[Intel-gfx] [PATCH i-g-t v2] pm_rps: Changes in waitboost scenario

[Chris Wilson]

Quoting Daniel Vetter (2017-08-18 21:28:08)
> On Fri, Aug 18, 2017 at 02:45:32PM +0100, Chris Wilson wrote:
> > Quoting Katarzyna Dec (2017-08-18 12:08:44)
> > > While I'm here let's also make sure that we're running as DRM
> > > master to catch the common case, where the test is being ran
> > > along with Xorg.
> > 
> > Semantic changes to the test though, that implies its needs to be master
> > to run.
> > 
> > If you want to impose the constraint that you are the only client
> > active, then add a lib function to check
> > /sys/kernel/debug/dri/$card/client that it only contains one client and
> > that client is you.
> 
> We should be doing that already, at least if you try to open the drm fd. I
> guess we could split that out, or maybe wrap it into a helper.

Only the first opening of the device is expected to be the singular
client. We don't do that as part of drm_open_diver(), but it did used to
be done for the shellscripts.
-Chris