[linux-cifs-client] Re: [PATCH 1/7] [CIFS] add local struct inode pointer to cifs_setattr

Jeff Layton jlayton at redhat.com
Sat May 10 10:18:20 GMT 2008


On Fri, 9 May 2008 21:08:27 -0500
"Steve French" <smfrench at gmail.com> wrote:

> On Fri, May 9, 2008 at 6:01 PM, Jeff Layton <jlayton at redhat.com> wrote:
> > On Fri, 9 May 2008 17:29:12 -0500
> > "Steve French" <smfrench at gmail.com> wrote:
> >
> >> Merged patches 1, 6 and 7 from this series.
> >>
> >> For the others since they change current behavior and would break some
> >> apps (including connectathon test cases) would like the behavior e.g.
> >> to be configurable by mount option
> >>
> >
> > Thanks...
> >
> > How about this? We make the new behavior the default, and add a
> > mount option ("fakeperm" or something) to allow for the legacy
> > behavior where permissions can be set in memory only.
> >
> > Sound reasonable? If so, I'll get to work on a respin of this set.
> 
> Yes - how about "dynamicperm" or "dynamicmode" ? The name of the
> option can be decided later.
> 

Ok. BTW, I tested this patchset with the cthon04 suite and there
was only one test that previously passed and then failed -- basic
test4. This is expected though since it does a "chmod" and checks the
result.

-- 
Jeff Layton <jlayton at redhat.com>


More information about the linux-cifs-client mailing list