On 2005-04-25T11:34:32, "goggin, edward" <egoggin@xxxxxxx> wrote: > Should the bi_error field of a failed bio clone completion be propagated > upwards into the parent of the cloned bio, possibly in dec_pending() of > dm.c? Doing so will allow target drivers higher than the one managing > the bottom-most mapped device to have access to enhanced bio sense > and error information. Otherwise, all they see is -EIO like now. Probably yes. Good point. Right now only dm-mpath knows about them of course, and we're very likely to be the lowest mapped device in the stack. But propagating the error upward might make sense. Sincerely, Lars Marowsky-Brée <lmb@xxxxxxx> -- High Availability & Clustering SUSE Labs, Research and Development SUSE LINUX Products GmbH - A Novell Business