At Thu, 27 Apr 2000 10:56:42 -0700,
Kip Matthew Macy <kip@xxxxxxxxxxxxxxx> wrote:
>
>
> You asked if I saw a corresponding vmalloc:
>
>
> prompt > mount -t xfs /dev/rd/c0d0 /xfs
> XFS Mode = NATIVE, arch = 1
> kmem_alloc doing a vmalloc 452224 size & PAGE_SIZE 0 rval=0xe089600
> start mounting filesystem: dac960(48, 0)
> Ending clean XFS mount for filesystme: dac960(48,0)
I don't think this is related to your problem then.
I just wanted to make sure the free code wasn't trying to release
something it shouldn't have.
We are probably going to need a backtrace of the unmount command.
from kdb do a ps, find the umount pid, then btp that pid.
We have known container size mismatches, I have a feeling you have
found one.
I need to start working on our type problems very soon. :-)
-Russell
|