Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <000000000000eb546f0570e84e90@google.com> <20180713145811.683ffd0043cac26a5a5af725@linux-foundation.org>
 <CACT4Y+b1HC5CtFSQJEDBJrP8u1brKxXaFcYKE=g+h3aOW6K3Kg@mail.gmail.com>
 <5f715c98-2b81-4eee-be3c-11cbc1bc36a8@googlegroups.com> <CACT4Y+ZCKFXK+9Bw1__ofUBLy6y8mQRoQHm5Qt135mByOrYk8g@mail.gmail.com>
 <CALMp9eRuaFnav=ZUJdb1AefkCPhzH9rjpSHis_4XNVFG4rRimg@mail.gmail.com>
 <CALMp9eTVzEMpGEOhOKxH8tjyggeXFJ=+3xxipzQ=yFB3cT85GA@mail.gmail.com> <CACT4Y+bCat++Gc5H2JYjsQQ2hzQyEa3H-VKmb93H7tLrTVS31A@mail.gmail.com>
In-Reply-To: <CACT4Y+bCat++Gc5H2JYjsQQ2hzQyEa3H-VKmb93H7tLrTVS31A@mail.gmail.com>
From:   Dmitry Vyukov <dvyukov@google.com>
Date:   Thu, 4 Mar 2021 12:25:50 +0100
Message-ID: <CACT4Y+Ze-AjTb2ri6D5-CyDtJvXVmww-DaT-cUBzWGdouwKo5Q@mail.gmail.com>
Subject: Re: unexpected kernel reboot (3)
To:     Jim Mattson <jmattson@google.com>
Cc:     syzkaller-bugs <syzkaller-bugs@googlegroups.com>,
        LKML <linux-kernel@vger.kernel.org>,
        syzbot <syzbot+cce9ef2dd25246f815ee@syzkaller.appspotmail.com>,
        Peter Shier <pshier@google.com>,
        Ben Gardon <bgardon@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Mar 13, 2020 at 8:43 AM Dmitry Vyukov <dvyukov@google.com> wrote:
>
> On Thu, Mar 12, 2020 at 9:36 PM Jim Mattson <jmattson@google.com> wrote:
> >
> > On Wed, Mar 11, 2020 at 1:35 PM Jim Mattson <jmattson@google.com> wrote:
> > >
> > > On Wed, Mar 11, 2020 at 1:18 PM Dmitry Vyukov <dvyukov@google.com> wrote:
> > >
> > > > This happened 10K+ times.
> > > > If GCE VM is rebooted by doing something with KVM subsystem, I assume
> > > > it's a GCE bug (?). +Jim
> > >
> > > The L1 guest may have done something that's unsupported, or it may
> > > just be a bug in the L0 hypervisor.
> >
> > I can't reproduce this in GCE with the collateral provided.
>
> As far as I understand, this may depend on GCE host kernel and/or host
> CPU, right? Do we have an ability to unit test on all combinations?
> Looking at the rate and reproducers on the dashboard I see interesting
> patterns. Perhaps the best time to try to reproduce is these periods
> when syzbot comes up with lots of reproducers. But it's not now...


These unexpected VM reboots continue to happen. I can reproduce them
on GCE, but not in qemu.
Jim, Peter, Ben, I can give you access to a VM I just used to
reproduce the reboot.
Here is the latest repro I used:
https://gist.github.com/dvyukov/e9ddb9c3f1117f8cfd85713cf44d81bd