https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115804
--- Comment #3 from Andrew Pinski ---
static struct builtin_function_entry list_of_builtins[] = {
This is not marked with GTY at all.
Though gm2_isfinite_node is.
static GTY (()) tree gm2_isfinite_node;
So that might not be GC'ed away.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115804
Andrew Pinski changed:
What|Removed |Added
Assignee|unassigned at gcc dot gnu.org |gaius at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115802
--- Comment #2 from Andrew Pinski ---
I am not 100% sure this is a bug since cond_wait is known not to touch val even
though it has an atomic in it. If you mark it as noipa, then the GCC trunk gets
it "correct".
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115802
--- Comment #1 from Andrew Pinski ---
Note clang has similar code generation here.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115782
Andrew Pinski changed:
What|Removed |Added
Summary|ICE on valid code at|[15 Regression] ICE on
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115783
Andrew Pinski changed:
What|Removed |Added
Ever confirmed|0 |1
Keywords|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115796
Andrew Pinski changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115796
Andrew Pinski changed:
What|Removed |Added
Target Milestone|--- |15.0
Summary|build failure
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
Andrew Pinski changed:
What|Removed |Added
Summary|GCC accepts [] throw () {} |throw() should be
|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115797
--- Comment #3 from Andrew Pinski ---
(In reply to Jonathan Wakely from comment #1)
> Doing this is undefined (and annoying).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=13071#c12
Is libstdc++ still allowing this for qoi? Or we going to
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #8 from Andrew Pinski ---
(In reply to 康桓瑋 from comment #7)
> (In reply to Andrew Pinski from comment #6)
> > So I looked into the change for clang:
> > https://github.com/llvm/llvm-project/commit/
> >
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #6 from Andrew Pinski ---
So I looked into the change for clang:
https://github.com/llvm/llvm-project/commit/0620e6f4b76a9725dbd82454d58c5a68a7e47074
And they didn't add a testcase for throw(). Only noexcept.
GCC add a testcase
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #5 from Andrew Pinski ---
Looks like clang misses exception specifiers as being optional before the parm.
Most likely because only `throw()` is valid for C++17+.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #4 from Andrew Pinski ---
(In reply to Andrew Pinski from comment #1)
> Note it looks like clang does not implement `parameter declaration before
> lambda declaration specifiers only optional with C++23` as clang rejects:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #3 from Andrew Pinski ---
"Exception specifications and noexcept" from the paper.
So yes GCC is correct.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115792
--- Comment #1 from Andrew Pinski ---
Note it looks like clang does not implement `parameter declaration before
lambda declaration specifiers only optional with C++23` as clang rejects:
```
auto l = [] static {};
```
Note all 3 accept:
```
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115791
--- Comment #3 from Andrew Pinski ---
So fixed for GCC 11. Note GCC 10.5.0 was the last release of GCC 10.x series
and there will be no other new 10.x.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115790
--- Comment #2 from Andrew Pinski ---
Also it is documented this way too:
-Wdeprecated-copy (C++ and Objective-C++ only)
Warn that the implicit declaration of a copy constructor or copy assignment
operator is deprecated if the class has a
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115790
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115780
Andrew Pinski changed:
What|Removed |Added
See Also||https://gcc.gnu.org/bugzill
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113116
--- Comment #5 from Andrew Pinski ---
This looks fixed since https://lnt.opensuse.org/db_default/v4/SPEC/49489 .
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
Bug 26163 depends on bug 114481, which changed state.
Bug 114481 Summary: 14% exec time slowdown of 433.milc on aarch64
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114481
What|Removed |Added
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114481
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656
Andrew Pinski changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
Bug 26163 depends on bug 114656, which changed state.
Bug 114656 Summary: [15 Regression] ~5% slowdown of 538.imagick_r on aarch64
since r14-9692
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656
What|Removed
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115463
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |FIXED
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
Bug 26163 depends on bug 115463, which changed state.
Bug 115463 Summary: [15 regression] 526.blender_r regressed 5% on Zen2 with
-Ofast -flto -march=native since r15-1058-gc989e59fc99d99
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115463
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112915
--- Comment #4 from Andrew Pinski ---
This looks like it was fixed in Jan 2024 around between 02948ced062b730e and
7f7d9c525c694e36 revision (https://lnt.opensuse.org/db_default/v4/SPEC/44304).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115777
Andrew Pinski changed:
What|Removed |Added
Last reconfirmed||2024-07-03
Keywords|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115778
Andrew Pinski changed:
What|Removed |Added
Known to fail||11.1.0
--- Comment #2 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101232
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever confirmed|0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115770
Andrew Pinski changed:
What|Removed |Added
Resolution|FIXED |INVALID
--- Comment #6 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115768
--- Comment #6 from Andrew Pinski ---
(In reply to biggs from comment #5)
> So the argument here is that C23's constexpr does not permit this
> optimization because it does not allow constexpr pointers other than nullptr?
No I am saying the
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103732
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115768
--- Comment #4 from Andrew Pinski ---
(In reply to biggs from comment #3)
>
> The unused names are optimized out here and string_view is simply a
> contiguous array of characters not pointers.
Nope, it is still an array of pointers (and
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103732
--- Comment #4 from Andrew Pinski ---
(In reply to Andrew Pinski from comment #3)
> >constexpr int y = s0.C;
>
>
> The above is now valid due to https://wg21.link/p2280r2 (which was acecpted
> as a defect report against all C++
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115770
Andrew Pinski changed:
What|Removed |Added
Resolution|FIXED |INVALID
--- Comment #4 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72756
--- Comment #7 from Andrew Pinski ---
Note the paper that was acepted in the end is https://wg21.link/p2280 .
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105696
--- Comment #4 from Andrew Pinski ---
(In reply to Andrew Pinski from comment #3)
> https://wg21.link/p2280r2 was accepted as a defect report against all
> versions of C++.
Note the correct link is https://wg21.link/p2280 .
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92171
Andrew Pinski changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105696
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103732
--- Comment #3 from Andrew Pinski ---
>constexpr int y = s0.C;
The above is now valid due to https://wg21.link/p2280r2 (which was acecpted as
a defect report against all C++ versions).
>constexpr int z = a[i]->C;
I think this is
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79083
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|NEW
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115768
--- Comment #2 from Andrew Pinski ---
For C++ (well with GCC extensions obvious) this would be valid:
```
static constexpr const char *names[] = {
[CE_RED] = "RED",
[CE_GREEN] = "GREEN",
[CE_BLUE] = "BLUE",
};
```
And gets
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115768
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106650
Andrew Pinski changed:
What|Removed |Added
CC||Shun.Yao at de dot bosch.com
---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55004
Bug 55004 depends on bug 115772, which changed state.
Bug 115772 Summary: static_assert rejected constexpr member function with
non-constexpr this
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115772
What|Removed
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115772
Andrew Pinski changed:
What|Removed |Added
Target Milestone|14.2|14.0
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106650
Andrew Pinski changed:
What|Removed |Added
Target Milestone|--- |14.0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115755
--- Comment #2 from Andrew Pinski ---
(In reply to Hongtao Liu from comment #1)
> mulx doesn't support imm operand, a register is still needed to put 123.
> mulq is used func/func1 should be ok.
Right, but mulx does not set the flags so it
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115771
Andrew Pinski changed:
What|Removed |Added
Known to work||15.0
Keywords|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115771
--- Comment #1 from Andrew Pinski ---
Note the testcase needs:
```
#include
```
for newer gcc versions.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115772
Andrew Pinski changed:
What|Removed |Added
Keywords||accepts-invalid,
|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115761
--- Comment #4 from Andrew Pinski ---
(In reply to Andrew Pinski from comment #2)
> "f int]::]::]:: >"
>
>
> vs
>
> "f(int):: >(f(int)::)::
> >(f(int)::
> >(f(int)::)::):: >"
> For the DW_AT_name
>
> Clang just uses "f" or "f<(lambda at
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115761
Andrew Pinski changed:
What|Removed |Added
Resolution|INVALID |WONTFIX
--- Comment #3 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115761
Andrew Pinski changed:
What|Removed |Added
Component|c++ |debug
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115762
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115761
--- Comment #1 from Andrew Pinski ---
You know what the difference in gcc 11 vs GCC 12 is actually dwarf 4 vs dwarf
5.
GCC now outputs a lot more debugging information in GCC 12 vs before.
You can get the same size back with `-gdwarf-4`
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107931
Andrew Pinski changed:
What|Removed |Added
CC||Changqing.Li at windriver dot
com
---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115679
Andrew Pinski changed:
What|Removed |Added
Status|WAITING |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115679
--- Comment #8 from Andrew Pinski ---
(In reply to Jan Hubicka from comment #2)
> > With -Og it's usually that the always-inline function is called indirectly -
> > that's an unsupported case.
> We can probably add CIF code for functions that
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115760
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115753
--- Comment #7 from Andrew Pinski ---
Here is a float/longlong that fails for both aarch64 and x86_64 (it works on
i686 though):
```
void f(float*);
void
foo1 ()
{
long long t0 = __LONG_LONG_MAX__;
long long t1 = __LONG_LONG_MAX__ - 1;
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115758
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |INVALID
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115753
--- Comment #5 from Andrew Pinski ---
Note the correct way to have a testcase that is able to handle float16 is to
do:
/* { dg-add-options float16 } */
/* { dg-require-effective-target float16 } */
This will allow it to work on 32bit x86 and
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115753
--- Comment #4 from Andrew Pinski ---
Here is one testcase that ICEs on aarch64 in the same way at `-O2
-frounding-math` since there is V4HF support there:
```
void f(_Float16*);
void
foo1 ()
{
int t0 = 3967;
int t1 = 3969;
int t2 = 3971;
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115753
--- Comment #3 from Andrew Pinski ---
(In reply to Hu Lin from comment #2)
> Created attachment 58572 [details]
> Untested fix.
>
> Confirmed, I need to check if TYPE_CODE is SSA_NAME before
> SSA_NAME_RANGE_INFO.
I think you could make the
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115751
Andrew Pinski changed:
What|Removed |Added
Summary|ICE building 521.wrf_r |[15 Regression] ICE
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115679
--- Comment #6 from Andrew Pinski ---
(In reply to Li, Changqing from comment #4)
> Created attachment 58570 [details]
> preprocessed source
This is not the preprocessed source.
Please read https://gcc.gnu.org/bugs/ and try again.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115743
Andrew Pinski changed:
What|Removed |Added
Status|WAITING |NEW
--- Comment #7 from Andrew Pinski
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115743
--- Comment #5 from Andrew Pinski ---
Can you attach the resulting `libstdc++.*-gdb.py` file (without the patch
included) and where it is located?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109130
--- Comment #5 from Andrew Pinski ---
This looks like it was fixed already, back to 355 which is GCC 13 .
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114412
--- Comment #6 from Andrew Pinski ---
Looks like the regression mentioned in comment #0 is fixed. And even the
regression that happened during GCC 14. There is still more from GCC 13 though.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114510
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |FIXED
Target Milestone|14.2
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115705
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |SUSPENDED
Ever confirmed|0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88670
Bug 88670 depends on bug 110473, which changed state.
Bug 110473 Summary: vec_convert for aarch64 seems to lower to something which
should be improved
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110473
What|Removed
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110473
Andrew Pinski changed:
What|Removed |Added
Target Milestone|--- |15.0
Status|ASSIGNED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115710
Andrew Pinski changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
Bug 53947 depends on bug 115710, which changed state.
Bug 115710 Summary: [11/12/13/14/15 Regression] complex abs does not vectorise
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115710
What|Removed |Added
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115743
--- Comment #3 from Andrew Pinski ---
So for a cross to aarch64-linux-gnu we get:
pythondir =
'/home/apinski/src/upstream-full-cross/install/share/gcc-15.0.0/python'
libdir =
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115743
--- Comment #2 from Andrew Pinski ---
Note for native, normally installs under
/home/apinski/upstream-gcc-new/share/gcc-14.0.0/python/libstdcxx/v6/printers.py
That is:
$prefix/share/gcc-$version/python/libstdcxx/v6/ .
For cross (to *-*-linux)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115743
Andrew Pinski changed:
What|Removed |Added
Last reconfirmed||2024-07-02
Ever confirmed|0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115749
--- Comment #8 from Andrew Pinski ---
(In reply to kim.walisch from comment #4)
> One possible explanation for why GCC's current integer division by a
> constant assembly sequence was chosen back in the day (I guess one or two
> decades ago) is
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115753
Andrew Pinski changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115749
Andrew Pinski changed:
What|Removed |Added
Last reconfirmed||2024-07-02
Depends on|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115756
Bug ID: 115756
Summary: default tuning for x86_64 produces shifts for `*240`
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115749
Andrew Pinski changed:
What|Removed |Added
Depends on||115755
--- Comment #6 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115755
Bug ID: 115755
Summary: mulx (with -mbmi2) does not show up with constant
multiply
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115749
--- Comment #5 from Andrew Pinski ---
For the original testcase, the imul vs shift can be reduced down to just:
```
unsigned long func(unsigned long x)
{
return x * 240;
}
```
GCC produces:
```
movq%rdi, %rax
salq$4,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115749
Andrew Pinski changed:
What|Removed |Added
Target||X86_64
--- Comment #2 from Andrew
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115742
Andrew Pinski changed:
What|Removed |Added
Last reconfirmed||2024-07-02
Ever confirmed|0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107432
Andrew Pinski changed:
What|Removed |Added
Target Milestone|--- |15.0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110473
Andrew Pinski changed:
What|Removed |Added
Assignee|unassigned at gcc dot gnu.org |pinskia at gcc dot
gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109009
Andrew Pinski changed:
What|Removed |Added
See Also||https://gcc.gnu.org/bugzill
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88403
--- Comment #3 from Andrew Pinski ---
aarch64 started doing the ifcvt in GCC 14.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109436
Andrew Pinski changed:
What|Removed |Added
CC||pinskia at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115724
Andrew Pinski changed:
What|Removed |Added
Keywords||diagnostic
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115740
Andrew Pinski changed:
What|Removed |Added
Resolution|--- |MOVED
Status|NEW
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115740
--- Comment #10 from Andrew Pinski ---
(In reply to Lockal from comment #6)
> 2) Specifically clang dislikes calls to host code from device code
THERE IS NO device or host code in libstdc++. That in itself a hack on how
cuda/ROCm are hacked
1 - 100 of 24043 matches
Mail list logo