[Bug lto/105600] New: ICE with LTO when building Mumble from git on arm64

[Bug lto/105600] New: ICE with LTO when building Mumble from git on arm64

[sam at gentoo dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

            Bug ID: 105600
           Summary: ICE with LTO when building Mumble from git on arm64
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sam at gentoo dot org
                CC: marxin at gcc dot gnu.org
  Target Milestone: ---

Reported this downstream at https://bugs.gentoo.org/843974 initially.

Apologies for not being able to give more information yet. I'm familiar with
reducing "normal" ICEs but I'm getting stuck with how to get more info because
of LTO.

One issue is it seems like I can't find a way to keep all the temporary files
required.

The bug seems similar to bug 102067 (same stacktrace?).

ICE occurs when building Mumble from their git repo
(https://github.com/mumble-voip/mumble) at
4f50172c5c8bc7c425efb350377106d3e83a7e79.

Needs LTO.

Occurs on arm64 with GCC 11.3.0 and GCC 12.1.0.

I've reduced it down to:
```
$ /usr/bin/c++ -freport-bug -flto=99 -o /dev/null -O2 -fsanitize=undefined -O3
-DNDEBUG -Wl,-z,relro -Wl,-z,now
/root/mumble/build/src/tests/TestSettingsJSONSerialization/CMakeFiles/TestSettingsJSONSerialization.dir/TestSettingsJSONSerialization.cpp.o
/root/mumble/build/src/tests/TestSettingsJSONSerialization/../../mumble/CMakeFiles/mumble_client_object_lib.dir/JSONSerialization.cpp.o
during IPA pass: icf
lto1: internal compiler error: Segmentation fault
0xb13af3 crash_signal
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/toplev.cc:322
0xdc3e64 varpool_node::get_constructor()
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/varpool.cc:300
0x15b8f53 ipa_icf::sem_variable::equals(ipa_icf::sem_item*,
hash_map<symtab_node*, ipa_icf::sem_item*,
simple_hashmap_traits<default_hash_traits<symtab_node*>, ipa_icf::sem_item*>
>&)
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/ipa-icf.cc:1695
0x15b9f67 ipa_icf::sem_item_optimizer::subdivide_classes_by_equality(bool)
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/ipa-icf.cc:2732
0x15c0683 ipa_icf::sem_item_optimizer::execute()
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/ipa-icf.cc:2464
0x15c1d6f ipa_icf_driver
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/ipa-icf.cc:3600
0x15c1d6f ipa_icf::pass_ipa_icf::execute(function*)
        /usr/src/debug/sys-devel/gcc-12.1.0/gcc-12.1.0/gcc/ipa-icf.cc:3647
Please submit a full bug report, with preprocessed source.
Please include the complete backtrace with any bug report.
See <https://bugs.gentoo.org/> for instructions.
lto-wrapper: fatal error: /usr/bin/c++ returned 1 exit status
compilation terminated.
/usr/lib/gcc/aarch64-unknown-linux-gnu/12.1.0/../../../../aarch64-unknown-linux-gnu/bin/ld:
error: lto-wrapper failed
collect2: error: ld returned 1 exit status
```

What more information do you need? I can give access to the environment if
required. Cheers.

[Bug lto/105600] ICE with LTO when building Mumble from git on arm64

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
>One issue is it seems like I can't find a way to keep all the temporary files required.

-save-temps is the best option there.
Also see
https://gcc.gnu.org/wiki/A_guide_to_testcase_reduction#Reducing_LTO_bugs for
maybe better help of reducing the testcase.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Alex Coplan <acoplan at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |acoplan at gcc dot gnu.org

--- Comment #2 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Hi Sam, can you reproduce this in (e.g.) an AArch64 Ubuntu environment? This
might make it easier for others to attempt to reproduce, as I suspect not many
have access to an AArch64 Gentoo environment.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #3 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Alternatively a Dockerfile to build the Gentoo environment needed to reproduce
the issue might help.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #4 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Created attachment 52988
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52988&action=edit
reduced testcase

I managed to reproduce the environment and reduce a testcase (see the attached
tarball):

$ cat build.sh
#!/bin/bash
set -e
CC=aarch64-linux-gnu-g++
$CC -fsanitize=undefined -flto -O2 -std=gnu++14 -o a.o -c a.cc
$CC -fsanitize=undefined -O2 -std=gnu++14 -o b.o -c b.cc
$CC b.o a.o
$ ./build.sh
a.cc: In instantiation of ‘nlohmann::basic_json< <template-parameter-1-1>,
<template-parameter-1-2>, <template-parameter-1-3>, <template-parameter-1-4>,
<template-parameter-1-5>, <template-parameter-1-6>, <template-parameter-1-7>,
<template-parameter-1-8>, <template-parameter-1-9>, <template-parameter-1-10>
>::reference nlohmann::basic_json< <template-parameter-1-1>,
<template-parameter-1-2>, <template-parameter-1-3>, <template-parameter-1-4>,
<template-parameter-1-5>, <template-parameter-1-6>, <template-parameter-1-7>,
<template-parameter-1-8>, <template-parameter-1-9>, <template-parameter-1-10>
>::operator[](T) [with T = const char*; ObjectType = std::map;
<template-parameter-1-2> = std::vector; <template-parameter-1-3> =
std::__cxx11::basic_string<char>; <template-parameter-1-4> = bool;
<template-parameter-1-5> = long int; <template-parameter-1-6> = long unsigned
int; <template-parameter-1-7> = double; <template-parameter-1-8> =
std::allocator; JSONSerializer = nlohmann::adl_serializer;
<template-parameter-1-10> = std::vector<unsigned char>; reference =
nlohmann::basic_json<>]’:
a.cc:87:65:   required from here
a.cc:83:3: warning: no return statement in function returning non-void
[-Wreturn-type]
   83 |   }
      |   ^
during IPA pass: icf
lto1: internal compiler error: Segmentation fault
0xd4b25f crash_signal
        /home/alecop01/toolchain/src/gcc/gcc/toplev.cc:322
0x11508b9 varpool_node::get_constructor()
        /home/alecop01/toolchain/src/gcc/gcc/varpool.cc:300
0x1bd2db6 ipa_icf::sem_variable::equals(ipa_icf::sem_item*,
hash_map<symtab_node*, ipa_icf::sem_item*,
simple_hashmap_traits<default_hash_traits<symtab_node*>, ipa_icf::sem_item*>
>&)
        /home/alecop01/toolchain/src/gcc/gcc/ipa-icf.cc:1697
0x1bd3e78 ipa_icf::sem_item_optimizer::subdivide_classes_by_equality(bool)
        /home/alecop01/toolchain/src/gcc/gcc/ipa-icf.cc:2731
0x1bdbce0 ipa_icf::sem_item_optimizer::execute()
        /home/alecop01/toolchain/src/gcc/gcc/ipa-icf.cc:2464
0x1bde1f2 ipa_icf_driver
        /home/alecop01/toolchain/src/gcc/gcc/ipa-icf.cc:3600
0x1bde1f2 ipa_icf::pass_ipa_icf::execute(function*)
        /home/alecop01/toolchain/src/gcc/gcc/ipa-icf.cc:3647
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
lto-wrapper: fatal error: aarch64-linux-gnu-g++ returned 1 exit status
compilation terminated.
/home/alecop01/toolchain/build-aarch64-linux-gnu/install/bin/../lib/gcc/aarch64-linux-gnu/13.0.0/../../../../aarch64-linux-gnu/bin/ld:
error: lto-wrapper failed
collect2: error: ld returned 1 exit status

Confirmed on trunk.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Alex Coplan <acoplan at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2022-05-18
      Known to work|                            |12.1.0, 13.0

--- Comment #5 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Confirmed.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #6 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Note the original bug also triggers on GCC 11 but the attached reduced testcase
doesn't, I'm attempting to reduce a testcase that crashes all versions.

[Bug ipa/105600] ICE with LTO when building Mumble from git on arm64

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |marxin at gcc dot gnu.org

--- Comment #7 from Martin Liška <marxin at gcc dot gnu.org> ---
I'm taking a look..

[Bug ipa/105600] [12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ICE with LTO when building  |[12/13 Regression] ICE with
                   |Mumble from git on arm64    |LTO when building Mumble
                   |                            |since
                   |                            |r12-5696-g53caa4723d8de73f
             Target|aarch64                     |aarch64, x86_64
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=103310
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |jason at gcc dot gnu.org

--- Comment #8 from Martin Liška <marxin at gcc dot gnu.org> ---
Can we seen also on x86_64-linux-gnu.
Started with r12-5696-g53caa4723d8de73f.

[Bug ipa/105600] [12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #9 from Martin Liška <marxin at gcc dot gnu.org> ---
So what happens in WPA:

1) we read IPA ICF summary and the variable has definition:

(gdb) p vnode->debug()
_ZN8nlohmann6detail12static_constINS0_10to_json_fnEE5valueE/10 (value)
@0x7ffff7fb3200
  Type: variable definition analyzed
  Visibility: externally_visible semantic_interposition preempted_reg public
weak comdat
comdat_group:_ZN8nlohmann6detail12static_constINS0_10to_json_fnEE5valueE
one_only
  References: 
  Referring:
_Z7to_jsonRN8nlohmann10basic_jsonISt3mapSt6vectorNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEblmiSaNS_14adl_serializerES2_IhSaIhEEEERK8Settings/1
(addr)
_Z7to_jsonRN8nlohmann10basic_jsonISt3mapSt6vectorNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEblmiSaNS_14adl_serializerES2_IhSaIhEEEERK8Settings/1
(addr) 
  Read from file: a.o
  Availability: not-ready
  Varpool flags: initialized read-only const-value-known
$1 = void

2) the symbols is removed in unreachable nodes:
#0  symbol_table::remove_unreachable_nodes (this=<optimized out>,
file=<optimized out>) at /home/marxin/Programming/gcc/gcc/ipa.cc:639
#1  0x00000000008e1277 in read_cgraph_and_symbols (nfiles=<optimized out>,
fnames=<optimized out>) at
/home/marxin/Programming/gcc/gcc/lto/lto-common.cc:2959
#2  0x00000000008c7540 in lto_main () at
/home/marxin/Programming/gcc/gcc/lto/lto.cc:626
#3  0x0000000000e229d4 in compile_file () at
/home/marxin/Programming/gcc/gcc/toplev.cc:452
#4  0x00000000008a0822 in do_compile (no_backend=false) at
/home/marxin/Programming/gcc/gcc/toplev.cc:2144
#5  toplev::main (this=this@entry=0x7fffffffd93e, argc=<optimized out>,
argc@entry=17, argv=<optimized out>, argv@entry=0x7fffffffda68) at
/home/marxin/Programming/gcc/gcc/toplev.cc:2296
#6  0x00000000008a1f32 in main (argc=17, argv=0x7fffffffda68) at
/home/marxin/Programming/gcc/gcc/main.cc:39

3) IPA ICF loads the body:
(gdb) p item->node->debug()
_ZN8nlohmann6detail12static_constINS0_10to_json_fnEE5valueE/10 (value)
@0x7ffff7fb3200
  Type: variable
  Body removed by symtab_remove_unreachable_nodes
  Visibility: externally_visible semantic_interposition preempted_reg external
public weak comdat
comdat_group:_ZN8nlohmann6detail12static_constINS0_10to_json_fnEE5valueE
one_only
  References: 
  Referring:
_Z7to_jsonRN8nlohmann10basic_jsonISt3mapSt6vectorNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEblmiSaNS_14adl_serializerES2_IhSaIhEEEERK8Settings/1
(addr)
_Z7to_jsonRN8nlohmann10basic_jsonISt3mapSt6vectorNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEblmiSaNS_14adl_serializerES2_IhSaIhEEEERK8Settings/1
(addr) 
  Read from file: a.o
  Availability: not_available
  Varpool flags: initialized read-only
$8 = void

So we need to filter our such symbols.

[Bug ipa/105600] [12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[sam at gentoo dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #10 from Sam James <sam at gentoo dot org> ---
Thanks Andrew. I spent a while trying to minimise it using that but I've got a
bit stuck. I'm used to reducing normal ICEs but link-time/LTO ones seem to be
harder to me.

I was going to upload the full lot with -save-temps and hope that might be
enough if I couldn't get further.

Thanks Alex - I've been setting up a Docker file for you to try out (although
was going to try clean up the env to give you access to the box too), but so
glad to see you guys have managed to get further.

Please tell me if I am needed to try and squeeze more out, although I'm likely
to need instructions given aforementioned.

[Bug ipa/105600] [12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #11 from Martin Liška <marxin at gcc dot gnu.org> ---
> Please tell me if I am needed to try and squeeze more out, although I'm
> likely to need instructions given aforementioned.

No, I've just sent a patch to mailing list that's going to fix it.

[Bug ipa/105600] [12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Alex Coplan <acoplan at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #52988|0                           |1
        is obsolete|                            |

--- Comment #12 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Created attachment 52994
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52994&action=edit
better reproducer (crashes GCC 9 through trunk)

Attaching a better reduced testcase that affects GCC versions all the way back
to GCC 9.

[Bug ipa/105600] [9/10/11/12/13 Regression] ICE with LTO when building Mumble since r12-5696-g53caa4723d8de73f

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |9.5

[Bug ipa/105600] [9/10/11/12/13 Regression] ICE with LTO when building Mumble

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Alex Coplan <acoplan at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11/12/13 Regression]  |[9/10/11/12/13 Regression]
                   |ICE with LTO when building  |ICE with LTO when building
                   |Mumble since                |Mumble
                   |r12-5696-g53caa4723d8de73f  |

--- Comment #13 from Alex Coplan <acoplan at gcc dot gnu.org> ---
FWIW, we started ICEing somewhere between
r9-856-g7ce4ce10cd0a6f332dd473bd4726f1a1adefda0e and
r9-870-ge062dc808950801a089ee6acd969393da2f39406, the build was broken between
these revisions.

[Bug ipa/105600] [9/10/11/12/13 Regression] ICE with LTO when building Mumble

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #14 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Alex Coplan from comment #13)
> FWIW, we started ICEing somewhere between
> r9-856-g7ce4ce10cd0a6f332dd473bd4726f1a1adefda0e and
> r9-870-ge062dc808950801a089ee6acd969393da2f39406, the build was broken
> between these revisions.

Sure, I've sent a fix that I'll backport later then.

[Bug ipa/105600] [10/11/12/13 Regression] ICE with LTO when building Mumble

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|9.5                         |10.4

--- Comment #15 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 9 branch is being closed

[Bug ipa/105600] [10/11/12/13 Regression] ICE with LTO when building Mumble

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #16 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Martin Liska <marxin@gcc.gnu.org>:

https://gcc.gnu.org/g:31ce821a790caec8a2849dd67a9847e78a33d14c

commit r13-1213-g31ce821a790caec8a2849dd67a9847e78a33d14c
Author: Martin Liska <mliska@suse.cz>
Date:   Wed May 18 15:07:53 2022 +0200

    ipa-icf: skip variables with body_removed

    Similarly to cgraph_nodes, it may happen that body_removed is set
    during merging of symbols.

            PR ipa/105600

    gcc/ChangeLog:

            * ipa-icf.cc (sem_item_optimizer::filter_removed_items):
            Skip variables with body_removed.

[Bug ipa/105600] [10/11/12 Regression] ICE with LTO when building Mumble

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |13.0
      Known to fail|13.0                        |
            Summary|[10/11/12/13 Regression]    |[10/11/12 Regression] ICE
                   |ICE with LTO when building  |with LTO when building
                   |Mumble                      |Mumble

--- Comment #17 from Martin Liška <marxin at gcc dot gnu.org> ---
Fixed on master so far.

[Bug ipa/105600] [10/11/12 Regression] ICE with LTO when building Mumble

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #18 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Martin Liska
<marxin@gcc.gnu.org>:

https://gcc.gnu.org/g:c600ff8dc5a870d952c9c63fe4fa4535b9f0a488

commit r12-8509-gc600ff8dc5a870d952c9c63fe4fa4535b9f0a488
Author: Martin Liska <mliska@suse.cz>
Date:   Wed May 18 15:07:53 2022 +0200

    ipa-icf: skip variables with body_removed

    Similarly to cgraph_nodes, it may happen that body_removed is set
    during merging of symbols.

            PR ipa/105600

    gcc/ChangeLog:

            * ipa-icf.cc (sem_item_optimizer::filter_removed_items):
            Skip variables with body_removed.

    (cherry picked from commit 31ce821a790caec8a2849dd67a9847e78a33d14c)

[Bug ipa/105600] [10/11/12 Regression] ICE with LTO when building Mumble

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

--- Comment #19 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Martin Liska
<marxin@gcc.gnu.org>:

https://gcc.gnu.org/g:3f1637f8b17ded9e08841abc1182dcde0e35e660

commit r11-10089-g3f1637f8b17ded9e08841abc1182dcde0e35e660
Author: Martin Liska <mliska@suse.cz>
Date:   Wed May 18 15:07:53 2022 +0200

    ipa-icf: skip variables with body_removed

    Similarly to cgraph_nodes, it may happen that body_removed is set
    during merging of symbols.

            PR ipa/105600

    gcc/ChangeLog:

            * ipa-icf.c (sem_item_optimizer::filter_removed_items):
            Skip variables with body_removed.

    (cherry picked from commit 31ce821a790caec8a2849dd67a9847e78a33d14c)

[Bug ipa/105600] [10/11/12 Regression] ICE with LTO when building Mumble

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105600

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #20 from Martin Liška <marxin at gcc dot gnu.org> ---
Fixed on gcc-11 and gcc-12 branches, skipping gcc-10 branch.