[Intel-gfx] [CI 08/14] drm/i915/selftests: Force a rewind if at first we don't succeed

[Chris Wilson]

Quoting Chris Wilson (2021-02-02 21:24:16)
> Quoting Chris Wilson (2021-02-02 21:14:35)
> > Quoting Chris Wilson (2021-02-02 17:43:53)
> > > Let's see how horrible it is to cycle elements on defer. (Curse the
> > > irqlock pollution.)
> > 
> > While that did work. I do not have a good idea on how to do list
> > rotation on an RCU list. I can see that it must require a pair of
> > synchronize_rcu, and that spells disaster (at least for handling it
> > inline).
> > 
> > Another way might be to randomize the deadlines along each branch to the
> > tree... Except we don't have deadlines at this point and we can't so
> > freely change the priorities.
> 
> Speaking of which, this is 'fixed' by the deadlines as there we will
> reorder ELSP as the test expects. (Which is why I didn't notice this for
> so long.)

And I thinks that's how I am going to handle this, by deferring the dfs
fix for defer_request until we are ready with deadlines.
-Chris