multicore test error: timed out

[syzbot]

Hello,

syzbot found the following crash on:

HEAD commit: 690d147b bump to 2.9.1
git tree: openbsd
kernel config: https://syzkaller.appspot.com/x/.config?x=fa145722143cbd64
dashboard link: https://syzkaller.appspot.com/bug?extid=91a56393697f34a7f11f

Unfortunately, I don't have any reproducer for this crash yet.

IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+91a563...@syzkaller.appspotmail.com



---
This bug 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 bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

[Dmitry Vyukov]

On Wed, Mar 27, 2019 at 7:45 AM syzbot
<syzbot+91a563...@syzkaller.appspotmail.com> wrote:
>
> Hello,
>
> syzbot found the following crash on:
>
> HEAD commit: 690d147b bump to 2.9.1
> git tree: openbsd
> kernel config: https://syzkaller.appspot.com/x/.config?x=fa145722143cbd64
> dashboard link: https://syzkaller.appspot.com/bug?extid=91a56393697f34a7f11f
>
> Unfortunately, I don't have any reproducer for this crash yet.
>
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+91a563...@syzkaller.appspotmail.com

There is a weird splash of these bugs across 5 build at roughly the
same time. And one was few days before on gVisor.
I am still puzzled why there is no output available at all. Looking at
the code again I think we should collect and attach any ssh/vm output.
It looks like we invoked ssh to start syz-fuzzer but ssh handed for 10
minutes producing no output whatsoever. Theoretically this can be due
to our use of preemptible VMs that can go away at any point without
notice. We do some attempts to detect this only during fuzzing, but
not anywhere else. I think for starters we should stop using
preemptible VMs for anything other then fuzzing.