On Fri, Apr 23, 2021 at 01:39:13PM +0200, Hannes Reinecke wrote:
> Introduce scsi_status_is_check_condition() and
That was added in the last patch, wasn't it?
> +++ b/block/bsg.c
> @@ -97,6 +97,8 @@ static int bsg_scsi_complete_rq(struct request *rq, struct sg_io_v4 *hdr)
> hdr->device_status = sreq->result & 0xff;
> hdr->transport_status = host_byte(sreq->result);
> hdr->driver_status = driver_byte(sreq->result);
> + if (scsi_status_is_check_condition(sreq->result))
> + hdr->driver_status |= DRIVER_SENSE;
I think hdr->driver_status also needs to be cleared to 0 first. A little
comment on the history would be nice as well.
> @@ -257,6 +257,8 @@ static int blk_complete_sghdr_rq(struct request *rq, struct sg_io_hdr *hdr,
> hdr->msg_status = msg_byte(req->result);
> hdr->host_status = host_byte(req->result);
> hdr->driver_status = driver_byte(req->result);
> + if (hdr->status == SAM_STAT_CHECK_CONDITION)
> + hdr->driver_status |= DRIVER_SENSE;
Same here. Also why the open coded check in one please and the
SAM_STAT_CHECK_CONDITION comparism in another?
Maybe we need a little helper instead of duplicating the logic?
> + } else
> + hp->driver_status |= DRIVER_SENSE;
Shouldn't this be where the previous driver_byte call was? And
maybe also use a proper helper?
> @@ -131,6 +131,8 @@ struct compat_sg_io_hdr {
> #define SG_INFO_DIRECT_IO 0x2 /* direct IO requested and performed */
> #define SG_INFO_MIXED_IO 0x4 /* part direct, part indirect IO */
>
> +/* Obsolete DRIVER_SENSE setting */
> +#define DRIVER_SENSE 0x08
I think this needs a better documentation on what it means and why it
is obsolete.
