[LU-9917] lnet_discover_peer_locked() must refresh lp after unlock and lock

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.11.0
• Labels: lnet multi-rail
• Severity: 3
• Assignee: Amir Shehata (Inactive)
• Reporter: John Hammond

Description

In lnet_discover_peer_locked() after the loop we unlock and relock the LNet cpt lock.

                lnet_net_lock(LNET_LOCK_EX);
                lnet_peer_decref_locked(lp);
                /* Peer may have changed */
                lp = lpni->lpni_peer_net->lpn_peer;
        }
        finish_wait(&lp->lp_dc_waitq, &wait);

        lnet_net_unlock(LNET_LOCK_EX);
        lnet_net_lock(cpt);

        if (signal_pending(current))
                rc = -EINTR;
        else if (the_lnet.ln_dc_state != LNET_DC_STATE_RUNNING)
                rc = -ESHUTDOWN;
        else if (lp->lp_dc_error)
                rc = lp->lp_dc_error;
        else if (!block)
                CDEBUG(D_NET, "non-blocking discovery\n");
        else if (!lnet_peer_is_uptodate(lp))
                goto again;
        CDEBUG(D_NET, "peer %s NID %s: %d. %s\n",
               (lp ? libcfs_nid2str(lp->lp_primary_nid) : "(none)"),
               libcfs_nid2str(lpni->lpni_nid), rc,
               (!block) ? "pending discovery" : "discovery complete");

        return rc;

After relocking lp may be invalid and we need to refresh it from lpni. Or move the unlock and lock down and adjust the again label. Do we need LNET_LOCK_EX to access lp?

Issue Links

• is related to: LU-10281 conf-sanity: test_54a hung at lnet_discover_peer_locked()