[PATCH 1/8] Bug 9395 - Samba fails the simple_nodelete test of smb2.rename tests from master.

[Jeremy Allison]

On Fri, Nov 16, 2012 at 10:18:35PM +0100, Stefan (metze) Metzmacher wrote:
> Am 16.11.2012 20:54, schrieb Jeremy Allison:
> > Make rename opens internal-only. They don't break oplocks.
> > 
> > Signed-off-by: Jeremy Allison <jra at samba.org>
> > ---
> >  source3/smbd/reply.c |    4 ++--
> >  1 files changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/source3/smbd/reply.c b/source3/smbd/reply.c
> > index 8db9c62..f1c7478 100644
> > --- a/source3/smbd/reply.c
> > +++ b/source3/smbd/reply.c
> > @@ -6597,7 +6597,7 @@ NTSTATUS rename_internals(TALLOC_CTX *ctx,
> >  			FILE_OPEN,			/* create_disposition*/
> >  			create_options,			/* create_options */
> >  			posix_pathnames ? FILE_FLAG_POSIX_SEMANTICS|0777 : 0, /* file_attributes */
> > -			0,				/* oplock_request */
> > +			INTERNAL_OPEN_ONLY,		/* oplock_request */
> >  			0,				/* allocation_size */
> >  			0,				/* private_flags */
> >  			NULL,				/* sd */
> > @@ -6743,7 +6743,7 @@ NTSTATUS rename_internals(TALLOC_CTX *ctx,
> >  			FILE_OPEN,			/* create_disposition*/
> >  			create_options,			/* create_options */
> >  			posix_pathnames ? FILE_FLAG_POSIX_SEMANTICS|0777 : 0, /* file_attributes */
> > -			0,				/* oplock_request */
> > +			INTERNAL_OPEN_ONLY,		/* oplock_request */
> >  			0,				/* allocation_size */
> >  			0,				/* private_flags */
> >  			NULL,				/* sd */
> 
> I'm not sure this is correct.
> 
> MS-FSA 2.1.4.12 Algorithm to Check for an Oplock Break
> says FileRenameInformation should break HANDLE_CACHING (and
> BATCH_OPLOCKS to none)
> 
> See also 2.1.5.14.11 FileRenameInformation

Yeah, never mind on this fix - it breaks raw.oplocks :-(.

That's what happens when you optimize for one set of tests.

Give me some more time, I'll re-submit. Ignore this patchset
for now :-).

Thanks !

Jeremy.