Hi,

On Mon, Jun 01, 2020 at 03:58:26PM +0800, Jian-Hong Pan wrote:
> Maxime Ripard <maxime@cerno.tech> 於 2020年5月28日 週四 下午3:30寫道：
> >
> > Hi Daniel,
> >
> > On Wed, May 27, 2020 at 05:15:12PM +0800, Daniel Drake wrote:
> > > On Wed, May 27, 2020 at 5:13 PM Maxime Ripard <maxime@cerno.tech> wrote:
> > > > I'm about to send a v3 today or tomorrow, I can Cc you (and Jian-Hong) if you
> > > > want.
> > >
> > > That would be great, although given the potentially inconsistent
> > > results we've been seeing so far it would be great if you could
> > > additionally push a git branch somewhere.
> > > That way we can have higher confidence that we are applying exactly
> > > the same patches to the same base etc.
> >
> > So I sent a new iteration yesterday, and of course forgot to cc you... Sorry for
> > that.
> >
> > I've pushed my current branch here:
> > https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux.git/log/?h=rpi4-kms
> 
> Thanks to Maxime!
> 
> I have tried your repository on branch rpi4-kms.  The DRM VC4 is used!
> But got some issues:
> 1. Some weird error message in dmesg.  Not sure it is related, or not
> [    5.219321] [drm:vc5_hdmi_init_resources] *ERROR* Failed to get
> HDMI state machine clock
> https://gist.github.com/starnight/3f317dca121065a361cf08e91225e389

That's a deferred probing. The first time the HDMI driver is being
probed, the firmware clock driver has not been probed yet. It's making
another attempt later on, which succeeds.

> 2. The screen flashes suddenly sometimes.
> 
> 3. The higher resolutions, like 1920x1080 ... are lost after hot
> re-plug HDMI cable (HDMI0)

I'm not sure on how to exactly reproduce those issues (or what they are)
though, can you expand on this?

Maxime