[dm-devel] Re: [PATCH-RFC] Bug in dm-raid1 as used by pvmove
Alasdair G Kergon
agk at redhat.com
Fri May 19 09:00:22 UTC 2006
On Thu, May 18, 2006 at 11:45:09PM -0500, Jon Brassow wrote:
> I've also check to make sure that it doesn't screw anything up when
> fault tolerance is included (see http://www.brassow.com/mirroring). We
> had to check to ensure that if bio_map was called (which alters the bio
> and takes off the ti->begin) that a later function would not then call
> the updated bio_to_region - effectively subtracting off ti->begin
> twice.
This is a small thing I want to consider for the 'next generation' of
device-mapper. The dm core should be responsible for subtracting the
offset. (Many target implementations originally forget to do this.)
Alasdair
--
agk at redhat.com
More information about the dm-devel
mailing list