Hi Sandra,
> 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
On Fri, Nov 3, 2017 at 6:38 AM, Andrew Pinski wrote:
> 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
Hi,
as explained in
https://gcc.gnu.org/ml/gcc-patches/2017-11/msg00105.html
we don't (necessarily) want to report non-executed blocks in Ada.
The couple of attached patches prevent this from happening by detecting
whether we are dealing with an Ada file; the first variant is optimal, the
Hi.
Honza can you please take a look at this, because Richi installed patch
that I originally suggested and you were not happy about: r251333.
Thanks,
Martin
On 3 November 2017 at 15:50, Richard Earnshaw
wrote:
>> 2017-10-27 Siddhesh Poyarekar
>> Jim Wilson
>>
>> gcc/
>> * config/aarch64/aarch64-cores.def (saphira): New.
>>
Richard Biener wrote:
> On Fri, Nov 3, 2017 at 6:38 AM, Andrew Pinski wrote:
> > 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
On Fri, Nov 3, 2017 at 9:52 AM, Martin Liška wrote:
> Hi.
>
> This is oneliner that fixes PGO bootstrap. I've discussed that with Richi and
> the core is correct.
> However we probably don't have an attribute that will ignore the warning?
I think
wide_int res = res;
might do
On Thu, Nov 2, 2017 at 10:28 PM, Martin Sebor wrote:
> 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,
On Fri, Nov 3, 2017 at 6:15 AM, Prathamesh Kulkarni
wrote:
> 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
On 03/11/17 09:29 +, Jonathan Wakely wrote:
On 02/11/17 18:00 -0400, Tim Song wrote:
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
On Fri, Nov 3, 2017 at 4:49 AM, Jeff Law wrote:
>
>
>
> 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
>
> I see no documentation here.
>
Ups, forgot this one :) Please find it attached. I'll merge it into the final
patch when everything is approved.
Thanks,
Claudiu
0001-ARC-DOC-Add-uncached-documentation.patch
Description: 0001-ARC-DOC-Add-uncached-documentation.patch
Hi,
When fixing PR70754, I thought the issue only happens for ZERO-length chains.
Well, that's apparently not true with PR82726.
The whole story is, with chain combination/re-association, new stmts may be
created/inserted at position not dominating following uses. This happens in
two scenarios:
Hi,
As described in message of previous patch:
This patch set fixes both PRs in the opposite way: Instead of find dominance
insertion position for root reference, we resort zero-distance references of
combined chain by their position information so that new root reference must
dominate others.
On Fri, Nov 3, 2017 at 2:06 AM, Marc Glisse wrote:
> 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.
-
On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
>>* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
>>(ix86_adjust_stack_and_probe_stack_clash): Use a push/pop sequence
>>to probe at the start of a noreturn function.
>>
>>
On Fri, Nov 3, 2017 at 11:14 AM, Richard Biener
wrote:
> On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
>>>* config/i386/i386.c (ix86_emit_restore_reg_using_pop):
>>> Prototype.
>>>
Status
==
The feature development phase of GCC 8, Stage 1, is coming to its ends
at Friday, Nov. 17th (as usual you can use your local timezone to your
own advantage).
This means that from Saturday, Nov. 18th we will be in Stage 3 which
allows for general bugfixing. All feature
Hi,
This is a simple patch exploiting more undefined pointer overflow behavior in
loop niter analysis. Originally, it only supports POINTER_PLUS_EXPR if the
offset part is IV. This patch also handles the case if pointer is IV. With
this patch, the while(true) loop in test can be removed by
Hi,
I ran into this memory leak issue in tree-predcom.c when investigating other
PRs.
This is the obvious fix by freeing reference of trivial component.
Bootstrap and test on x86_64. Is it OK?
Thanks,
bin
2017-11-02 Bin Cheng
* tree-predcom.c
On 02/11/17 18:00 -0400, Tim Song wrote:
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."
Because I forgot to implement
On 27/10/17 13:43, 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
> - Fix to unwinding. Now is is possible to unwind from syscall
> wrappers, signals and functions with dynamic stack allocation.
>
> - Patch also fixes millicode. Although millicode save and restore functions
> would change blink, the calls to those functions were not clobbering blink.
>
On 10/24/2017 04:19 PM, Jason Merrill wrote:
> On 10/18/2017 09:07 AM, Martin Liška wrote:
>> @@ -1182,7 +1182,13 @@ cxx_eval_builtin_function_call (const constexpr_ctx
>> *ctx, tree t, tree fun,
>> {
>> new_call = build_call_array_loc (EXPR_LOCATION (t), TREE_TYPE (t),
>>
On Thu, Nov 02, 2017 at 02:50:17PM -0400, Jason Merrill wrote:
> We probably want to call them something like default_is_empty_type and
> default_is_empty_record, though.
Done.
> > For one thing I thought we should be consistent in treating these two
> > structs, regarding being empty:
> >
> >
Tested as usual, committing to trunk.
Segher
2017-11-03 Segher Boessenkool
* config/rs6000/rs6000.md (move_from_CR_ov_bit): Change condition to
TARGET_PAIRED_FLOAT.
---
gcc/config/rs6000/rs6000.md | 2 +-
1 file changed, 1 insertion(+), 1
On 11/03/2017 04:05 AM, Richard Biener wrote:
On Fri, Nov 3, 2017 at 4:49 AM, Jeff Law wrote:
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
On Thu, Nov 02, 2017 at 06:41:58PM +, 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
On Fri, 3 Nov 2017, Jakub Jelinek wrote:
> On Fri, Nov 03, 2017 at 02:14:39PM +0100, Richard Biener wrote:
> > > +/* Return true if stmts in between FIRST (inclusive) and LAST (exclusive)
> > > + may clobber REF. FIRST and LAST must be in the same basic block and
> > > + have non-NULL vdef.
Someone noticed I'd not updated the error message when cloneing
scan-tree-dump into scan-lang-dump.
Fixed thusly and applied
nathan
--
Nathan Sidwell
2017-11-03 Nathan Sidwell
* lib/scanlang.exp: Fix error message to refer to scan-lang-dump.
Index: lib/scanlang.exp
On Fri, Nov 3, 2017 at 10:25 AM, Martin Liška wrote:
> On 10/27/2017 09:44 PM, Nathan Sidwell wrote:
>> On 10/27/2017 02:34 PM, Jakub Jelinek wrote:
>>
>>> But when singly inheriting a polymorphic base and thus mapped to the same
>>> vptr all but the last dtor will not be in
On Thu, Oct 26, 2017 at 8:14 AM, Martin Liška wrote:
> On 10/24/2017 04:39 PM, Jason Merrill wrote:
>> On 10/18/2017 08:48 AM, Martin Liška wrote:
>>> This is second patch that addresses test-suite fallout. All these tests
>>> fail because -Wreturn-type is
>>> now on by default.
On Fri, Nov 03, 2017 at 03:25:25PM +0100, Martin Liška wrote:
> On 10/27/2017 09:44 PM, Nathan Sidwell wrote:
> > On 10/27/2017 02:34 PM, Jakub Jelinek wrote:
> >
> >> But when singly inheriting a polymorphic base and thus mapped to the same
> >> vptr all but the last dtor will not be in charge,
A recent change to remove the movdi_vfp_cortexa8 meant that ldrd was used in
ITs block even when arm_restrict_it was enabled. Rather than just fixing this
latent issue, change the default of predicable_short_it to "no" so that only
16-bit instructions need to be marked with it. As a result there
In this PR we tried to create a widening multiply of two 3-bit numbers,
but that isn't a widening multiply at the optab/rtl level, since both
the input and output still have the same mode.
We could trap this either in is_widening_mult_p or (as the patch does)
in the routines that actually ask for
On Fri, Nov 3, 2017 at 1:36 PM, Bin Cheng wrote:
> Hi,
> I ran into this memory leak issue in tree-predcom.c when investigating other
> PRs.
> This is the obvious fix by freeing reference of trivial component.
> Bootstrap and test on x86_64. Is it OK?
Ok.
Thanks,
Richard.
Hi,
On 02/11/2017 15:42, Jason Merrill wrote:
This is a good suggestion. I have attached the revised patch. Thanks.
OK, thanks!
Thanks Jason.
I was about to volunteer committing the patch but then noticed that the
testcase includes quite a lot, eg, too, which we never
include in the
Committed.
On Thu, 02 Nov 2017 09:03:19 PDT (-0700), Palmer Dabbelt wrote:
> 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
On 03/11/17 12:45, Siddhesh Poyarekar wrote:
> On 3 November 2017 at 15:50, Richard Earnshaw
> wrote:
>>> 2017-10-27 Siddhesh Poyarekar
>>> Jim Wilson
>>>
>>> gcc/
>>> *
The rs6000 port currently has an *lt0_disi define_insn, setting the DI
result to whether the SI argument is negative or not. It turns out the
generic optimisers cannot always figure out in the other cases either
that this is just a shift for us. This patch adds patterns for all
four SI/DI
This series of patches generalises the vector testsuite and makes
it cope better with arbitrary vector lengths. It also adds some
target selectors needed for SVE.
Tested on aarch64-linux-gnu without SVE, with various fixed-length
SVE modes, and with the default variable-length SVE mode. Also
This patch adds a target selector for targets whose
preferred_vector_alignment is the alignment of one element. We'll never
peel in that case, and the step of a loop that operates on normal (as
opposed to packed) elements will always divide the preferred alignment.
2017-11-03 Richard Sandiford
This patch adds a target selector that says whether we can ever
generate an "unaligned" accesses, where "unaligned" is relative
to the target's preferred vector alignment. This is already true if:
vect_no_align && { ! vect_hw_misalign }
i.e. if the target doesn't have any alignment mechanism
A general TARGET_MEM_REF is:
BASE + STEP * INDEX + INDEX2 + OFFSET
After classifying the address in this way, the code that builds
TARGET_MEM_REFs tries to simplify the address until it's valid
for the current target and for the mode of memory being addressed.
It does this in a fixed order:
extract_bit_field_1 tries to use vec_extract to extract part of a
vector. However, if that pattern isn't defined or if the operands
aren't suitable, another good approach is to try a direct subreg
reference. This is particularly useful for multi-vector modes on
SVE (e.g. when extracting one
PR jit/82826 reports a crash when running jit.dg/test-benchmark.c,
introduced by r254140
(aka "Extend ipa-pure-const pass to propagate malloc attribute.")
I see the crash on the 28th of 400 in-process iterations of the
compiler; on turning on GCC_JIT_BOOL_OPTION_SELFCHECK_GC, it shows up
on the
On Fri, Nov 03, 2017 at 10:22:12AM -0700, Andi Kleen wrote:
>
> It looks like some non POSIX symbols got removed from the header
> files, which breaks the libsanitizer build.
nm, looks like i was on a old checkout. Seems to be already fixed
in current trunk.
-Andi
Several vector tests are sensitive to the vector size. This patch adds
a VECTOR_BITS macro to tree-vect.h to select the expected vector size
and uses it to influence iteration counts and array sizes. The tests
keep the original values if the vector size is small enough.
For now VECTOR_BITS is
From: Michael Clark
This changes makes GCC asm output use instruction names that are
consistent with the RISC-V ISA manual. The assembler accepts
immediate-operand instructions without the "i" suffix, so this all
worked before, it's just a bit cleaner to match the ISA
This patch adds a routine that lists the available vector sizes
for a target and uses it for some existing target conditions.
Later patches add more uses.
The cases are taken from multiple_sizes.
2017-11-03 Richard Sandiford
Alan Hayward
Hi,
On Thu, Oct 26, 2017 at 02:43:02PM +0200, Richard Biener wrote:
> On Thu, Oct 26, 2017 at 2:18 PM, Martin Jambor wrote:
> >
> > Nevertheless, I still intend to experiment with the limit, I sent out
> > this RFC exactly so that I don't spend a lot of time benchmarking
> >
It looks like some non POSIX symbols got removed from the header
files, which breaks the libsanitizer build.
struct sigaltstack now only exists as stack_t (which is the offical
POSIX name)
__res_state typedef is now only struct __res_state
This fixes bootstrap of trunk on a current opensuse
The vectoriser tests used a combination of:
1) if (impossible condition) abort ();
2) volatile int x; ... *x = ...;
3) asm volatile ("" ::: "memory");
to prevent vectorisation of a set-up loop. The problem with 1) is that
the compiler can often tell that the condition is false and optimise
it
From: Andrew Waterman
2017-11-03 Andrew Waterman
* config/riscv/riscv.c (riscv_option_override): Conditionally set
TARGET_STRICT_ALIGN based upon -mtune argument.
---
gcc/config/riscv/riscv.c | 10 +++---
1 file changed, 7
This patch adds a target selector to say whether it's possible to
align a local variable to the target's preferred vector alignment.
This can be false for large vectors if the alignment is only
a preference and not a hard requirement (and thus if there is no
need to support a stack realignment
On 11/03/2017 09:01 AM, Jeff Law wrote:
On 11/03/2017 04:05 AM, Richard Biener wrote:
On Fri, Nov 3, 2017 at 4:49 AM, Jeff Law wrote:
Several passes which perform dominator walks want to identify when
block has
a single incoming edge, ignoring loop backedges.
I'm aware
On 11/03/2017 04:46 AM, Uros Bizjak wrote:
On Fri, Nov 3, 2017 at 11:14 AM, Richard Biener
wrote:
On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
On Fri, Nov 3, 2017 at 6:13 PM, Jeff Law wrote:
> On 11/03/2017 04:46 AM, Uros Bizjak wrote:
>>
>> On Fri, Nov 3, 2017 at 11:14 AM, Richard Biener
>> wrote:
>>>
>>> On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
>
>
This is a tiny feature, small but perfectly formed. And easy to
implement.
* include/std/type_traits (endian): Define new enumeration type.
* testsuite/20_util/endian/1.cc: New test.
Tested powerpc64le-linux, committed to trunk.
commit ad69c5875cdefc5c699b50975ee6016a424d450b
SLP load permutation fails if any individual permutation requires more
than two vector inputs. For 128-bit vectors, it's possible to permute
3 contiguous loads of 32-bit and 8-bit elements, but not 16-bit elements
or 64-bit elements. The results are reversed for 256-bit vectors,
and so on for
On Fri, Nov 3, 2017 at 9:55 AM, Marek Polacek wrote:
> + TYPE_EMPTY_P (t) = targetm.calls.empty_record_p (t);
I think we want to set this in finalize_type_size; since the point of
all this is becoming compliant with the psABI (and compatible with the
C front end), I wouldn't
This patch allows us to recognise:
... = bool1 != bool2 ? x : y
as equivalent to:
bool tmp = bool1 ^ bool2;
... = tmp ? x : y
For the latter we were already able to find the natural number
of vector units for tmp based on the types that feed bool1 and
bool2, whereas with the former
This patch improves the ivopts address cost calculcation for modes
in which an index must be scaled rather than unscaled. Previously
we would only try the scaled form if the unscaled form was valid.
Many of the SVE tests rely on this when matching scaled indices.
Tested on aarch64-linux-gnu,
Once SVE is enabled, a general AArch64 spill slot offset will be
A + B * VL
where A is a constant and B is a multiple of the SVE vector length.
The offsets in SVE load and store instructions are a multiple of VL
(and so can encode some values of B), while offsets for base AArch64
load and
Hello!
This testcase uses undefined values. Attached patch fixes the
testcase; the scan for message in the RTL dump assures us that the fix
for PR 70263 still performs its magic.
2017-11-03 Uros Bizjak
PR testsuite/82828
PR rtl-optimization/70263
*
On 11/03/2017 04:14 AM, Richard Biener wrote:
On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
(ix86_adjust_stack_and_probe_stack_clash): Use a push/pop sequence
to probe
Some tests assumed that there would only be 2 vector sizes if
vect_multiple_sizes, whereas for SVE there are three (SVE, 128-bit
and 64-bit). This patch replaces scan-tree-dump-times with
scan-tree-dump for vect_multiple_sizes but keeps it for
!vect_multiple_sizes.
2017-11-03 Richard Sandiford
This patch adds a target selector for variable-length vectors.
Initially it's always false, but the SVE patch provides a case
in which it's true.
2017-11-03 Richard Sandiford
Alan Hayward
David Sherwood
This patch adds a target selector that says whether the target
supports IFN_MASK_STORE.
2017-11-03 Richard Sandiford
Alan Hayward
David Sherwood
gcc/
* doc/sourcebuild.texi
James Greenhalgh wrote:
>
> This caused:
>
> Failures:
> gcc.target/aarch64/test_frame_4.c
> gcc.target/aarch64/test_frame_2.c
> gcc.target/aarch64/test_frame_7.c
> gcc.target/aarch64/test_frame_10.c
Sorry, I missed that in testing. I've reverted part of the patch
Hi,
this ICE on valid (given GNU's designated initializers) is rather simple
to analyze: for the testcase, the gcc_assert in
process_init_constructor_array triggers because at that time INDEX1 is
still a CONST_DECL, not an INTEGER_CST. As regards fixing the problem, I
immediately noticed
Instead of calling rs6000_emit_sISEL, call rs6000_emit_int_cmove
directly, in the one place it is used.
Tested as usual; committing to trunk.
Segher
2017-11-03 Segher Boessenkool
* config/rs6000/rs6-protos.h (rs6000_emit_sISEL): Delete.
From: Andrew Waterman
When implementing the RISC-V port, I took the name of this macro at
face value. It appears we were mistaken in what this means, here's a
quote from the SPARC port that better describes what SLOW_BYTE_ACCESS
does
/* Nonzero if access to memory by
On 11/03/2017 02:38 AM, Uros Bizjak wrote:
* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
(ix86_adjust_stack_and_probe_stack_clash): Use a push/pop sequence
to probe at the start of a noreturn function.
*
Almost all targets add an explict -fomit-frame-pointer in the target specific
options. Rather than doing this in a target-specific way, do this in the
generic options so it works identically across all targets. In many cases the
target no longer needs to define TARGET_OPTION_OPTIMIZATION_TABLE,
On 11/03/2017 04:46 AM, Uros Bizjak wrote:
On Fri, Nov 3, 2017 at 11:14 AM, Richard Biener
wrote:
On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak wrote:
* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
On Thu, 2 Nov 2017, David Malcolm wrote:
> +{"offsetof", {"", ""} },
offsetof is in stddef.h for C, not stdalign.h.
--
Joseph S. Myers
jos...@codesourcery.com
The Changelog entry and patch are somewhat self-explanatory.
2017-11-3 Steven G. Kargl
* gfortran.dg/large_real_kind_2.F90: Test passes on FreeBSD. Remove
dg-xfail-if directive.
Index: gcc/testsuite/gfortran.dg/large_real_kind_2.F90
This was already fixed by the patch for PR82375 on trunk, albeit less
well than the patch applied here. I will correct trunk tomorrow.
Committed as 'obvious' on 6-branch(r254393) and 7-branch(r254389)
after bootstrapping and regtesting.
Cheers
Paul
2017-11-03 Paul Thomas
On 11/03/2017 11:38 AM, Uros Bizjak wrote:
On Fri, Nov 3, 2017 at 6:13 PM, Jeff Law wrote:
On 11/03/2017 04:46 AM, Uros Bizjak wrote:
On Fri, Nov 3, 2017 at 11:14 AM, Richard Biener
wrote:
On Fri, Nov 3, 2017 at 9:38 AM, Uros Bizjak
This patch adds new target selectors for SVE and updates existing
selectors accordingly. It also XFAILs some tests that don't yet
work for some SVE modes; most of these go away with follow-on
vectorisation enhancements.
2017-11-03 Richard Sandiford
This patch adds gcc.target/aarch64 tests for SVE, and forces some
existing Advanced SIMD tests to use -march=armv8-a.
2017-11-03 Richard Sandiford
Alan Hayward
David Sherwood
On Fri, 3 Nov 2017, Wilco Dijkstra wrote:
> Almost all targets add an explict -fomit-frame-pointer in the target specific
> options. Rather than doing this in a target-specific way, do this in the
Which targets do not? You should explicitly list them and CC their
maintainers and seek
This series adds support for ARM's Scalable Vector Extension.
More details on SVE can be found here:
https://developer.arm.com/products/architecture/a-profile/docs/arm-architecture-reference-manual-supplement-armv8-a
There are four parts for ease of review, but it probably makes
sense to
On Fri, 2017-11-03 at 17:49 +, Joseph Myers wrote:
> On Thu, 2 Nov 2017, David Malcolm wrote:
>
> > +{"offsetof", {"", ""} },
>
> offsetof is in stddef.h for C, not stdalign.h.
Thanks. Here's an updated version of the patch which fixes that.
OK for trunk? (assuming bootstrap and
On 11/03/2017 05:22 AM, Claudiu Zissulescu wrote:
I see no documentation here.
Ups, forgot this one :) Please find it attached. I'll merge it into the final
patch when everything is approved.
Thanks,
Claudiu
+@node ARC Type Attributes
+@subsection ARC Type Attributes
+
+@cindex
On November 3, 2017 8:17:30 PM GMT+01:00, Jakub Jelinek
wrote:
>On Fri, Nov 03, 2017 at 03:04:18PM +0100, Jakub Jelinek wrote:
>> single-use vs. multiple uses is something I've thought about, but
>don't
>> know whether it is better to require single-use or not (or sometimes,
>>
3. nov. 2017 20:30 je oseba "Jeff Law" napisala:
>
> On 11/03/2017 11:38 AM, Uros Bizjak wrote:
>>
>> On Fri, Nov 3, 2017 at 6:13 PM, Jeff Law wrote:
>>>
>>> On 11/03/2017 04:46 AM, Uros Bizjak wrote:
On Fri, Nov 3, 2017 at 11:14 AM, Richard
Here is the solution I propose:
gcc/
* common/config/i386/i386-common.c
(OPTION_MASK_ISA_GENERAL_REGS_ONLY_UNSET): Remove MPX from flag.
(ix86_handle_option): Move MPX to isa_flags2 and GFNI to isa_flags.
* config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
This patch adds support for ARM's Scalable Vector Extension.
The patch just contains the core features that work with the
current vectoriser framework; later patches will add extra
capabilities to both the target-independent code and AArch64 code.
The patch doesn't include:
- support for
On Wed, Nov 1, 2017 at 9:39 AM, H.J. Lu wrote:
> On Wed, Nov 1, 2017 at 9:32 AM, Rich Felker wrote:
>> On Sun, Oct 15, 2017 at 06:16:57AM -0700, H.J. Lu wrote:
>>> crt1.o is used to create dynamic and non-PIE static executables. Static
>>> PIE needs to link
On 11/03/2017 10:54 AM, Wilco Dijkstra wrote:
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index
71b2445f70fd5b832c68c08e69e71d8ecad37a4a..1c56f4b12495fe97c604200ef245c9fa02684b0f
100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -7436,16 +7436,17 @@ machine-description
This patch adds support for unwinding frames that use the SVE
pseudo VG register. We want this register to act like a normal
register if the CFI explicitly sets it, but want to provide a
default value otherwise. Computing the default value requires
an SVE target, so we only want to compute it on
On Thu, Oct 26, 2017 at 6:17 AM, Paolo Carlini wrote:
> Hi again,
>
> On 24/10/2017 20:58, Jason Merrill wrote:
>>
>> This seems like an odd place to add the complete_type call. What
>> happens if we change the COMPLETE_TYPE_P (type) in
>> cp_apply_type_quals_to_decl to
On Fri, Nov 03, 2017 at 03:04:18PM +0100, Jakub Jelinek wrote:
> single-use vs. multiple uses is something I've thought about, but don't
> know whether it is better to require single-use or not (or sometimes,
> under some condition?). Say if we have:
So, here is what I've committed in the end
> But what do you think about adding AVX/SSE flags to this special set instead?
Ok, was wrong, it is impossible to add SSE, because it is used in normal "or"
way. Then I'll add GFNI/VAES instead.
There is also another problem there: GFNI belongs to isa_flags2, while
AVX512VL/AVX/SSE belong to
>* config/i386/i386.c (ix86_emit_restore_reg_using_pop): Prototype.
>(ix86_adjust_stack_and_probe_stack_clash): Use a push/pop sequence
>to probe at the start of a noreturn function.
>
>* gcc.target/i386/stack-check-12.c: New test
-
Hello!
-ENOCHANGELOG
diff --git a/gcc/testsuite/gcc.c-torture/execute/pr82788.c
b/gcc/testsuite/gcc.c-torture/execute/pr82788.c
new file mode 100644
index 000..ceaa25f
--- /dev/null
+++ b/gcc/testsuite/gcc.c-torture/execute/pr82788.c
@@ -0,0 +1,2 @@
+
+int main() { int a[1442]; return 0;}
Hi.
This is oneliner that fixes PGO bootstrap. I've discussed that with Richi and
the core is correct.
However we probably don't have an attribute that will ignore the warning?
Only option is to push/pop Wuninitialized warning.
Ready for trunk?
Thanks,
Martin
gcc/ChangeLog:
2017-11-03 Martin
On 11/03/2017 04:22 PM, Daniel Santos wrote:
> ...
> How does this patch look? (Also, I've updated comments for
> choose_baseaddr.) Currently re-running tests.
>
> Thanks,
> Daniel
>
> @@ -13110,10 +13125,26 @@ ix86_expand_prologue (void)
>target. */
>if (TARGET_SEH)
>
1 - 100 of 116 matches
Mail list logo