Re: Once again

Jamie Lokier (lkd@tantalophile.demon.co.uk)
Tue, 29 Jun 1999 04:25:29 +0200


Dale Amon wrote:
> > I don't think slow NFS links is an issue.
>
> Only in the sense that if there were just one serial number,
> different browsers would not be in sync (which probably isn't
> important unless they had to communicate, for which I can't
> think of a good reason).

There's always good old stat() for browsers that want to synchronise.
Not 100% reliable, but considered acceptable for this sort of thing.
The poll() tests are essentially only a hint anyway.

> And if there are many fd's on
> many hosts, how do they get their Modified and Serial
> numbers synced to the remote NFS files?

The numbers and flags are local. Remote modification is detected using
the normal NFS ->revalidate() checks just like any other fs.

It's only for ext2 that I'd propose _storing_ the data.

-- Jamie

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/