On 20/12/12 10:26 PM, Thierry Vignaud wrote: > On 16 December 2012 22:01, Alan Coopersmith <[email protected]> > wrote: >>>> When EXA was added to mach64, it included UploadToScreen and >>>> DownloadFromScreen hooks that did the same thing that EXA would do in a >>>> software fallback. Now they are out of date and cause crashes. The >>>> slight benefit of reimplementing those hooks (if there is any) is not >>>> worth the maintenance burden of having to rewrite them every time there >>>> is a change to the way EXA handles pixmaps. >>>> Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=51137 >>> Confirmed that replacing your previous patch with this one I can still >>> login to GNOME and bring up Firefox on the Sun Ultra 20 test system we >>> have here (the original model had a RageXL chip onboard, before moving >>> to ES-1000 for RoHS compliance in later revs). > (...) > >> Unless anyone speaks up soon, I'll probably cut a 6.9.4 release in the >> next couple days so I can ship it with our Xorg 1.13.1 packages without >> having to add a bunch more patches. > I still got report of mach64 crashing if Accel is "on" with 6.9.4: > https://bugs.mageia.org/show_bug.cgi?id=8349 > https://bugs.mageia.org/attachment.cgi?id=3269 (too sad xorg doesn't use > mini debuginfo in order to get better traces in xorg.log...) > > At least one of the crashing card is Proliant Rage XL according to its pci > ids.
Does it still crash with Option "ExaNoComposite" "true"? It only takes very minimal edits to determine which hook causes the crash. You can make CheckComposite return FALSE right away. If that's okay, try leaving CheckComposite alone but making PrepareComposite return FALSE right away and so on.
signature.asc
Description: OpenPGP digital signature
_______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
