Bug#280384: XFree crashing on kernel 2.4.28
On Wed, Dec 01, 2004 at 09:52:48AM -0800, [EMAIL PROTECTED] wrote: A couple of issues come to mind looking at the transcript of our bug #280384: 1) Why couldn't Peter De Schrijver experience this problem for himself on the Debian 2.6.8.1 kernel? 2) Why does Admar Schoonen's work-around (using the debug server) function properly? I think it's supremely ironic that the debug build doesn't have the bug. LOL. Well, I have to admit that this bug does occur on my Sun Blade 100 (with 2 almost identical ati mach 64 cards), but not on my Sun Ultra 5 (with only 1 ati mach 64 card that is slightly different than the one in the blade 100). The bug also doesn't occur on a Sun Ultra 5 at my friend's place, but iirc, it does occur on an Ultra 5 or Ultra 10 from someone else, so I'm a bit sceptical to say that U5's are not affected. The fact that the debug server doesn't show the bug could indicate that it's a timing issue, but to be honest, I am a terrible programmer and I wouldn't know how to start debugging this issue. I hope this could be useful for someone, and if he/she needs more information/testing, I'd happy volunteer. Admar
Bug#280384: XFree crashing on kernel 2.4.28
Hello, Branden Robinson of the Debian's X Strike Force (XFS) mentioned the bug #225526, which might be the same problem, according to him. Presumably, this bug should be fixed by the following commit to the XFS' SVN repository: * Apply patch from David Mosberger that replaces the fix for #225526 with one that works on systems that do not have a PCI bus numbered 0. Thanks, David! (Closes: #279436) The ultimate test would be to build the packages from the SVN source and test it on the machines, which are affected. I'll try to arrange the build, but it can take a while, since I do not have access to any decent Ultra hardware. Best regards, Jurij Smakov[EMAIL PROTECTED] Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC
Bug#280384: XFree crashing on kernel 2.4.28
Jurij Smakov wrote: Hello, Branden Robinson of the Debian's X Strike Force (XFS) mentioned the bug #225526, which might be the same problem, according to him. Presumably, this bug should be fixed by the following commit to the XFS' SVN repository: * Apply patch from David Mosberger that replaces the fix for #225526 with one that works on systems that do not have a PCI bus numbered 0. Thanks, David! (Closes: #279436) The ultimate test would be to build the packages from the SVN source and test it on the machines, which are affected. I'll try to arrange the build, but it can take a while, since I do not have access to any decent Ultra hardware. Colour me doubtful about this as a fix. Its bug report has an XFree86.log that actually appears to scan the PCI bus, then does lots of other things before reporting no screens found. In contrast, both the logs from the originator of this thread and the XFree86.log in bug #280384 show the crash occurring immediately after loading the pcidata module, with no attempt to scan the PCI bus. That is also my experience, as evidenced by the log in my own post to bug #280384. They don't look like the same problem to me at all. That said, if somebody can tell me how to extract sources with this patch, I'm willing to try compiling it. Only takes the machine six hours these days :-) .Ron -- Ron Murray ([EMAIL PROTECTED]) http://www.rjmx.net/~ron GPG Public Key Fingerprint: F2C1 FC47 5EF7 0317 133C D66B 8ADA A3C4 D86C 74DE