On Wed, 26 Feb 2020, Jakub Jelinek wrote:
> Hi!
>
> The following testcase is miscompiled in 8+.
> The problem is that check_no_overlap has a special case for INTEGER_CST
> marked stores (i.e. stores of constants), if both all currenly merged stores
> and the one under consideration for merging w
On 2020/2/18 17:57, Richard Biener wrote:
> On Tue, 18 Feb 2020, Xionghu Luo wrote:
>
>> Store-merging pass should run twice, the reason is pass fre/pre will do
>> some kind of optimizations to instructions by:
>>1. Converting the load from address to load from function arguments
>>(store_
On Tue, Feb 25, 2020 at 07:43:09PM -0600, Bill Schmidt wrote:
> The reason that homogeneous aggregates matter (at least somewhat) is that
> the ABI ^H^H^H^HAPI requires establishing a calling convention and a name-
> mangling formula that includes the length of parameters and return values.
> Since
On Wed, Feb 26, 2020 at 09:09:11AM +0100, Richard Biener wrote:
> > Note, as I said in the PR, for GCC11 we could consider performing some kind
> > of cheap DSE during the store merging (perhaps guarded with flag_tree_dse).
> > And another thing to consider is only consider as problematic non-merge
On Wed, 26 Feb 2020, luoxhu wrote:
> On 2020/2/18 17:57, Richard Biener wrote:
> > On Tue, 18 Feb 2020, Xionghu Luo wrote:
> >
> >> Store-merging pass should run twice, the reason is pass fre/pre will do
> >> some kind of optimizations to instructions by:
> >>1. Converting the load from addre
On Wed, Feb 26, 2020 at 04:09:16PM +0800, luoxhu wrote:
> Thanks Richard, not sure about my understanding and please correct if any.
>
> I tried Jukub's latest patch of "sccvn: Handle bitfields in push_partial_def".
> Got to know fre pass checks the load instruction's vuse chain and do the
> cons
On Wed, 26 Feb 2020, Richard Biener wrote:
> On Wed, 26 Feb 2020, luoxhu wrote:
>
> > On 2020/2/18 17:57, Richard Biener wrote:
> > > On Tue, 18 Feb 2020, Xionghu Luo wrote:
> > >
> > >> Store-merging pass should run twice, the reason is pass fre/pre will do
> > >> some kind of optimizations to
On Wed, Feb 26, 2020 at 12:30 AM Richard Biener wrote:
>
> On Wed, 26 Feb 2020, luoxhu wrote:
>
> > On 2020/2/18 17:57, Richard Biener wrote:
> > > On Tue, 18 Feb 2020, Xionghu Luo wrote:
> > >
> > >> Store-merging pass should run twice, the reason is pass fre/pre will do
> > >> some kind of optim
Hi!
Based on IRC discussions, I've added another test with __GIMPLE next to the
one with -fno-tree-dse.
2020-02-26 Jakub Jelinek
PR tree-optimization/93820
* gcc.dg/pr93820-2.c: New test.
--- gcc/testsuite/gcc.dg/pr93820-2.c.jj 2020-02-26 10:50:18.796137652 +0100
+++ gcc/test
Hi!
The following patch does the same check the tree-ssa-strlen.c pass does
to punt on broken builtin redeclarations.
I agree the ultimate fix is change the C FE, but don't feel I know that part
of the FE good enough to know all the consequences and so it might be too
risky for stage4.
Bootstrapp
On Wed, 26 Feb 2020, Jakub Jelinek wrote:
> Hi!
>
> The following patch does the same check the tree-ssa-strlen.c pass does
> to punt on broken builtin redeclarations.
> I agree the ultimate fix is change the C FE, but don't feel I know that part
> of the FE good enough to know all the consequenc
Hi,
This patch solves problem caused by the disagreement between FRE and access
path orracle.
FRE is checking stores for equivalence based on their address, value and
base+ref alias sets. Because ref alias set is not always the alias set of
innermost type, but it may be one of refs in the access
On Wed, 26 Feb 2020, Jan Hubicka wrote:
> Hi,
> This patch solves problem caused by the disagreement between FRE and access
> path orracle.
>
> FRE is checking stores for equivalence based on their address, value and
> base+ref alias sets. Because ref alias set is not always the alias set of
> i
In this PR we had a conversion between two integer vectors that
both had scalar integer modes. We then tried to implement the
conversion using the scalar optab for those modes, instead of
doing the conversion elementwise.
I wondered about letting through scalar modes for single-element
vectors, b
This patch adds various information to -fdump-analyzer and
-fdump-analyzer-stderr to make it easier to track down
problems with state explosions in the exploded_graph.
It logs the number of unprocessed nodes in the worklist, for
the case where the upper limit on exploded nodes is reached.
It prin
On Wed, Feb 26, 2020 at 12:43 PM Richard Sandiford
wrote:
>
> In this PR we had a conversion between two integer vectors that
> both had scalar integer modes. We then tried to implement the
> conversion using the scalar optab for those modes, instead of
> doing the conversion elementwise.
>
> I w
> > Bootstrapped/regtested x86_64-linux, OK?
>
> OK and thanks for the elaborate write-up and comments in the code ;)
Unforutnately our discussion on IRC let me construct another wrong code
testcase based on mixing up base alias sets
union U { long long i; long f; };
struct a {union U u;};
struc
On Wed, Feb 26, 2020 at 11:43:10AM +, Richard Sandiford wrote:
> In this PR we had a conversion between two integer vectors that
> both had scalar integer modes. We then tried to implement the
> conversion using the scalar optab for those modes, instead of
> doing the conversion elementwise.
>
Jakub Jelinek writes:
> Can you please explain how do you want to pass the
> void sincos (double, double *, double *);
> arguments? I must say it isn't entirely clear from the document.
> You talk there about double[2], but sincos certainly doesn't have such an
> argument.
The plan [1] is to re
We are calling _OuterIter::__current from _InnerIter::operator==, but the former
is private within this non-member friend. Fix this by calling
_OuterIter::operator== instead, which does the right thing here.
libstdc++-v3/ChangeLog:
PR libstdc++/93936
* include/std/ranges (split_v
This is the second in the series to bring the GCC implementation into line
with the current standard.
@JunMa
I believe that this should solve the problems you were looking at in
“[PATCH Coroutines] Fix issue with unused corutine function parameters”
and supercedes that patch (since we needed
On Wed, Feb 26, 2020 at 10:32:17AM -0300, Tulio Magno Quites Machado Filho
wrote:
> Jakub Jelinek writes:
>
> > Can you please explain how do you want to pass the
> > void sincos (double, double *, double *);
> > arguments? I must say it isn't entirely clear from the document.
> > You talk ther
On Wed, Feb 26, 2020 at 2:46 PM Jakub Jelinek wrote:
>
> On Wed, Feb 26, 2020 at 10:32:17AM -0300, Tulio Magno Quites Machado Filho
> wrote:
> > Jakub Jelinek writes:
> >
> > > Can you please explain how do you want to pass the
> > > void sincos (double, double *, double *);
> > > arguments? I
This adjusts dumping as proved useful in debugging.
Bootstrapped / tested on x86_64-unknown-linux-gnu, applied.
Richard.
2020-02-26 Richard Biener
* tree-vect-slp.c (vect_print_slp_tree): Also dump ref count
and load permutation.
---
gcc/tree-vect-slp.c | 13 ++---
1
On 2/26/20 2:18 AM, Jakub Jelinek wrote:
On Tue, Feb 25, 2020 at 07:43:09PM -0600, Bill Schmidt wrote:
The reason that homogeneous aggregates matter (at least somewhat) is that
the ABI ^H^H^H^HAPI requires establishing a calling convention and a name-
mangling formula that includes the length of
On 2/26/20 8:43 AM, Iain Sandoe wrote:
This is the second in the series to bring the GCC implementation into line
with the current standard.
2020-02-26 Iain Sandoe
* coroutines.cc (struct param_info): Keep track of params
that are references, and cache the original type an
On 26/02/20 08:37 -0500, Patrick Palka wrote:
We are calling _OuterIter::__current from _InnerIter::operator==, but the former
is private within this non-member friend. Fix this by calling
_OuterIter::operator== instead, which does the right thing here.
libstdc++-v3/ChangeLog:
PR libst
On 25/02/20 15:36 -0500, Patrick Palka wrote:
This adds constexpr to 11 algorithms defined in as per P1645R1.
Tested on x86_64-pc-linux-gnu, OK to commit?
libstdc++-v3/ChangeLog:
P1645R1 constexpr for algorithms
* include/bits/stl_numeric.h (iota, accumulate, inner_product,
On Wed, Feb 26, 2020 at 07:55:53AM -0600, Bill Schmidt wrote:
> The hope is that we can create a vectorized version that returns values
> in registers rather than the by-ref parameters, and add code to GCC to
> copy things around correctly following the call. Ideally the signature of
> the vectori
Early expand ADDDI3 and SUBDI3 for better code gen.
gcc/
-xx-xx Claudiu Zissulescu
* config/arc/arc.md (adddi3): Early expand the 64bit operation into
32bit ops.
(subdi3): Likewise.
(adddi3_i): Remove pattern.
(subdi3_i): Likewise.
---
gcc/config/ar
Add length attribute to eh_return pattern.
gcc/
-xx-xx Claudiu Zissulescu
* config/arc/arc.md (eh_return): Add length info.
---
gcc/config/arc/arc.md | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/gcc/config/arc/arc.md b/gcc/config/arc/arc.md
index fb25aafb
All ARC's small data adressing is using address scaling feature of the
load/store instructions (i.e., the address is made of a general
pointer plus a shifted offset. The shift amount depends on the
addressing mode). This patch is checking the offset of an address if
it fits the scalled constraint.
With the refurbish of ARC600' accumulator support, the mlo_operand
doesn't reflect the proper low accumulator register for the newer
ARCv2 accumulator register used by the fma instructions, replace
it with accl_operand predicate.
gcc/
-xx-xx Claudiu Zissulescu
* config/arc/arc.md (
On 23/02/2020 21:25, Vladimir Makarov wrote:
The following patch is for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93564
The patch was successfully bootstrapped on x86-64 and benchmarked on
SPEC2000.
Since this patch I get an ICE with checking enabled, for amdgcn-amdhsa:
during RTL pa
This introduces a couple of convenience alias templates to be used for
some repeated patterns using std::conditional_t.
* include/std/ranges (__detail::__maybe_empty_t): Define new helper
alias.
(__detail::__maybe_const_t): Likewise.
(__adaptor::_RangeAdaptor): Use
* include/bits/ranges_algo.h (__lexicographical_compare_fn): Declare
variables in smaller scope and avoid calling ranges::distance when we
know they are pointers. Remove statically-unreachable use of
__builtin_unreachable().
* include/bits/stl_algobase.h (__l
On Wed, 26 Feb 2020, Jonathan Wakely wrote:
> On 25/02/20 15:36 -0500, Patrick Palka wrote:
> > This adds constexpr to 11 algorithms defined in as per P1645R1.
> >
> > Tested on x86_64-pc-linux-gnu, OK to commit?
> >
> > libstdc++-v3/ChangeLog:
> >
> > P1645R1 constexpr for algorithms
> >
Hi,
this is and TBAA stat for building cc1 with -flto-partition=none.
From:
Alias oracle query stats:
refs_may_alias_p: 46099243 disambiguations, 55677716 queries
ref_maybe_used_by_call_p: 124351 disambiguations, 46883813 queries
call_may_clobber_ref_p: 12673 disambiguations, 17133 queries
On 24/02/20 21:53 -0500, Thomas Rodgers wrote:
+bool
+_S_futex_wait_until(int* __addr, int __val,
+ bool __has_timeout = false,
+ std::chrono::seconds __s = std::chrono::seconds::zero(),
+ std::chrono::nanoseconds __ns =
s
On Tue, 11 Feb 2020, Patrick Palka wrote:
> This patch adds memoization for these four views so that their begin() has the
> required constant time amortized complexity.
>
> In the general case we use std::optional to cache the result. When the
> underlying range is a random_access_range then we
Hi!
On Tue, Feb 25, 2020 at 07:43:09PM -0600, Bill Schmidt wrote:
> On 2/25/20 12:45 PM, Segher Boessenkool wrote:
> >I don't agree we should have a new ABI, and an API (which this *is* as
> >far as I can tell) works fine on *any* ABI. Homogeneous aggregates has
> >nothing to do with anything eit
On 2/25/20 4:25 PM, Segher Boessenkool wrote:
> Well, you now get an extra mask instruction (rldicr for example) as well,
> right? While that mask usually isn't needed.
I believe the mask is implicit by the pattern used by the vec_st()
builtin, which normally gets mapped to the Altivec stvx insn,
On 2/26/20 11:29 AM, Peter Bergner wrote:
>> You want
>> {\m(stvx|stxv)\M}
As we discussed offline, the regex above double counts everything,
so I went with {\m(?:stvx|stxv|stxvx)\M} which you pointed me to.
Pushed to master. I'll push to the gcc-9 branch tomorrow after
Bill's regression script
Am Mi., 26. Feb. 2020 um 16:20 Uhr schrieb Jonathan Wakely :
>
> This introduces a couple of convenience alias templates to be used for
> some repeated patterns using std::conditional_t.
I find it a bit confusing/inconsistent to define __maybe_const_t such
that the bool argument says "is const", w
Patch ping.
On 2/19/20 3:30 PM, Alexey Neyman wrote:
Hi all,
Attached is a patch adjusted per discussion in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93751
- The manual is corrected to reflect that DIEs for external variables
are not generated when using DWARF
- A new option is introduce
On 2/26/20 2:02 PM, Alexey Neyman wrote:
Patch ping.
On 2/19/20 3:30 PM, Alexey Neyman wrote:
Hi all,
Attached is a patch adjusted per discussion in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93751
- The manual is corrected to reflect that DIEs for external variables
are not generated wh
On Wed, 26 Feb 2020, Jason Merrill wrote:
> Don't we want to fix the DWARF behavior to match the documentation?
+1 - I think Alexey correctly pointed out in the Bugzilla that debuginfo
growth from this change would be minimal (usually the number of global
variables is very small compared to numbe
On February 26, 2020 8:26:06 PM GMT+01:00, Alexander Monakov
wrote:
>On Wed, 26 Feb 2020, Jason Merrill wrote:
>
>> Don't we want to fix the DWARF behavior to match the documentation?
>
>+1 - I think Alexey correctly pointed out in the Bugzilla that
>debuginfo
>growth from this change would be mi
This patch updates libgo to the final Go1.14 release. Bootstrapped
and ran Go testsuite on x86_64-pc-linux-gnu. Committed to mainline.
Ian
0f9e2b205ccb34be93aafa2669c6c9b5a8c557ec
diff --git a/gcc/go/gofrontend/MERGE b/gcc/go/gofrontend/MERGE
index 7dcb9ecb4fd..a62c8292e0a 100644
--- a/gcc/go/go
r7-2111 introduced maybe_constant_value in cp_fully_fold.
maybe_constant_value uses cxx_eval_outermost_constant_expr, which
can clear TREE_CONSTANT:
6510 else if (non_constant_p && TREE_CONSTANT (r))
[...]
6529 TREE_CONSTANT (r) = false;
In this test the array size is '(long int) "h"'. Th
On Tue, 2020-02-25 at 17:45 -0300, Alexandre Oliva wrote:
> Including limits.h on vxworks pollutes the global namespace with SH
> and many other platform names; work around it.
>
> Tested on the affected platform. Ok to install?
>
>
> for gcc/testsuite/ChangeLog
>
> * gcc.target/aarch64
On Tue, 2020-02-25 at 17:47 -0300, Alexandre Oliva wrote:
> There was a mistake in forward-porting and contributing some
> vxworks7r2 changes, that caused a conditional to be dropped around a
> couple of preprocessor directives, needed only up to vxworks6, that
> change the compiler's behavior WRT
On 2/25/20 4:09 PM, Nathan Sidwell wrote:
We consider all TYPE_ARGUMENT_PACKS distinct types, leading to problems
with redeclarations.
This patch fixes things by:
a) marking all such types for structural comparison
b) teaching structural_comptypes how to compare them.
1) It appears that NONTY
On Wed, 2020-02-19 at 17:26 -0700, Martin Sebor wrote:
> The buffer overflow detection for multi-char stores uses the size
> of a source array even when what's actually being accessed (read
> and stored) is a pointer to the array. That leads to incorrect
> warnings in some cases.
>
> The attached
Status
==
The GCC 8 branch is now frozen for blocking regressions and documentation
fixes only, all changes to the branch require a RM approval now.
Quality Data
Priority # Change from last report
--- ---
P10 -
As part of improving the vector extract built-ins for the future processor, and
also looking at optimizing all of the cases with vector extracts and sign,
zero, float extension for GCC 11 (PR target/93230), I noticed that the code
generated for some of the tests had regressed from the GCC 8 time fr
Hi!
When doing the 8.4-rc1, I've noticed (probably also because of the dying
disk on sourceware) that git clone is extremely slow, and furthermore when
all of us have some local snapshots, it is a waste of resources to download
everything again. Especially for the -f runs when we'll need to wait
On Sat, 2020-02-15 at 07:26 -0800, H.J. Lu wrote:
> On x86, when AVX and AVX512 are enabled, vector move instructions can
> be encoded with either 2-byte/3-byte VEX (AVX) or 4-byte EVEX (AVX512):
>
>0: c5 f9 6f d1 vmovdqa %xmm1,%xmm2
>4: 62 f1 fd 08 6f d1 vmovdqa64 %xmm1,
GCC accepts attribute format even on functions without a prototype.
However, when such a function is subsequently redeclared with
a prototype the attribute validation runs into trouble. To avoid
the ICE I considered a) dropping the attribute on functions declared
without a prototype (like Clang d
On 2/26/20 3:44 PM, Marek Polacek wrote:
r7-2111 introduced maybe_constant_value in cp_fully_fold.
maybe_constant_value uses cxx_eval_outermost_constant_expr, which
can clear TREE_CONSTANT:
6510 else if (non_constant_p && TREE_CONSTANT (r))
[...]
6529 TREE_CONSTANT (r) = false;
In this t
On 2/26/20 3:09 PM, Jeff Law wrote:
On Wed, 2020-02-19 at 17:26 -0700, Martin Sebor wrote:
The buffer overflow detection for multi-char stores uses the size
of a source array even when what's actually being accessed (read
and stored) is a pointer to the array. That leads to incorrect
warnings i
On Wed, 26 Feb 2020 at 18:31, Daniel Krügler wrote:
>
> Am Mi., 26. Feb. 2020 um 16:20 Uhr schrieb Jonathan Wakely
> :
> >
> > This introduces a couple of convenience alias templates to be used for
> > some repeated patterns using std::conditional_t.
>
> I find it a bit confusing/inconsistent to d
On Wed, Feb 26, 2020 at 2:42 PM Jeff Law wrote:
>
> On Sat, 2020-02-15 at 07:26 -0800, H.J. Lu wrote:
> > On x86, when AVX and AVX512 are enabled, vector move instructions can
> > be encoded with either 2-byte/3-byte VEX (AVX) or 4-byte EVEX (AVX512):
> >
> >0: c5 f9 6f d1 vmovdqa
On Wed, 2020-02-26 at 16:02 -0800, H.J. Lu wrote:
> On Wed, Feb 26, 2020 at 2:42 PM Jeff Law wrote:
> > On Sat, 2020-02-15 at 07:26 -0800, H.J. Lu wrote:
> > > On x86, when AVX and AVX512 are enabled, vector move instructions can
> > > be encoded with either 2-byte/3-byte VEX (AVX) or 4-byte EVEX
On 2/26/20 1:44 PM, Marek Polacek wrote:
r7-2111 introduced maybe_constant_value in cp_fully_fold.
maybe_constant_value uses cxx_eval_outermost_constant_expr, which
can clear TREE_CONSTANT:
6510 else if (non_constant_p && TREE_CONSTANT (r))
[...]
6529 TREE_CONSTANT (r) = false;
In this t
On Wed, 26 Feb 2020, Jakub Jelinek wrote:
> Hi!
>
> When doing the 8.4-rc1, I've noticed (probably also because of the dying
> disk on sourceware) that git clone is extremely slow, and furthermore when
> all of us have some local snapshots, it is a waste of resources to download
> everything agai
PR analyzer/93947 reports an ICE at -O1 when attempting to analyze a
call that has been optimized away as unreachable.
The root cause is a NULL dereference due to the fndecl having a NULL
cgraph_node: the cgraph_node was created by
pass_build_cgraph_edges::execute, but was later removed by
symbol_
PR analyzer/93950 reports an ICE when pruning the path of a
-Wanalyzer-null-dereference diagnostic.
The root cause is a bug in the state-tracking code, in which the
variable of interest is tracked from the callee to a "nullptr" param
at the caller, whereupon we have an INTEGER_CST "variable", and
在 2020/2/26 下午9:43, Iain Sandoe 写道:
This is the second in the series to bring the GCC implementation into line
with the current standard.
@JunMa
I believe that this should solve the problems you were looking at in
“[PATCH Coroutines] Fix issue with unused corutine function parameters”
and
On Wed, Feb 26, 2020 at 06:01:30PM -0700, Martin Sebor wrote:
> On 2/26/20 1:44 PM, Marek Polacek wrote:
> > r7-2111 introduced maybe_constant_value in cp_fully_fold.
> > maybe_constant_value uses cxx_eval_outermost_constant_expr, which
> > can clear TREE_CONSTANT:
> > 6510 else if (non_constant_
在 2020/2/10 下午7:42, JunMa 写道:
Ping~
Regards
JunMa
Kindly ping.
Regards
JunMa
在 2020/2/5 下午5:17, JunMa 写道:
在 2020/2/5 下午2:14, JunMa 写道:
Hi
This patch builds co_await/yield_expr with unknown_type when we can not
know the promise type in processing_template_decl phase. it avoid to
confuse compi
在 2020/2/11 上午10:50, JunMa 写道:
Hi
kindly ping~
Regards
JunMa
Hi
As title. in maybe_promote_captured_temps, we promote captured
temporaries
and co_await_expr into a new BIND_EXPR. As the BIND_EXPR contains
co_await_expr and maybe other function calls, the side effects flag
should
be set.
Thi
在 2020/2/11 上午10:14, JunMa 写道:
Kindly ping
Regards
JunMa
Hi
In maybe_promote_captured_temps, the cleanup_point_stmt has been
stripped when handle temporaries captured by reference. However, maybe
there are non-reference temporaries in current stmt which cause ice in
gimpilify pass.
This patch f
在 2020/2/12 下午3:23, JunMa 写道:
Kindly ping
Regards
JunMa
Hi
In captures_temporary, the current implementation fails to handle
component_ref. This causes ice with case co_await A while
operator co_await is defined in base class of A. Also it is necessary
to capture the object of base class as if i
On Wed, Feb 26, 2020 at 05:54:03PM -0500, Jason Merrill wrote:
> On 2/26/20 3:44 PM, Marek Polacek wrote:
> > r7-2111 introduced maybe_constant_value in cp_fully_fold.
> > maybe_constant_value uses cxx_eval_outermost_constant_expr, which
> > can clear TREE_CONSTANT:
> > 6510 else if (non_constant
On 2/26/20 9:31 PM, Marek Polacek wrote:
On Wed, Feb 26, 2020 at 05:54:03PM -0500, Jason Merrill wrote:
On 2/26/20 3:44 PM, Marek Polacek wrote:
r7-2111 introduced maybe_constant_value in cp_fully_fold.
maybe_constant_value uses cxx_eval_outermost_constant_expr, which
can clear TREE_CONSTANT:
6
On Wed, Feb 26, 2020 at 09:41:14PM -0500, Jason Merrill wrote:
> On 2/26/20 9:31 PM, Marek Polacek wrote:
> > On Wed, Feb 26, 2020 at 05:54:03PM -0500, Jason Merrill wrote:
> > > On 2/26/20 3:44 PM, Marek Polacek wrote:
> > > > r7-2111 introduced maybe_constant_value in cp_fully_fold.
> > > > maybe
I've checked in this patch to address PR90467, which complained "many
warning options that are enabled by default are documented in the
-Woption form, not -Wno-option". While I was in there I also found
- the option summary lists, which are supposed to be (mostly)
alphabetized, had gotten
Hi all!
Does anybody considered backporting this?
Given https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92410#c3
"Confirmed also with r247015 (GCC 7 branch point)."
For example, we may apply it without an assertion in release branches.
Roman
12.11.2019 12:11, Matthew Malcomson wrote:
> In schedul
79 matches
Mail list logo
