On Mon, Sep 17, 2001 at 11:01:17PM +0200, Michel Dänzer wrote:
> Gerd Knorr wrote:
> >
> > > > to the Rage128's CCE engine. Although it only accelerates SolidFill and
> > > > ScreenToScreenCopies it should help a lot.
> > >
> > > Hmm, maybe it's easy for someone who doesn't touch the dri code for the
> > > first time, but not for me. I've tried, the code builds and doesn't
> > > crash, but doesn't work. Can someone have a look at it please?
> >
> > Uhm, well, with r128_cce_indirect not implemented yet in th kernel it
> > can't work. New version + kernel patch, working this time.
>
> Great work, this seems to fix the hardware lockups with DMA for Xv! :) But
> there are a few problems:
>
> - didn't work on big endian systems. The calls to cpu_to_le32() in the kernel
> code are there for a reason. ;) I just committed the one liner fix.
>
> - the indirect buffers aren't flushed often enough. When I move a window with
> amiwm, the contents don't move at first. Only when I move it again do the
> contents appear in the current place. A quick comparison with the Radeon code
> didn't reveal any obvious difference, so I wonder what's wrong?
>
> - visual artifacts when moving a window which contains a DRI client. Maybe
> caused by the previous problem.
>
>
> Another problem isn't directly related to this code. While both the X server
> and the client library code expect DRM version 3.0.0, it's still at 2.1.6 . I
> fixed it locally by bumping the DRM version to 3.0.0, is that correct?
>
Jeff bumped them, then rolled the versions back to work with 4.1.0.
They should be bumped for 4.2.0 - so I've just made that change. As you
can see it's already done in the ddx and client side drivers.
Alan.
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel