcases that shouldn't be
optimized for the INT_MIN / unsigned case would be nice.
New version, that does a bit more.
Bootstrap+regtest on powerpc64le-unknown-linux-gnu.
2017-11-03 Marc Glisse
gcc/
Here's an updated version of the patch.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
OK for trunk?
Changed in v4:
- updated for changes of "inform_at_rich_loc" to "inform"
- added #define INCLUDE_UNIQUE_PTR to known-headers.cc
Changed in v3:
- fixed WINT_MAX and WINT_MIN
Chang
Jeff: You previously had concerns about the refcounting used in v1
of this patch; this avoids that in favor of using gnu::unique_ptr.
Joseph already approved the C frontend parts of v2 of this
patch.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
OK for trunk?
Changed in v3:
- W
This adds some extra debug info to the dump file for combine: print
the insns that are input to try_combine. I was worried printing more
will make the dump file only harder to read, but especially the info
from the REG_DEAD notes is invaluable.
Committing to trunk.
Segher
2017-11-02 Segher B
> Can be also seen in GCC PGO:
>
> checking for ssize_t... ../../libdecnumber/decNumber.c: In function
> ‘decDecap’:
> ../../libdecnumber/decNumber.c:7640:25: error: corrupted value profile:
> interval profile counter (0 out of 0) inconsistent with basic-block count
> (4356)
>if (cut!=DECDP
Hi Carl,
On Thu, Nov 02, 2017 at 12:54:57PM -0700, Carl Love wrote:
> 2017-11-01 Carl Love
>
> * config/rs6000/rs6000-c.c (P8V_BUILTIN_VEC_REVB): Add power 8
> definitions.
> (P9V_BUILTIN_VEC_REVB): Remove the power 9 instance definitions.
> * config/rs6000/altivec.h (v
On 10/31/2017 04:31 AM, Uros Bizjak wrote:
> On Tue, Oct 31, 2017 at 3:09 AM, Daniel Santos
> wrote:
>> When we are realigning the stack pointer, making an ms_abi to sysv_abi
>> call and alllocating 2GiB or more on the stack we end up with an invalid
>> INSN due to a non-immediate offset. This o
When adding the call to gen_vec_duplicate, I failed to notice that
code further down modified the VEC_DUPLICATE in place. That isn't
safe if gen_vec_duplicate returned a const_vector.
Tested on x86_64-linux-gnu. OK to install?
Thanks,
Richard
2017-11-02 Richard Sandiford
gcc/
PR t
On 10/26/2017 12:23 PM, Nathan Sidwell wrote:
On 10/26/2017 02:12 PM, Eric Gallager wrote:
On 10/26/17, Nathan Sidwell wrote:
On 10/26/2017 10:34 AM, David Malcolm wrote:
Possibly a silly question, but is it OK to have a formatted string
call in which some of the arguments aren't consumed?
Um, why are those member get's there at all (and with an index mapping
that doesn't agree with the member order)? [container.insert.return]
says that "It has no base classes or members other than those
specified."
On 11/02/2017 02:51 PM, David Malcolm wrote:
We currently identify our diagnostics via command-line options, and by
the text of the option itself.
This patch adds a way to supply metadata with a diagnostic, classifying
the problem being reported, according to one of the software problem
taxonomi
Apart from that I also see this GCC PGO:
../../libiberty/pex-unix.c:789:1: warning: Missing counts for called function
pex_child_error.isra.1/69
}
^
during IPA pass: inline
../../libiberty/pex-unix.c: In function ‘pex_child_error.isra.1’:
../../libiberty/pex-unix.c:373:1: internal compiler er
Can be also seen in GCC PGO:
checking for ssize_t... ../../libdecnumber/decNumber.c: In function ‘decDecap’:
../../libdecnumber/decNumber.c:7640:25: error: corrupted value profile:
interval profile counter (0 out of 0) inconsistent with basic-block count (4356)
if (cut!=DECDPUN) *msu%=powers[
We currently identify our diagnostics via command-line options, and by
the text of the option itself.
This patch adds a way to supply metadata with a diagnostic, classifying
the problem being reported, according to one of the software problem
taxonomies e.g. "INT15-C" within the CERT C Secure Codi
On 10/26/2017 10:34 AM, David Malcolm wrote:
[CCing Rainer and Mike for the gcc-dg.exp part]
My Tcl skills aren't great, so hopefully someone else can review this;
CCing Rainer and Mike.
Ping?
https://gcc.gnu.org/ml/gcc-patches/2017-10/msg01911.html
nathan
--
Nathan Sidwell
On 11/02/2017 08:06 PM, Jan Hubicka wrote:
Sorry, I must have used older diff file, because it is one of unfinished
chnages I made today.
I am attaching correct diff.
Thank you. This one works for me, however I see various errors for postgres PGO:
cd src/backend/replication/
marxin@marxinbox
This patch adds some selftesting of diagnostic_get_location_text,
since Nathan was looking at rewriting the insides.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
I also tested it with:
LANG=de_DE.UTF-8 make selftest
to avoid recurrences of PR bootstrap/71481 comment #8.
Commit
Dear All,
Please find attached the revised version of the patch following my
late realizations in yesterday's submission.
Cheers
Paul
On 1 November 2017 at 18:22, Paul Richard Thomas
wrote:
> Dear All,
>
> This patch is adequately described by the comment in the second chunk
> applied to reso
It's useful to not rely on global_dc in selftests, so this patch
moves class selftest::test_diagnostic_context from
diagnostic-show-locus.c to a new header and source file.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Committed to trunk as r254354.
gcc/ChangeLog:
* Make
OK.
On Thu, Nov 2, 2017 at 3:44 PM, Paolo Carlini wrote:
> Hi,
>
> On 02/11/2017 17:54, Jason Merrill wrote:
>>
>> Do we really need to add tsubst_flags_t to template_parm_to_arg and
>> friends? It should never produce an error.
>
> I see. I'm finishing testing the below, everything looks good s
GCC Maintainers:
I have addressed the issues that Segher mentioned. Per our discussions,
I have added new iterator VSX_XXBR to cover the needed integer and float
values with the exception of V16QI. The define_expand for revb uses the
new iterator to generate the needed Power 8 code. The V16QI i
Hi all,
The attached patch fixes the PR 82635. I tested it on all active
branches including the gcc-6 branch where we make extensive use of this
MD_FALLBACK_FRAME_STATE_FOR macro. Namely in the libjava test suite.
(Found no other suitable tests which do make use of this macro).
No regressions
Hi,
On 02/11/2017 17:54, Jason Merrill wrote:
Do we really need to add tsubst_flags_t to template_parm_to_arg and
friends? It should never produce an error.
I see. I'm finishing testing the below, everything looks good so far.
Looks ok?
Thanks,
Paolo.
/
/cp
2017-11-02 P
FT32B is a new FT32 architecture type. Ft32B has a code compression
scheme which uses linker relaxations. It also has a security option to
prevent reads from program memory.
[gcc]
2017-11-02 James Bowman
* config/ft32/ft32.c (ft32_addr_space_legitimate_address_p): increase
off
Ping: https://gcc.gnu.org/ml/gcc-patches/2017-10/msg01834.html
On 10/25/17, Eric Gallager wrote:
> On Sat, Sep 30, 2017 at 8:05 PM, Eric Gallager
> wrote:
>> On Fri, Sep 29, 2017 at 11:15 AM, David Malcolm
>> wrote:
>>> On Sun, 2017-09-17 at 20:00 -0400, Eric Gallager wrote:
Attached is a
On Thu, Nov 2, 2017 at 1:08 PM, Marek Polacek wrote:
> On Thu, Nov 02, 2017 at 09:53:33AM -0400, Jason Merrill wrote:
>> On Thu, Nov 2, 2017 at 8:21 AM, Richard Biener wrote:
>> > On Wed, 1 Nov 2017, Marek Polacek wrote:
>> >
>> >> On Fri, Oct 27, 2017 at 12:46:12PM +0200, Richard Biener wrote:
>
The Arm backend sets the default sched-pressure algorithm to
SCHED_PRESSURE_MODEL. Benchmarking on AArch64 shows this
speeds up floating point performance on SPEC - eg. CactusBSSN improves
by ~16%. The gains are mostly due to less spilling, so enable this on AArch64
by default.
OK for commit?
This patch fixes pr82710, were we erroneously warn on something like:
friend class_X (::other::name (...));
the parens are needed, otherwise the '::other' is taken to be a
qualified lookup inside the class_X.
Unfortunately, at the point we can check, we've lost information that
'::' was use
On 09/17/2017 11:44 AM, Tom Tromey wrote:
+@code{@w{__VA_OPT__}} is also available in GNU C and GNU C++.
+{ /* c99 c++ xnum xid c11 std digr ulit rlit udlit bincst digsep
trig u8chlit vaopt */
+ /* GNUC89 */ { 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,0, 0,
0, 0
On Thu, Nov 02, 2017 at 03:38:45PM +, Kyrill Tkachov wrote:
> this looks great! I have a couple of comments.
> * Can you please extend file comments for gimple-ssa-store-merging.c ?
> Currently it mostly describes how we merge constants together. Once we start
> accepting non-constant members
>
Ping!
Siddhesh
On 27 October 2017 at 18:13, Siddhesh Poyarekar wrote:
> From: Siddhesh Poyarekar
>
> This patch adds an mcpu option for the Qualcomm saphira server part.
> Tested on aarch64 and did not find any regressions resulting from this
> patch.
>
> Siddhesh
>
> 2017-10-27 Siddhe
Hi Honza.
Thanks for the huge patch. I'm willing to help you with testing, but I can't
apply the patch on top of r254348:
../../gcc/profile.c: In function ‘void compute_branch_probabilities(unsigned
int, unsigned int)’:
../../gcc/profile.c:794:11: error: ‘flag_guess_branch_probability’ was not
On Thu, Nov 02, 2017 at 09:53:33AM -0400, Jason Merrill wrote:
> On Thu, Nov 2, 2017 at 8:21 AM, Richard Biener wrote:
> > On Wed, 1 Nov 2017, Marek Polacek wrote:
> >
> >> On Fri, Oct 27, 2017 at 12:46:12PM +0200, Richard Biener wrote:
> >> > On Fri, 27 Oct 2017, Jakub Jelinek wrote:
> >> >
> >>
On Tue, Oct 31, 2017 at 5:45 AM, Paolo Carlini wrote:
> this ICE on valid seems rather easy to fix, one of those bugs where we
> aren't propagating the tsubst_flags_t argument. In this case, we aren't
> propagating from tsubst_pack_expansion to make_pack_expansion. Doing it,
> fixes the ICE and we
From: Kito Cheng
GCC may generate non-legitimate address due to we allow some
load/store with non-legitimate address in pic.md.
gcc/ChangeLog
2017-11-02 Kito Cheng
* config/riscv/riscv.c (riscv_legitimize_move): Handle
non-legitimate address.
---
gcc/config/riscv/riscv.c |
Tom> [ __VA_OPT__ ]
Tom> Here's v3.
Tom> Ping.
Tom> Ping #2.
Ping #3.
Tom
On Thu, Nov 02, 2017 at 03:38:45PM +, Kyrill Tkachov wrote:
> this looks great! I have a couple of comments.
> * Can you please extend file comments for gimple-ssa-store-merging.c ?
> Currently it mostly describes how we merge constants together. Once we start
> accepting non-constant members
>
On 11/02/2017 09:33 AM, Markus Trippelsdorf wrote:
On 2017.11.02 at 08:55 -0600, Jeff Law wrote:
As has been discussed on-list. This patch adds a virtual destructor to
the new classes in tree-ssa-propagate.h per our coding conventions and
what are considered best practices. It doesn't matter
Hi Jakub,
On 02/11/17 14:10, Jakub Jelinek wrote:
Hi!
The following patch improves store merging, so that it doesn't handle
just constant stores into adjacent memory, but also adjacent memory copying
and simple bitwise logical ops where at least one argument is a load
from adjacent memory and t
On 11/02/2017 09:31 AM, Richard Biener wrote:
On Thu, Nov 2, 2017 at 3:55 PM, Jeff Law wrote:
As has been discussed on-list. This patch adds a virtual destructor to the
new classes in tree-ssa-propagate.h per our coding conventions and what are
considered best practices. It doesn't matter fo
On 2017.11.02 at 08:55 -0600, Jeff Law wrote:
>
> As has been discussed on-list. This patch adds a virtual destructor to
> the new classes in tree-ssa-propagate.h per our coding conventions and
> what are considered best practices. It doesn't matter for any code I'm
> aware of today -- it's a
> It's possible to have a line of code that has a non-zero coverage.
> However, it can contain unexecuted blocks and I hope adding a
> notification can be usefull. LLVM also does that:
>
> -:0:Source:ternary.c
> -:0:Graph:ternary.gcno
> -:0:Data:ternary.gcda
>
On Thu, Nov 2, 2017 at 3:55 PM, Jeff Law wrote:
>
> As has been discussed on-list. This patch adds a virtual destructor to the
> new classes in tree-ssa-propagate.h per our coding conventions and what are
> considered best practices. It doesn't matter for any code I'm aware of
> today -- it's a
On Thu, Nov 2, 2017 at 2:11 PM, Marc Glisse wrote:
> On Thu, 2 Nov 2017, Richard Biener wrote:
>
>> On Wed, Nov 1, 2017 at 12:47 PM, Marc Glisse wrote:
>>>
>>> Hello,
>>>
>>> just a little tweak to that transformation. There is some overlap between
>>> the 2 versions, but it seemed easier to hand
As has been discussed on-list. This patch adds a virtual destructor to
the new classes in tree-ssa-propagate.h per our coding conventions and
what are considered best practices. It doesn't matter for any code I'm
aware of today -- it's a defensive measure.
This also drops the "virtual" key
On Thu, Nov 02, 2017 at 01:21:17PM +0100, Richard Biener wrote:
> On Wed, 1 Nov 2017, Marek Polacek wrote:
>
> > On Fri, Oct 27, 2017 at 12:46:12PM +0200, Richard Biener wrote:
> > > On Fri, 27 Oct 2017, Jakub Jelinek wrote:
> > >
> > > > On Fri, Oct 27, 2017 at 12:31:46PM +0200, Richard Biener w
On Tue, Oct 31, 2017 at 1:57 PM, H.J. Lu wrote:
> On Tue, Oct 24, 2017 at 8:26 PM, Jason Merrill wrote:
>> On Thu, Jul 27, 2017 at 3:50 PM, H.J. Lu wrote:
>>> execute_dwarf2_frame is called for each funtion. But create_cie_data
>>> is called only once to initialize cie_cfi_row for all functions
On Tue, Oct 31, 2017 at 5:09 PM, David Malcolm wrote:
> On Tue, 2017-10-24 at 09:53 -0400, Jason Merrill wrote:
>> On Fri, Oct 20, 2017 at 5:53 PM, David Malcolm
>> wrote:
>> > Design questions:
>> >
>> > * The patch introduces a new kind of tree node, currently called
>> > DECL_WRAPPER_EXPR (a
On Wed, Nov 1, 2017 at 4:45 PM, Mukesh Kapoor wrote:
> On 11/1/2017 1:02 PM, Jason Merrill wrote:
>>
>> On Tue, Oct 31, 2017 at 12:17 PM, Mukesh Kapoor
>> wrote:
>>>
>>> On 10/25/2017 6:44 PM, Mukesh Kapoor wrote:
On 10/25/2017 4:20 AM, Nathan Sidwell wrote:
>
> On 10/25/2017 12
On Thu, Nov 02, 2017 at 11:48:36AM +, Paul Richard Thomas wrote:
> Hi Steve,
>
> I read the correspondence on clf and your earlier posting here. With
> those in mind, the patch looks to be OK to commit.
>
> Thanks
>
Thanks. The c.l.f posting is addressing a different issue.
This patch simp
On 11/01/2017 04:08 PM, Jason Merrill wrote:
On Wed, Nov 1, 2017 at 2:29 PM, Nathan Sidwell wrote:
-&& IDENTIFIER_NEWDEL_OP_P (unqualified_id)))
+&& IDENTIFIER_OVL_OP_P (unqualified_id)
+&& (IDENTIFIER_OVL_OP_FLAGS (unqualified_id) & OVL_OP_FLAG_ALLOC)))
Why not keep the name ID
Hi!
The following patch improves store merging, so that it doesn't handle
just constant stores into adjacent memory, but also adjacent memory copying
and simple bitwise logical ops where at least one argument is a load
from adjacent memory and the other argument as well or a constant.
The loads ar
On 11/01/2017 03:40 PM, Jakub Jelinek wrote:
On Wed, Nov 01, 2017 at 03:24:56PM -0400, Nathan Sidwell wrote:
Thomas Koenig noticed -fdump-lang-raw gives an immediate ICE. I broke it
with the DECL_ASSEMBLER_NAME changes last month. Fixed thusly, applied as
obvious.
Shouldn't there be a testca
Hi,
core2 used to have quite large penalty for partial flag registers store
done by INCDEC. This was improved on Sandybridge where extra merging uop
is produced and more at Haswell where there is no extra uop unless there
is instruction accessing both. For this reason we can use inc/dec again
on
On Thu, Nov 2, 2017 at 8:21 AM, Richard Biener wrote:
> On Wed, 1 Nov 2017, Marek Polacek wrote:
>
>> On Fri, Oct 27, 2017 at 12:46:12PM +0200, Richard Biener wrote:
>> > On Fri, 27 Oct 2017, Jakub Jelinek wrote:
>> >
>> > > On Fri, Oct 27, 2017 at 12:31:46PM +0200, Richard Biener wrote:
>> > > >
PING^1
On 10/19/2017 01:36 PM, Martin Liška wrote:
> On 09/20/2017 10:15 AM, Jakub Jelinek wrote:
>> On Wed, Sep 20, 2017 at 09:50:32AM +0200, Martin Liška wrote:
>>> Hello.
>>>
>>> Following patch handles UBSAN (overflow) in dce.c.
>>
>> dse.c ;)
>>
>>> --- a/gcc/dse.c
>>> +++ b/gcc/dse.c
>>> @@
On 10/19/2017 01:58 PM, Richard Biener wrote:
> In the end I don't think we want to slow down code just for the sake of
> UBSAN. IMHO for code invoking undefined behavior (object too large)
> it's reasonable for the compiler to invoke undefined behavior
Agree that it's very artificial test-case.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2017-11-02 Richard Biener
PR tree-optimization/82795
* tree-if-conv.c (predicate_mem_writes): Remove bogus assert.
* gcc.target/i386/pr82795.c: New testcase.
Index: gcc/tree-if-conv.c
===
On Thu, 2 Nov 2017, Richard Biener wrote:
On Wed, Nov 1, 2017 at 12:47 PM, Marc Glisse wrote:
Hello,
just a little tweak to that transformation. There is some overlap between
the 2 versions, but it seemed easier to handle the NOP case (including the
case without convert and the vector case) s
On Wed, Nov 1, 2017 at 7:18 PM, Alexandre Oliva wrote:
> On Oct 31, 2017, Jeff Law wrote:
>
>> On 09/30/2017 03:08 AM, Alexandre Oliva wrote:
>>> This API change will enable final_start_function() to "consume"
>>> initial insns, and choose the first insn to be passed to final().
>>>
>>> Many port
On Wed, Nov 1, 2017 at 12:47 PM, Marc Glisse wrote:
> Hello,
>
> just a little tweak to that transformation. There is some overlap between
> the 2 versions, but it seemed easier to handle the NOP case (including the
> case without convert and the vector case) separately from the narrowing /
> sign
From: claziss
The 'aux' variable attribute is used to directly access the auxiliary
register space from C.
gcc/
2017-07-25 Claudiu Zissulescu
* config/arc/arc.c (arc_handle_aux_attribute): New function.
(arc_attribute_table): Add 'aux' attribute.
(arc_in_small_data_p)
The ARCv2 ISA provides the JLI instruction, which is two-byte instructions
that can be used to reduce code size in an application. To make use of it,
we provide two new function attributes 'jli_always' and 'jli_fixed' which
will force the compiler to call the indicated function using a jli_s
instru
gcc/
2017-02-20 Claudiu Zissulescu
* config/arc/arc-protos.h: Add arc_is_secure_call_p proto.
* config/arc/arc.c (arc_handle_secure_attribute): New function.
(arc_attribute_table): Add 'secure_call' attribute.
(arc_print_operand): Print secure call operand.
From: claziss
The _Uncached type qualifier can be used to bypass the cache without
resorting to declaring variables as volatile.
gcc/
2017-07-12 Claudiu Zissulescu
* config/arc/arc-protos.h (arc_is_uncached_mem_p): Function proto.
* config/arc/arc.c (arc_handle_uncached_attri
gcc/
2017-03-20 Claudiu Zissulescu
* config/arc/arc-arches.def: Option mrf16 valid for all
architectures.
* config/arc/arc-c.def (__ARC_RF16__): New predefined macro.
* config/arc/arc-cpus.def (em_mini): New cpu with rf16 on.
* config/arc/arc-options.def
From: claziss
Delegitimize address is used to undo the obfuscating effect of PIC
addresses, returning the address in a way which is understood by the
compiler.
gcc/
2017-04-25 Claudiu Zissulescu
* config/arc/arc.c (arc_delegitimize_address_0): Refactored to
recognize new pic
From: claziss
Hi,
This set of patches are adding support for:
- JLI/SJLI instructions;
- reduced register file 16 specific for particular ARC configurations;
- 'uncached' attribute, used to toggle .di flag on variable basis;
- 'aux' attribute, used to easily access auxiliary
On Wed, 1 Nov 2017, Marek Polacek wrote:
> On Fri, Oct 27, 2017 at 12:46:12PM +0200, Richard Biener wrote:
> > On Fri, 27 Oct 2017, Jakub Jelinek wrote:
> >
> > > On Fri, Oct 27, 2017 at 12:31:46PM +0200, Richard Biener wrote:
> > > > I fear it doesn't work at all with LTO (you'll always get the
This is patch which solves the ARC issues with exception handling support.
Ok to apply?
Claudiu
2016-06-09 Claudiu Zissulescu
Andrew Burgess
* config/arc/arc-protos.h (arc_compute_frame_size): Delete
declaration.
(arc_return_slot_offset): Likewise.
The documentation is right, I was wrong not adding SSE/AVX flags in these
builtin declaratuin.
> The exceptions are
> MMX, AVX512VL and 64BIT is also special.
> So, shall GFNI be added to that set?
Turns out only GFNI and VAES(haven't sent those yet, they are from the same
Icelake pdf) are lik
Richard Earnshaw wrote:
> On 01/11/17 17:40, James Greenhalgh wrote:
>> As far as I understand it, because we have 128-bit types, a malloc of
>> anything greater than 16 bytes would require 16-byte alignment. So, assuming
>> this macro isn't required to desribe possibly unaligned smaller allocatio
Hi Steve,
I read the correspondence on clf and your earlier posting here. With
those in mind, the patch looks to be OK to commit.
Thanks
Paul
On 2 November 2017 at 01:09, Steve Kargl
wrote:
> The attached patch fixes a regression where gfortran was
> issuing an error for an entity in a common
On Mon, 30 Oct 2017, Martin Sebor wrote:
> On 10/30/2017 02:56 PM, Richard Biener wrote:
> > On October 30, 2017 9:13:04 PM GMT+01:00, Martin Sebor
> > wrote:
> > > On 10/30/2017 01:53 PM, Richard Biener wrote:
> > > > On October 30, 2017 4:19:25 PM GMT+01:00, Martin Sebor
> > > wrote:
> > > > >
On Mon, 30 Oct 2017, Jeff Law wrote:
> On 10/30/2017 05:29 PM, Martin Sebor wrote:
> > On 10/30/2017 03:48 PM, Jeff Law wrote:
> >> On 10/30/2017 09:19 AM, Martin Sebor wrote:
> >>> On 10/30/2017 05:45 AM, Richard Biener wrote:
> On Sun, 29 Oct 2017, Martin Sebor wrote:
>
> > In my w
On 01/11/17 17:40, James Greenhalgh wrote:
> On Tue, Oct 31, 2017 at 05:07:54PM +, Wilco Dijkstra wrote:
>> The AArch64 backend currently doesn't set MALLOC_ABI_ALIGNMENT, so
>> add this to enable alignment optimizations on malloc pointers.
>>
>> OK for commit?
>
> As far as I understand it, b
On 31/10/17 16:53, Steve Ellcey wrote:
> On Tue, 2017-10-31 at 09:57 +, Richard Earnshaw (lists) wrote:
>>
>> This is looking better...
>>
>> I may have missed some discussion on this topic, but what's the
>> reasoning behind changing the quoting around the 'str' parameter
>> value in
>>
>> -
There's some potential to cleanup Solaris linker version checks:
* gcc_AC_INITFINI_ARRAY in acinclude.m4 is unused, but still serves as
documentation for the various variants of ld version strings. This is
now moved to configure.ac.
* Solaris 11.4 has (or shortly will have) some changes to t
Hi Sandra,
> On 10/27/2017 05:48 AM, Rainer Orth wrote:
>> I happened to notice that install.texi still contains references to the
>> Tru64 UNIX and IRIX ports I've removed in GCC 4.8. I believe it's time
>> now to get rid of those completely.
>>
>> Tested with make doc/gccinstall.info and doc/gc
On Mon, Jan 14, 2013 at 10:51:25PM +0100, Thomas Koenig wrote:
> Am 14.01.2013 14:29, schrieb Mikael Morin:
> > Le 13/01/2013 23:14, Thomas Koenig a écrit :
>
> > > OK for trunk?
> > >
> > OK with the changes suggested above. Thanks.
>
> Committed as rev. 195179 with your changes.
s/intrnisics/
Hi Jeff,
> On 10/27/2017 05:48 AM, Rainer Orth wrote:
>> I happened to notice that install.texi still contains references to the
>> Tru64 UNIX and IRIX ports I've removed in GCC 4.8. I believe it's time
>> now to get rid of those completely.
>>
>> Tested with make doc/gccinstall.info and doc/gcc
> > gcc/
> > 2017-09-01 Claudiu Zissulescu
> >
> > * config/arc/arc.c (hwloop_optimize): Account for empty
> > body loops.
>
> Looks good to me.
>
> Thanks,
> Andrew
>
>
Pushed. Thank you for your review,
Claudiu
Hi,
This patch makes "prefer-avx256" option as default tuning for "skylake-avx512".
This is due to better performance of 256-bit code for some of the cases. In
case of
Skylake Server the Optimization Manual has following "Since port 0 and port 1
are 256-bits wide,
Intel AVX-512 operations that
On Fri, Oct 27, 2017 at 02:31:35PM +0100, Richard Sandiford wrote:
> This patch records the integer value of the address offset in
> aarch64_address_info, so that it doesn't need to be re-extracted
> from the rtx. The SVE port will make more use of this. The patch
> also uses poly_int64 routines
On Fri, Oct 27, 2017 at 02:29:30PM +0100, Richard Sandiford wrote:
> This patch passes the number of units to aarch64_expand_vec_perm
> and aarch64_expand_vec_perm_const, which avoids a to_constant ()
> once GET_MODE_NUNITS is variable.
OK.
Reviewed-by: James Greenhalgh
Thanks,
James
> 2017-10
On Fri, Oct 27, 2017 at 02:28:57PM +0100, Richard Sandiford wrote:
> This patch passes the number of units to aarch64_simd_vect_par_cnst_half,
> which avoids a to_constant () once GET_MODE_NUNITS is variable.
OK.
Reviewed-by: James GReenhalgh
Thanks,
James
> 2017-10-27 Richard Sandiford
>
On Fri, Oct 27, 2017 at 02:28:27PM +0100, Richard Sandiford wrote:
> This patch passes the number of units to aarch64_reverse_mask,
> which avoids a to_constant () once GET_MODE_NUNITS is variable.
OK
Reviewed-by: James Greenhalgh
Thanks,
James
>
>
> 2017-10-26 Richard Sandiford
>
On Fri, Oct 27, 2017 at 02:27:50PM +0100, Richard Sandiford wrote:
> Later patches turn the number of vector units into a poly_int.
> We deliberately don't support applying GEN_INT to those (except
> in target code that doesn't disguish between poly_ints and normal
> constants); gen_int_mode needs
[ was: Re: [PATCH, i386]: Implement attribute ((naked)) ]
On 07/30/2017 10:14 PM, Uros Bizjak wrote:
Index: testsuite/gcc.target/i386/naked-1.c
+/* { dg-final { scan-assembler "ud2" } } */
+/* { dg-final { scan-assembler-not "ret" } } */
Index: testsuite/gcc.target/i386/naked-2.c
+/* {
James Greenhalgh writes:
> On Fri, Oct 27, 2017 at 02:22:39PM +0100, Richard Sandiford wrote:
>> This patch replaces switch statements that call specific generator
>> functions with code that constructs the rtl pattern directly.
>> This seemed to scale better to SVE and also seems less error-prone
[ was: Re: [Patch, fortran] PR35339 Optimize implied do loops in io
statements ]
On 06/03/2017 03:48 PM, Nicolas Koenig wrote:
! { dg-final { scan-tree-dump-times "while" 7 "original" } }
Hi,
I build and tested gcc in a directory with "while" in the path. That
made this scan fail (with 25
99 matches
Mail list logo
