Re: [git pull] drm fixes for 5.7-rc8/final

[Dave Airlie]

On Fri, 29 May 2020 at 12:02, Dave Airlie <airlied@xxxxxxxxx> wrote:
>
> On Fri, 29 May 2020 at 11:49, Linus Torvalds
> <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Thu, May 28, 2020 at 5:21 PM Dave Airlie <airlied@xxxxxxxxx> wrote:
> > >
> > > Seems to have wound down nicely, a couple of i915 fixes, amdgpu fixes
> > > and minor ingenic fixes.
> >
> > Dave, this doesn't even build. WTF?
> >
> > In drivers/gpu/drm/i915/gt/selftest_lrc.c, there's a
> > engine_heartbeat_disable() function that takes two arguments, but the
> > new "live_timeslice_nopreempt()" gives it only one.
> >
> > I'd blame a merge problem, since the failure is in new code, but the
> > problem exists in your top-of-tree, not just my merge.
> >
> > In fact, it's not even your merge of the i915 tree that is the source
> > of the problem (although the fact that you clearly didn't _test_ the
> > end result most definitely is _part_ of the problem!).
> >
> > Because the problem exists in the commit that introduced that thing:
> > commit 1f65efb624c4 ("drm/i915/gt: Prevent timeslicing into
> > unpreemptable requests").
> >
> > It's garbage, and never compiled.
>
> I thought I'd dropped the ball completely. but I see it's a selftest
> failure, I must not have selftests built in my config here, I don't do
> exhaustive builds randconfig
>
> This has also been built by Intel, but I'm assuming they missed their
> selftest bits as well.
>
> I'll revert that and resend.

I did drop the ball in one way, I see sfr reported it broken this morning

I normally expect stuff coming from Intel has been through their CI,
even their fixes tree generally gets pushed through that system before
I get it, and it usually catches these things.

I might have to push back on intel fixes this late in the day, as
maybe the land on next and cherry-pick to fixes model has made them a
bit lax on how much stuff goes through CI.

I've just drop the whole i915 fixes from the tree and will resend with
it removed.

Dave.