Jens,=20

On Oct 25, 2018, at 3:18 PM, Jens Axboe [off-list ref] wrote:
=20
External Email
=20
On 10/25/18 3:36 PM, Bart Van Assche wrote:

quoted

On Thu, 2018-10-25 at 15:10 -0600, Jens Axboe wrote:

quoted

@@ -3265,25 +3261,17 @@ qla2x00_probe_one(struct pci_dev *pdev, const s=

truct pci_device_id *id)

quoted

quoted

        base_vha->mgmt_svr_loop_id, host->sg_tablesize);
=20
    if (ha->mqenable) {
-            bool mq =3D false;
            bool startit =3D false;
=20
-            if (QLA_TGT_MODE_ENABLED()) {
-                    mq =3D true;
+            if (QLA_TGT_MODE_ENABLED())
                    startit =3D false;
-            }
=20
-            if ((ql2x_ini_mode =3D=3D QLA2XXX_INI_MODE_ENABLED) &&
-                shost_use_blk_mq(host)) {
-                    mq =3D true;
+            if (ql2x_ini_mode =3D=3D QLA2XXX_INI_MODE_ENABLED)
                    startit =3D true;
-            }
=20
-            if (mq) {
-                    /* Create start of day qpairs for Block MQ */
-                    for (i =3D 0; i < ha->max_qpairs; i++)
-                            qla2xxx_create_qpair(base_vha, 5, 0, start=

it);

quoted

quoted

-            }
+            /* Create start of day qpairs for Block MQ */
+            for (i =3D 0; i < ha->max_qpairs; i++)
+                    qla2xxx_create_qpair(base_vha, 5, 0, startit);
    }
=20
    if (ha->flags.running_gold_fw)

=20
(+Himanshu)
=20
Since I'm not sure that "mq" in the above code refers to "scsi-mq" nor t=

hat it

quoted

refers to "blk-mq", I'm not sure the above changes should be included in=

 this

quoted

patch. Himanshu, can you have a look?

=20
There's literally a comment there that says "for Block MQ" :-)

This change is Good.

We were using mq boolean to determine if we are in Target mode or Initiator=
 with BLK-MQ on
to create qpair in firmware.

Since this patch is removing shost_use_blk_mq)(), so the mq boolean becomes=
 no-op.=20

quoted

quoted

+static ssize_t
+show_use_blk_mq(struct device *dev, struct device_attribute *attr, cha=

r *buf)

quoted

quoted

+{
+    return snprintf(buf, 20, "1\n");
+}
+static DEVICE_ATTR(use_blk_mq, S_IRUGO, show_use_blk_mq, NULL);

=20
The pattern "snprintf(buf, 20, ...)" looks like cargo-cult programming t=

o me.

quoted

Please consider changing "snprintf(buf, 20, " into "sprintf(buf, ".

=20
Good point, I'll do that.
=20
--
Jens Axboe
=20

Thanks,
- Himanshu