On Fri, Jun 24, 2011 at 04:13:14PM -0400, Thor Lancelot Simon wrote: > On Fri, Jun 24, 2011 at 10:00:26PM +0200, J. Hannken-Illjes wrote: > > ... this time with a diff attached: > > > > With the attached diff vnd(4) devices will work on sparse files. > > - Make the strategy decision a device flag and set to VNF_USE_VN_RDWR for > > files known to be sparse. > > What will the consumer of the vnd (in the important case, Xen, I suppose) > see happen if there are no free blocks available when it writes into a > hole?
the block device backen response can only be: BLKIF_RSP_EOPNOTSUPP, BLKIF_RSP_ERROR or BLKIF_RSP_OKAY. BLKIF_RSP_EOPNOTSUPP is not allowed for a write, so the only possible solution is BLKIF_RSP_ERROR (assuming the write to the vnd will return an error - which I really hope it will do !) -- Manuel Bouyer <[email protected]> NetBSD: 26 ans d'experience feront toujours la difference --
