[AFS3-std] Locking, ACL's, and Capabilities

[Jeffrey Hutzelman]

On Friday, July 21, 2006 02:01:46 PM -0700 Russ Allbery <rra@stanford.edu> 
wrote:

> Jeffrey Hutzelman <jhutz@cmu.edu> writes:
>
>> Discussion revealed that ideally, we'd like for the bit to be able to be
>> set distinctly for every directory, by the same users who can set the
>> ACL for that directory.  However, this is difficult to do, requiring
>> changes to both the protocol and the format of the large vnode index.
>> It also has the issue that it's an awful lot of bits to set for sites
>> where ACL's have traditionally been sane.
>
> This is perhaps a dumb idea, precisely because these bits already exist
> and may already be used for other things, but AFS *does* have the
> application ACL bits A-H.  Using one of them for this purpose would break
> the existing semantics, though (which say that they have no meaning to AFS
> server processes).  I don't have a good feeling on how many sites actually
> use them for something now, only that I've never seen them mentioned in
> the context of actually being used in any of the public mailing list
> traffic about AFS I've read.

I've seen more than one reference to sites using these for things.  I think 
it's best to consider them reserved for local use, and figure out how to 
extend ACL's when we need to.

FWIW, I don't think using an ACL bit for this particular purpose is ideal 
anyway, because the flag we need is a property of the entire directory ACL, 
rather than each individual ACE.

-- Jeff