> Do maybe the newest NVIDIA modules break XFS again?
Axel, maybe you should add that the bug you have reported to me has been
initiated by vmware. Also it would be helpful to note what kind of video card
you use. Even more helpful it would be to describe the steps in detail how to
reproduce it. Vmware is a very powerful software with many dependencies, and
maybe there are certain actions you took when the bug occured.
I for myself don't use vmware, but I still encounter random crashes with the
nvidia driver on a Geforce 2 Go since I installed the nvidia driver many
versions before, so I'm getting used to it, it's getting old. It shouldn't be
concerned too much to XFS. Most it's due to dependencies between the modified
kernel memory management patches in the -jp kernel patch set I guess, and
that concerns both nvidia and XFS modules, and much more. The XFree86 4.2.0
server is probably concerned, too.
What I observe are infrequent freezes, sometimes NULL pointers in page_alloc,
sometimes panicking. I don't have much time to spend on debugging. I comfort
myself that it's absolutely normal in a bleeding edge development kernel
under stress testing conditions.
The hard one to deal with is the nvidia driver, because I'm bound to it, but
it's impossible to trace the source code, and I know most kernel hackers wish
they could do so. I wish I could help to make the nvidia driver more stable,
or adjustable to patches, but I can't.
The file system corruption you reported is "only" a side effect of the crash.
Here is the nvidia driver version I currently use:
NVRM version: NVIDIA NVdriver Kernel Module 1.0-2960 Tue May 14 07:41:42 PDT
2002
GCC version: gcc version 2.95.3 20010315 (SuSE)
Best Regards,
Jörg
|