Mythtv playing way to fast using DRM VBlank ioctl

[Alex Deucher]

On 30 Mar 2006 13:57:12 -0500, Greg Stark <gsstark at mit.edu> wrote:
>
> "Alex Deucher" <alexdeucher at gmail.com> writes:
>
> > On 30 Mar 2006 11:58:59 -0500, Greg Stark <gsstark at mit.edu> wrote:
> > >
> > > I'm still having trouble getting mythtv running displaying on my R350 Radeon.
> > > I've defeated the bogus mergedfb dot clocks with an LD_PRELOAD hack but now I
> > > have a new problem.
> > >
> >
> > Once again if you are using mergedfb the drm vblank won't work for you
> > on crtc2 since it only applies to the first crtc.  AFAIK, there's no
> > drm support for vblank on crtc2.
>
> Does it work even on crtc1? The problem seems to occur regardless of which
> screen the window is on.

presumably, although I haven't tested it directly in a long time.

>
> For that matter how would the DRM even know which crtc the query was for? It's
> not like it gets a window id or anything like that with the ioctl.

It doesn't know.  it's only implemented for crtc1.  the drm generally
assumes on head.

>
> It seems like it would be better to fail in some way, perhaps printk'ing a
> message, than to silently return immediately if that's what it's doing.

something weird may be happening because of mergedfb.  I'm not that
familiar with the drm code.

Alex

>
> --
> greg
>
>