Merge tag 'nfs-for-3.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Features include:

   - Remove CONFIG_EXPERIMENTAL dependency from NFSv4.1
     Aside from the issues discussed at the LKS, distros are shipping
     NFSv4.1 with all the trimmings.
   - Fix fdatasync()/fsync() for the corner case of a server reboot.
   - NFSv4 OPEN access fix: finally distinguish correctly between
     open-for-read and open-for-execute permissions in all situations.
   - Ensure that the TCP socket is closed when we're in CLOSE_WAIT
   - More idmapper bugfixes
   - Lots of pNFS bugfixes and cleanups to remove unnecessary state and
     make the code easier to read.
   - In cases where a pNFS read or write fails, allow the client to
     resume trying layoutgets after two minutes of read/write-
     through-mds.
   - More net namespace fixes to the NFSv4 callback code.
   - More net namespace fixes to the NFSv3 locking code.
   - More NFSv4 migration preparatory patches.
     Including patches to detect network trunking in both NFSv4 and
     NFSv4.1
   - pNFS block updates to optimise LAYOUTGET calls."

* tag 'nfs-for-3.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (113 commits)
  pnfsblock: cleanup nfs4_blkdev_get
  NFS41: send real read size in layoutget
  NFS41: send real write size in layoutget
  NFS: track direct IO left bytes
  NFSv4.1: Cleanup ugliness in pnfs_layoutgets_blocked()
  NFSv4.1: Ensure that the layout sequence id stays 'close' to the current
  NFSv4.1: Deal with seqid wraparound in the pNFS return-on-close code
  NFSv4 set open access operation call flag in nfs4_init_opendata_res
  NFSv4.1: Remove the dependency on CONFIG_EXPERIMENTAL
  NFSv4 reduce attribute requests for open reclaim
  NFSv4: nfs4_open_done first must check that GETATTR decoded a file type
  NFSv4.1: Deal with wraparound when updating the layout "barrier" seqid
  NFSv4.1: Deal with wraparound issues when updating the layout stateid
  NFSv4.1: Always set the layout stateid if this is the first layoutget
  NFSv4.1: Fix another refcount issue in pnfs_find_alloc_layout
  NFSv4: don't put ACCESS in OPEN compound if O_EXCL
  NFSv4: don't check MAY_WRITE access bit in OPEN
  NFS: Set key construction data for the legacy upcall
  NFSv4.1: don't do two EXCHANGE_IDs on mount
  NFS: nfs41_walk_client_list(): re-lock before iterating
  ...

3 files changed, 74 insertions, 17 deletions

diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c
index 7ef14b3c5bee..e4fb3ba5a58a 100644
--- a/fs/lockd/mon.c
+++ b/fs/lockd/mon.c
@@ -7,7 +7,6 @@
  */
 
 #include <linux/types.h>
-#include <linux/utsname.h>
 #include <linux/kernel.h>
 #include <linux/ktime.h>
 #include <linux/slab.h>
@@ -19,6 +18,8 @@
 
 #include <asm/unaligned.h>
 
+#include "netns.h"
+
 #define NLMDBG_FACILITY         NLMDBG_MONITOR
 #define NSM_PROGRAM             100024
 #define NSM_VERSION             1
@@ -40,6 +41,7 @@ struct nsm_args {
         u32                     proc;
 
         char                    *mon_name;
+        char                    *nodename;
 };
 
 struct nsm_res {
@@ -70,7 +72,7 @@ static struct rpc_clnt *nsm_create(struct net *net)
         };
         struct rpc_create_args args = {
                 .net                    = net,
-                .protocol               = XPRT_TRANSPORT_UDP,
+                .protocol               = XPRT_TRANSPORT_TCP,
                 .address                = (struct sockaddr *)&sin,
                 .addrsize               = sizeof(sin),
                 .servername             = "rpc.statd",
@@ -83,10 +85,54 @@ static struct rpc_clnt *nsm_create(struct net *net)
         return rpc_create(&args);
 }
 
-static int nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res,
-                         struct net *net)
+static struct rpc_clnt *nsm_client_get(struct net *net)
 {
+        static DEFINE_MUTEX(nsm_create_mutex);
         struct rpc_clnt *clnt;
+        struct lockd_net *ln = net_generic(net, lockd_net_id);
+
+        spin_lock(&ln->nsm_clnt_lock);
+        if (ln->nsm_users) {
+                ln->nsm_users++;
+                clnt = ln->nsm_clnt;
+                spin_unlock(&ln->nsm_clnt_lock);
+                goto out;
+        }
+        spin_unlock(&ln->nsm_clnt_lock);
+
+        mutex_lock(&nsm_create_mutex);
+        clnt = nsm_create(net);
+        if (!IS_ERR(clnt)) {
+                ln->nsm_clnt = clnt;
+                smp_wmb();
+                ln->nsm_users = 1;
+        }
+        mutex_unlock(&nsm_create_mutex);
+out:
+        return clnt;
+}
+
+static void nsm_client_put(struct net *net)
+{
+        struct lockd_net *ln = net_generic(net, lockd_net_id);
+        struct rpc_clnt *clnt = ln->nsm_clnt;
+        int shutdown = 0;
+
+        spin_lock(&ln->nsm_clnt_lock);
+        if (ln->nsm_users) {
+                if (--ln->nsm_users)
+                        ln->nsm_clnt = NULL;
+                shutdown = !ln->nsm_users;
+        }
+        spin_unlock(&ln->nsm_clnt_lock);
+
+        if (shutdown)
+                rpc_shutdown_client(clnt);
+}
+
+static int nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res,
+                         struct rpc_clnt *clnt)
+{
         int             status;
         struct nsm_args args = {
                 .priv           = &nsm->sm_priv,
@@ -94,31 +140,24 @@ static int nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res,
                 .vers           = 3,
                 .proc           = NLMPROC_NSM_NOTIFY,
                 .mon_name       = nsm->sm_mon_name,
+                .nodename       = clnt->cl_nodename,
         };
         struct rpc_message msg = {
                 .rpc_argp       = &args,
                 .rpc_resp       = res,
         };
 
-        clnt = nsm_create(net);
-        if (IS_ERR(clnt)) {
-                status = PTR_ERR(clnt);
-                dprintk("lockd: failed to create NSM upcall transport, "
-                                "status=%d\n", status);
-                goto out;
-        }
+        BUG_ON(clnt == NULL);
 
         memset(res, 0, sizeof(*res));
 
         msg.rpc_proc = &clnt->cl_procinfo[proc];
-        status = rpc_call_sync(clnt, &msg, 0);
+        status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFTCONN);
         if (status < 0)
                 dprintk("lockd: NSM upcall RPC failed, status=%d\n",
                                 status);
         else
                 status = 0;
-        rpc_shutdown_client(clnt);
- out:
         return status;
 }
 
@@ -138,6 +177,7 @@ int nsm_monitor(const struct nlm_host *host)
         struct nsm_handle *nsm = host->h_nsmhandle;
         struct nsm_res  res;
         int             status;
+        struct rpc_clnt *clnt;
 
         dprintk("lockd: nsm_monitor(%s)\n", nsm->sm_name);
 
@@ -150,7 +190,15 @@ int nsm_monitor(const struct nlm_host *host)
          */
         nsm->sm_mon_name = nsm_use_hostnames ? nsm->sm_name : nsm->sm_addrbuf;
 
-        status = nsm_mon_unmon(nsm, NSMPROC_MON, &res, host->net);
+        clnt = nsm_client_get(host->net);
+        if (IS_ERR(clnt)) {
+                status = PTR_ERR(clnt);
+                dprintk("lockd: failed to create NSM upcall transport, "
+                                "status=%d, net=%p\n", status, host->net);
+                return status;
+        }
+
+        status = nsm_mon_unmon(nsm, NSMPROC_MON, &res, clnt);
         if (unlikely(res.status != 0))
                 status = -EIO;
         if (unlikely(status < 0)) {
@@ -182,9 +230,11 @@ void nsm_unmonitor(const struct nlm_host *host)
 
         if (atomic_read(&nsm->sm_count) == 1
          && nsm->sm_monitored && !nsm->sm_sticky) {
+                struct lockd_net *ln = net_generic(host->net, lockd_net_id);
+
                 dprintk("lockd: nsm_unmonitor(%s)\n", nsm->sm_name);
 
-                status = nsm_mon_unmon(nsm, NSMPROC_UNMON, &res, host->net);
+                status = nsm_mon_unmon(nsm, NSMPROC_UNMON, &res, ln->nsm_clnt);
                 if (res.status != 0)
                         status = -EIO;
                 if (status < 0)
@@ -192,6 +242,8 @@ void nsm_unmonitor(const struct nlm_host *host)
                                         nsm->sm_name);
                 else
                         nsm->sm_monitored = 0;
+
+                nsm_client_put(host->net);
         }
 }
 
@@ -430,7 +482,7 @@ static void encode_my_id(struct xdr_stream *xdr, const struct nsm_args *argp)
 {
         __be32 *p;
 
-        encode_nsm_string(xdr, utsname()->nodename);
+        encode_nsm_string(xdr, argp->nodename);
         p = xdr_reserve_space(xdr, 4 + 4 + 4);
         *p++ = cpu_to_be32(argp->prog);
         *p++ = cpu_to_be32(argp->vers);
diff --git a/fs/lockd/netns.h b/fs/lockd/netns.h
index 4eee248ba96e..5010b55628b4 100644
--- a/fs/lockd/netns.h
+++ b/fs/lockd/netns.h
@@ -12,6 +12,10 @@ struct lockd_net {
         struct delayed_work grace_period_end;
         struct lock_manager lockd_manager;
         struct list_head grace_list;
+
+        spinlock_t nsm_clnt_lock;
+        unsigned int nsm_users;
+        struct rpc_clnt *nsm_clnt;
 };
 
 extern int lockd_net_id;
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 31a63f87b806..7e355870d519 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -596,6 +596,7 @@ static int lockd_init_net(struct net *net)
 
         INIT_DELAYED_WORK(&ln->grace_period_end, grace_ender);
         INIT_LIST_HEAD(&ln->grace_list);
+        spin_lock_init(&ln->nsm_clnt_lock);
         return 0;
 }