Small fix.
gcc/
* config.gcc (icelake-client, icelake-server): New.
(icelake): Remove.
* config/i386/i386.c (initial_ix86_tune_features): Extend to 64 bit.
(initial_ix86_arch_features): Ditto.
(PTA_SKYLAKE): Add SGX.
(PTA_ICELAKE): Remove.
(P
Gentle ping.
> -Original Message-
> From: Koval, Julia
> Sent: Monday, February 12, 2018 10:57 AM
> To: Kirill Yukhin
> Cc: 'GCC Patches'
> Subject: RE: [x86,avx] Fix __builtin_cpu_supports for icelake and cannonlake
> isa
>
> Hi,
>
> There is no PR for this. This builtin was just mis
On Mar 12, 2018, "Bin.Cheng" wrote:
> internal compiler error: in aarch64_classify_address, at
> config/aarch64/aarch64.c:5678
> 0xfe3c29 aarch64_classify_address
> /.../build/src/gcc/gcc/config/aarch64/aarch64.c:5677
> 0xfe8be8 aarch64_legitimate_address_hook_p
> /.../build/src/gcc/gcc/c
On 8 February 2018 at 14:54, Richard Sandiford
wrote:
> One advantage of the new permute handling compared to the old way is
> that we can now easily take advantage of the vectoriser's divmod patterns
> for SVE.
>
> I realise we're in stage 4, but this is entirely SVE-specific.
>
> Tested on aarch
On Tue, Mar 13, 2018 at 9:30 PM, Jakub Jelinek wrote:
> Hi!
>
> As mentioned in bugzilla, when e.g. sel-sched queries (indirectly) before
> reload
> some attributes like get_attr_type that depend on alternatives, GCC attempts
> to constrain the operands in non-strict mode, which implies that if
>
On Wed, Mar 14, 2018 at 9:36 AM, Uros Bizjak wrote:
> On Tue, Mar 13, 2018 at 9:30 PM, Jakub Jelinek wrote:
>> Hi!
>>
>> As mentioned in bugzilla, when e.g. sel-sched queries (indirectly) before
>> reload
>> some attributes like get_attr_type that depend on alternatives, GCC attempts
>> to const
Hi!
The following testcase ICEs on aarch64-linux, because combiner matches the
(insn 25 24 26 2 (set (reg:DI 114)
(rotatert:DI (reg:DI 115 [ d ])
(subreg:QI (neg:SI (and:SI (reg:SI 112)
(const_int 65535 [0x]))) 0))) "pr84845.c":8 664
{*aarch64_reg_d
On Tue, 13 Mar 2018, Jakub Jelinek wrote:
> Hi!
>
> The sanopt maybe_optimize_ubsan_ptr_ifn optimization behaves differently
> on 32-bit and on 64-bit targets when using similar arguments maximum or
> minimum of ptrdiff_t or values close to it.
>
> The problem is that UHWI is 64-bit, regardless
On Wed, Mar 14, 2018 at 8:44 AM, Koval, Julia wrote:
> Small fix.
>
> gcc/
> * config.gcc (icelake-client, icelake-server): New.
> (icelake): Remove.
> * config/i386/i386.c (initial_ix86_tune_features): Extend to 64 bit.
> (initial_ix86_arch_features): Ditto.
>
I noticed that data-sym-pool.c fails on -O0 flags.
-O0 output :
-cut--
frob:
.frame $17,8,$31 # vars= 0, regs= 1/0, args= 0, gp= 0
.mask 0x0002,0
.fmask 0x,0
addiu $sp,-8
sd $17,0($sp)
Christophe Lyon writes:
> On 8 February 2018 at 14:54, Richard Sandiford
> wrote:
>> Index: gcc/testsuite/gcc.target/aarch64/sve/mul_highpart_1_run.c
>> ===
>> --- /dev/null 2018-02-08 11:17:10.862716283 +
>> +++ gcc/testsuite/
Hi.
Current code that tracks memory statistics wrongly decremented a counter
based on t_kind, which is an enum value and thus a constant:
- tree_node_counts[(int) t_kind]--;
That's obviously a mistake.
May I install the patch after reg tests&bootstrap?
Thanks,
Martin
gcc/ChangeLog:
2018-
On Wed, Mar 14, 2018 at 10:51 AM, Martin Liška wrote:
> Hi.
>
> Current code that tracks memory statistics wrongly decremented a counter
> based on t_kind, which is an enum value and thus a constant:
>
> - tree_node_counts[(int) t_kind]--;
>
> That's obviously a mistake.
>
> May I install the
Hi
This patch fixes PR82989 so that we avoid NEON instructions when
-mneon-for-64bits is not enabled. This is more of a short term fix for
the real deeper problem of making and early decision of choosing or
rejecting NEON instructions. There is now a new ticket PR84467 to deal
with the longer
Bootstrapped on x86_64-unknown-linux-gnu, applied.
Richard.
2018-03-14 Richard Biener
* tree-ssa-pre.c (compute_antic_aux): Remove code that asserts
that the value-set of ANTIC_IN doesn't grow.
Revert
* tree-ssa-pre.c (struct bb_bitmap_sets): Add visited_with
On Wed, Mar 14, 2018 at 8:03 AM, Alexandre Oliva wrote:
> On Mar 12, 2018, "Bin.Cheng" wrote:
>
>> internal compiler error: in aarch64_classify_address, at
>> config/aarch64/aarch64.c:5678
>> 0xfe3c29 aarch64_classify_address
>> /.../build/src/gcc/gcc/config/aarch64/aarch64.c:5677
>> 0xfe8be8
On Wed, Feb 21, 2018 at 3:02 AM, H.J. Lu wrote:
> On Wed, Oct 18, 2017 at 5:25 PM, H.J. Lu wrote:
>> config/plugins.m4 has
>>
>> if test "$plugins" = "yes"; then
>> AC_SEARCH_LIBS([dlopen], [dl])
>> fi
>>
>> Plugin uses dlsym, but libasan.so only intercepts dlopen, not dlsym:
>>
>> [hjl@gn
Hi,
I have made changes to this patch.
I attached a new version.
14.03. 2018 Olga Makhotina
gcc/
* config/i386/sgxintrin.h (_enclv_u32): New intrinsic.
(__enclv_bc, __enclv_cd, __enclv_generic): New definitions.
(ERDINFO, ETRACKC, ELDBC, ELDUC): New leaves.
gcc/testsu
On 03/13/2018 04:23 PM, Jakub Jelinek wrote:
> On Tue, Mar 13, 2018 at 04:19:21PM +0100, Martin Liška wrote:
>>> Yes, see e.g. TARGET_LIBC_HAS_FUNCTION target hook,
>>> where in particular linux_libc_has_function deals with various C libraries.
>>> Of course, in this case you need another target ho
On Wed, Mar 14, 2018 at 5:54 AM, Martin Liška wrote:
> On 03/13/2018 04:23 PM, Jakub Jelinek wrote:
>> On Tue, Mar 13, 2018 at 04:19:21PM +0100, Martin Liška wrote:
Yes, see e.g. TARGET_LIBC_HAS_FUNCTION target hook,
where in particular linux_libc_has_function deals with various C librar
On 03/14/2018 01:57 PM, H.J. Lu wrote:
> On Wed, Mar 14, 2018 at 5:54 AM, Martin Liška wrote:
>> On 03/13/2018 04:23 PM, Jakub Jelinek wrote:
>>> On Tue, Mar 13, 2018 at 04:19:21PM +0100, Martin Liška wrote:
> Yes, see e.g. TARGET_LIBC_HAS_FUNCTION target hook,
> where in particular linux_
On Wed, Mar 14, 2018 at 01:54:39PM +0100, Martin Liška wrote:
> --- a/gcc/builtins.c
> +++ b/gcc/builtins.c
> @@ -3651,13 +3651,24 @@ expand_builtin_memory_copy_args (tree dest, tree src,
> tree len,
>src_mem = get_memory_rtx (src, len);
>set_mem_align (src_mem, src_align);
>
> + bool i
On Wed, Mar 14, 2018 at 02:08:07PM +0100, Martin Liška wrote:
> > diff --git a/gcc/testsuite/gcc.c-torture/execute/builtins/mempcpy.c
> > b/gcc/testsuite/gcc.c-torture/execute/builtins/mempcpy.c
> > index d82e2232d7b..91e1c87f83f 100644
> > --- a/gcc/testsuite/gcc.c-torture/execute/builtins/mempcpy
distribute_links tries to place a log_link for whatever the destination
of the modified instruction is. It shouldn't do that when that dest
is pc_rtx, which isn't actually a register.
Committing to trunk.
Segher
2018-03-14 Segher Boessenkool
* combine.c (distribute_links): Don't m
This fixes more than half of our testcase failures on BE.
Committing.
Segher
2018-03-14 Segher Boessenkool
libsanitizer/
* sanitizer_common/sanitizer_stacktrace.cc
(BufferedStackTrace::FastUnwindStack): Use the correct frame offset
for PowerPC SYSV ABI.
---
libsan
On 03/14/2018 02:07 PM, Jakub Jelinek wrote:
> On Wed, Mar 14, 2018 at 01:54:39PM +0100, Martin Liška wrote:
>> --- a/gcc/builtins.c
>> +++ b/gcc/builtins.c
>> @@ -3651,13 +3651,24 @@ expand_builtin_memory_copy_args (tree dest, tree
>> src, tree len,
>>src_mem = get_memory_rtx (src, len);
>>
PR c/84852 reports an ICE inside diagnostic_show_locus when printing
a diagnostic for a source file with a #line >= 2^31:
#line 77
int foo (void) { return strlen(""); }
where we're attempting to print a fix-it hint at the top of the file
and underline the "strlen" (two "line spans").
On 14/03/18 14:07, Jakub Jelinek wrote:
On Fri, Mar 02, 2018 at 10:34:22AM -0700, Jeff Law wrote:
2018-02-12 Kyrylo Tkachov
PR target/84164
* simplify-rtx.c (simplify_truncation): Simplify truncation of
masking
operation.
* config/aarch64/aarch64.md (*aarch64_reg_3_neg_m
On Tue, Mar 13, 2018 at 7:22 PM, Siddhesh Poyarekar
wrote:
> Avoid issuing a bogus warning when the source of strncpy is bound by a
> constant known to be less than the minimum size of the destination.
>
> Changes from v1:
>
> - Use range-info instead of the MIN_EXPR hack
> - Get the minimum size
Segher:
I have addressed the following comments in the patch
> >
> > 2018-03-12 Carl Love
> >
> > * config/rs6000/rs6000-c.c: Add macro definitions for
> > ALTIVEC_BUILTIN_VEC_PERM.
>
> ALTIVEC_BUILTIN_VEC_PERMXOR.
Fixed typo.
>
> > * config/rs6000/altivec.md: Add define_ins
GCC Maintainers:
The following patch fixes an ICE when compiling the test case
gcc -mcpu=power7 builtin-fctid-fctiw-runnable.c
The GCC compiler now gives a message
"error: builtin function ‘__builtin_fctiw’ requires the ‘-mpower8-vector’
option"
and exits without generating an interna
GCC Maintainers:
The following patch fixes an ICE when compiling the test case
gcc -mcpu=power7 sse2-pmuludq-1.c
The GCC compiler now gives a message
"error: builtin function ‘__builtin_altivec_vmuleuw’ requires the
‘-mpower8-vector’ option"
and exits without generating an internal er
cxx_constant_value doesn't understand template codes, and neither it
understands OVERLOADs, so if we pass an OVERLOAD to it, we crash. Here
instantiate_non_dependent_expr got an OVERLOAD, but since it calls
is_nondependent_constant_expression which checks type_unknown_p, it left the
expression as
Hi,
here's an attempt to document some of the inline asm operand modifiers
which make sense and which get used so that people can find what they
mean in the docs.
This is my first gcc patch so there might be clumsiness ahead. :)
Thanks,
Boris.
gcc/Changelog:
2018-03-14 Borislav Petkov
Hi!
On Wed, Mar 14, 2018 at 08:17:42AM -0700, Carl Love wrote:
> 2018-03-14 Carl Love
>
> * config/rs6000/rs6000-c.c: Add macro definitions for
> ALTIVEC_BUILTIN_VEC_PERMXOR.
> * config/rs6000/rs6000.h: Add #define for vec_permxor builtin.
> * config/rs6000/rs6000-built
Hi.
This is a new test-case isolated from libreoffice.
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
I'm going to install it.
Martin
gcc/testsuite/ChangeLog:
2018-03-14 Martin Liska
PR ipa/8480
* g++.dg/lto/pr84805_0.C: New test.
* g++.d
On Fri, Mar 02, 2018 at 10:34:22AM -0700, Jeff Law wrote:
> >>> 2018-02-12 Kyrylo Tkachov
> >>>
> >>> PR target/84164
> >>> * simplify-rtx.c (simplify_truncation): Simplify truncation of
> >>> masking
> >>> operation.
> >>> * config/aarch64/aarch64.md (*aarch64_reg_3_neg_mask2):
Hi
This patch is another partial fix for PR 84521. This is adding a
definition to one of the target hooks used in the SJLJ implemetation so
that AArch64 defines the hard_frame_pointer_rtx as the
TARGET_BUILTIN_SETJMP_FRAME_VALUE. As pointed out by Wilco there is
still a lot more work to be do
On Wed, Mar 14, 2018 at 11:59 AM, Marek Polacek wrote:
> cxx_constant_value doesn't understand template codes, and neither it
> understands OVERLOADs, so if we pass an OVERLOAD to it, we crash. Here
> instantiate_non_dependent_expr got an OVERLOAD, but since it calls
> is_nondependent_constant_ex
Hello Julia!
> On 14 Mar 2018, at 10:48, Koval, Julia wrote:
>
> Gentle ping.
Your patch is OK for main trunk.
—
Thanks, K
>> -Original Message-
>> From: Koval, Julia
>> Sent: Monday, February 12, 2018 10:57 AM
>> To: Kirill Yukhin
>> Cc: 'GCC Patches'
>> Subject: RE: [x86,avx] Fi
In the wg21.link/P0522 change to template template parameter matching,
we ran into trouble with this testcase: when we try to form
TTC, we need to convert TA to TC, and try to tsubst TC to
produce the real desired type. But trying to look up TC in the args
for TTC fails, and indeed makes no sense.
Thank you for feedback and discussion regarding my first draft of this
patch with Segher Boessenkool and Bill Schmidt. This revision of the
patch differs from the first in the following regards:
1. I have also removed the vector_altivec_load_ and
vectore_altivec_store_ expansions from vector.md.
Ping
On Fri, Mar 2, 2018 at 1:23 PM, Jason Merrill wrote:
> As I mentioned in the PR, the problem here is that we're replacing a
> constructor with a dispatcher function which doesn't look much like a
> constructor. This patch adjusts make_dispatcher_decl to make it look
> more like the function
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
Please make sure to test it on more platforms (see PR ipa/83983 for details).
--
Eric Botcazou
Hi Carl,
On Wed, Mar 14, 2018 at 08:32:54AM -0700, Carl Love wrote:
> 2018-03-13 Carl Love
>
> PR 84422 - ICE on various builtin test functions when compiled with
> -mcpu=power7.
This should be _exactly_
PR target/84422
(one such line per PR, and no blank line after it)
Hi
Following PR 78420 patch I realized that we can use similar
technique to have assertions in string_view implementations.
I also rename testsuite files expected to XFAIL as they should have
a trailing '_neg'. It fixes 4 XPASS when run in debug mode
Note that I also try to use
During code review, it was discovered that the implementation of
__builtin_altivec_lvx_v1ti is not complete. The constant
ALTIVEC_BUILTINLVX_V1TI is introduced and is bound to the function
__builtin_altivec_lvx_v1ti. However, this function's implementation is
incomplete because there is no call t
Hi!
prepare_cmp_insn in some cases changes both the passed in comparison and the
comparison mode, e.g. by promoting the arguments from SFmode to DFmode etc.
In emit_conditional_move we call this in a loop, for (pass = 0; pass < 2;
pass++)
and in each case construct comparison arguments as well a
Hi!
As the following testcase shows, if we have a deprecated class, we warn
about any uses, including e.g. arguments of methods of that class (how can
one e.g. declare or define a copy ctor without warnings?).
The following patch changes it, so that we don't warn about deprecated uses
in methods
Hi!
The change for better warnings on vector shifts unfortunately started
accepting even shifts of floating point vectors by int shift count, which is
something that shouldn't be supported and ICEs later on.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
trunk?
2018-
Hi!
The following patch is an attempt to fix PLACEHOLDER_EXPR handling.
As e.g.
struct Y
{
static Y bar (Y y) { return y; }
int i;
int n = bar (Y{2,i}).m + bar {Y{2,i,i}).n;
int m = i;
};
is rejected - one can't use incomplete ctors which would need NSDMIs
until the class is defined, I bel
(Resending from a different account, sorry for the duplicate).
On 14/03/18 22:12 +0100, François Dumont wrote:
> constexpr const _CharT&
> operator[](size_type __pos) const noexcept
> {
>- // TODO: Assert to restore in a way compatible with the constexpr.
>- // __glibcx
I am starting to work on cleaning up the memory addressing support in the GCC 9
time frame. At the moment, I am working on upgrading the infrastructure to
allow in the future to prevent splitting memory on 64-bit LE systems too early,
rework the fusion support, and provide a pathway for future pro
Here's a very different patch. This gets rid of the __ptr_rel_ops and
just puts the special handling for pointers directly in the
std::less<_Tp*> etc. specializations. Then std::less uses
std::less for some pointer type P*. I've also added a ton of ugly
metaprogramming to detect the "if the call op
On 14 March 2018 at 22:52, Jonathan Wakely wrote:
> (Resending from a different account, sorry for the duplicate).
>
> On 14/03/18 22:12 +0100, François Dumont wrote:
>> constexpr const _CharT&
>> operator[](size_type __pos) const noexcept
>> {
>>- // TODO: Assert to restore
In rare situations, pa_emit_move_sequence needs to handle a paradoxical
SUBREG in a
floating-point load or store. There is existing code to adjust the
operand's address but
we fail to emit the reload if we don't need to load the address to a
register. We just
fall through pa_emit_move_sequenc
On 14 March 2018 at 23:27, Jonathan Wakely wrote:
> Here's one way to generalize this idea. We could potentially replace
> most of the lightweight __glibcxx_assert checks with this, to get
> zero-overhead static checking at compile-time whenever possible (even
> in constexpr functions) and have opt
On Wed, 14 Mar 2018, Jakub Jelinek wrote:
> Hi!
>
> The change for better warnings on vector shifts unfortunately started
> accepting even shifts of floating point vectors by int shift count, which is
> something that shouldn't be supported and ICEs later on.
>
> Fixed thusly, bootstrapped/regte
On Wed, Mar 14, 2018 at 6:26 PM, Jakub Jelinek wrote:
> As the following testcase shows, if we have a deprecated class, we warn
> about any uses, including e.g. arguments of methods of that class (how can
> one e.g. declare or define a copy ctor without warnings?).
>
> The following patch changes
On Wed, Mar 14, 2018 at 6:33 PM, Jakub Jelinek wrote:
> The following patch is an attempt to fix PLACEHOLDER_EXPR handling.
> As e.g.
> struct Y
> {
> static Y bar (Y y) { return y; }
> int i;
> int n = bar (Y{2,i}).m + bar {Y{2,i,i}).n;
> int m = i;
> };
> is rejected - one can't use inco
The attachedi patch detects situations where the sum of
an array's rank and corank exceeds the maximum allowed
by the Standard. Regression tested on x86_64-*-freebsd.
2018-03-14 Steven G. Kargl
PR fortran/69395
* decl.c (merge_array_spec): Limit the merging to maximum allowed
Hi:
The fixits-pr84852-1.c fails on mips64el target.
FAIL: gcc.dg/fixits-pr84852-1.c (test for excess errors)
FAIL: gcc.dg/fixits-pr84852-1.c dg-regexp 25 not found:
".*fixits-pr84852.c:-812156810:25:"
see this patch:
diff --git a/gcc/testsuite/gcc.dg/fixits-pr84852-1.c
b/gcc/testsuite/gcc.dg/f
On Tue, Aug 29, 2017 at 5:37 PM, Jason Merrill wrote:
> We could approach this by extending that change to all generic
> lambdas, and that might be appropriate for GCC 7, but it seems to me
> that this approach will just mean any problems with doing all the
> normal processing in a template will r
Tested x86_64-pc-linux-gnu, applying to trunk.
commit c977f3b7668e49eb8f8fd3fdabb6cc7a47e37adc
Author: Jason Merrill
Date: Wed Mar 14 23:29:39 2018 -0400
PR c++/84801 - ICE with unexpanded pack in lambda.
* pt.c (check_for_bare_parameter_packs): Don't return early f
The problem here was that although cp_parser_declarator returned
cp_error_declarator, we then wrapped it in another declarator, hiding
the problem.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 44b23ac745e9f249ff3971fcb4021facbbf8741c
Author: Jason Merrill
Date: Thu Mar 15 00:14:04 2018
On Wednesday 14 March 2018 08:40 PM, Richard Biener wrote:
> Instead of building a tree from max you should use
>
> if (wi::to_widest (max) < wi::to_widest (wi::to_wide (dstsize)))
> return;
>
> given compute_objsize is somewhat confused about the type it returns
> a widest_int compare
66 matches
Mail list logo
