[PATCH] Fix bug #Bug 12831 - smbcacls got error NT_STATUS_NETWORK_NAME_DELETED
jra at samba.org
Fri Jun 16 18:56:12 UTC 2017
On Fri, Jun 16, 2017 at 07:38:06PM +0200, Ralph Böhme via samba-technical wrote:
> On Fri, Jun 16, 2017 at 09:20:53AM -0700, Jeremy Allison wrote:
> > On Wed, Jun 14, 2017 at 08:50:57AM -0700, Jeremy Allison wrote:
> > > It turns out that smbclient, smbcacls and smbtorture3 depend
> > > on the ability to temporarily replace the client tcon connection
> > > struct internally to the client with a new connection to IPC$,
> > > do some calls on that new connection (usually lookupnames/lookupsids),
> > > then replace the old values of the tcon and continue.
> > >
> > > This has some problems. (a). It only works for SMB1 (for SMB2 we
> > > overwrote the tcon pointer) and (b). It didn't really work for SMB1
> > > either - we ended up with a bastardized cli->smb1.tcon pointer that
> > > contains type and string from the IPC$ connection whilst being
> > > connected to the share connection. We only got away with this
> > > due to the fact we don't use the type and string for anything
> > > on the wire.
> > >
> > > This patch unifies the cli->smb[1|2].tcon pointer handling and
> > > allows temporary replacement of the pointer, and also SMB2 access
> > > to the get/set the 32-bit tid value.
> > >
> > > Most of this patch is fixing up the smbclient/smbcacls/smbtorture3
> > > code so everything keeps working with make test, so I know this
> > > is being tested (it took me a day or so to keep all our tests
> > > passing :-).
> > >
> > > Please review and push if you're happy !
> > Ping ! Can I get a review on this please ?
> currently sick in bed, will review when I'm back up, probably on Monday, if
> noone else does it.
Ouch, hope you get better soon Ralph, was missing you on the lists :-).
FYI, I'll be up in Redmond at the Microsoft Interop event all next
week, so ping me if there's anything you need testing !
More information about the samba-technical