http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=320828 is very similar, if not the same bug. As seen in the backtraces, DstH is 0 on entry to ATIMach64PutImage(). I can trigger this very consistently with mplayer and xine.
Created attachment 10084 [details] [review] a workaround
As the problem only seems related to DstH, why test DstW as well? Also, somebody followed up to the Debian bug report saying that it already occurs with DstH < 16. If the rest of the code can't be fixed to work with DstH < 16, you should probably return Success instead of an error to avoid breaking clients.
I didn't have to resize mplayer window -- just ran it, and there was roughly 1/10 chance that X would crash. As for DstW check, I added it just to be sure.
Created attachment 10793 [details] [review] new workaround, with DstH<16 and no check on DstW This patch summarizes what has been discussed here and in the Debian bug so far. I talked to George Sapountzis about it, he didn't have a better idea at that time (and no hardware at home to look any deeper). FWIW, a Debian user reported that this patch worked fine, so we'll probably include it in the next Debian package.
committed: 882fe7631586b0a7919f808588a2ea4fb555f7e8
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.