On Thu, 2014-06-05 at 23:30 +0000, Nicholas A. Bellinger wrote:
> From: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx>
>
> This patch changes ft_queue_status() to set SAM_STAT_TASK_SET_FULL
> status upon lport->tt.seq_send( failure, and return -EAGAIN to notify
> target-core to attempt to requeue the response.
>
> It also does the same for a fc_frame_alloc() failures, in order to
> signal the initiator that it should try to reduce it's current
> queue_depth, to lower the number of outstanding I/Os on the wire.
>
> Reported-by: Vasu Dev <vasu.dev@xxxxxxxxxxxxxxx>
> Cc: Vasu Dev <vasu.dev@xxxxxxxxxxxxxxx>
> Cc: Jun Wu <jwu@xxxxxxxxxxxx>
> Signed-off-by: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx>
> ---
> drivers/target/tcm_fc/tfc_cmd.c | 19 ++++++++++++++++---
> 1 file changed, 16 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/target/tcm_fc/tfc_cmd.c b/drivers/target/tcm_fc/tfc_cmd.c
> index f5fd515..5585038 100644
> --- a/drivers/target/tcm_fc/tfc_cmd.c
> +++ b/drivers/target/tcm_fc/tfc_cmd.c
> @@ -128,6 +128,7 @@ int ft_queue_status(struct se_cmd *se_cmd)
> struct fc_lport *lport;
> struct fc_exch *ep;
> size_t len;
> + int rc;
>
> if (cmd->aborted)
> return 0;
> @@ -137,9 +138,10 @@ int ft_queue_status(struct se_cmd *se_cmd)
> len = sizeof(*fcp) + se_cmd->scsi_sense_length;
> fp = fc_frame_alloc(lport, len);
> if (!fp) {
> - /* XXX shouldn't just drop it - requeue and retry? */
> - return 0;
> + se_cmd->scsi_status = SAM_STAT_TASK_SET_FULL;
> + return -ENOMEM;
> }
> +
> fcp = fc_frame_payload_get(fp, len);
> memset(fcp, 0, len);
> fcp->resp.fr_status = se_cmd->scsi_status;
> @@ -170,7 +172,18 @@ int ft_queue_status(struct se_cmd *se_cmd)
> fc_fill_fc_hdr(fp, FC_RCTL_DD_CMD_STATUS, ep->did, ep->sid, FC_TYPE_FCP,
> FC_FC_EX_CTX | FC_FC_LAST_SEQ | FC_FC_END_SEQ, 0);
>
> - lport->tt.seq_send(lport, cmd->seq, fp);
> + rc = lport->tt.seq_send(lport, cmd->seq, fp);
> + if (rc) {
> + pr_err_ratelimited("%s: Failed to send response frame %p, "
> + "xid <0x%x>\n", __func__, fp, ep->xid);
May be should be pr_info_ratelimited since this is not really error and
IO could get aborted due to timeouts in any setups or stressed
workloads.
> + /*
> + * Generate a TASK_SET_FULL status to notify the initiator
> + * to reduce it's queue_depth after the se_cmd response has
> + * been re-queued by target-core.
> + */
> + se_cmd->scsi_status = SAM_STAT_TASK_SET_FULL;
> + return -ENOMEM;
> + }
> lport->tt.exch_done(cmd->seq);
> return 0;
> }
Reviewed-by: Vasu Dev <vasu.dev@xxxxxxxxx>
--
To unsubscribe from this list: 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
