Hello Mike,
You're right. That extra checking for !hba was added recently as
defensive programming and it should never hit. I'll remove it in v2.
Thanks,
Eddie
On Tue, 2010-06-29 at 22:53 -0700, Mike Christie wrote:
> On 06/25/2010 08:39 PM, Eddie Wai wrote:
> > +/**
> > + * bnx2i_ep_disconnect - executes TCP connection teardown process
> > + * @ep: TCP connection (iscsi endpoint) handle
> > + *
> > + * executes TCP connection teardown process
> > + */
> > +static void bnx2i_ep_disconnect(struct iscsi_endpoint *ep)
> > +{
> > + struct bnx2i_endpoint *bnx2i_ep;
> > + struct bnx2i_conn *bnx2i_conn = NULL;
> > + struct iscsi_conn *conn = NULL;
> > + struct bnx2i_hba *hba;
> > +
> > + bnx2i_ep = ep->dd_data;
> > +
> > + /* driver should not attempt connection cleanup until TCP_CONNECT
> > + * completes either successfully or fails. Timeout is 9-secs, so
> > + * wait for it to complete
> > + */
> > + while ((bnx2i_ep->state == EP_STATE_CONNECT_START)&&
> > + !time_after(jiffies, bnx2i_ep->timestamp + (12 * HZ)))
> > + msleep(250);
> > +
> > + if (bnx2i_ep->conn) {
> > + bnx2i_conn = bnx2i_ep->conn;
> > + conn = bnx2i_conn->cls_conn->dd_data;
> > + iscsi_suspend_queue(conn);
> > + }
> > + hba = bnx2i_ep->hba;
> > + if (!hba)
> > + goto out;
>
>
> Here you have a ep, so if you goto out you will not release it. Need to
> call iscsi_destroy_endpoint at least. Can you hit this "if" even?
>
--
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
