sample 5.3 based trusted os ;-)

[Chris Wright]

* Robert Watson (rwatson at FreeBSD.org) wrote:
> Since we have your ear -- one of the ongoing concerns in the FreeBSD port
> of the Darwin audit parts has been the reliable generation of path names
> for file operations.  Path names seem like an unreliable thing at the best
> of times on most UNIX systems -- files with 0, 1, or more names, covered
> mount points, hard links, non-canonical names, etc.  How hard does Solaris
> try to generate "reliable" path names, and are the caveats and design
> choices in path name generation for Audit currently publically documented
> somewhere?  Find "a name" is typically fairly straight forward, but
> presumably the real value in a path name is how consistently available and
> useful it is for post-processing -- knowing what applications consuming
> BSM audit streams expect would be very helpful in knowing just how hard to
> make the FreeBSD kernel try to generate a useful name.

I'm interested as well.  The Linux effort currently records the name
used to get at the file object, as well as inode number and device that
it resolves to.

thanks,
-chris
-- 
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net

To Unsubscribe: send mail to majordomo at trustedbsd.org
with "unsubscribe trustedbsd-discuss" in the body of the message