[PATCH] Better handling of stale filehandles in open() in the NFS client

[Rick Macklem]

On Thu, 20 May 2010, John Baldwin wrote:

>
> It doesn't change the RPC count because of changes that Mohan added to the
> NFS client a while ago so that nfs_open() doesn't invalide the attribute
> cache during nfs_open() if it was already updated via nfs_lookup() during
> the same system call.  With Mohan's changes in place, all this change does
> is move the GETATTR/ACCESS RPC earlier in the case of a namecache hit.
>
I tried another kernel build with and without the patch and an about 6%
increase in Getattr RPCs seems real. (The 1% increase in Access RPCs
was not related to the patch and appears to have happened because I
didn't do an unmount/mount between runs, which results in RPC cnts
changing by up to 1%.)

I have no idea whether or not an approx. 6% increase in Getattr RPCs
for this case is enough of a concern w.r.t. patch that it needs to be
looked at further?

Btw, I'm doing the tests on a pretty recent -current kernel which
appears to do shared vnode locks for read opens, which might explain
why the Getattr cnt changes? (You're patch had quite different line #s,
so I assume it wasn't against -current?)

Have fun with it, rick