[SCM] Samba Shared Repository - branch master updated - release-4-0-0alpha7-1743-gd4d06a4

Jeremy Allison jra at samba.org
Fri May 22 00:57:24 GMT 2009


On Fri, May 22, 2009 at 10:49:32AM +1000, Andrew Bartlett wrote:
> On Thu, 2009-05-21 at 19:27 -0500, Jeremy Allison wrote:
> > The branch, master has been updated
> >        via  d4d06a4ef921e8fa543b86de002829ba0e7f77d3 (commit)
> >       from  a91bcbccf8a2243dac57cacec6fdfc9907580f69 (commit)
> > 
> > http://gitweb.samba.org/?p=samba.git;a=shortlog;h=master
> > 
> > 
> > - Log -----------------------------------------------------------------
> > commit d4d06a4ef921e8fa543b86de002829ba0e7f77d3
> > Author: Jeremy Allison <jra at samba.org>
> > Date:   Thu May 21 17:27:25 2009 -0700
> > 
> >     Attempt to fix bug #6386 - Samba Panic triggered by Sophos Control Centre.
> >     Don't indirect a potentially null pointer.
> >     Jeremy.
> 
> Shouldn't the fix be to the ldb_search() function in in Samba 3.2?  
> 
> If ldb_search() returns LDB_SUCCESS, then res *must* be a valid,
> non-NULL pointer.  I've confirmed that the current LDB in Samba4 (and
> even the Samba3 copy of ldb in master) does not suffer from this bug.

I also have a patch for the 3.2 version, which I've attached to the
bug report. But this is clear "belt and braces" paranoia before indirecting
through a pointer explicitly set to null before calling the function.

Actually I really should make it "|| !res ||", so I'll do that now.

Jeremy.


More information about the samba-technical mailing list