[PATCH] Add tests for our NFSv4 ACL code

[Jeremy Allison]

On Sun, Apr 14, 2013 at 10:00:45PM +1000, Andrew Bartlett wrote:
> I've been asked to look into some bugs in our NFSv4 ACL code, and before
> I started, I wanted to have some tests to try and ensure I didn't make
> any unfortunate regressions while fixing things. 
> 
> These patches (on master) are a starting point.  We now have a share on
> s3dc in make test that will run NFSv4 ACLs, and has raw.acls run against
> it.  Currently most (all?) if raw.acls fails, but at least it does not
> segfault, which is a really good start.
> 
> >From here, I want to learn more about the failures, work out at what
> layer we should be doing various inheritance operations (adding
> emulation if required), and possibly patch raw.acls to optionally skip
> SACLs in the comparisons. 
> 
> In the future, we may wish to move more of our NFSv4 ACL handling to use
> these structures, so we can (when combined with acl_xattr) hash the
> NFSv4 ACL using this IDL, where we cannot obtain something like the
> 'backup blob' we get from GPFS. 
> 
> Please review (and in the unlikely event there are no comments, feel
> free to push). 
> 
> Ambi:  As you have easier access to AIX and GPFS than I, can you check
> I've not missed anything in the changes to those files.  I know I've
> missed that before. 

Thanks Andrew, this looks like a great improvement and
is a sorely missed part of our ACL testing (I don't know
many people using NFSv4 ACLs with Samba) !

I'll try and review asap, although I'm at the Linux
Foundation Collaboration Summit followed by the
Filesystem Developers Summit in San Francisco all
next week so reviewing time will be scarce :-).

Cheers,

Jeremy.