segfault in smbclient

[Andrew Bartlett]

Andrew Bartlett wrote:
> 
> Alexander Bokovoy wrote:
> >
> > I'm experiencing segfaults in smbclient with latest (today) CVS HEAD.
> > GDB backtracing states that error occurs at either
> > source/libsmb/cliconnect.c:442 or 447. Unfortunately, I can't track
> > more exactly as GDB generates wrong source references in this case:
> 
> Could you ensure you have done a 'make clean'?  I changed the definition
> of a DATA_BLOB and that could cause all sorts of confusion for code
> still compiled with the old definition.
> 
> And yes, samba needs proper dependency support...

OK, sorry about that.

I didn't realise the consequences of my DATA_BLOB change, and didn't
realise that some of the ntlmssp/spngeo code didn't use the data_blob()
interface when creating the DATA_BLOB.  This meant that the 'free'
function pointer was complete garbage.

I missed this all because I forgot I had 'use spnego = false' in my
smb.conf, and as such missed testing that section of the code.

And a big thankyou to tridge for tracking it down and for coming up with
such a simple fix.

So, sorry for all that mess...

Andrew Bartlett
-- 
Andrew Bartlett                                 abartlet at pcug.org.au
Manager, Authentication Subsystems, Samba Team  abartlet at samba.org
Student Network Administrator, Hawker College   abartlet at hawkerc.net
http://samba.org     http://build.samba.org     http://hawkerc.net