upstream build error (12)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: e609571b Merge tag 'nfs-for-5.11-2' of git://git.linux-nfs..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=15965a00d00000
kernel config: https://syzkaller.appspot.com/x/.config?x=6157970d0a91b812
dashboard link: https://syzkaller.appspot.com/bug?extid=76880518931d755473cf
compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project.git ca2dcbd030eadbf0aa9b660efe864ff08af6e18b)

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+768805...@syzkaller.appspotmail.com

clang-11: error: unable to execute command: Aborted (core dumped)
clang-11: error: clang frontend command failed due to signal (use -v to see invocation)

---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

[Nathan Chancellor]

Would it be possible for clang-bu...@googlegroups.com to be CC'd
when there is a build error and the compiler is clang? Especially if
clang is hitting an assertion.

Cheers,
Nathan

[Nathan Chancellor]

I cannot reproduce this with clang 11.0.1. I did a reverse bisect on
LLVM and found commit 4b0aa5724fea ("Change the INLINEASM_BR MachineInstr
to be a non-terminating instruction.") as the fix (which makes sense,
that commit has direct kernel implications). It is probably worth
upgrading syzkaller's clang to 11.0.1, where I have not seen any
assertions across all of the configurations that I test.

Ccing Dmitry and Marco directly.

Cheers,
Nathan

[Marco Elver]

I'll upgrade syzbot's clang 11. I think the version we use right now
was still a pre-release.

Thanks,
-- Marco

[Dmitry Vyukov]

Hi Nathan,

I am adding functionality to CC specific emails on build errors on
specific instances:
https://github.com/google/syzkaller/pull/2388

[Dmitry Vyukov]

I've updated configs to CC clang-bu...@googlegroups.com. This
should take effect on the next build failure (hopefully).

[Nick Desaulniers]

Thanks Dmitry, (and thanks Marco for updating syzbot's clang).

Marco, can you update:
https://github.com/google/syzkaller/blob/master/docs/syzbot.md#crash-does-not-reproduce

I wasn't able to reproduce the reported compiler crash with ToT llvm
and the reported config, so I assume it was a mid-release of llvm-11
bug that has since been fixed. Do we need to report to syzbot to
close the issue? I did a quick skim of
https://github.com/google/syzkaller/blob/master/docs/syzbot.md, but it
looks like communication with syzbot is based on commit messages?
--
Thanks,
~Nick Desaulniers

[Marco Elver]

On Thu, 14 Jan 2021 at 19:07, Nick Desaulniers <ndesau...@google.com> wrote:
> On Thu, Jan 14, 2021 at 4:38 AM 'Dmitry Vyukov' via Clang Built Linux
> <clang-bu...@googlegroups.com> wrote:
> >
> > On Thu, Jan 14, 2021 at 12:32 PM Dmitry Vyukov <dvy...@google.com> wrote:
> > >
> > > On Tue, Jan 12, 2021 at 8:41 PM Nathan Chancellor
> > > <natecha...@gmail.com> wrote:
> > > >
> > > > Would it be possible for clang-bu...@googlegroups.com to be CC'd
> > > > when there is a build error and the compiler is clang? Especially if
> > > > clang is hitting an assertion.
> > >
> > > Hi Nathan,
> > >
> > > I am adding functionality to CC specific emails on build errors on
> > > specific instances:
> > > https://github.com/google/syzkaller/pull/2388
> >
> > I've updated configs to CC clang-bu...@googlegroups.com. This
> > should take effect on the next build failure (hopefully).
>
> Thanks Dmitry, (and thanks Marco for updating syzbot's clang).
>
> Marco, can you update:
> https://github.com/google/syzkaller/blob/master/docs/syzbot.md#crash-does-not-reproduce

We have:

> compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project.git ca2dcbd030eadbf0aa9b660efe864ff08af6e18b)

Which is in the list of compilers:
https://storage.googleapis.com/syzkaller/clang-11-prerelease-ca2dcbd030e.tar.xz

> I wasn't able to reproduce the reported compiler crash with ToT llvm
> and the reported config, so I assume it was a mid-release of llvm-11
> bug that has since been fixed. Do we need to report to syzbot to
> close the issue? I did a quick skim of
> https://github.com/google/syzkaller/blob/master/docs/syzbot.md, but it
> looks like communication with syzbot is based on commit messages?

Given we this is likely due to a prerelease compiler, we can just use:

#syz invalid

^^ it should pick this up.

Thanks,
-- Marco