[EMAIL PROTECTED] (Niels M�ller) writes:

> How is that used? I was thinking of some "signal" that underlying
> device (or store) can send to the file system translator. After all,
> the device is probably the component that have the best chance of
> detecting changes.

Hrm, --update is the traditional "remount" option.  It's more a
top-down thing than a bottom-up thing.  Still, all the functionality
is there in diskfs now.

> Or perhaps the general change notification mechanism can handle that
> already?

Ah, this is probably the best thing.  Ideally, read-only diskfs could
register a change-notification on the underlying store, and if it gets
changes, it could do the remount automatically.  There might be some
problems making it all work.  The read-write version is pretty
fundamentally unable to DTRT, of course.

Still, I think that it's a minor issue to worry about.  I'm content
with just saying "no, of course you can't run two different
filesystems on the same store", and then leaving it at that.  

Thomas

_______________________________________________
Help-hurd mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/help-hurd

Reply via email to