[Labeled-nfs] Current development prototype patches.

[Matthew N. Dodd]

Between traveling and recovering from the flu I've been out a good bit 
of this week and the last part of last.  I'll try and respond to points 
in previous messages this weekend.

Attached are the current patches I'm running relative to 2.6.21.3.

I'm passing around a 32 bit sid variable in iattr/nfs_fattr structures 
to simplify things during development.  I expect to be using a opaque 
"blob" as things progress but currently I'm more interested in figuring 
out where code needs to be added and getting to the point where I can 
verify that change notification will function correctly.

There have been many good comments directed at the patches and I'll 
incorporate them once I'm out of "quick and dirty" mode.

Anyhow, the current status is that labels are passed on the wire both to 
and from the server.  I've not hooked things up on the server so that 
the label is saved to the local filesystem because there is still some 
ambiguity as to the code paths (at least in my mind.)  I've also not 
made any attempt so far to support a "dumb" server, though its really a 
matter of code.

Only explicit SETATTR or file creation will result in a label being 
pushed to the server.  mknod/mkdir/mkfifo etc. are next on my list.

I would like to ask opinions on a less EA centric mechanism for setting 
labels from userland.  As we're pursuing a labeling solution that does 
not rely on EAs for persistent storage (from the client's point of view) 
it becomes difficult to shoehorn things so that userland tools work as 
expected.

-- 
Matthew N. Dodd <Matthew.Dodd at sparta.com>
Principal Engineer, ISSO, SRD
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: labeled-nfs.patch
Url: http://linux-nfs.org/pipermail/labeled-nfs/attachments/20070803/30e23459/attachment.ksh