LU-12678 lnet: fix small race in unloading klnd modules.

[fs/lustre-release.git] / lnet / lnet / api-ni.c

diff --git a/lnet/lnet/api-ni.c b/lnet/lnet/api-ni.c
index d298f46..399acdb 100644 (file)
--- a/lnet/lnet/api-ni.c
+++ b/lnet/lnet/api-ni.c
@@ -550,11 +550,6 @@ lnet_init_locks(void)
        mutex_init(&the_lnet.ln_lnd_mutex);
 }
 
-static void
-lnet_fini_locks(void)
-{
-}
-
 struct kmem_cache *lnet_mes_cachep;       /* MEs kmem_cache */
 struct kmem_cache *lnet_small_mds_cachep;  /* <= LNET_SMALL_MD_SIZE bytes
                                            *  MDs kmem_cache */
@@ -643,8 +638,6 @@ lnet_destroy_locks(void)
                cfs_percpt_lock_free(the_lnet.ln_net_lock);
                the_lnet.ln_net_lock = NULL;
        }
-
-       lnet_fini_locks();
 }
 
 static int
@@ -800,9 +793,9 @@ static void lnet_assert_wire_constants(void)
        BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_ni) != 0);
 }
 
-static struct lnet_lnd *lnet_find_lnd_by_type(__u32 type)
+static const struct lnet_lnd *lnet_find_lnd_by_type(__u32 type)
 {
-       struct lnet_lnd *lnd;
+       const struct lnet_lnd *lnd;
 
        /* holding lnd mutex */
        if (type >= NUM_LNDS)
@@ -821,7 +814,7 @@ lnet_get_lnd_timeout(void)
 EXPORT_SYMBOL(lnet_get_lnd_timeout);
 
 void
-lnet_register_lnd(struct lnet_lnd *lnd)
+lnet_register_lnd(const struct lnet_lnd *lnd)
 {
        mutex_lock(&the_lnet.ln_lnd_mutex);
 
@@ -837,7 +830,7 @@ lnet_register_lnd(struct lnet_lnd *lnd)
 EXPORT_SYMBOL(lnet_register_lnd);
 
 void
-lnet_unregister_lnd(struct lnet_lnd *lnd)
+lnet_unregister_lnd(const struct lnet_lnd *lnd)
 {
        mutex_lock(&the_lnet.ln_lnd_mutex);
 
@@ -2069,7 +2062,14 @@ lnet_clear_zombies_nis_locked(struct lnet_net *net)
                islo = ni->ni_net->net_lnd->lnd_type == LOLND;
 
                LASSERT(!in_interrupt());
+               /* Holding the mutex makes it safe for lnd_shutdown
+                * to call module_put(). Module unload cannot finish
+                * until lnet_unregister_lnd() completes, and that
+                * requires the mutex.
+                */
+               mutex_lock(&the_lnet.ln_lnd_mutex);
                (net->net_lnd->lnd_shutdown)(ni);
+               mutex_unlock(&the_lnet.ln_lnd_mutex);
 
                if (!islo)
                        CDEBUG(D_LNI, "Removed LNI %s\n",
@@ -2274,7 +2274,7 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun)
        int                     rc;
        int                     ni_count = 0;
        __u32                   lnd_type;
-       struct lnet_lnd *lnd;
+       const struct lnet_lnd  *lnd;
        int                     peer_timeout =
                net->net_tunables.lct_peer_timeout;
        int                     maxtxcredits =
@@ -3916,9 +3916,7 @@ LNetCtl(unsigned int cmd, void *arg)
                return 0;
 
        case IOC_LIBCFS_TESTPROTOCOMPAT:
-               lnet_net_lock(LNET_LOCK_EX);
                the_lnet.ln_testprotocompat = data->ioc_flags;
-               lnet_net_unlock(LNET_LOCK_EX);
                return 0;
 
        case IOC_LIBCFS_LNET_FAULT: