On Tue, Apr 14, 2009 at 7:35 AM, Michael Sullivan <msulli1...@gmail.com> wrote: > On Tue, 2009-04-14 at 07:43 +0100, Neil Bothwick wrote: >> On Mon, 13 Apr 2009 17:14:06 -0700, Mark Knecht wrote: >> >> > > (II) LoadModule: "i810" >> > > >> > > (WW) Warning, couldn't open module i810 >> > > (II) UnloadModule: "i810" >> > > (EE) Failed to load module "i810" (module does not exist, 0) >> > > (EE) No drivers available. >> > <SNIP> >> > >> > OK, that's pretty serious. Is the i810 module loaded? That's the video >> > driver, or used to be >> >> Used to be, it's now called intel. You need to edit xorg.conf and change >> it. This is one of the situations that XOrg are trying to avoid with >> autodetection. If you had to driver specified in xorg.conf (or no >> xorg.conf) it would have loaded the right module for you. >> >> > > OK I've successfully booted with the old config, modified with the > changes suggested, and after re-merging xorg-server with > VIDEO_CARDS="intel". My mythfrontend is still segfaulting, and I've got > a couple of errors in my Xorg.0.log file, which I am attaching. It > gives me errors on modules freetype and type1. Here is my current > xorg.conf file: >
Micheal, OK, now your machine and my wife's machine are probably in about the exact same situation. Can you clarify exactly *when* mythfrontend segfaults for you? On my wife's machine the GUI comes up fine, and I can look at all the text screens, set up new recordings, etc., but the frontend segfaults only when I actually try to watch video or live TV. On my wife's machine it actually completely blows away X and dumps me back to the gdm login screen. I am coming to the conclusion that this may not be an xorg-server problem per se, but rather a problem with the Intel driver. In my Xorg.0.log.old file, after logging back into X, I sometimes see a message "(EE) intel(0): Failed to pin xv buffer" which is causing X to crash for other people also. As best I can figure out with my limited understanding is this xv buffer is normally some part of graphics memory that needs to be locked down but the new Intel driver isn't doing it right. People are seeing this problem on mplayer and apps other than Myth and there are some threads out there from developer types with potential code changes to fix this but they seem to be in development at this time. I haven't looked yet but possibly there's a newer version of the Intel driver that's not stable but has a fix in it? Or possibly we can get a test version somewhere. I think I'm not seeing it on my AMD64 machine because that's Radeon based. I suggest you don't waste too much time on this right now. I'm going to look for a new driver and test it out if I can find one but won't have time until later this afternoon. Hope this helps, Mark
