[pnfs] [PATCH 0/5] cb_sequence patches for the sessions branch

[Benny Halevy]

I'll send more patches for the pnfs part of the mini-project,
ETA Sunday...

Benny

Benny Halevy wrote:
> The following patchset implements basic support for cb_sequence
> both on the client and on the server end.
> 
> There's still work to be done to fully support referring call lists
> and use it for layout recall (on both ends of the wire).
> 
> cb_sequence.cachethis is always 0.
> 
> only one slot is used and a per-client mutex is serializing all
> callbacks for that client.  we may want to improve that later.
> 
> with this patchset connectathon test are passing as well as
> getting a delegation and recalling it (triggered by writing into
> the delegated file over a local mount on the server)
> I see a few seconds delay when a cb_recall is sent in this case
> and I need to investigate why.
> 
> The patches are against the 4.1-sessions branch.
> --
>  fs/nfs/callback.h          |   32 +++++
>  fs/nfs/callback_proc.c     |   17 +++
>  fs/nfs/callback_xdr.c      |  146 ++++++++++++++++++++++-
>  fs/nfsd/nfs4callback.c     |  281 ++++++++++++++++++++++++++++++++++++++-----
>  fs/nfsd/nfs4state.c        |    6 +
>  include/linux/nfs4.h       |   34 +++++-
>  include/linux/nfsd/state.h |   14 ++-
>  7 files changed, 493 insertions(+), 37 deletions(-)
> 
> [PATCH 1/5] cb_sequence: client support for minorversion1 callbacks
> [PATCH 2/5] define nfs4.1 (draft-10) protocol constants
> [PATCH 3/5] cb_sequence: client implementation of cb_sequence
> [PATCH 4/5] cb_sequence: nfsd callback minorversion1 cleanup
> [PATCH 5/5] cb_sequence: nfsd initial implementation
> 
> _______________________________________________
> pNFS mailing list
> pNFS at linux-nfs.org
> http://linux-nfs.org/cgi-bin/mailman/listinfo/pnfs