When I manage to get DRI working with the latest i810 driver and my i915 hardware, I get problems with XVideo. The frame rate of e.g. tvtime becomes extremely slow, only a few frames per second. I can work around the problem by 1. disabling DRI (which made me experience bug 177417) or 2. downgrading to 1.7.4 (which solves bug 176501 as well) Another thing I observed so far only when using the 2.0.0 driver with DRI enabled were black margins in the video image at the right and at the bottom. They were visible only if the video image exceeded a certain size, and the part of the image that was supposed to be in these parts was no longer visible. The problem is that I could not reproduce this issue immediately after a restart, and I still don't know the factors that caused it to occur later on. However whenever it happened once, the problem was reproducible till I restarted X. I could reproduce this problem with tvtime and mplayer, and I guess the problem its XVideo as well.
I confirm that XVideo is broken in the i810 driver version 2.0.0, running on a 82852/855GM chipset. MPlayer and Xine show a blue screen instead of the video, but do not report any errors.
(In reply to comment #1) > I confirm that XVideo is broken in the i810 driver version 2.0.0, running on a > 82852/855GM chipset. MPlayer and Xine show a blue screen instead of the video, > but do not report any errors. Makes me remember this here: http://intellinuxgraphics.org/documentation.html There it says that i915 and later will no longer use such color-keyed frames but "textured video" instead. Maybe the support for textured video is buggy, and the driver relies on this and won't work without it either? Interesting line from my xvoutput in this context: Adaptor #1: "Intel(R) Textured Video" That is running 1.7.4, but I believe it was the same for 2.0.0. What does your xvoutput say to this?
(In reply to comment #1) > I confirm that XVideo is broken in the i810 driver version 2.0.0, running on a > 82852/855GM chipset. MPlayer and Xine show a blue screen instead of the video, > but do not report any errors. I guess I found at least this colorkey symptom on i8xx hardware in an upstream report: https://bugs.freedesktop.org/show_bug.cgi?id=10732 Doesn't sound a lot like my own problem with slow Xv, though.
(In reply to comment #3) > I guess I found at least this colorkey symptom on i8xx hardware in an upstream > report: https://bugs.freedesktop.org/show_bug.cgi?id=10732 > Doesn't sound a lot like my own problem with slow Xv, though. > I'll mark this bug upstream with that URL. Feel free to find or create another upstream bug for the slow Xv problems.
I had masked 2.0.0 and completely forgotten about this, until 2.1.0 came along with the same problem again. Now I searched upstream again and, finding nothing, created https://bugs.freedesktop.org/show_bug.cgi?id=11690 for the slowness issue.
