[PATCHES] SMB3 lock sequence (replay detection)

[Michael Adam]

On 2018-09-18 at 19:04 +0200, Michael Adam via samba-technical wrote:
> On 2018-09-18 at 16:31 +0200, Stefan Metzmacher wrote:
> > Hi Michael,
> > 
> > thanks for driving this forward again!
> > 
> > I'd like to see a few minor changes before this goes upstream.
> > 
> > - Can we have an additional test the requires SMB3
> >   and tries to open a persistent handle, but allows a fallback
> >   just a durable handle?
> 
> Good point, we'll do that!
> 
> > - Please only activate this when "server multi channel support" is
> >   activated for now.
> 
> "This" being lock sequence number checking in the server?

If that's what you mean, then we do actually have a possible
problem with enabling this in the current master: In selftest,
multi-channel does not work fully right now, because the
socket-wrapper work for supporting fd-passing is not quite
complete yet. And since connection passing is not only
triggered by session binds but also by negotiate requests
with a client GUID that is already connected to the server,
enabling this has side effects.

We can of course create a new test env that will only be
used for this set of lock replay detection tests, if that's
what you want...

Additional thoughts would be appreciated!

Cheers - Michael



> Thanks for the feed-back!
> 
> Michael
> 
> 
> 
> > Thanks!
> > metze
> > 
> > Am 18.09.2018 um 10:23 schrieb Michael Adam via samba-technical:
> > > Hi all,
> > > 
> > > reviving this slightly older topic:
> > > 
> > > SMB2.1 (for resilient file handles) and SMB3.X implement
> > > the so called lock sequence number and based on that,
> > > the lock replay detection in the server.
> > > 
> > > This was done as part of the original efforts of
> > > implementing multi-channel, and is now popping up
> > > again as we are cleaning up the remaining bits and
> > > finishing up the last parts of multi-channel.
> > > 
> > > The patches were originally not merged because
> > > windows activates them only for resilient or
> > > persistent file hanldes, but as Guenther points
> > > out in one of his patches, the SMB2 doc actually
> > > says, the server SHOULD actually implement them
> > > for all of smb 3.x.
> > > 
> > > Hence re-proposing this patcheset to be merged
> > > to prevent bit-rot and have the feature ready
> > > and useable in the official samba master.
> > > 
> > > Cheers - Michael
> > > 
> > 
> > 
> 
> 
> 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 163 bytes
Desc: not available
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20180920/e904da75/signature.sig>