Questions about smbd option "strict rename"

[Ralph Boehme]

On Mon, Nov 23, 2015 at 12:59:35PM +0100, Stefan Metzmacher wrote:
> Am 22.11.2015 um 13:49 schrieb Ralph Boehme:
> > On Fri, Nov 20, 2015 at 01:45:08PM -0800, Jeremy Allison wrote:
> >> On Fri, Nov 20, 2015 at 10:01:46AM +0100, Ralph Boehme wrote:
> >>> - "strict rename = no": doesn't work, opens are always checked,
> >>>   regardless of the setting of "strict rename".  can_rename(), the
> >>>   function where we do this check when renaming a directory, is
> >>>   missing a check for lp_strict_rename() or similar.
> >>
> >> Yep. That's how it was supposed to work. The code in
> >> can_rename() should probably be the same as the code in
> >> source/smbd/dir.c which is:
> >>
> >>         if (!lp_posix_pathnames() &&
> >>             lp_strict_rename(SNUM(conn)) &&
> >>             have_file_open_below(fsp->conn, fsp->fsp_name))
> >>         {
> >>                 return NT_STATUS_ACCESS_DENIED;
> >>         }
> > 
> > Ok, thanks for clarifying. Does everybody agree? Metze?
> 
> I don't agree, sorry.
> 
> We should provide the semantics the client asked for
> and try to behave like a windows server.
> 
> I think we should fix the manpage and use file_find_subpath()
> if lp_strict_rename() is false.

I'm leaning towards that interpretation of "strict rename" too.

> Can we use fsp->posix_open instead of lp_posix_pathnames() ?
> 
> We fruit module could set fsp->posix_open in that case.

posix_open is too broad and implies other things like POSIX brl
behaviour.

-Ralph

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen
http://www.sernet.de,mailto:kontakt@sernet.de