No subject

OPENMODE, and it's not obvious to me why it's hitting that printk.  I'll
have to experiment with it later today.

I suspect the leak is because once the seqid gets out of sync, there's
no way for the client to close the file, so the openowner just gets left
dangling.  I'm not sure what to do about that.  At some point we need to
figure out how to impose some limits on simultaneous opens.  And we may
also want special handling for the case of a client that's producing a
lot of bad_stateid's and leaving opens around.

Thanks for the report!

--b.