On Mon, Oct 25, 2004 at 10:10:49PM +0200, Per Blomqvist wrote: > >I'm sorry to tell you this, but that delta (4.1.0-16woody4 to > >4.3.0.dfsg.1-8) is so huge as to be useless for determining what precisely > >may have caused the regression. > > > >I'd like to request a couple more pieces of information: > > > >1) Does turning XAA acceleration off prevent the problem from recurring? > > (Search for "NoAccel" in the XF86Config-4(5x) manual page.) > > NoAccel is in the "nv" manual, I use the default installed config file, > with no options on that device. Are there any other bugreport that > suspect exclusion of this option cause problem? > (is it an option or not?) I don't understand your question. The nv(4x) manual page does document the NoAccel option, as does the XF86Config-4(5x) manual page. I am asking for this information because it will help obtain an answer to my question: "Does turning XAA acceleration off prevent the problem from recurring?". If you add the specified option to the "Device" section of your XF86Config-4 file, XAA acceleration will be turned off. > >2) Can you try various earlier versions in the 4.3.0.dfsg.1 release series > > to see if any of them rectify the problem? You can find these at > > snapshot.debian.net. Please start with the most recent prior to yours > > (-7) and work your way backwards. > > I often update my system, so.. > the problem is in the latest version (4.3.0.dfsg.1-8). > > A little more description of the problem: > If X didn't froze at first apps, it did froze within an hour, a stress > test with an open aterm and just scroll dselect could cause it, also to > froze. > > The problem have not reapeard since: > I copied "one single" file "nv_drv.o" from the woody-relise > (xserver-xfree86_4.1.0-16woody4_i386.deb), too the same path as the > latest debian-sarge xserver-xfree86-4.3.0.dfsg.1-8 (that I still have > installed). > > I didn't know about http://snapshot.debian.net though, I used > http://packages.debian.org.. Okay. But as I said: > >I'm sorry to tell you this, but that delta (4.1.0-16woody4 to > >4.3.0.dfsg.1-8) is so huge as to be useless for determining what precisely > >may have caused the regression. Even narrowing it down to just the one file (nv_drv.o), the nv driver itself has greatly changed in the interim from 4.1.0-16woody4 to 4.3.0.dfsg.1-8. If you could use versions of the xserver-xfree86 package from snapshot.debian.net to identify where the regression specifically took place, it would help us to track down this bug. Otherwise, we might not be able to resolve it. -- G. Branden Robinson | The best place to hide something is Debian GNU/Linux | in documentation. branden@debian.org | -- Ethan Benson http://people.debian.org/~branden/ |
Attachment:
signature.asc
Description: Digital signature