[git pull] drm fixes round two for 5.13-rc1

[Dave Airlie]

On Mon, 10 May 2021 at 07:08, Linus Torvalds
<torvalds at linux-foundation.org> wrote:
>
> On Sun, May 9, 2021 at 11:16 AM Dave Airlie <airlied at gmail.com> wrote:
> >
> > Bit later than usual, I queued them all up on Friday then promptly
> > forgot to write the pull request email. This is mainly amdgpu fixes,
> > with some radeon/msm/fbdev and one i915 gvt fix thrown in.
>
> Hmm. Gcc seems ok with this, but clang complains:
>
>    drivers/video/fbdev/core/fbmem.c:736:21: warning: attribute
> declaration must precede definition [-Wignored-attributes]
>    static const struct __maybe_unused seq_operations proc_fb_seq_ops = {
>                        ^
>
> but I noticed it only after I had already pushed out the pull.
>
> I'm actually surprised that gcc accepted that horrid mess: putting
> "__maybe_unused" between the "struct" and the struct name is very very
> wrong.
>
> I fixed it up after the merge due to not noticing earlier..
>
> Maybe the drm test robots should start testing with clang too?

My current build, test + sign machine " 09:23:22 up 419 days", running
fc29, has clang 7 which is my current blocker on implementing a clang
bit in my build cycles.

The machine is in an office, that due to RH not being able to fathom a
country where COVID isn't rampant, I have to fill insane amounts of
paperwork and high management intervention to visit, I'm afraid to
remote upgrade or reboot it unless I really have to.

I've some rough ideas to update it to f33/34 which would give me a
good enough clang, I suppose I could just build my own llvm/clang I've
done so for years for other reasons, if I can't come up with an update
plan soon I might consider it.

Dave.