On Thursday January 30, neilb@cse.unsw.edu.au wrote:
> On Wednesday January 29, mikpe@csd.uu.se wrote:
> > Neil Brown writes:
> > > On Wednesday January 29, mikpe@csd.uu.se wrote:
> > > > Kernel 2.5.59. A local ext2 file system is mounted at $MNTPNT
> > > > and exported through NFS V3. A client mounts and unmounts it,
> > > > w/o any I/O in between. The NFS server is shut down. Nothing in
> > > > user-space refers to $MNTPNT.
> > > >
> > > > The bug is that $MNTPNT now can't be unmounted. umount fails with
> > > > "device is busy". A forced umount at shutdown fails with "device
> > > > or resource busy" and "illegal seek", and leaves the underlying
> > > > fs marked dirty.
....
>
> Ok, it defaintely sounds like a leak. I'll be back at my desk on
> Monday and I will try to reproduce it and explore the situation then.
>
> NeilBrown
Found and fixed... the following patch should make it work for you.
Thanks,
NeilBrown
-----------------------------
Fix problem where knfsd wouldn't release filesystem on unexport.
Problem was that the cache was being updated inplace, rather
than swapping in a new entry, so old filesystem pointers
were overwritten without being released.
----------- Diffstat output ------------
./fs/nfsd/export.c | 2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
diff ./fs/nfsd/export.c~current~ ./fs/nfsd/export.c
--- ./fs/nfsd/export.c~current~ 2003-02-03 16:49:49.000000000 +1100
+++ ./fs/nfsd/export.c 2003-02-03 16:49:52.000000000 +1100
@@ -691,7 +691,7 @@ exp_export(struct nfsctl_export *nxp)
new.ex_anon_gid = nxp->ex_anon_gid;
new.ex_fsid = nxp->ex_dev;
- exp = svc_export_lookup(&new, 2);
+ exp = svc_export_lookup(&new, 1);
if (exp == NULL)
goto finish;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Feb 07 2003 - 22:00:13 EST