Merge git://git.linux-nfs.org/pub/linux/nfs-2.6

* git://git.linux-nfs.org/pub/linux/nfs-2.6: (103 commits)
  SUNRPC,RPCSEC_GSS: spkm3--fix config dependencies
  SUNRPC,RPCSEC_GSS: spkm3: import contexts using NID_cast5_cbc
  LOCKD: Make nlmsvc_traverse_shares return void
  LOCKD: nlmsvc_traverse_blocks return is unused
  SUNRPC,RPCSEC_GSS: fix krb5 sequence numbers.
  NFSv4: Dont list system.nfs4_acl for filesystems that don't support it.
  SUNRPC,RPCSEC_GSS: remove unnecessary kmalloc of a checksum
  SUNRPC: Ensure rpc_call_async() always calls tk_ops->rpc_release()
  SUNRPC: Fix memory barriers for req->rq_received
  NFS: Fix a race in nfs_sync_inode()
  NFS: Clean up nfs_flush_list()
  NFS: Fix a race with PG_private and nfs_release_page()
  NFSv4: Ensure the callback daemon flushes signals
  SUNRPC: Fix a 'Busy inodes' error in rpc_pipefs
  NFS, NLM: Allow blocking locks to respect signals
  NFS: Make nfs_fhget() return appropriate error values
  NFSv4: Fix an oops in nfs4_fill_super
  lockd: blocks should hold a reference to the nlm_file
  NFSv4: SETCLIENTID_CONFIRM should handle NFS4ERR_DELAY/NFS4ERR_RESOURCE
  NFSv4: Send the delegation stateid for SETATTR calls
  ...

15 files changed, 330 insertions, 69 deletions

diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
index 8d6f1a1..55163af 100644
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -64,14 +64,26 @@ rpcauth_create(rpc_authflavor_t pseudoflavor, struct rpc_clnt *clnt)
 	struct rpc_authops	*ops;
 	u32			flavor = pseudoflavor_to_flavor(pseudoflavor);
 
-	if (flavor >= RPC_AUTH_MAXFLAVOR || !(ops = auth_flavors[flavor]))
-		return ERR_PTR(-EINVAL);
+	auth = ERR_PTR(-EINVAL);
+	if (flavor >= RPC_AUTH_MAXFLAVOR)
+		goto out;
+
+	/* FIXME - auth_flavors[] really needs an rw lock,
+	 * and module refcounting. */
+#ifdef CONFIG_KMOD
+	if ((ops = auth_flavors[flavor]) == NULL)
+		request_module("rpc-auth-%u", flavor);
+#endif
+	if ((ops = auth_flavors[flavor]) == NULL)
+		goto out;
 	auth = ops->create(clnt, pseudoflavor);
 	if (IS_ERR(auth))
 		return auth;
 	if (clnt->cl_auth)
 		rpcauth_destroy(clnt->cl_auth);
 	clnt->cl_auth = auth;
+
+out:
 	return auth;
 }
 
diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index bb46efd..900ef31 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -721,6 +721,8 @@ gss_destroy(struct rpc_auth *auth)
 
 	gss_auth = container_of(auth, struct gss_auth, rpc_auth);
 	rpc_unlink(gss_auth->path);
+	dput(gss_auth->dentry);
+	gss_auth->dentry = NULL;
 	gss_mech_put(gss_auth->mech);
 
 	rpcauth_free_credcache(auth);
diff --git a/net/sunrpc/auth_gss/gss_krb5_seal.c b/net/sunrpc/auth_gss/gss_krb5_seal.c
index d0dfdfd..f433112 100644
--- a/net/sunrpc/auth_gss/gss_krb5_seal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_seal.c
@@ -70,15 +70,19 @@
 # define RPCDBG_FACILITY        RPCDBG_AUTH
 #endif
 
+spinlock_t krb5_seq_lock = SPIN_LOCK_UNLOCKED;
+
 u32
 gss_get_mic_kerberos(struct gss_ctx *gss_ctx, struct xdr_buf *text,
 		struct xdr_netobj *token)
 {
 	struct krb5_ctx		*ctx = gss_ctx->internal_ctx_id;
 	s32			checksum_type;
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	unsigned char		*ptr, *krb5_hdr, *msg_start;
 	s32			now;
+	u32			seq_send;
 
 	dprintk("RPC:     gss_krb5_seal\n");
 
@@ -133,16 +137,15 @@ gss_get_mic_kerberos(struct gss_ctx *gss_ctx, struct xdr_buf *text,
 		BUG();
 	}
 
-	kfree(md5cksum.data);
+	spin_lock(&krb5_seq_lock);
+	seq_send = ctx->seq_send++;
+	spin_unlock(&krb5_seq_lock);
 
 	if ((krb5_make_seq_num(ctx->seq, ctx->initiate ? 0 : 0xff,
-			       ctx->seq_send, krb5_hdr + 16, krb5_hdr + 8)))
+			       seq_send, krb5_hdr + 16, krb5_hdr + 8)))
 		goto out_err;
 
-	ctx->seq_send++;
-
 	return ((ctx->endtime < now) ? GSS_S_CONTEXT_EXPIRED : GSS_S_COMPLETE);
 out_err:
-	kfree(md5cksum.data);
 	return GSS_S_FAILURE;
 }
diff --git a/net/sunrpc/auth_gss/gss_krb5_unseal.c b/net/sunrpc/auth_gss/gss_krb5_unseal.c
index db055fd..0828cf6 100644
--- a/net/sunrpc/auth_gss/gss_krb5_unseal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_unseal.c
@@ -79,7 +79,8 @@ gss_verify_mic_kerberos(struct gss_ctx *gss_ctx,
 	int			signalg;
 	int			sealalg;
 	s32			checksum_type;
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	s32			now;
 	int			direction;
 	s32			seqnum;
@@ -176,6 +177,5 @@ gss_verify_mic_kerberos(struct gss_ctx *gss_ctx,
 
 	ret = GSS_S_COMPLETE;
 out:
-	kfree(md5cksum.data);
 	return ret;
 }
diff --git a/net/sunrpc/auth_gss/gss_krb5_wrap.c b/net/sunrpc/auth_gss/gss_krb5_wrap.c
index af777cf..89d1f3e 100644
--- a/net/sunrpc/auth_gss/gss_krb5_wrap.c
+++ b/net/sunrpc/auth_gss/gss_krb5_wrap.c
@@ -121,12 +121,14 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
 {
 	struct krb5_ctx		*kctx = ctx->internal_ctx_id;
 	s32			checksum_type;
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	int			blocksize = 0, plainlen;
 	unsigned char		*ptr, *krb5_hdr, *msg_start;
 	s32			now;
 	int			headlen;
 	struct page		**tmp_pages;
+	u32			seq_send;
 
 	dprintk("RPC:     gss_wrap_kerberos\n");
 
@@ -205,23 +207,22 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
 		BUG();
 	}
 
-	kfree(md5cksum.data);
+	spin_lock(&krb5_seq_lock);
+	seq_send = kctx->seq_send++;
+	spin_unlock(&krb5_seq_lock);
 
 	/* XXX would probably be more efficient to compute checksum
 	 * and encrypt at the same time: */
 	if ((krb5_make_seq_num(kctx->seq, kctx->initiate ? 0 : 0xff,
-			       kctx->seq_send, krb5_hdr + 16, krb5_hdr + 8)))
+			       seq_send, krb5_hdr + 16, krb5_hdr + 8)))
 		goto out_err;
 
 	if (gss_encrypt_xdr_buf(kctx->enc, buf, offset + headlen - blocksize,
 									pages))
 		goto out_err;
 
-	kctx->seq_send++;
-
 	return ((kctx->endtime < now) ? GSS_S_CONTEXT_EXPIRED : GSS_S_COMPLETE);
 out_err:
-	if (md5cksum.data) kfree(md5cksum.data);
 	return GSS_S_FAILURE;
 }
 
@@ -232,7 +233,8 @@ gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
 	int			signalg;
 	int			sealalg;
 	s32			checksum_type;
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	s32			now;
 	int			direction;
 	s32			seqnum;
@@ -358,6 +360,5 @@ gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
 
 	ret = GSS_S_COMPLETE;
 out:
-	if (md5cksum.data) kfree(md5cksum.data);
 	return ret;
 }
diff --git a/net/sunrpc/auth_gss/gss_spkm3_mech.c b/net/sunrpc/auth_gss/gss_spkm3_mech.c
index 5840080..5bf11cc 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_mech.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_mech.c
@@ -102,6 +102,12 @@ get_key(const void *p, const void *end, struct crypto_tfm **res, int *resalg)
 			alg_mode = CRYPTO_TFM_MODE_CBC;
 			setkey = 1;
 			break;
+		case NID_cast5_cbc:
+			/* XXXX here in name only, not used */
+			alg_name = "cast5";
+			alg_mode = CRYPTO_TFM_MODE_CBC;
+			setkey = 0; /* XXX will need to set to 1 */
+			break;
 		case NID_md5:
 			if (key.len == 0) {
 				dprintk("RPC: SPKM3 get_key: NID_md5 zero Key length\n");
diff --git a/net/sunrpc/auth_gss/gss_spkm3_seal.c b/net/sunrpc/auth_gss/gss_spkm3_seal.c
index 86fbf7c..18c7862 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_seal.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_seal.c
@@ -57,7 +57,8 @@ spkm3_make_token(struct spkm3_ctx *ctx,
 {
 	s32			checksum_type;
 	char			tokhdrbuf[25];
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	struct xdr_netobj	mic_hdr = {.len = 0, .data = tokhdrbuf};
 	int			tokenlen = 0;
 	unsigned char		*ptr;
@@ -115,13 +116,11 @@ spkm3_make_token(struct spkm3_ctx *ctx,
 		dprintk("RPC: gss_spkm3_seal: SPKM_WRAP_TOK not supported\n");
 		goto out_err;
 	}
-	kfree(md5cksum.data);
 
 	/* XXX need to implement sequence numbers, and ctx->expired */
 
 	return  GSS_S_COMPLETE;
 out_err:
-	kfree(md5cksum.data);
 	token->data = NULL;
 	token->len = 0;
 	return GSS_S_FAILURE;
diff --git a/net/sunrpc/auth_gss/gss_spkm3_unseal.c b/net/sunrpc/auth_gss/gss_spkm3_unseal.c
index 96851b0..8537f58 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_unseal.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_unseal.c
@@ -56,7 +56,8 @@ spkm3_read_token(struct spkm3_ctx *ctx,
 {
 	s32			code;
 	struct xdr_netobj	wire_cksum = {.len =0, .data = NULL};
-	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	char			cksumdata[16];
+	struct xdr_netobj	md5cksum = {.len = 0, .data = cksumdata};
 	unsigned char		*ptr = (unsigned char *)read_token->data;
 	unsigned char           *cksum;
 	int			bodysize, md5elen;
@@ -120,7 +121,6 @@ spkm3_read_token(struct spkm3_ctx *ctx,
 	/* XXX: need to add expiration and sequencing */
 	ret = GSS_S_COMPLETE;
 out:
-	kfree(md5cksum.data);
 	kfree(wire_cksum.data);
 	return ret;
 }
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index d784797..aa8965e 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -28,12 +28,11 @@
 #include <linux/mm.h>
 #include <linux/slab.h>
 #include <linux/utsname.h>
+#include <linux/workqueue.h>
 
 #include <linux/sunrpc/clnt.h>
-#include <linux/workqueue.h>
 #include <linux/sunrpc/rpc_pipe_fs.h>
-
-#include <linux/nfs.h>
+#include <linux/sunrpc/metrics.h>
 
 
 #define RPC_SLACK_SPACE		(1024)	/* total overkill */
@@ -71,8 +70,15 @@ rpc_setup_pipedir(struct rpc_clnt *clnt, char *dir_name)
 	static uint32_t clntid;
 	int error;
 
+	clnt->cl_vfsmnt = ERR_PTR(-ENOENT);
+	clnt->cl_dentry = ERR_PTR(-ENOENT);
 	if (dir_name == NULL)
 		return 0;
+
+	clnt->cl_vfsmnt = rpc_get_mount();
+	if (IS_ERR(clnt->cl_vfsmnt))
+		return PTR_ERR(clnt->cl_vfsmnt);
+
 	for (;;) {
 		snprintf(clnt->cl_pathname, sizeof(clnt->cl_pathname),
 				"%s/clnt%x", dir_name,
@@ -85,6 +91,7 @@ rpc_setup_pipedir(struct rpc_clnt *clnt, char *dir_name)
 		if (error != -EEXIST) {
 			printk(KERN_INFO "RPC: Couldn't create pipefs entry %s, error %d\n",
 					clnt->cl_pathname, error);
+			rpc_put_mount();
 			return error;
 		}
 	}
@@ -147,6 +154,7 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname,
 	clnt->cl_vers     = version->number;
 	clnt->cl_prot     = xprt->prot;
 	clnt->cl_stats    = program->stats;
+	clnt->cl_metrics  = rpc_alloc_iostats(clnt);
 	rpc_init_wait_queue(&clnt->cl_pmap_default.pm_bindwait, "bindwait");
 
 	if (!clnt->cl_port)
@@ -175,7 +183,11 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname,
 	return clnt;
 
 out_no_auth:
-	rpc_rmdir(clnt->cl_pathname);
+	if (!IS_ERR(clnt->cl_dentry)) {
+		rpc_rmdir(clnt->cl_pathname);
+		dput(clnt->cl_dentry);
+		rpc_put_mount();
+	}
 out_no_path:
 	if (clnt->cl_server != clnt->cl_inline_name)
 		kfree(clnt->cl_server);
@@ -240,11 +252,15 @@ rpc_clone_client(struct rpc_clnt *clnt)
 	new->cl_autobind = 0;
 	new->cl_oneshot = 0;
 	new->cl_dead = 0;
+	if (!IS_ERR(new->cl_dentry)) {
+		dget(new->cl_dentry);
+		rpc_get_mount();
+	}
 	rpc_init_rtt(&new->cl_rtt_default, clnt->cl_xprt->timeout.to_initval);
 	if (new->cl_auth)
 		atomic_inc(&new->cl_auth->au_count);
 	new->cl_pmap		= &new->cl_pmap_default;
-	rpc_init_wait_queue(&new->cl_pmap_default.pm_bindwait, "bindwait");
+	new->cl_metrics         = rpc_alloc_iostats(clnt);
 	return new;
 out_no_clnt:
 	printk(KERN_INFO "RPC: out of memory in %s\n", __FUNCTION__);
@@ -314,6 +330,12 @@ rpc_destroy_client(struct rpc_clnt *clnt)
 	if (clnt->cl_server != clnt->cl_inline_name)
 		kfree(clnt->cl_server);
 out_free:
+	rpc_free_iostats(clnt->cl_metrics);
+	clnt->cl_metrics = NULL;
+	if (!IS_ERR(clnt->cl_dentry)) {
+		dput(clnt->cl_dentry);
+		rpc_put_mount();
+	}
 	kfree(clnt);
 	return 0;
 }
@@ -473,15 +495,16 @@ rpc_call_async(struct rpc_clnt *clnt, struct rpc_message *msg, int flags,
 	int		status;
 
 	/* If this client is slain all further I/O fails */
+	status = -EIO;
 	if (clnt->cl_dead) 
-		return -EIO;
+		goto out_release;
 
 	flags |= RPC_TASK_ASYNC;
 
 	/* Create/initialize a new RPC task */
 	status = -ENOMEM;
 	if (!(task = rpc_new_task(clnt, flags, tk_ops, data)))
-		goto out;
+		goto out_release;
 
 	/* Mask signals on GSS_AUTH upcalls */
 	rpc_task_sigmask(task, &oldset);		
@@ -496,7 +519,10 @@ rpc_call_async(struct rpc_clnt *clnt, struct rpc_message *msg, int flags,
 		rpc_release_task(task);
 
 	rpc_restore_sigmask(&oldset);		
-out:
+	return status;
+out_release:
+	if (tk_ops->rpc_release != NULL)
+		tk_ops->rpc_release(data);
 	return status;
 }
 
@@ -993,6 +1019,8 @@ call_timeout(struct rpc_task *task)
 	}
 
 	dprintk("RPC: %4d call_timeout (major)\n", task->tk_pid);
+	task->tk_timeouts++;
+
 	if (RPC_IS_SOFT(task)) {
 		printk(KERN_NOTICE "%s: server %s not responding, timed out\n",
 				clnt->cl_protname, clnt->cl_server);
@@ -1045,6 +1073,11 @@ call_decode(struct rpc_task *task)
 		return;
 	}
 
+	/*
+	 * Ensure that we see all writes made by xprt_complete_rqst()
+	 * before it changed req->rq_received.
+	 */
+	smp_rmb();
 	req->rq_rcv_buf.len = req->rq_private_buf.len;
 
 	/* Check that the softirq receive buffer is valid */
@@ -1194,8 +1227,8 @@ call_verify(struct rpc_task *task)
 			task->tk_action = call_bind;
 			goto out_retry;
 		case RPC_AUTH_TOOWEAK:
-			printk(KERN_NOTICE "call_verify: server requires stronger "
-			       "authentication.\n");
+			printk(KERN_NOTICE "call_verify: server %s requires stronger "
+			       "authentication.\n", task->tk_client->cl_server);
 			break;
 		default:
 			printk(KERN_WARNING "call_verify: unknown auth error: %x\n", n);
diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c
index 8139ce6..d25b054 100644
--- a/net/sunrpc/pmap_clnt.c
+++ b/net/sunrpc/pmap_clnt.c
@@ -82,6 +82,7 @@ rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt)
 	rpc_call_setup(child, &msg, 0);
 
 	/* ... and run the child task */
+	task->tk_xprt->stat.bind_count++;
 	rpc_run_child(task, child, pmap_getport_done);
 	return;
 
@@ -103,6 +104,11 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot)
 		.pm_prot	= prot,
 		.pm_port	= 0
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &pmap_procedures[PMAP_GETPORT],
+		.rpc_argp	= &map,
+		.rpc_resp	= &map.pm_port,
+	};
 	struct rpc_clnt	*pmap_clnt;
 	char		hostname[32];
 	int		status;
@@ -116,7 +122,7 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot)
 		return PTR_ERR(pmap_clnt);
 
 	/* Setup the call info struct */
-	status = rpc_call(pmap_clnt, PMAP_GETPORT, &map, &map.pm_port, 0);
+	status = rpc_call_sync(pmap_clnt, &msg, 0);
 
 	if (status >= 0) {
 		if (map.pm_port != 0)
@@ -161,16 +167,27 @@ pmap_getport_done(struct rpc_task *task)
 int
 rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
 {
-	struct sockaddr_in	sin;
-	struct rpc_portmap	map;
+	struct sockaddr_in	sin = {
+		.sin_family	= AF_INET,
+		.sin_addr.s_addr = htonl(INADDR_LOOPBACK),
+	};
+	struct rpc_portmap	map = {
+		.pm_prog	= prog,
+		.pm_vers	= vers,
+		.pm_prot	= prot,
+		.pm_port	= port,
+	};
+	struct rpc_message msg = {
+		.rpc_proc	= &pmap_procedures[port ? PMAP_SET : PMAP_UNSET],
+		.rpc_argp	= &map,
+		.rpc_resp	= okay,
+	};
 	struct rpc_clnt		*pmap_clnt;
 	int error = 0;
 
 	dprintk("RPC: registering (%d, %d, %d, %d) with portmapper.\n",
 			prog, vers, prot, port);
 
-	sin.sin_family = AF_INET;
-	sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
 	pmap_clnt = pmap_create("localhost", &sin, IPPROTO_UDP, 1);
 	if (IS_ERR(pmap_clnt)) {
 		error = PTR_ERR(pmap_clnt);
@@ -178,13 +195,7 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
 		return error;
 	}
 
-	map.pm_prog = prog;
-	map.pm_vers = vers;
-	map.pm_prot = prot;
-	map.pm_port = port;
-
-	error = rpc_call(pmap_clnt, port? PMAP_SET : PMAP_UNSET,
-					&map, okay, 0);
+	error = rpc_call_sync(pmap_clnt, &msg, 0);
 
 	if (error < 0) {
 		printk(KERN_WARNING
@@ -260,6 +271,8 @@ static struct rpc_procinfo	pmap_procedures[] = {
 	  .p_decode		= (kxdrproc_t) xdr_decode_bool,
 	  .p_bufsiz		= 4,
 	  .p_count		= 1,
+	  .p_statidx		= PMAP_SET,
+	  .p_name		= "SET",
 	},
 [PMAP_UNSET] = {
 	  .p_proc		= PMAP_UNSET,
@@ -267,6 +280,8 @@ static struct rpc_procinfo	pmap_procedures[] = {
 	  .p_decode		= (kxdrproc_t) xdr_decode_bool,
 	  .p_bufsiz		= 4,
 	  .p_count		= 1,
+	  .p_statidx		= PMAP_UNSET,
+	  .p_name		= "UNSET",
 	},
 [PMAP_GETPORT] = {
 	  .p_proc		= PMAP_GETPORT,
@@ -274,6 +289,8 @@ static struct rpc_procinfo	pmap_procedures[] = {
 	  .p_decode		= (kxdrproc_t) xdr_decode_port,
 	  .p_bufsiz		= 4,
 	  .p_count		= 1,
+	  .p_statidx		= PMAP_GETPORT,
+	  .p_name		= "GETPORT",
 	},
 };
 
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index ad9d9fc..aa4158b 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -91,7 +91,8 @@ rpc_queue_upcall(struct inode *inode, struct rpc_pipe_msg *msg)
 		res = 0;
 	} else if (rpci->flags & RPC_PIPE_WAIT_FOR_OPEN) {
 		if (list_empty(&rpci->pipe))
-			schedule_delayed_work(&rpci->queue_timeout,
+			queue_delayed_work(rpciod_workqueue,
+					&rpci->queue_timeout,
 					RPC_UPCALL_TIMEOUT);
 		list_add_tail(&msg->list, &rpci->pipe);
 		rpci->pipelen += msg->len;
@@ -132,7 +133,7 @@ rpc_close_pipes(struct inode *inode)
 		if (ops->release_pipe)
 			ops->release_pipe(inode);
 		cancel_delayed_work(&rpci->queue_timeout);
-		flush_scheduled_work();
+		flush_workqueue(rpciod_workqueue);
 	}
 	rpc_inode_setowner(inode, NULL);
 	mutex_unlock(&inode->i_mutex);
@@ -434,14 +435,17 @@ static struct rpc_filelist authfiles[] = {
 	},
 };
 
-static int
-rpc_get_mount(void)
+struct vfsmount *rpc_get_mount(void)
 {
-	return simple_pin_fs("rpc_pipefs", &rpc_mount, &rpc_mount_count);
+	int err;
+
+	err = simple_pin_fs("rpc_pipefs", &rpc_mount, &rpc_mount_count);
+	if (err != 0)
+		return ERR_PTR(err);
+	return rpc_mount;
 }
 
-static void
-rpc_put_mount(void)
+void rpc_put_mount(void)
 {
 	simple_release_fs(&rpc_mount, &rpc_mount_count);
 }
@@ -451,12 +455,13 @@ rpc_lookup_parent(char *path, struct nameidata *nd)
 {
 	if (path[0] == '\0')
 		return -ENOENT;
-	if (rpc_get_mount()) {
+	nd->mnt = rpc_get_mount();
+	if (IS_ERR(nd->mnt)) {
 		printk(KERN_WARNING "%s: %s failed to mount "
 			       "pseudofilesystem \n", __FILE__, __FUNCTION__);
-		return -ENODEV;
+		return PTR_ERR(nd->mnt);
 	}
-	nd->mnt = mntget(rpc_mount);
+	mntget(nd->mnt);
 	nd->dentry = dget(rpc_mount->mnt_root);
 	nd->last_type = LAST_ROOT;
 	nd->flags = LOOKUP_PARENT;
@@ -593,7 +598,6 @@ __rpc_mkdir(struct inode *dir, struct dentry *dentry)
 	d_instantiate(dentry, inode);
 	dir->i_nlink++;
 	inode_dir_notify(dir, DN_CREATE);
-	rpc_get_mount();
 	return 0;
 out_err:
 	printk(KERN_WARNING "%s: %s failed to allocate inode for dentry %s\n",
@@ -614,7 +618,6 @@ __rpc_rmdir(struct inode *dir, struct dentry *dentry)
 	if (!error) {
 		inode_dir_notify(dir, DN_DELETE);
 		d_drop(dentry);
-		rpc_put_mount();
 	}
 	return 0;
 }
@@ -668,7 +671,7 @@ rpc_mkdir(char *path, struct rpc_clnt *rpc_client)
 out:
 	mutex_unlock(&dir->i_mutex);
 	rpc_release_path(&nd);
-	return dentry;
+	return dget(dentry);
 err_depopulate:
 	rpc_depopulate(dentry);
 	__rpc_rmdir(dir, dentry);
@@ -732,7 +735,7 @@ rpc_mkpipe(char *path, void *private, struct rpc_pipe_ops *ops, int flags)
 out:
 	mutex_unlock(&dir->i_mutex);
 	rpc_release_path(&nd);
-	return dentry;
+	return dget(dentry);
 err_dput:
 	dput(dentry);
 	dentry = ERR_PTR(-ENOMEM);
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index dff0779..b9969b9 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -65,7 +65,7 @@ static LIST_HEAD(all_tasks);
  */
 static DEFINE_MUTEX(rpciod_mutex);
 static unsigned int		rpciod_users;
-static struct workqueue_struct *rpciod_workqueue;
+struct workqueue_struct *rpciod_workqueue;
 
 /*
  * Spinlock for other critical sections of code.
@@ -182,6 +182,7 @@ static void __rpc_add_wait_queue(struct rpc_wait_queue *queue, struct rpc_task *
 	else
 		list_add_tail(&task->u.tk_wait.list, &queue->tasks[0]);
 	task->u.tk_wait.rpc_waitq = queue;
+	queue->qlen++;
 	rpc_set_queued(task);
 
 	dprintk("RPC: %4d added to queue %p \"%s\"\n",
@@ -216,6 +217,7 @@ static void __rpc_remove_wait_queue(struct rpc_task *task)
 		__rpc_remove_wait_queue_priority(task);
 	else
 		list_del(&task->u.tk_wait.list);
+	queue->qlen--;
 	dprintk("RPC: %4d removed from queue %p \"%s\"\n",
 				task->tk_pid, queue, rpc_qname(queue));
 }
@@ -816,6 +818,9 @@ void rpc_init_task(struct rpc_task *task, struct rpc_clnt *clnt, int flags, cons
 
 	BUG_ON(task->tk_ops == NULL);
 
+	/* starting timestamp */
+	task->tk_start = jiffies;
+
 	dprintk("RPC: %4d new task procpid %d\n", task->tk_pid,
 				current->pid);
 }
@@ -917,8 +922,11 @@ struct rpc_task *rpc_run_task(struct rpc_clnt *clnt, int flags,
 {
 	struct rpc_task *task;
 	task = rpc_new_task(clnt, flags, ops, data);
-	if (task == NULL)
+	if (task == NULL) {
+		if (ops->rpc_release != NULL)
+			ops->rpc_release(data);
 		return ERR_PTR(-ENOMEM);
+	}
 	atomic_inc(&task->tk_count);
 	rpc_execute(task);
 	return task;
diff --git a/net/sunrpc/stats.c b/net/sunrpc/stats.c
index 4979f22..790941e 100644
--- a/net/sunrpc/stats.c
+++ b/net/sunrpc/stats.c
@@ -21,6 +21,7 @@
 #include <linux/seq_file.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/svcsock.h>
+#include <linux/sunrpc/metrics.h>
 
 #define RPCDBG_FACILITY	RPCDBG_MISC
 
@@ -106,6 +107,120 @@ void svc_seq_show(struct seq_file *seq, const struct svc_stat *statp) {
 	}
 }
 
+/**
+ * rpc_alloc_iostats - allocate an rpc_iostats structure
+ * @clnt: RPC program, version, and xprt
+ *
+ */
+struct rpc_iostats *rpc_alloc_iostats(struct rpc_clnt *clnt)
+{
+	unsigned int ops = clnt->cl_maxproc;
+	size_t size = ops * sizeof(struct rpc_iostats);
+	struct rpc_iostats *new;
+
+	new = kmalloc(size, GFP_KERNEL);
+	if (new)
+		memset(new, 0 , size);
+	return new;
+}
+EXPORT_SYMBOL(rpc_alloc_iostats);
+
+/**
+ * rpc_free_iostats - release an rpc_iostats structure
+ * @stats: doomed rpc_iostats structure
+ *
+ */
+void rpc_free_iostats(struct rpc_iostats *stats)
+{
+	kfree(stats);
+}
+EXPORT_SYMBOL(rpc_free_iostats);
+
+/**
+ * rpc_count_iostats - tally up per-task stats
+ * @task: completed rpc_task
+ *
+ * Relies on the caller for serialization.
+ */
+void rpc_count_iostats(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_iostats *stats = task->tk_client->cl_metrics;
+	struct rpc_iostats *op_metrics;
+	long rtt, execute, queue;
+
+	if (!stats || !req)
+		return;
+	op_metrics = &stats[task->tk_msg.rpc_proc->p_statidx];
+
+	op_metrics->om_ops++;
+	op_metrics->om_ntrans += req->rq_ntrans;
+	op_metrics->om_timeouts += task->tk_timeouts;
+
+	op_metrics->om_bytes_sent += task->tk_bytes_sent;
+	op_metrics->om_bytes_recv += req->rq_received;
+
+	queue = (long)req->rq_xtime - task->tk_start;
+	if (queue < 0)
+		queue = -queue;
+	op_metrics->om_queue += queue;
+
+	rtt = task->tk_rtt;
+	if (rtt < 0)
+		rtt = -rtt;
+	op_metrics->om_rtt += rtt;
+
+	execute = (long)jiffies - task->tk_start;
+	if (execute < 0)
+		execute = -execute;
+	op_metrics->om_execute += execute;
+}
+
+void _print_name(struct seq_file *seq, unsigned int op, struct rpc_procinfo *procs)
+{
+	if (procs[op].p_name)
+		seq_printf(seq, "\t%12s: ", procs[op].p_name);
+	else if (op == 0)
+		seq_printf(seq, "\t        NULL: ");
+	else
+		seq_printf(seq, "\t%12u: ", op);
+}
+
+#define MILLISECS_PER_JIFFY	(1000 / HZ)
+
+void rpc_print_iostats(struct seq_file *seq, struct rpc_clnt *clnt)
+{
+	struct rpc_iostats *stats = clnt->cl_metrics;
+	struct rpc_xprt *xprt = clnt->cl_xprt;
+	unsigned int op, maxproc = clnt->cl_maxproc;
+
+	if (!stats)
+		return;
+
+	seq_printf(seq, "\tRPC iostats version: %s  ", RPC_IOSTATS_VERS);
+	seq_printf(seq, "p/v: %u/%u (%s)\n",
+			clnt->cl_prog, clnt->cl_vers, clnt->cl_protname);
+
+	if (xprt)
+		xprt->ops->print_stats(xprt, seq);
+
+	seq_printf(seq, "\tper-op statistics\n");
+	for (op = 0; op < maxproc; op++) {
+		struct rpc_iostats *metrics = &stats[op];
+		_print_name(seq, op, clnt->cl_procinfo);
+		seq_printf(seq, "%lu %lu %lu %Lu %Lu %Lu %Lu %Lu\n",
+				metrics->om_ops,
+				metrics->om_ntrans,
+				metrics->om_timeouts,
+				metrics->om_bytes_sent,
+				metrics->om_bytes_recv,
+				metrics->om_queue * MILLISECS_PER_JIFFY,
+				metrics->om_rtt * MILLISECS_PER_JIFFY,
+				metrics->om_execute * MILLISECS_PER_JIFFY);
+	}
+}
+EXPORT_SYMBOL(rpc_print_iostats);
+
 /*
  * Register/unregister RPC proc files
  */
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 8ff2c8a..4dd5b3c 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -44,13 +44,13 @@
 #include <linux/random.h>
 
 #include <linux/sunrpc/clnt.h>
+#include <linux/sunrpc/metrics.h>
 
 /*
  * Local variables
  */
 
 #ifdef RPC_DEBUG
-# undef  RPC_DEBUG_DATA
 # define RPCDBG_FACILITY	RPCDBG_XPRT
 #endif
 
@@ -548,6 +548,7 @@ void xprt_connect(struct rpc_task *task)
 
 		task->tk_timeout = xprt->connect_timeout;
 		rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
+		xprt->stat.connect_start = jiffies;
 		xprt->ops->connect(task);
 	}
 	return;
@@ -558,6 +559,8 @@ static void xprt_connect_status(struct rpc_task *task)
 	struct rpc_xprt	*xprt = task->tk_xprt;
 
 	if (task->tk_status >= 0) {
+		xprt->stat.connect_count++;
+		xprt->stat.connect_time += (long)jiffies - xprt->stat.connect_start;
 		dprintk("RPC: %4d xprt_connect_status: connection established\n",
 				task->tk_pid);
 		return;
@@ -601,16 +604,14 @@ static void xprt_connect_status(struct rpc_task *task)
 struct rpc_rqst *xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
 {
 	struct list_head *pos;
-	struct rpc_rqst	*req = NULL;
 
 	list_for_each(pos, &xprt->recv) {
 		struct rpc_rqst *entry = list_entry(pos, struct rpc_rqst, rq_list);
-		if (entry->rq_xid == xid) {
-			req = entry;
-			break;
-		}
+		if (entry->rq_xid == xid)
+			return entry;
 	}
-	return req;
+	xprt->stat.bad_xids++;
+	return NULL;
 }
 
 /**
@@ -646,7 +647,12 @@ void xprt_complete_rqst(struct rpc_task *task, int copied)
 	dprintk("RPC: %5u xid %08x complete (%d bytes received)\n",
 			task->tk_pid, ntohl(req->rq_xid), copied);
 
+	task->tk_xprt->stat.recvs++;
+	task->tk_rtt = (long)jiffies - req->rq_xtime;
+
 	list_del_init(&req->rq_list);
+	/* Ensure all writes are done before we update req->rq_received */
+	smp_wmb();
 	req->rq_received = req->rq_private_buf.len = copied;
 	rpc_wake_up_task(task);
 }
@@ -723,7 +729,6 @@ void xprt_transmit(struct rpc_task *task)
 
 	dprintk("RPC: %4d xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
 
-	smp_rmb();
 	if (!req->rq_received) {
 		if (list_empty(&req->rq_list)) {
 			spin_lock_bh(&xprt->transport_lock);
@@ -744,12 +749,19 @@ void xprt_transmit(struct rpc_task *task)
 	if (status == 0) {
 		dprintk("RPC: %4d xmit complete\n", task->tk_pid);
 		spin_lock_bh(&xprt->transport_lock);
+
 		xprt->ops->set_retrans_timeout(task);
+
+		xprt->stat.sends++;
+		xprt->stat.req_u += xprt->stat.sends - xprt->stat.recvs;
+		xprt->stat.bklog_u += xprt->backlog.qlen;
+
 		/* Don't race with disconnect */
 		if (!xprt_connected(xprt))
 			task->tk_status = -ENOTCONN;
 		else if (!req->rq_received)
 			rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
+
 		xprt->ops->release_xprt(xprt, task);
 		spin_unlock_bh(&xprt->transport_lock);
 		return;
@@ -848,6 +860,7 @@ void xprt_release(struct rpc_task *task)
 
 	if (!(req = task->tk_rqstp))
 		return;
+	rpc_count_iostats(task);
 	spin_lock_bh(&xprt->transport_lock);
 	xprt->ops->release_xprt(xprt, task);
 	if (xprt->ops->release_request)
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index c458f8d..4b4e7df 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -382,6 +382,7 @@ static int xs_tcp_send_request(struct rpc_task *task)
 		/* If we've sent the entire packet, immediately
 		 * reset the count of bytes sent. */
 		req->rq_bytes_sent += status;
+		task->tk_bytes_sent += status;
 		if (likely(req->rq_bytes_sent >= req->rq_slen)) {
 			req->rq_bytes_sent = 0;
 			return 0;
@@ -1114,6 +1115,8 @@ static void xs_tcp_connect_worker(void *args)
 	}
 
 	/* Tell the socket layer to start connecting... */
+	xprt->stat.connect_count++;
+	xprt->stat.connect_start = jiffies;
 	status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr,
 			sizeof(xprt->addr), O_NONBLOCK);
 	dprintk("RPC: %p  connect status %d connected %d sock state %d\n",
@@ -1177,6 +1180,50 @@ static void xs_connect(struct rpc_task *task)
 	}
 }
 
+/**
+ * xs_udp_print_stats - display UDP socket-specifc stats
+ * @xprt: rpc_xprt struct containing statistics
+ * @seq: output file
+ *
+ */
+static void xs_udp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
+{
+	seq_printf(seq, "\txprt:\tudp %u %lu %lu %lu %lu %Lu %Lu\n",
+			xprt->port,
+			xprt->stat.bind_count,
+			xprt->stat.sends,
+			xprt->stat.recvs,
+			xprt->stat.bad_xids,
+			xprt->stat.req_u,
+			xprt->stat.bklog_u);
+}
+
+/**
+ * xs_tcp_print_stats - display TCP socket-specifc stats
+ * @xprt: rpc_xprt struct containing statistics
+ * @seq: output file
+ *
+ */
+static void xs_tcp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
+{
+	long idle_time = 0;
+
+	if (xprt_connected(xprt))
+		idle_time = (long)(jiffies - xprt->last_used) / HZ;
+
+	seq_printf(seq, "\txprt:\ttcp %u %lu %lu %lu %ld %lu %lu %lu %Lu %Lu\n",
+			xprt->port,
+			xprt->stat.bind_count,
+			xprt->stat.connect_count,
+			xprt->stat.connect_time,
+			idle_time,
+			xprt->stat.sends,
+			xprt->stat.recvs,
+			xprt->stat.bad_xids,
+			xprt->stat.req_u,
+			xprt->stat.bklog_u);
+}
+
 static struct rpc_xprt_ops xs_udp_ops = {
 	.set_buffer_size	= xs_udp_set_buffer_size,
 	.reserve_xprt		= xprt_reserve_xprt_cong,
@@ -1191,6 +1238,7 @@ static struct rpc_xprt_ops xs_udp_ops = {
 	.release_request	= xprt_release_rqst_cong,
 	.close			= xs_close,
 	.destroy		= xs_destroy,
+	.print_stats		= xs_udp_print_stats,
 };
 
 static struct rpc_xprt_ops xs_tcp_ops = {
@@ -1204,6 +1252,7 @@ static struct rpc_xprt_ops xs_tcp_ops = {
 	.set_retrans_timeout	= xprt_set_retrans_timeout_def,
 	.close			= xs_close,
 	.destroy		= xs_destroy,
+	.print_stats		= xs_tcp_print_stats,
 };
 
 /**