Random guest crashes since 5c34d002dcc7 ("virtio_pci: use shared interrupts for virtqueues")

From: Thorsten Leemhuis
Date: Thu Mar 23 2017 - 11:09:31 EST


Hi Christoph! Hi Michael!

(Mail roughly based on text from
https://bugzilla.kernel.org/show_bug.cgi?id=194911 )

I'm seeing random crashes during boot every few boot attempts when
running Linux 4.11-rc/mainline in a Fedora 26 guest under a CentOS7 host
(CPU: Intel(R) Pentium(R) CPU G3220) using KVM. Sometimes when the guest
actually booted the network did not work. To get some impressions of the
crashes I got see this gallery:
https://plus.google.com/+ThorstenLeemhuis/posts/FjyyGjNtrrG

Richard W.M. Jones and Adam Williamson see the same problems. See above
bug for details. It seems they ran into the problem in the past few
days, so I assume it's still present in mainline (I'm travelling
currently and haven't had time for proper tests since last last Friday
(pre-rc3); but I thought it's time to get the problem to the lists).

Long story short: Richard and I did bisections and we both found that
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=07ec51480b5e
("virtio_pci: use shared interrupts for virtqueues") is the first bad
commit. Any idea what might be wrong? Do you need more details from us
to fix this?

Ciao, Thorsten

P.S.: Sorry, I should have written this mail a few days ago after filing
above bug report, but I didn't get around to it :-/