Hello! On Thu, 05 Jan 2006 17:46:24 +0100, Luca Capello wrote: > On Thu, 05 Jan 2006 17:39:29 +0100, Luca Capello wrote: >> On Thu, 05 Jan 2006 16:50:49 +0100, Michel Dänzer wrote: >>> On Thu, 2006-01-05 at 15:58 +0100, Luca Capello wrote: >>>> >>>> So, I'd say that I'm experiencing two bugs: the first is a >>>> general X.Org freezing (withouth DRI enabled or with DRI enabled >>>> and no "radeon" kernel module), >>> >>> Benjamin Herrenschmidt's patch attached to >>> https://bugs.freedesktop.org/show_bug.cgi?id=3911 (note that he >>> only posted later, improved versions to the xorg mailing list) >>> seems to help with some such problems. >> >> I hope you're referring to [1], because I'm going to applying it >> and locally compiling a custom Debian xserver-xorg-6.9 to test what >> happens ;-) >> >> I'll report my tests ASAP. > > Sorry, missing [1]: > > http://lists.freedesktop.org/archives/xorg/2005-December/011679.html Ok, here we are! I applied the patch and compiled the whole X.Org with revision 1.gismo. Then, I installed only xserver-xorg and xserver-common (because xserver-xorg depends on at least the same version of it). It seems that with Benjamin's patch I don't have any freezes *without* DRI for about 2h (xorg.conf included). As soon as I enable the DRI support, X.Org freezes. But I have some good news about DRI... :-D I think I found the second bug I was experiencing, i.e. the fact that with DRI enabled I cannot see the pointer at all (so the screen goes black and freezes). This behaviour is still present with Benjamin's patch and is triggered by DRI with the radeon option "AGPFastWrite" (Xorg.0.logs attached): - when it's on, X.Org 6.9.0 (patched or not) starts and freezes the machine before any output is drawn - when it's off, X.Org 6.9.0 (patched or not) at least starts As you can see, I wrote 'at least', because the situation is actually more complex with "AGPFastWrite" off (this is the default option): - vanilla 6.9.0 freezes with or without DRI - patched 6.9.0 freezes only with DRI Now, I think that this bug and bug #345929 [1] requires Benjamin's patch, but what should I do for the DRI bug? Should I open a new Debian bug? And I think one test should be to try with the latest Mesa release (6.4.1), because the one present in Debian is a bit old (6.3.2). I wanted to file a new-upstream bug for mesa, but I found bug #326466 [2] :-( BTW, if someone wants to test Benjamin's patch, my packages are available at [3] (or directly at [4]). Thx, bye, Gismo / Luca [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=345929 [2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=326466 [3] http://luca.pca.it/debian/ [4] http://luca.pca.it/debian/deb/xorg/
Attachment:
xorg.conf_no-dri_20060106.gz
Description: xorg.conf, dri disabled
Attachment:
Xorg.0.logs.tar.gz
Description: Xorg.0.logs, fastwrite and dri on/off
Attachment:
pgprlsyEUG4sB.pgp
Description: PGP signature