NFSD: limit the number of v4 clients to 1024 per 1GB of system memory

Currently there is no limit on how many v4 clients are supported
by the system. This can be a problem in systems with small memory
configuration to function properly when a very large number of
clients exist that creates memory shortage conditions.

This patch enforces a limit of 1024 NFSv4 clients, including courtesy
clients, per 1GB of system memory.  When the number of the clients
reaches the limit, requests that create new clients are returned
with NFS4ERR_DELAY and the laundromat is kicked start to trim old
clients. Due to the overhead of the upcall to remove the client
record, the maximun number of clients the laundromat removes on
each run is limited to 128. This is done to ensure the laundromat
can still process the other tasks in a timely manner.

Since there is now a limit of the number of clients, the 24-hr
idle time limit of courtesy client is no longer needed and was
removed.

Signed-off-by: Dai Ngo <dai.ngo@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

fs/nfsd/netns.h

@@ -191,6 +191,7 @@ struct nfsd_net {
 	siphash_key_t		siphash_key;
 
 	atomic_t		nfs4_client_count;
+	int			nfs4_max_clients;
 };
 
 /* Simple check to find out if a given net was properly initialized */

fs/nfsd/nfs4state.c

@@ -2059,6 +2059,10 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name,
 	struct nfs4_client *clp;
 	int i;
 
+	if (atomic_read(&nn->nfs4_client_count) >= nn->nfs4_max_clients) {
+		mod_delayed_work(laundry_wq, &nn->laundromat_work, 0);
+		return NULL;
+	}
 	clp = kmem_cache_zalloc(client_slab, GFP_KERNEL);
 	if (clp == NULL)
 		return NULL;
@@ -4336,6 +4340,9 @@ out:
 
 void nfsd4_init_leases_net(struct nfsd_net *nn)
 {
+	struct sysinfo si;
+	u64 max_clients;
+
 	nn->nfsd4_lease = 90;	/* default lease time */
 	nn->nfsd4_grace = 90;
 	nn->somebody_reclaimed = false;
@@ -4346,6 +4353,10 @@ void nfsd4_init_leases_net(struct nfsd_net *nn)
 	nn->s2s_cp_cl_id = nn->clientid_counter++;
 
 	atomic_set(&nn->nfs4_client_count, 0);
+	si_meminfo(&si);
+	max_clients = (u64)si.totalram * si.mem_unit / (1024 * 1024 * 1024);
+	max_clients *= NFS4_CLIENTS_PER_GB;
+	nn->nfs4_max_clients = max_t(int, max_clients, NFS4_CLIENTS_PER_GB);
 }
 
 static void init_nfs4_replay(struct nfs4_replay *rp)
@@ -5811,9 +5822,12 @@ static void
 nfs4_get_client_reaplist(struct nfsd_net *nn, struct list_head *reaplist,
 				struct laundry_time *lt)
 {
+	unsigned int maxreap, reapcnt = 0;
 	struct list_head *pos, *next;
 	struct nfs4_client *clp;
 
+	maxreap = (atomic_read(&nn->nfs4_client_count) >= nn->nfs4_max_clients) ?
+			NFSD_CLIENT_MAX_TRIM_PER_RUN : 0;
 	INIT_LIST_HEAD(reaplist);
 	spin_lock(&nn->client_lock);
 	list_for_each_safe(pos, next, &nn->client_lru) {
@@ -5824,14 +5838,15 @@ nfs4_get_client_reaplist(struct nfsd_net *nn, struct list_head *reaplist,
 			break;
 		if (!atomic_read(&clp->cl_rpc_users))
 			clp->cl_state = NFSD4_COURTESY;
-		if (!client_has_state(clp) ||
-				ktime_get_boottime_seconds() >=
-				(clp->cl_time + NFSD_COURTESY_CLIENT_TIMEOUT))
+		if (!client_has_state(clp))
 			goto exp_client;
-		if (nfs4_anylock_blockers(clp)) {
+		if (!nfs4_anylock_blockers(clp))
+			if (reapcnt >= maxreap)
+				continue;
 exp_client:
-			if (!mark_client_expired_locked(clp))
-				list_add(&clp->cl_lru, reaplist);
+		if (!mark_client_expired_locked(clp)) {
+			list_add(&clp->cl_lru, reaplist);
+			reapcnt++;
 		}
 	}
 	spin_unlock(&nn->client_lock);

fs/nfsd/nfsd.h

@@ -341,6 +341,8 @@ void		nfsd_lockd_shutdown(void);
 
 #define NFSD_LAUNDROMAT_MINTIMEOUT      1   /* seconds */
 #define	NFSD_COURTESY_CLIENT_TIMEOUT	(24 * 60 * 60)	/* seconds */
+#define	NFSD_CLIENT_MAX_TRIM_PER_RUN	128
+#define	NFS4_CLIENTS_PER_GB		1024
 
 /*
  * The following attributes are currently not supported by the NFSv4 server: