[pve-devel] kernel BUG at ./include/linux/swapops.h:129!

Fabian Grünbichler f.gruenbichler at proxmox.com
Mon May 8 10:12:45 CEST 2017


On Wed, Apr 26, 2017 at 08:48:09AM +0200, Fabian Grünbichler wrote:
> On Wed, Apr 19, 2017 at 10:39:32PM +0200, Michael Rasmussen wrote:
> > On Wed, 19 Apr 2017 22:29:23 +0200 (CEST)
> > Fabian Grünbichler <f.gruenbichler at proxmox.com> wrote:
> > 
> > > 
> > > we are aware of this issue (and have been chasing it for around two weeks already, it is most likely affecting all of the Ubuntu Zesty based 4.10 kernels we have released in the PVE 5 Beta so far). unfortunately there is no reliable reproducer so far - we've just seen three separate spurios occurrences on our dev machines.
> > > 
> > Could this eventually postpone the release of pve-5?
> > I for one will not do any upgrades to pve-5 before this issue has been
> > solved.
> > 
> 
> just as a short follow-up, while we still don't have a reproducer, it
> seems like I found the culprit (and missing followup fix)[1] based on
> the git history. initial feedback by affected Ubuntu 17.04 users is
> positive at least[2], so I guess this will be fixed in the next Ubuntu
> kernel release (and then in the PVE kernel based upon it).
> 
> 1: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1674838/comments/52
> 2: e.g., https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1674838/comments/67
> 

and another follow-up, the latest kernel from pvetest
(pve-kernel-4.10.8-1-pve in version 4.10.8-7) contains the proposed
upstream fix (reverting the unneeded, problematic commits, and instead
backporting the originally cherry-picked bug fix commit). so far none of
the affected Ubuntu users reported encountering the issue anymore with a
test kernel with this fix, so hopefully this saga has an end now ;)




More information about the pve-devel mailing list