Re: Q. drm/i915 shrinker, synchronize_rcu_expedited() from handlers

From: J. R. Okajima
Date: Fri May 05 2017 - 19:29:11 EST

Joonas Lahtinen:
> Filing a bug in with all the details is the fastest way
> of getting help. Without the bug (and with such little information as
> the previous e-mail) it's hard to estimate the extent and nature of the
> bug.

My original report was

The report contained
- kernel command line options
- lockdep msg
- call trace
but it didn't look drm/i915 shrinker is related. It was git-bisect which
lead me to drm/i915 shrinker.

> I've anyway gone and prepared a patch to drop the RCU sync completely
> from shrinker phase, as discussed originally with Chris.

Thank you.
I don't know whether the fix is good to me or not yet. I will test your
fix, but I am busy now and my test will be a few weeks later. Other
people may want the fix soon. So I'd suggest you to reproduce the
problem on your side. I guess "mem=1G" or "mem=512M" will make it easier
to reproduce the problem.
Of course, if you are sure the fix is correct, then you don't have to
wait for my test. Release it soon for other people.

J. R. Okajima