Re: Crashes with 874bbfe600a6 in 3.18.25

On Fri, 05 Feb 2016 03:40:46 +0100
Mike Galbraith [off-list ref] wrote:

On Thu, 2016-02-04 at 17:39 +0100, Daniel Bilik wrote:

quoted

On Thu, 4 Feb 2016 12:20:44 +0100
Jan Kara [off-list ref] wrote:

quoted

Thanks for backport Thomas and to Mike for persistence :). I've
asked my friend seeing crashes with 3.18.25 to try whether this
patch fixes the issues. It may take some time so stay tuned...

Patch tested and it really fixes the crash we were experiencing on
3.18.25 with commit 874bbfe+. But it seem to introduce (rather scary)
regression. Tested host shows abnormal cpu usage in both kernel and
userland under the same load and traffic pattern. One picture is worth
a thousand words, so I've taken snapshots of our graphs, see here:
http://neosystem.cz/test/linux-3.18.25/
The host was running 3.18.25 with commit 874bbfe+ (1e7af29+ on
3.18-stable) reverted. With this commit included, it crashed within
minutes. Around 13:30 we booted 3.18.25 with commit 874bbfe+ included
and with the patch from Thomas. And around 15:40 we've booted the host
with previous kernel, just to ensure this abnormal behaviour was
really caused by the test kernel.
Also interesting, in addition to high cpu usage, there is abnormally
high number of zombie processes reported by the system.

IMHO you should restore the CC list and re-post.  (If I were the
maintainer of either the workqueue code or 3.18-stable, I'd be highly
interested in this finding).

Sorry, I haven't realized tha patch proposed by Thomas is already on its
way to stable. CC restored and re-posting.

--
						Daniel Bilik
						neosystem.cz