LU-2726 procfs: Invalid assertion for counter_{add,sub}

[fs/lustre-release.git] / lnet / lnet / peer.c

diff --git a/lnet/lnet/peer.c b/lnet/lnet/peer.c
index b257624..27be863 100644 (file)
--- a/lnet/lnet/peer.c
+++ b/lnet/lnet/peer.c
@@ -26,6 +26,8 @@
 /*
  * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
  * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, Intel Corporation.
  */
 /*
  * This file is part of Lustre, http://www.lustre.org/
@@ -186,8 +188,7 @@ lnet_find_peer_locked(struct lnet_peer_table *ptable, lnet_nid_t nid)
        cfs_list_t      *peers;
        lnet_peer_t     *lp;
 
-       if (the_lnet.ln_shutdown)
-               return NULL;
+       LASSERT(!the_lnet.ln_shutdown);
 
        peers = &ptable->pt_hash[lnet_nid2peerhash(nid)];
        cfs_list_for_each_entry(lp, peers, lp_hashlist) {
@@ -207,16 +208,21 @@ lnet_nid2peer_locked(lnet_peer_t **lpp, lnet_nid_t nid, int cpt)
        lnet_peer_t             *lp = NULL;
        lnet_peer_t             *lp2;
        int                     cpt2;
+       int                     rc = 0;
+
+       *lpp = NULL;
+       if (the_lnet.ln_shutdown) /* it's shutting down */
+               return -ESHUTDOWN;
 
        /* cpt can be LNET_LOCK_EX if it's called from router functions */
-       cpt2 = cpt != LNET_LOCK_EX ? cpt : lnet_cpt_of_nid(nid);
+       cpt2 = cpt != LNET_LOCK_EX ? cpt : lnet_cpt_of_nid_locked(nid);
 
        ptable = the_lnet.ln_peer_tables[cpt2];
        lp = lnet_find_peer_locked(ptable, nid);
-        if (lp != NULL) {
-                *lpp = lp;
-                return 0;
-        }
+       if (lp != NULL) {
+               *lpp = lp;
+               return 0;
+       }
 
        if (!cfs_list_empty(&ptable->pt_deathrow)) {
                lp = cfs_list_entry(ptable->pt_deathrow.next,
@@ -224,21 +230,26 @@ lnet_nid2peer_locked(lnet_peer_t **lpp, lnet_nid_t nid, int cpt)
                cfs_list_del(&lp->lp_hashlist);
        }
 
+       /*
+        * take extra refcount in case another thread has shutdown LNet
+        * and destroyed locks and peer-table before I finish the allocation
+        */
+       ptable->pt_number++;
        lnet_net_unlock(cpt);
 
        if (lp != NULL)
                memset(lp, 0, sizeof(*lp));
        else
-               LIBCFS_ALLOC(lp, sizeof(*lp));
+               LIBCFS_CPT_ALLOC(lp, lnet_cpt_table(), cpt2, sizeof(*lp));
 
        if (lp == NULL) {
-                *lpp = NULL;
-                LNET_LOCK();
-                return -ENOMEM;
-        }
+               rc = -ENOMEM;
+               lnet_net_lock(cpt);
+               goto out;
+       }
 
        CFS_INIT_LIST_HEAD(&lp->lp_txq);
-        CFS_INIT_LIST_HEAD(&lp->lp_rtrq);
+       CFS_INIT_LIST_HEAD(&lp->lp_rtrq);
        CFS_INIT_LIST_HEAD(&lp->lp_routes);
 
         lp->lp_notify = 0;
@@ -250,51 +261,47 @@ lnet_nid2peer_locked(lnet_peer_t **lpp, lnet_nid_t nid, int cpt)
         lp->lp_last_alive = cfs_time_current(); /* assumes alive */
         lp->lp_last_query = 0; /* haven't asked NI yet */
         lp->lp_ping_timestamp = 0;
-       lp->lp_ping_version = LNET_PROTO_PING_UNKNOWN;
+       lp->lp_ping_feats = LNET_PING_FEAT_INVAL;
        lp->lp_nid = nid;
        lp->lp_cpt = cpt2;
-       lp->lp_refcount = 2;                    /* 1 for caller; 1 for hash */
+       lp->lp_refcount = 2;    /* 1 for caller; 1 for hash */
        lp->lp_rtr_refcount = 0;
 
        lnet_net_lock(cpt);
 
+       if (the_lnet.ln_shutdown) {
+               rc = -ESHUTDOWN;
+               goto out;
+       }
+
        lp2 = lnet_find_peer_locked(ptable, nid);
        if (lp2 != NULL) {
-               cfs_list_add(&lp->lp_hashlist, &ptable->pt_deathrow);
-
-                if (the_lnet.ln_shutdown) {
-                        lnet_peer_decref_locked(lp2);
-                        *lpp = NULL;
-                        return -ESHUTDOWN;
-                }
-
-                *lpp = lp2;
-                return 0;
-        }
+               *lpp = lp2;
+               goto out;
+       }
 
        lp->lp_ni = lnet_net2ni_locked(LNET_NIDNET(nid), cpt2);
        if (lp->lp_ni == NULL) {
-               cfs_list_add(&lp->lp_hashlist, &ptable->pt_deathrow);
-
-                *lpp = NULL;
-                return the_lnet.ln_shutdown ? -ESHUTDOWN : -EHOSTUNREACH;
-        }
-
-        lp->lp_txcredits    =
-        lp->lp_mintxcredits = lp->lp_ni->ni_peertxcredits;
-        lp->lp_rtrcredits    =
-        lp->lp_minrtrcredits = lnet_peer_buffer_credits(lp->lp_ni);
+               rc = -EHOSTUNREACH;
+               goto out;
+       }
 
-        /* can't add peers after shutdown starts */
-        LASSERT (!the_lnet.ln_shutdown);
+       lp->lp_txcredits    =
+       lp->lp_mintxcredits = lp->lp_ni->ni_peertxcredits;
+       lp->lp_rtrcredits    =
+       lp->lp_minrtrcredits = lnet_peer_buffer_credits(lp->lp_ni);
 
        cfs_list_add_tail(&lp->lp_hashlist,
                          &ptable->pt_hash[lnet_nid2peerhash(nid)]);
        ptable->pt_version++;
-       ptable->pt_number++;
-
        *lpp = lp;
+
        return 0;
+out:
+       if (lp != NULL)
+               cfs_list_add(&lp->lp_hashlist, &ptable->pt_deathrow);
+       ptable->pt_number--;
+       return rc;
 }
 
 void