> From: Maximilian Pichler <maxim.pich...@gmail.com> > Date: Mon, 5 Feb 2018 21:29:11 +0100 > > On Sat, Sep 23, 2017 at 6:56 AM, Niels Kobschaetzki > <ni...@kobschaetzki.net> wrote: > > Sep 22 22:38:41 netcat /bsd: drm:pid71340:i915_drm_suspend *ERROR* GEM idle > > failed, resume might fail > > Not sure if this helps, but since I'm seeing exactly the same thing on > a Zenbook 3 I added some logging to the kernel. Turns out this happens > if and only if i915_gem_suspend returns -ERESTARTSYS. > > So it looks like the call is interrupted by a signal, which would > explain why this occurs only sometimes. > > What's the proper way to recover from this sort of interruption? (Both > repeating the call in a loop and ignoring the error appear to make it > "work", but obviously I have no idea what I'm doing...)
I'm a bit busy right now. I don't think i915_gem_suspend() is supposed to return -ERESTARTSYS. It would help we if you could track down which function up the call chain return -ERESTARTSYS.