--- Arjan van de Ven <arjan@xxxxxxxxxxxxx> wrote:
> custom error handling was supposed to have gone away already, but then
> libata came around for now... (until that moves to be a block driver),
> so.. I don't see the problem really.
Shouldn't one want more "custom error handling"? That is, per protocol.
So in effect what you call "custom error handling" actually means
per protocol error handling, be it FC, SATA, iSCSI, USB, SAS, etc.
So SCSI Core offloads recovery to the protocol stack, which depending
on the transport/interconnect may offload some (but not all) functionality
to the port driver. (Case in point: tunneling protocols.)
Then you can have SCSI Core only decide whether to retry to command.
(If so asked by the submitter from uppper layer.)
Luben
-
: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
