Re: SMB2 implementation

Jeff Layton <[email protected]> Sun, 28 Feb 2010 20:55:08 -0500

Newsgroups	gmane.linux.file-systems.cifs
Message-ID	<[email protected]>

On Sun, 28 Feb 2010 18:15:43 -0600
Steve French <[email protected]> wrote:

> 1) Agreed
> 
> 2) agreed
> 
> 3) I was waiting to look at this until we start looking at batch
> oplock, assuming that that code would have to change a lot to handle
> batch oplock
> 
> 4) I think the general idea of the change to the "need_reconnect" to a
> "serial number" (a glorified counter) which you discussed at
> Connectathon makes sense.
> 

Excellent on all of the above. I may be able to hack up a patchset for
#4 when I get time.

> 5) It has to have a pointer to the smb ses info to do the signing
> 

Do we? The signing key is tracked in the TCP session:

    ses->server->mac_signing_key

Cheers,
-- 
Jeff Layton <[email protected]>