[Intel-gfx] ✗ Fi.CI.BAT: warning for kernfs: Move faulting copy_user operations outside of the mutex (rev2)

Joonas Lahtinen joonas.lahtinen at linux.intel.com
Tue Mar 29 12:45:48 UTC 2016


On ti, 2016-03-29 at 13:38 +0100, Chris Wilson wrote:
> On Tue, Mar 29, 2016 at 03:31:24PM +0300, Joonas Lahtinen wrote:
> > 
> > On to, 2016-03-24 at 17:03 +0000, Chris Wilson wrote:
> > > 
> > > On Thu, Mar 24, 2016 at 04:34:01PM -0000, Patchwork wrote:
> > > > 
> > > > 
> > > > == Series Details ==
> > > > 
> > > > Series: kernfs: Move faulting copy_user operations outside of the mutex (rev2)
> > > > URL   : https://patchwork.freedesktop.org/series/3722/
> > > > State : warning
> > > > 
> > > > == Summary ==
> > > > 
> > > > Series 3722v2 kernfs: Move faulting copy_user operations outside of the mutex
> > > > http://patchwork.freedesktop.org/api/1.0/series/3722/revisions/2/mbox/
> > > > 
> > > > Test pm_rpm:
> > > >         Subgroup basic-rte:
> > > >                 pass       -> DMESG-WARN (bsw-nuc-2)
> > This WARN is about Unclaimed access detected, which we have an open bug
> > for.
> > 
> > > 
> > > > 
> > > >                 dmesg-warn -> PASS       (byt-nuc) UNSTABLE
> > > Something is notable by its absence here!
> > Other tests were fine in the run.
> > 
> > > 
> > > 
> > > Would someone do the honours and see if the suspend tests pass without
> > > lockdep WARNs on Brasweel?
> > To me it seems it worked just fine. OK if I merge this (and rebase if
> > needed)?
> I was expecting to see a fair few dmesg-warn -> PASS (bsw) since the
> purpose of this patch is to cut down on the CI noise. If they are
> passing without the patch, this patch is not required, right?

The problem with the lockdep issues that are plaguing the CI is they
only appear every now and then. So we would only notice after applying,
when the amount of noise reduces. I would use the above as a reference
that the patch will not make things worse.

You can take a look at /archive/results/CI_IGT_test/bsw-nuc-2.html to
see for yourself a long-term history.

Regards, Joonas

> -Chris
> 
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation


More information about the Intel-gfx mailing list