This is a note to let you know that I've just added the patch titled
scsi: qla2xxx: Fix hang in task management
to the 5.15-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
scsi-qla2xxx-fix-hang-in-task-management.patch
and it can be found in the queue-5.15 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
commit 92c8ae8a606aee9eeeb8e31aa58d646a6665bbfb
Author: Quinn Tran <qutran@xxxxxxxxxxx>
Date: Fri Apr 28 00:53:36 2023 -0700
scsi: qla2xxx: Fix hang in task management
[ Upstream commit 9ae615c5bfd37bd091772969b1153de5335ea986 ]
Task management command hangs where a side
band chip reset failed to nudge the TMF
from it's current send path.
Add additional error check to block TMF
from entering during chip reset and along
the TMF path to cause it to bail out, skip
over abort of marker.
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Quinn Tran <qutran@xxxxxxxxxxx>
Signed-off-by: Nilesh Javali <njavali@xxxxxxxxxxx>
Link: https://lore.kernel.org/r/20230428075339.32551-5-njavali@xxxxxxxxxxx
Reviewed-by: Himanshu Madhani <himanshu.madhani@xxxxxxxxxx>
Signed-off-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index 2a2bca2fb57ee..83228ce822af3 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -5481,4 +5481,8 @@ struct ql_vnd_tgt_stats_resp {
_fp->disc_state, _fp->scan_state, _fp->loop_id, _fp->deleted, \
_fp->flags
+#define TMF_NOT_READY(_fcport) \
+ (!_fcport || IS_SESSION_DELETED(_fcport) || atomic_read(&_fcport->state) != FCS_ONLINE || \
+ !_fcport->vha->hw->flags.fw_started)
+
#endif
diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 2c42ecb2a64a5..a97872b6350ca 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -1998,6 +1998,11 @@ qla2x00_tmf_iocb_timeout(void *data)
int rc, h;
unsigned long flags;
+ if (sp->type == SRB_MARKER) {
+ complete(&tmf->u.tmf.comp);
+ return;
+ }
+
rc = qla24xx_async_abort_cmd(sp, false);
if (rc) {
spin_lock_irqsave(sp->qpair->qp_lock_ptr, flags);
@@ -2025,6 +2030,7 @@ static void qla_marker_sp_done(srb_t *sp, int res)
sp->handle, sp->fcport->d_id.b24, sp->u.iocb_cmd.u.tmf.flags,
sp->u.iocb_cmd.u.tmf.lun, sp->qpair->id);
+ sp->u.iocb_cmd.u.tmf.data = res;
complete(&tmf->u.tmf.comp);
}
@@ -2041,6 +2047,11 @@ static void qla_marker_sp_done(srb_t *sp, int res)
} while (cnt); \
}
+/**
+ * qla26xx_marker: send marker IOCB and wait for the completion of it.
+ * @arg: pointer to argument list.
+ * It is assume caller will provide an fcport pointer and modifier
+ */
static int
qla26xx_marker(struct tmf_arg *arg)
{
@@ -2050,6 +2061,14 @@ qla26xx_marker(struct tmf_arg *arg)
int rval = QLA_FUNCTION_FAILED;
fc_port_t *fcport = arg->fcport;
+ if (TMF_NOT_READY(arg->fcport)) {
+ ql_dbg(ql_dbg_taskm, vha, 0x8039,
+ "FC port not ready for marker loop-id=%x portid=%06x modifier=%x lun=%lld qp=%d.\n",
+ fcport->loop_id, fcport->d_id.b24,
+ arg->modifier, arg->lun, arg->qpair->id);
+ return QLA_SUSPENDED;
+ }
+
/* ref: INIT */
sp = qla2xxx_get_qpair_sp(vha, arg->qpair, fcport, GFP_KERNEL);
if (!sp)
@@ -2076,11 +2095,19 @@ qla26xx_marker(struct tmf_arg *arg)
if (rval != QLA_SUCCESS) {
ql_log(ql_log_warn, vha, 0x8031,
- "Marker IOCB failed (%x).\n", rval);
+ "Marker IOCB send failure (%x).\n", rval);
goto done_free_sp;
}
wait_for_completion(&tm_iocb->u.tmf.comp);
+ rval = tm_iocb->u.tmf.data;
+
+ if (rval != QLA_SUCCESS) {
+ ql_log(ql_log_warn, vha, 0x8019,
+ "Marker failed hdl=%x loop-id=%x portid=%06x modifier=%x lun=%lld qp=%d rval %d.\n",
+ sp->handle, fcport->loop_id, fcport->d_id.b24,
+ arg->modifier, arg->lun, sp->qpair->id, rval);
+ }
done_free_sp:
/* ref: INIT */
@@ -2093,6 +2120,8 @@ static void qla2x00_tmf_sp_done(srb_t *sp, int res)
{
struct srb_iocb *tmf = &sp->u.iocb_cmd;
+ if (res)
+ tmf->u.tmf.data = res;
complete(&tmf->u.tmf.comp);
}
@@ -2106,6 +2135,14 @@ __qla2x00_async_tm_cmd(struct tmf_arg *arg)
fc_port_t *fcport = arg->fcport;
+ if (TMF_NOT_READY(arg->fcport)) {
+ ql_dbg(ql_dbg_taskm, vha, 0x8032,
+ "FC port not ready for TM command loop-id=%x portid=%06x modifier=%x lun=%lld qp=%d.\n",
+ fcport->loop_id, fcport->d_id.b24,
+ arg->modifier, arg->lun, arg->qpair->id);
+ return QLA_SUSPENDED;
+ }
+
/* ref: INIT */
sp = qla2xxx_get_qpair_sp(vha, arg->qpair, fcport, GFP_KERNEL);
if (!sp)
@@ -2180,7 +2217,9 @@ int qla_get_tmf(fc_port_t *fcport)
msleep(1);
spin_lock_irqsave(&ha->tgt.sess_lock, flags);
- if (fcport->deleted) {
+ if (TMF_NOT_READY(fcport)) {
+ ql_log(ql_log_warn, vha, 0x802c,
+ "Unable to acquire TM resource due to disruption.\n");
rc = EIO;
break;
}
@@ -2206,7 +2245,10 @@ qla2x00_async_tm_cmd(fc_port_t *fcport, uint32_t flags, uint64_t lun,
struct scsi_qla_host *vha = fcport->vha;
struct qla_qpair *qpair;
struct tmf_arg a;
- int i, rval;
+ int i, rval = QLA_SUCCESS;
+
+ if (TMF_NOT_READY(fcport))
+ return QLA_SUSPENDED;
a.vha = fcport->vha;
a.fcport = fcport;
@@ -2225,6 +2267,14 @@ qla2x00_async_tm_cmd(fc_port_t *fcport, uint32_t flags, uint64_t lun,
qpair = vha->hw->queue_pair_map[i];
if (!qpair)
continue;
+
+ if (TMF_NOT_READY(fcport)) {
+ ql_log(ql_log_warn, vha, 0x8026,
+ "Unable to send TM due to disruption.\n");
+ rval = QLA_SUSPENDED;
+ break;
+ }
+
a.qpair = qpair;
a.flags = flags|TCF_NOTMCMD_TO_TARGET;
rval = __qla2x00_async_tm_cmd(&a);
@@ -2233,10 +2283,14 @@ qla2x00_async_tm_cmd(fc_port_t *fcport, uint32_t flags, uint64_t lun,
}
}
+ if (rval)
+ goto bailout;
+
a.qpair = vha->hw->base_qpair;
a.flags = flags;
rval = __qla2x00_async_tm_cmd(&a);
+bailout:
if (a.modifier == MK_SYNC_ID_LUN)
qla_put_tmf(fcport);
