On Thu, 3 Oct 2019, Prathamesh Kulkarni wrote:
> On Wed, 2 Oct 2019 at 12:28, Richard Biener wrote:
> >
> > On Wed, 2 Oct 2019, Prathamesh Kulkarni wrote:
> >
> > > On Wed, 2 Oct 2019 at 01:08, Jeff Law wrote:
> > > >
> > > > On 10/1/19 12:40 AM, Richard Biener wrote:
> > > > > On Mon, 30 Sep 20
On Thu, Oct 03, 2019 at 05:22:49PM -0500, Aaron Sawdey wrote:
> --- gcc/expr.c(revision 276516)
> +++ gcc/expr.c(working copy)
> @@ -1624,9 +1624,8 @@
>set_mem_size (y, const_size);
> }
>
> - bool pieces_ok = false;
> - if (CONST_INT_P (size))
> -pieces_ok = can_
On 9/27/19 1:00 PM, Jonathan Wakely wrote:
On 19/07/19 23:37 +0200, François Dumont wrote:
It sounds reasonable to overload std::copy_n for istreambuf_iterator.
I wonder whether it's worth doing:
#if __cplusplus >= 201703L
if constexpr (is_same_v<_OutputIterator, _CharT*>)
return __res
On 10/3/19 4:57 PM, Jakub Jelinek wrote:
On Thu, Oct 03, 2019 at 03:57:17PM -0400, Jason Merrill wrote:
That sounds better.
Ok.
And, second comment, not 100% sure about the unions, e.g. whether we
shouldn't somehow try to figure what is the active union member and only use
the active one rat
Hi
May I add this convenient function ? I'll also use it in coming
patches.
Note that I removed a template parameter in __is_random_access_iter
in C++11.
* include/bits/stl_iterator_base_types.h (__iterator_category_t):
Define
for C++11.
(__is_random_access_iter): Ada
Hi
We are missing obvious debug checks in std::copy_n. Moreover I'll
need them when I'll remove the Debug layer in a coming patch.
Tested under Linux x86_64.
* include/bits/stl_algo.h (copy_n): Add
__glibcxx_requires_can_increment
debug checks.
* testsuite/25_algorithms
On Wed, 2 Oct 2019 at 12:28, Richard Biener wrote:
>
> On Wed, 2 Oct 2019, Prathamesh Kulkarni wrote:
>
> > On Wed, 2 Oct 2019 at 01:08, Jeff Law wrote:
> > >
> > > On 10/1/19 12:40 AM, Richard Biener wrote:
> > > > On Mon, 30 Sep 2019, Prathamesh Kulkarni wrote:
> > > >
> > > >> On Wed, 25 Sep 2
On 10/3/19 5:59 AM, Vladislav Ivanishin wrote:
> This patch removes the old implementation of the warning on RTL.
>
> Note: regstat_get_setjmp_crosses() and the associated data seems to be
> needed for the RA itself, so I am not touching that.
>
>
> 0002-Remove-the-old-implementation-of-Wclobber
On 10/3/19 5:58 AM, Vladislav Ivanishin wrote:
> This refactoring makes gsi_next_nonvirtual_phi() always advance the
> gimple phi iterator (regardless of whether the current phi is virtual or
> not). It matches the behavior of other gsi_next functions.
>
> It can be applied independently of the o
On 10/3/19 9:16 AM, Bernd Edlinger wrote:
> Hi,
>
> I've noticed that the documentation of -Wshadow=x has some
> missing bits, and I want to add an negative example to
> -Wshadow=compatble-local.
>
>
> Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
> Is it OK for trunk?
>
>
> Thanks
> Ber
On 10/3/19 5:28 PM, Martin Sebor wrote:
> The count_nonzero_bytes() function that computes the range of bytes
> stored by a multi-byte assignment is overly conservative and fails
> to determine the number of bytes to store for expressions involving
> MEM_REF with DECL operands without a constant in
On 9/28/19 8:14 PM, Oleg Endo wrote:
> Hi,
>
> I've been dragging this patch along with me for a while.
> At the moment, I don't have the resources to fully test it as requested
> by Ian in the PR discussion.
>
> So I would like to ask for general comments on this one and hope that
> folks with b
On 10/3/19 6:47 AM, Andrea Corallo wrote:
>
> Jeff Law writes:
>
>> On 10/1/19 4:11 AM, Andrea Corallo wrote:
>>> Martin Jambor writes:
>>>
Hi,
On Mon, Sep 30 2019, Andrea Corallo wrote:
> Hi all,
> I'd like to submit this patch.
> It release the ipa cp transformation s
On 10/2/19 5:15 PM, Martin Sebor wrote:
> The attached patch adds an optional argument to
> compute_builtin_object_size to let it return the DECL of the object
> whose size it's called to compute. This lets -Wstringop-overflow
> point to the object in more instances of the warning than it does
> n
On 10/3/19 6:01 AM, Vladislav Ivanishin wrote:
> What it does
>
>
> A fundamental limitation of the new approach is that it requires phi
> nodes for variables to perform the analysis it needs to issue the
> warning for them. No phis - no warning. In particular, it doesn't deal
> wit
On Thu, Oct 3, 2019 at 8:20 AM Bernd Edlinger wrote:
>
> this would normally be a trivial changes, but unless I am mistaken,
> I cannot simply commit to the go subtree. Except the -Wshadow=local
> warnings this fixes also a few minor memory leaks.
>
>
> Bootstrapped and reg-tested on x86_64-pc-li
We do so non standard stuff in __canonicalize_funcptr_for_compare() to handle
function pointer canonicalization. The updated -Warray-bounds now causes
warnings
about negative array offsets. This patch disables the warning in this code.
Tested on hppa-unknown-linux-gnu. Committed to trunk.
Dav
Here is the patch to update GCC 9 release note. Is it OK?
--
H.J.
---
Index: htdocs/gcc-9/changes.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-9/changes.html,v
retrieving revision 1.77
diff -u -p -r1.77 changes.html
--- htdocs/gc
There are still a few packages that fail to build on Debian linux because we
end up
with branches that can't reach their target. It turns out that the default
maximum
stub group size in binutils is smaller than what we allow in gcc. This patch
reduces
the maximum offset in gcc to match that in
On Thu, Oct 03, 2019 at 02:08:16PM -0300, Tulio Magno Quites Machado Filho
wrote:
> Remove the extra -L directory so that user directories specified at
> build time have higher preference over the advance toolchain libraries.
>
> 2019-10-03 Tulio Magno Quites Machado Filho
>
> * config.
The count_nonzero_bytes() function that computes the range of bytes
stored by a multi-byte assignment is overly conservative and fails
to determine the number of bytes to store for expressions involving
MEM_REF with DECL operands without a constant initializer. Even
though it's not possible to de
Various built-in functions that GCC has as extensions are now standard
functions in C2x. This patch adds DEF_C2X_BUILTIN and uses it to mark
them as such. Some of the so-marked functions were previously
DEF_EXT_LIB_BUILTIN, while some DFP ones were DEF_GCC_BUILTIN
(i.e. __builtin_* only); both se
64-bit sibling calls are disabled on all PA-RISC targets. This is because the
argument pointer
handling is incompatible with doing sibcalls. Thus, we can remove the
non-local sibcall sequence
that was used in testing.
Tested on hppa64-hp-hpux11.11.
Dave
2019-10-03 John David Anglin
On 10/2/19 5:44 PM, Aaron Sawdey wrote:
> On 10/2/19 5:35 PM, Jakub Jelinek wrote:
>> On Wed, Oct 02, 2019 at 09:21:23AM -0500, Aaron Sawdey wrote:
> 2019-09-27 Aaron Sawdey
>
> * builtins.c (expand_builtin_memory_copy_args): Add might_overlap parm.
> (expand_builtin_memcpy):
On 10/3/19 3:39 PM, Ian Lance Taylor wrote:
> On Wed, Sep 25, 2019 at 10:27 AM Eduard-Mihai Burtescu wrote:
>>
>> 2019-09-25 Eduard-Mihai Burtescu
>> libiberty/ChangeLog:
>> * rust-demangle.c (looks_like_rust): Remove.
>> (rust_is_mangled): Don't check escapes.
>> (is_pr
On Wed, Sep 25, 2019 at 10:27 AM Eduard-Mihai Burtescu wrote:
>
> 2019-09-25 Eduard-Mihai Burtescu
> libiberty/ChangeLog:
> * rust-demangle.c (looks_like_rust): Remove.
> (rust_is_mangled): Don't check escapes.
> (is_prefixed_hash): Allow 0-9a-f permutations.
> (
On Thu, Oct 03, 2019 at 03:57:17PM -0400, Jason Merrill wrote:
> That sounds better.
Ok.
> > And, second comment, not 100% sure about the unions, e.g. whether we
> > shouldn't somehow try to figure what is the active union member and only use
> > the active one rather than trying all.
>
> Hmm, g
Hi,
For compatibility with xcrun (and the behaviour of the clang driver), make use
of the setting of the SDKROOT environment variable when it is available.
This applies to both finding headers and libraries (i.e. it is also passed to
ld64).
Priority:
1. User's command-line specified --sysroot= o
On 10/3/19 2:38 PM, Jakub Jelinek wrote:
On Tue, Oct 01, 2019 at 05:56:00PM -0400, Jason Merrill wrote:
+ else if (DECL_NAME (var) == heap_deleted_identifier)
+ {
+ if (!ctx->quiet)
+ error_at (loc, "deallocation of alre
On 10/3/19 1:46 PM, Jan Hubicka wrote:
>>
>>
>>> Tests that now fail, but worked before (15 tests):
>>>
>>> g++.dg/tree-ssa/pr61034.C -std=gnu++14 scan-tree-dump-times fre3 ";;
>>> Function" 1
>>> g++.dg/tree-ssa/pr61034.C -std=gnu++14 scan-tree-dump-times fre3 "free" 14
>>> g++.dg/tree-ssa/pr
On 10/3/19 5:54 AM, Vladislav Ivanishin wrote:
> Hi!
>
> This series implements the -Wclobbered warning on GIMPLE in place of the
> old RTL implementation.
>
> This cover letter contains a high-level explanation of what is going on
> and why. The new implementation itself is in the patch 3. I w
On 10/3/19 2:25 PM, Jakub Jelinek wrote:
Hi!
On Tue, Oct 01, 2019 at 05:56:00PM -0400, Jason Merrill wrote:
I think we want to factor this function more, so we don't have the same code
in multiple places for handling an array, and an array member, and a pointer
to array. Do you want to take a
>
>
> > Tests that now fail, but worked before (15 tests):
> >
> > g++.dg/tree-ssa/pr61034.C -std=gnu++14 scan-tree-dump-times fre3 ";;
> > Function" 1
> > g++.dg/tree-ssa/pr61034.C -std=gnu++14 scan-tree-dump-times fre3 "free" 14
> > g++.dg/tree-ssa/pr61034.C -std=gnu++14 scan-tree-dump-
On 10/3/19 12:16 PM, Jakub Jelinek wrote:
C++17 and later say
"The postfix-expression is sequenced before each expression in the
expression-list
and any default argument."
As the following testcase shows, we don't honor that. Either
cp_gimplify_expr gimplifies some arguments (if CALL_EXPR_REVER
On Tue, Oct 01, 2019 at 05:56:00PM -0400, Jason Merrill wrote:
> > + else if (DECL_NAME (var) == heap_deleted_identifier)
> > + {
> > + if (!ctx->quiet)
> > + error_at (loc, "deallocation of already deallocated "
> > +
Hi!
On Tue, Oct 01, 2019 at 05:56:00PM -0400, Jason Merrill wrote:
> I think we want to factor this function more, so we don't have the same code
> in multiple places for handling an array, and an array member, and a pointer
> to array. Do you want to take a look at bug 71504 while you're touchin
On Thu, Oct 03, 2019 at 09:35:03AM -0700, Julian Brown wrote:
> This patch provides some macros naming special offset field values used
> in libgomp/target.c. It was previously posted as part of the series
> supporting OpenACC 2.6 manual deep copy here:
>
> https://gcc.gnu.org/ml/gcc-patches/20
Remove the extra -L directory so that user directories specified at
build time have higher preference over the advance toolchain libraries.
2019-10-03 Tulio Magno Quites Machado Filho
* config.gcc: Remove an extra -L from --with-advance-toolchain.
---
gcc/config.gcc | 6 ++
1 file
Steve,
The attach patch silences -Wconversion and -Wconversion-extra
warnings that had previously been issued for explicit conversions
(see testcase for examples).
The patch has been tested on x86-*-freebsd. OK to commit?
OK.
Thanks a lot for the patch!
Regards
Thomas
This patch provides self-checking for consistency of the OpenACC
reference-counting implementation in libgomp. Tested alongside (and
dependent on) the patch posted adjacent to this one that overhauls said
reference-counting implementation.
Tested (with RC_CHECKING enabled) with offloading to NVPT
This patch has been broken out of the patch supporting OpenACC 2.6 manual
deep copy last posted here:
https://gcc.gnu.org/ml/gcc-patches/2018-12/msg01084.html
As part of developing that patch, libgomp's OpenACC reference counting
implementation proved to be somewhat inconsistent, especially whe
This patch provides some macros naming special offset field values used
in libgomp/target.c. It was previously posted as part of the series
supporting OpenACC 2.6 manual deep copy here:
https://gcc.gnu.org/ml/gcc-patches/2018-11/msg00825.html
Parts of that patch were committed already by Thoma
On Thu, Oct 03, 2019 at 04:26:55PM +0100, Richard Earnshaw (lists) wrote:
> Well if that's how tightly ALL_REGS is defined, then really no back-end
> should be defining it at all. Instead the build system (or the run-time
> initialization code) should be constructing it on the fly. Similarly
>
Hi!
C++17 and later say
"The postfix-expression is sequenced before each expression in the
expression-list
and any default argument."
As the following testcase shows, we don't honor that. Either
cp_gimplify_expr gimplifies some arguments (if CALL_EXPR_REVERSE_ARGS
or CALL_EXPR_ORDERED_ARGS or fo
On Thu, Oct 03, 2019 at 03:19:15PM +, Bernd Edlinger wrote:
>
> This fixes two -Wshadow=local warnings, and a undefined behavior that is
> used in the loop termination logic here:
>
> for (p = c - 1; p >= buffer; p--)
> {
> if (*p == '.')
> continue;
>
> In order to terminate the
> On Thu, 2019-10-03 at 17:29 +0200, Jan Hubicka wrote:
>
> [...]
>
> > In this patch I however untangle it from the hinting mechanism.
> >
> [...]
>
> > Bootstrapped/regtested x86_64-linux.
>
> If I'm reading this right it looks like the patch is missing.
Hi,
patch is attached.
* pa
On Thu, 2019-10-03 at 17:29 +0200, Jan Hubicka wrote:
[...]
> In this patch I however untangle it from the hinting mechanism.
>
[...]
> Bootstrapped/regtested x86_64-linux.
If I'm reading this right it looks like the patch is missing.
Dave
On 10/3/19 5:25 PM, Jakub Jelinek wrote:
> On Thu, Oct 03, 2019 at 03:17:47PM +, Bernd Edlinger wrote:
>> Hi,
>>
>> this fixes -Wshadow=local warnings in the RTL_FLAG_CHECKx macros,
>> which happen when this macro is used recursively in a macro
>> argument. The __typeof (RTX) const _rtx in the
Hi,
the previous patch (which enabled -finline-functions with limited
parameters at -O2) had bit unexpected effect on trashing tramp3d and dlv
performance which as not caught by LNT testing but is clearly visible in
legacy testers. I did know about the tramp3d problems but since LNT was
happy I tou
On 03/10/2019 15:15, Segher Boessenkool wrote:
On Thu, Oct 03, 2019 at 09:26:45AM -0400, Paul Koning wrote:
On Oct 3, 2019, at 9:12 AM, Richard Earnshaw (lists)
wrote:
On 03/10/2019 10:48, Segher Boessenkool wrote:
...
But ALL_REGS should contain *all* registers. The union of any two regis
On Thu, Oct 03, 2019 at 03:17:47PM +, Bernd Edlinger wrote:
> Hi,
>
> this fixes -Wshadow=local warnings in the RTL_FLAG_CHECKx macros,
> which happen when this macro is used recursively in a macro
> argument. The __typeof (RTX) const _rtx in the inner macro
> expansions shadows the outer mac
Hi,
This fixes two -Wshadow=local warnings, and a undefined behavior that is
used in the loop termination logic here:
for (p = c - 1; p >= buffer; p--)
{
if (*p == '.')
continue;
In order to terminate the loop p is decremented until it points
to buffer-1 (buffer is alloca'd, so that
Hi,
this would normally be a trivial changes, but unless I am mistaken,
I cannot simply commit to the go subtree. Except the -Wshadow=local
warnings this fixes also a few minor memory leaks.
Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
Is it OK for trunk?
Ian, can you please commit thi
Hi,
this fixes a few -Wshadow=local warnings in hash-table.h.
Since values of type size_t are assigned here to int variables
an overflow may happen resulting in memory leaks or malfunction.
Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
Is it OK for trunk?
Thanks
Bernd.
2019-10-03 Bern
Hi,
this fixes a few -Wshadow=local warnings in expr.c,
and a minor bug in expand_assignment.
The problem is the value if to_mode is usually GET_MODE (to_rtx),
concat_store_slow:;
rtx temp = assign_stack_temp (to_mode,
GET_MOD
Hi,
this fixes a -Wshadow=local warning when using AUTO_DUMP_SCOPE in
nested blocks. Since NAME i a string I cannot use it to create
a unique name for the auto_dump_scope object.
So I used XCONCAT2 from good old symcat.h, to mangle the __LINE__
macro into the scope object name.
Bootstrapped an
Hi,
this fixes a -Wshadow=local warning in the FOR_EACH_ALIAS macro
that happens when it is used in lto/lto-partition.c in a nested
block.
For now to keep the patch simple, using the fact that the ALIAS
parameter is always a simple name, concatenate _iter_ to make
the loop variable a reserved nam
Hi,
this fixes -Wshadow=local warnings in passes.c.
The non-trivial part is due to the PUSH_INSERT_PASSES_WITHIN
is used recursively, and shadows the local value p
in each invocation.
Fixed by using a helper class that restores the saved content
of p at the end of the block.
The shadowing variab
Hi,
this fixes -Wshadow=local warnings in genmatch.c itself and in generated
files as well.
The non-trivial part of this patch is renaming the generated local variables
in the gimple-match.c, to use different names for variables in inner blocks,
and use a depth index to access the right value. A
Hi,
this fixes -Wshadow=local warnings in the RTL_FLAG_CHECKx macros,
which happen when this macro is used recursively in a macro
argument. The __typeof (RTX) const _rtx in the inner macro
expansions shadows the outer macro expansions.
So reworked the macro to not use statement expressions but
u
Hi,
this adds -Wshadow=local to the GCC build rules.
It is to be applied after all other patches in this series
including the trivial ones are applied.
Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
Is it OK for trunk?
Thanks
Bernd.
2019-10-03 Bernd Edlinger
* configure.ac (ACX_PROG
Hi,
I've noticed that the documentation of -Wshadow=x has some
missing bits, and I want to add an negative example to
-Wshadow=compatble-local.
Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
Is it OK for trunk?
Thanks
Bernd.
2019-10-03 Bernd Edlinger
* doc/invoke.texi (-Wshadow=globa
Oh, and this old documentation bug. (AFAICT this is a historical
artifact from the pre-tree-ssa times when both -Wclobbered and
-Wuninitialized were implemented on RTL.)
This one is actually independent of the others and, I believe, is
obvious enough. I am going to install tomorrow.
>From 1ddc8
My motivation for the patch below comes from GDB. The binutils-gdb
project maintains a copy of the contrib directory that it keeps in
sync with upstream GCC, and patches to contrib/ are ideally first
applied to GCC then backported to binutils-gdb.
This patch extends the dg-extract-results.* scrip
On 03/10/19 16:24 +0200, Rainer Orth wrote:
_C is used in . Fixed as in other similar instances, tested on
i386-pc-solaris2.11. Ok for mainline?
Yes, thanks.
Hi Jonathan,
> * include/parallel/algo.h: Replace non-reserved names.
> * include/parallel/multiway_merge.h: Likewise.
> * include/parallel/multiway_mergesort.h: Likewise.
> * include/parallel/numericfwd.h: Likewise.
> * testsuite/17_intro/names.cc: Add RAI to test ma
On Thu, Oct 03, 2019 at 09:26:45AM -0400, Paul Koning wrote:
> > On Oct 3, 2019, at 9:12 AM, Richard Earnshaw (lists)
> > wrote:
> >
> > On 03/10/2019 10:48, Segher Boessenkool wrote:
> >>> ...
> >> But ALL_REGS should contain *all* registers. The union of any two register
> >> classes has to b
The attach patch silences -Wconversion and -Wconversion-extra
warnings that had previously been issued for explicit conversions
(see testcase for examples).
The patch has been tested on x86-*-freebsd. OK to commit?
2019-10-03 Steven G. Kargl
PR fortran/91497
* simplify.c (gfc
Branch created for Arm GCC releases.
- Joey
I found g++.dg/cpp0x/gen-attrs-67.C to FAIL on Solaris 11 when
configuring with --disable-initfini-array:
+FAIL: g++.dg/cpp0x/gen-attrs-67.C -std=c++14 (test for excess errors)
+FAIL: g++.dg/cpp0x/gen-attrs-67.C -std=c++17 (test for excess errors)
Excess errors:
/vol/gcc/src/hg/trunk/local/gcc/
> On Oct 3, 2019, at 9:12 AM, Richard Earnshaw (lists)
> wrote:
>
> On 03/10/2019 10:48, Segher Boessenkool wrote:
>>> ...
>> But ALL_REGS should contain *all* registers. The union of any two register
>> classes has to be contained in some register class, so every register class
>> has to be
On 03/10/2019 10:48, Segher Boessenkool wrote:
On Thu, Oct 03, 2019 at 10:35:15AM +0100, Richard Earnshaw (lists) wrote:
On 03/10/2019 00:01, Segher Boessenkool wrote:
The documentation says
@findex ALL_REGS
@findex NO_REGS
In general, each register will belong to several classes. In fact, on
Kyrill Tkachov writes:
> Hi Richard,
>
> On 10/3/19 1:45 PM, Richard Sandiford wrote:
>> Noticed while debugging the arm bootstrap failure.
>>
>> Tested on arm-linux-gnueabihf. OK to install?
>>
>> Richard
>>
>>
>> 2019-10-03 Richard Sandiford
>>
>> gcc/
>> * config/arm/arm.c (arm_print
On Mon, Sep 30, 2019 at 3:25 PM Paolo Carlini wrote:
>
> On 30/09/19 20:46, Jason Merrill wrote:
> > On Mon, Sep 30, 2019 at 6:54 AM Paolo Carlini
> > wrote:
> >> On 28/09/19 04:05, Jason Merrill wrote:
> >>> On 9/26/19 3:39 PM, Paolo Carlini wrote:
> +template void foo( // { dg-error "15:
Hi Richard,
On 10/3/19 1:45 PM, Richard Sandiford wrote:
Noticed while debugging the arm bootstrap failure.
Tested on arm-linux-gnueabihf. OK to install?
Richard
2019-10-03 Richard Sandiford
gcc/
* config/arm/arm.c (arm_print_value): Use real_to_decimal
to print CONST_DOU
Jeff Law writes:
> On 10/1/19 4:11 AM, Andrea Corallo wrote:
>> Martin Jambor writes:
>>
>>> Hi,
>>>
>>> On Mon, Sep 30 2019, Andrea Corallo wrote:
Hi all,
I'd like to submit this patch.
It release the ipa cp transformation summary after functions being
expanded.
This i
Noticed while debugging the arm bootstrap failure.
Tested on arm-linux-gnueabihf. OK to install?
Richard
2019-10-03 Richard Sandiford
gcc/
* config/arm/arm.c (arm_print_value): Use real_to_decimal
to print CONST_DOUBLEs.
Index: gcc/config/arm/arm.c
Ilya Leoshkevich writes:
> z13 supports only non-signaling vector comparisons. This means we
> cannot vectorize LT, LE, GT, GE and LTGT when compiling for z13.
> However, we cannot express this restriction today: the code only checks
> whether vcond$a$b optab exists, which does not contain inform
Hi,
On Wed, Oct 02 2019, Andreas Schwab wrote:
> FAIL: gcc.dg/ipa/ipa-sra-19.c (test for excess errors)
> Excess errors:
> /daten/gcc/gcc-20191001/gcc/testsuite/gcc.dg/ipa/ipa-sra-19.c:19:3: error:
> AltiVec argument passed to unprototyped function
> /daten/gcc/gcc-20191001/gcc/testsuite/gcc.dg/i
What it does
A fundamental limitation of the new approach is that it requires phi
nodes for variables to perform the analysis it needs to issue the
warning for them. No phis - no warning. In particular, it doesn't deal
with register asm variables (see gcc.target/i386/attr-returns_tw
This patch removes the old implementation of the warning on RTL.
Note: regstat_get_setjmp_crosses() and the associated data seems to be
needed for the RA itself, so I am not touching that.
>From c86a8fdc3f91d3cd33d73c7063bb9be2a7ae7e1f Mon Sep 17 00:00:00 2001
From: Vladislav Ivanishin
Date: Thu
This refactoring makes gsi_next_nonvirtual_phi() always advance the
gimple phi iterator (regardless of whether the current phi is virtual or
not). It matches the behavior of other gsi_next functions.
It can be applied independently of the other patches (but patch 3
depends on this one).
I regstr
Hi!
This series implements the -Wclobbered warning on GIMPLE in place of the
old RTL implementation.
This cover letter contains a high-level explanation of what is going on
and why. The new implementation itself is in the patch 3. I will post
the details in the accompanying blurb.
The most obv
On Thu, Oct 03, 2019 at 12:16:46PM +0100, Maciej W. Rozycki wrote:
> On Thu, 3 Oct 2019, Segher Boessenkool wrote:
> > When a (conditional) branch executes, it is detected if the branch was
> > mispredicted, and if so it (and everything after it) is flushed. This is
> > similar to how it was done
On Thu, 3 Oct 2019, Segher Boessenkool wrote:
> > I don't think this is right. As I have just mentioned in a related
> > discussion elsewhere, the NVAX and NVAX+ implementations include a branch
> > predictor in their microarchitecture[1], so obviously they do execute
> > speculatively. I th
On Thu, Oct 03, 2019 at 10:52:38AM +0100, Richard Sandiford wrote:
> /* Certain fixed registers might be of the class NO_REGS. This means
>that not only can they not be allocated by the compiler, but
>they cannot be used in substitutions or canonicalizations
>either.
Jakub Jelinek writes:
> On Thu, Oct 03, 2019 at 11:14:25AM +0100, Richard Sandiford wrote:
>> Confusingly, it needs to be --enable-checking=yes,extra,rtl for rtl
>> to be a pure addition to the default.
>>
>> I guess it's time for a new entry in the semi-regular series
>> "should rtl checking be
On Thu, Oct 03, 2019 at 03:12:34AM +0100, Maciej W. Rozycki wrote:
> On Fri, 20 Sep 2019, Jeff Law wrote:
>
> > > According to Bob Supnik (who is a very authoritative source on VAX),
> > >
> > >> Funny you should ask. The short answer is no. No VAX ever did
> > >> speculative or out of order exec
Ping: https://gcc.gnu.org/ml/gcc-patches/2019-09/msg01469.html
On Wed, Sep 25, 2019, at 8:25 PM, Eduard-Mihai Burtescu wrote:
> The main change here is in the treatment of $...$ escapes.
> I've relaxed the treatment of unknown escapes, during
> unescaping, to continue processing the input string,
On Thu, Oct 03, 2019 at 11:14:25AM +0100, Richard Sandiford wrote:
> Confusingly, it needs to be --enable-checking=yes,extra,rtl for rtl
> to be a pure addition to the default.
>
> I guess it's time for a new entry in the semi-regular series
> "should rtl checking be enabled by default for develop
Aaron Sawdey writes:
> On 10/2/19 5:35 PM, Jakub Jelinek wrote:
>> On Wed, Oct 02, 2019 at 09:21:23AM -0500, Aaron Sawdey wrote:
> 2019-09-27 Aaron Sawdey
>
> * builtins.c (expand_builtin_memory_copy_args): Add might_overlap parm.
> (expand_builtin_memcpy): Use might_overlap
Segher Boessenkool writes:
> On Wed, Oct 02, 2019 at 10:22:22PM +0100, Richard Sandiford wrote:
>> My change to the -fipa-ra bookkeeping used ALL_REGS as the supposedly
>> safe default assumption, but ALL_REGS isn't literally all registers,
>> just a close approximation.
>>
>> This caused a boots
On Thu, Oct 03, 2019 at 10:35:15AM +0100, Richard Earnshaw (lists) wrote:
> On 03/10/2019 00:01, Segher Boessenkool wrote:
> >The documentation says
> >
> >@findex ALL_REGS
> >@findex NO_REGS
> >In general, each register will belong to several classes. In fact, one
> >class must be named @code{ALL
On 03/10/2019 00:01, Segher Boessenkool wrote:
On Wed, Oct 02, 2019 at 10:22:22PM +0100, Richard Sandiford wrote:
My change to the -fipa-ra bookkeeping used ALL_REGS as the supposedly
safe default assumption, but ALL_REGS isn't literally all registers,
just a close approximation.
This caused a
On Thu, Oct 3, 2019 at 12:46 AM Fangrui Song wrote:
>
>
> On 2019-10-03, Andrew Pinski wrote:
> >On Wed, Oct 2, 2019@9:52 PM Fangrui Song wrote:
> >>
> >> On 2019-09-24, Martin Liška wrote:
> >> >On 9/19/19 10:33 AM, Martin Liška wrote:
> >> >> - One needs modified binutils and I that would proba
On 2019-10-03, Andrew Pinski wrote:
On Wed, Oct 2, 2019@9:52 PM Fangrui Song wrote:
On 2019-09-24, Martin Liška wrote:
>On 9/19/19 10:33 AM, Martin Liška wrote:
>> - One needs modified binutils and I that would probably require a configure
detection. The only way
>> which I see is based o
On Wed, Oct 2, 2019 at 9:52 PM Fangrui Song wrote:
>
> On 2019-09-24, Martin Liška wrote:
> >On 9/19/19 10:33 AM, Martin Liška wrote:
> >> - One needs modified binutils and I that would probably require a
> >> configure detection. The only way
> >> which I see is based on ld --version. I'm plan
97 matches
Mail list logo
