On Tue, Jun 14, 2005 at 11:42:24AM -0400, Thor Lancelot Simon wrote:
> On Tue, Jun 14, 2005 at 03:32:42PM +0200, Manuel Bouyer wrote:
> > 
> > I think we can deal with this by sending a partial xfer to the drive, it
> > shouldn't break anything. But obviously the completion message has to be
> > sent only one the whole request has been handled.
> 
> What if the domain 0 kernel crashes, or the system is powered down,
> between the partial transfer and the remaining part?  Now neither the
> old nor the new data is actually intact on disk.  This could have a
> rather negative impact on, for example, filesystem metadata...

Isn't it something that can already happen inside the disk ? It's not clear if
multi-sector writes are atomics.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--