From: Roland Dreier <roland@xxxxxxxxxxxxxxx>
In the two xmit_response functions, the test of the return value from
qla_tgt_check_reserve_free_req() was messed up: if QLA_TGT_XMIT_DATA
was set, then even if qla_tgt_check_reserve_free_req() failed, the
code would go ahead and submit an IOCB to the request queue. This
overruns the request queue and confuses the HBA hardware.
Fix this so we properly return -EAGAIN and trigger QUEUE FULL handling.
Signed-off-by: Roland Dreier <roland@xxxxxxxxxxxxxxx>
---
drivers/scsi/qla2xxx/qla_target.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
index 1d3b3b8..bef1a21 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -2214,7 +2214,7 @@ static int __qla2xxx_xmit_response(struct qla_tgt_cmd *cmd, int xmit_type, uint8
/* Does F/W have an IOCBs for this request */
res = qla_tgt_check_reserve_free_req(vha, full_req_cnt);
- if (unlikely(res != 0) && (xmit_type & QLA_TGT_XMIT_DATA))
+ if (unlikely(res))
goto out_unmap_unlock;
qla2xxx_build_ctio_pkt(&prm, cmd->vha);
@@ -2471,7 +2471,7 @@ static int __qla24xx_xmit_response(struct qla_tgt_cmd *cmd, int xmit_type, uint8
/* Does F/W have an IOCBs for this request */
res = qla_tgt_check_reserve_free_req(vha, full_req_cnt);
- if (unlikely(res != 0) && (xmit_type & QLA_TGT_XMIT_DATA))
+ if (unlikely(res))
goto out_unmap_unlock;
res = qla24xx_build_ctio_pkt(&prm, vha);
--
1.7.5.4
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
