On Sat, Mar 20, 2021 at 07:48:44AM +1100, NeilBrown wrote: > On Fri, Mar 19 2021, J. Bruce Fields wrote: > > > On Fri, Mar 19, 2021 at 02:36:24PM +1100, NeilBrown wrote: > >> On Mon, Mar 01 2021, J. Bruce Fields wrote: > >> > >> > On Tue, Mar 02, 2021 at 02:01:36PM +1100, NeilBrown wrote: > >> >> On Mon, Mar 01 2021, J. Bruce Fields wrote: > >> >> > >> >> > I've gotten requests for similar functionality, and intended to > >> >> > implement it using directory notifications on /proc/fs/nfsd/clients. > >> >> > >> >> I've been exploring this a bit. > >> >> When I mount a filesystem, 2 clients get created. > >> >> With NFSv4.0, the second client is immediately deleted, and the first > >> >> client is deleted one grace period after the filesystem is unmounted. > >> >> With NFSv4.1 and 4.2, the first client is immediately deleted, and the > >> >> second client is deleted immediately after the unmount. > >> > > >> > Yeah, internally it's creating an "unconfirmed client" on SETCLIENTID > >> > (or EXCHANGE_ID) and then a new "confirmed client" on > >> > SETCLIENTID_CONFIRM (or CREATE_SESSION). > >> > > >> > I'm not sure why the ordering's a little different between the 4.0/4.1+ > >> > cases. > >> > >> The multiple clients are not really nfsd's "fault". The Linux NFS > >> client sends multiple EXCHANGE_ID or SET_CLIENT_ID requests, so NFSD > >> really does need to create multiple clients. > >> > >> For NFSv4.0, when nfsd gets a repeat SET_CLIENT_ID, it keeps the old one > >> and discards the new. > >> For NFSv4.1, the spec requires that it keep the new one and discard the > >> old. > >> This explains the different ordering. > > > > Hm, is this the client's trunking-detection logic? > > Yes. > > > > > In which case, it's not just unconfirmed clients. > > For NFSv4.1, only the EXCHANGE_ID is duplicate. There is only one > CREATE_SESSION, and that is where the client is confirmed. So only one > confirmed client. > > For NFSv4.0 bother SETCLIENTID and SETCLIENDID_CONFIRM are duplicate. > So maybe both clients get confirmed. I should check that. Drifting off topic, but I don't see how this client behavior makes sense. Mount is chatty enough without the unnecessary duplication. Looking at the code.... --b.
