[PATCH] copy-chunk / FSRVP (remote snapshots) changes

[Jeremy Allison]

On Sat, Oct 06, 2012 at 03:40:36PM +0200, Volker Lendecke wrote:
> On Fri, Oct 05, 2012 at 05:27:28PM -0700, Jeremy Allison wrote:
> > On Tue, Sep 18, 2012 at 02:09:17AM +0200, David Disseldorp wrote:
> > > Hi,
> > > 
> > > I've cleaned up, collapsed and rebased the copy-chunk and FSRVP (remote
> > > snapshot RPC server) changes presented at SambaXP. Additionally:
> > > - copy-chunk request processing takes locks into account
> > > - permission checking is performed when processing FSRVP requests
> > > - The FSRVP server retains state across restart
> > > - Extra smbtorture tests are added
> > > - vfs_btrfs is now documented
> > > 
> > > The code can be found at:
> > > git://git.samba.org/ddiss/samba.git async_fsrvp_srv_wip_sxp2012_btrfs_enums_collapsed_ioctl
> > > http://gitweb.samba.org/?p=ddiss/samba.git;a=shortlog;h=refs/heads/async_fsrvp_srv_wip_sxp2012_btrfs_enums_collapsed_ioctl
> > > 
> > > I'll be running this code at the SDC plugfest, so stop by if you'd like
> > > to see it in action. I'd also much appreciate review of any or all of
> > > the changes.
> > 
> > FYI. I haven't forgotten about this. I'm currently doing
> > some cleanup work on the second patch :
> > 
> > s3-rpc: convert process_complete_pdu and callers async
> > 
> > In lots of error cases you're doing things like:
> > 
> >                         if (data_left > 0) {
> >                                 DEBUG(0, ("Unacceptable, more than one PDU\n"));
> >                                 state->data_processed = -1;
> >                                 tevent_req_done(req);
> >                                 return tevent_req_post(req, ev);
> >                         }
> > 
> > and I think this should look like:
> > 
> >                         if (data_left > 0) {
> >                                 DEBUG(0, ("Unacceptable, more than one PDU\n"));
> >                                 state->data_processed = -1;
> >                                 tevent_req_error(req, EINVAL);
> >                                 return tevent_req_post(req, ev);
> >                         }
> 
> Also, at least I try to avoid DEBUG(0 unless it's a really
> bad emergency that needs admin intervention. In typical
> setups this ends up in syslog.  DEBUG(1 shows up in log.smbd
> only.

Oh I already removed that :-).