On Mon, Mar 12, 2001 at 10:58:29AM -0500, Alexander Viro wrote:
> What I mean is quite simple - objects in cred.cache can be
> shared, but to change the contents of object you need to make sure
> that you are the sole owner of that object. I.e.
>
> modify_cred(cred, ...)
> {
> if ((*cred)->count > 1) {
> old_cred = *cred;
> *cred = copy_cred(old_cred);
> put_cred(old_cred);
> }
> /* modify *cred */
> }
>
> IOW, no third-part changes of contents. Simple, easy to prove
> the correctness. Pthreads folks can just do modify_cred() for process
> credentials of all memebers of thread group. IMO trying to be smart
> and physically share the credentials for all threads is wrong.
This sounds like the kind of thing which the read-copy update strategy
was designed to solve.
modify_cred(&cred, ...)
{
new_cred = copy_cred(cred);
/* modify cred */
*cred = new_cred;
add_idle_free(*cred);
}
as i understand the read-copy update strategy anyway.
--
Revolutions do not require corporate support.
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [EMAIL PROTECTED]
