Re: drm pull for v5.3-rc1

[Dave Airlie]

On Tue, 16 Jul 2019 at 03:38, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> On Mon, Jul 15, 2019 at 12:08 AM Dave Airlie <airlied@xxxxxxxxx> wrote:
> >
> > VMware had some mm helpers go in via my tree (looking back I'm not
> > sure Thomas really secured enough acks on these, but I'm going with it
> > for now until I get push back).
>
> Yeah, this is the kind of completely unacceptable stuff that I was
> _afraid_ I'd get from the hmm tree, but didn't.

Looks like we were all focused on making sure hmm tree was good, I
really dropped the ball watching the other ball.

I pulled stuff in from Thomas quite a while ago, and his pull request
did say it had been looked at by mm devs, I looked back
a week or so ago before the flu hit me badly and went hey this isn't
as good, but removing it is a mess I better ping some people, then I
promptly fell into a hole.

It's bad though so I'm just going to revert it all out.

I'll send a new PR today with it reverted, rebuilding the tree might
be possible, but you'd lose a lot of testing confidence in the rest of
it.

> I'm not pulling this. Why did you merge it into your tree, when
> apparently you were aware of how questionable it is judging by the drm
> pull request.

I totally over trusted Thomas on this, I glanced at the helpers when I
merged them and went they seemed reasonable for the vmware address
space coherency model, and they'd been posted to linux-mm a few times
and had some feedback,

I caught it more last week when I was re-reviewing all the stuff in my
tree and I was like hey that isn't right, but removing it might be
tricky, then I spent a week with a couch and no brain.

So I'm totally responsible for this crap landing in my tree, and
Thomas will be getting a lot more push back in future.

Dave.