On Thu, Jun 01, 2006 at 01:06:44AM +0300, Dmitry Pryanishnikov wrote:
>
> Hello!
>
> On Thu, 25 May 2006, Konstantin Belousov wrote:
> > KASSERT(!(debug_mpsafenet == 1 && mtx_owned(&Giant)),
> > ("nfssvc_nfsd(): debug.mpsafenet=1 && Giant"));
> >
> >from nfsserver/nfs_syscalls.c, line 570.
> >
> >As I understand the problem, kern/vfs_lookup.c:lookup() could
> >aquire additional locks on Giant, indicating this by GIANTHELD
> >flag in nd. All processing in nfsserver already goes with Giant held,
> >so, I just dropped that excessive locks after return from lookup.
> >System with patch applied survived smoke test (client did
> >du on mounted dir, patch was generated from exported fs, etc.).
> >nfsd eats no more than 25% of CPU (with INVARIANTS).
> >
> >Please, users who reported the problem and willing to help,
> >try the patch (generated against STABLE) and give the feedback.
>
> Thank you very much. Your patch actually fixes "nfssvc_nfsd():
> debug.mpsafenet=1 && Giant" panic during NFS mount of server's "/usr".
> Oddly enough, NFS mount of server's "/" doesn't panic the server.
> My kernel config contains "options QUOTA", however quotas are not enabled.
> Please commit the fix, IMHO long-term breakage of such a basic functionality
> (NFS server + quotas) in -STABLE branch isn't a Good Thing (TM).FYI, if you're not using quotas then you should remove the option from your kernel config to avoid trashing your performance. Kris
pgpHaBFWdNItK.pgp
Description: PGP signature
