Merge tag 'nfsd-5.2' of git://linux-nfs.org/~bfields/linux

Pull nfsd updates from Bruce Fields:
 "This consists mostly of nfsd container work:

  Scott Mayhew revived an old api that communicates with a userspace
  daemon to manage some on-disk state that's used to track clients
  across server reboots. We've been using a usermode_helper upcall for
  that, but it's tough to run those with the right namespaces, so a
  daemon is much friendlier to container use cases.

  Trond fixed nfsd's handling of user credentials in user namespaces. He
  also contributed patches that allow containers to support different
  sets of NFS protocol versions.

  The only remaining container bug I'm aware of is that the NFS reply
  cache is shared between all containers. If anyone's aware of other
  gaps in our container support, let me know.

  The rest of this is miscellaneous bugfixes"

* tag 'nfsd-5.2' of git://linux-nfs.org/~bfields/linux: (23 commits)
  nfsd: update callback done processing
  locks: move checks from locks_free_lock() to locks_release_private()
  nfsd: fh_drop_write in nfsd_unlink
  nfsd: allow fh_want_write to be called twice
  nfsd: knfsd must use the container user namespace
  SUNRPC: rsi_parse() should use the current user namespace
  SUNRPC: Fix the server AUTH_UNIX userspace mappings
  lockd: Pass the user cred from knfsd when starting the lockd server
  SUNRPC: Temporary sockets should inherit the cred from their parent
  SUNRPC: Cache the process user cred in the RPC server listener
  nfsd: Allow containers to set supported nfs versions
  nfsd: Add custom rpcbind callbacks for knfsd
  SUNRPC: Allow further customisation of RPC program registration
  SUNRPC: Clean up generic dispatcher code
  SUNRPC: Add a callback to initialise server requests
  SUNRPC/nfs: Fix return value for nfs4_callback_compound()
  nfsd: handle legacy client tracking records sent by nfsdcld
  nfsd: re-order client tracking method selection
  nfsd: keep a tally of RECLAIM_COMPLETE operations when using nfsdcld
  nfsd: un-deprecate nfsdcld
  ...

17 files changed, 781 insertions, 172 deletions

diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index 802993d8912f..baa01956a5b3 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -570,13 +570,13 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen)
 		err = get_int(&mesg, &an_int);
 		if (err)
 			goto out3;
-		exp.ex_anon_uid= make_kuid(&init_user_ns, an_int);
+		exp.ex_anon_uid= make_kuid(current_user_ns(), an_int);
 
 		/* anon gid */
 		err = get_int(&mesg, &an_int);
 		if (err)
 			goto out3;
-		exp.ex_anon_gid= make_kgid(&init_user_ns, an_int);
+		exp.ex_anon_gid= make_kgid(current_user_ns(), an_int);
 
 		/* fsid */
 		err = get_int(&mesg, &an_int);
@@ -1170,15 +1170,17 @@ static void show_secinfo(struct seq_file *m, struct svc_export *exp)
 static void exp_flags(struct seq_file *m, int flag, int fsid,
 		kuid_t anonu, kgid_t anong, struct nfsd4_fs_locations *fsloc)
 {
+	struct user_namespace *userns = m->file->f_cred->user_ns;
+
 	show_expflags(m, flag, NFSEXP_ALLFLAGS);
 	if (flag & NFSEXP_FSID)
 		seq_printf(m, ",fsid=%d", fsid);
-	if (!uid_eq(anonu, make_kuid(&init_user_ns, (uid_t)-2)) &&
-	    !uid_eq(anonu, make_kuid(&init_user_ns, 0x10000-2)))
-		seq_printf(m, ",anonuid=%u", from_kuid(&init_user_ns, anonu));
-	if (!gid_eq(anong, make_kgid(&init_user_ns, (gid_t)-2)) &&
-	    !gid_eq(anong, make_kgid(&init_user_ns, 0x10000-2)))
-		seq_printf(m, ",anongid=%u", from_kgid(&init_user_ns, anong));
+	if (!uid_eq(anonu, make_kuid(userns, (uid_t)-2)) &&
+	    !uid_eq(anonu, make_kuid(userns, 0x10000-2)))
+		seq_printf(m, ",anonuid=%u", from_kuid_munged(userns, anonu));
+	if (!gid_eq(anong, make_kgid(userns, (gid_t)-2)) &&
+	    !gid_eq(anong, make_kgid(userns, 0x10000-2)))
+		seq_printf(m, ",anongid=%u", from_kgid_munged(userns, anong));
 	if (fsloc && fsloc->locations_count > 0) {
 		char *loctype = (fsloc->migrated) ? "refer" : "replicas";
 		int i;
diff --git a/fs/nfsd/netns.h b/fs/nfsd/netns.h
index 32cb8c027483..789abc4dd1d2 100644
--- a/fs/nfsd/netns.h
+++ b/fs/nfsd/netns.h
@@ -104,6 +104,9 @@ struct nfsd_net {
 	time_t nfsd4_grace;
 	bool somebody_reclaimed;
 
+	bool track_reclaim_completes;
+	atomic_t nr_reclaim_complete;
+
 	bool nfsd_net_up;
 	bool lockd_up;
 
@@ -131,10 +134,18 @@ struct nfsd_net {
 	u32		s2s_cp_cl_id;
 	struct idr	s2s_cp_stateids;
 	spinlock_t	s2s_cp_lock;
+
+	/*
+	 * Version information
+	 */
+	bool *nfsd_versions;
+	bool *nfsd4_minorversions;
 };
 
 /* Simple check to find out if a given net was properly initialized */
 #define nfsd_netns_ready(nn) ((nn)->sessionid_hashtbl)
 
+extern void nfsd_netns_free_versions(struct nfsd_net *nn);
+
 extern unsigned int nfsd_net_id;
 #endif /* __NFSD_NETNS_H__ */
diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c
index 8d789124ed3c..fcf31822c74c 100644
--- a/fs/nfsd/nfs3xdr.c
+++ b/fs/nfsd/nfs3xdr.c
@@ -96,7 +96,7 @@ decode_filename(__be32 *p, char **namp, unsigned int *lenp)
 }
 
 static __be32 *
-decode_sattr3(__be32 *p, struct iattr *iap)
+decode_sattr3(__be32 *p, struct iattr *iap, struct user_namespace *userns)
 {
 	u32	tmp;
 
@@ -107,12 +107,12 @@ decode_sattr3(__be32 *p, struct iattr *iap)
 		iap->ia_mode = ntohl(*p++);
 	}
 	if (*p++) {
-		iap->ia_uid = make_kuid(&init_user_ns, ntohl(*p++));
+		iap->ia_uid = make_kuid(userns, ntohl(*p++));
 		if (uid_valid(iap->ia_uid))
 			iap->ia_valid |= ATTR_UID;
 	}
 	if (*p++) {
-		iap->ia_gid = make_kgid(&init_user_ns, ntohl(*p++));
+		iap->ia_gid = make_kgid(userns, ntohl(*p++));
 		if (gid_valid(iap->ia_gid))
 			iap->ia_valid |= ATTR_GID;
 	}
@@ -165,12 +165,13 @@ static __be32 *
 encode_fattr3(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp,
 	      struct kstat *stat)
 {
+	struct user_namespace *userns = nfsd_user_namespace(rqstp);
 	struct timespec ts;
 	*p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]);
 	*p++ = htonl((u32) (stat->mode & S_IALLUGO));
 	*p++ = htonl((u32) stat->nlink);
-	*p++ = htonl((u32) from_kuid(&init_user_ns, stat->uid));
-	*p++ = htonl((u32) from_kgid(&init_user_ns, stat->gid));
+	*p++ = htonl((u32) from_kuid_munged(userns, stat->uid));
+	*p++ = htonl((u32) from_kgid_munged(userns, stat->gid));
 	if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) {
 		p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN);
 	} else {
@@ -325,7 +326,7 @@ nfs3svc_decode_sattrargs(struct svc_rqst *rqstp, __be32 *p)
 	p = decode_fh(p, &args->fh);
 	if (!p)
 		return 0;
-	p = decode_sattr3(p, &args->attrs);
+	p = decode_sattr3(p, &args->attrs, nfsd_user_namespace(rqstp));
 
 	if ((args->check_guard = ntohl(*p++)) != 0) { 
 		struct timespec time; 
@@ -455,7 +456,7 @@ nfs3svc_decode_createargs(struct svc_rqst *rqstp, __be32 *p)
 	switch (args->createmode = ntohl(*p++)) {
 	case NFS3_CREATE_UNCHECKED:
 	case NFS3_CREATE_GUARDED:
-		p = decode_sattr3(p, &args->attrs);
+		p = decode_sattr3(p, &args->attrs, nfsd_user_namespace(rqstp));
 		break;
 	case NFS3_CREATE_EXCLUSIVE:
 		args->verf = p;
@@ -476,7 +477,7 @@ nfs3svc_decode_mkdirargs(struct svc_rqst *rqstp, __be32 *p)
 	if (!(p = decode_fh(p, &args->fh)) ||
 	    !(p = decode_filename(p, &args->name, &args->len)))
 		return 0;
-	p = decode_sattr3(p, &args->attrs);
+	p = decode_sattr3(p, &args->attrs, nfsd_user_namespace(rqstp));
 
 	return xdr_argsize_check(rqstp, p);
 }
@@ -491,7 +492,7 @@ nfs3svc_decode_symlinkargs(struct svc_rqst *rqstp, __be32 *p)
 	if (!(p = decode_fh(p, &args->ffh)) ||
 	    !(p = decode_filename(p, &args->fname, &args->flen)))
 		return 0;
-	p = decode_sattr3(p, &args->attrs);
+	p = decode_sattr3(p, &args->attrs, nfsd_user_namespace(rqstp));
 
 	args->tlen = ntohl(*p++);
 
@@ -519,7 +520,7 @@ nfs3svc_decode_mknodargs(struct svc_rqst *rqstp, __be32 *p)
 
 	if (args->ftype == NF3BLK  || args->ftype == NF3CHR
 	 || args->ftype == NF3SOCK || args->ftype == NF3FIFO)
-		p = decode_sattr3(p, &args->attrs);
+		p = decode_sattr3(p, &args->attrs, nfsd_user_namespace(rqstp));
 
 	if (args->ftype == NF3BLK || args->ftype == NF3CHR) {
 		args->major = ntohl(*p++);
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 9b93e7a9a26d..397eb7820929 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -1123,10 +1123,11 @@ static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
 		rpc_restart_call_prepare(task);
 		return;
 	case 1:
-		break;
-	case -1:
-		/* Network partition? */
-		nfsd4_mark_cb_down(clp, task->tk_status);
+		switch (task->tk_status) {
+		case -EIO:
+		case -ETIMEDOUT:
+			nfsd4_mark_cb_down(clp, task->tk_status);
+		}
 		break;
 	default:
 		BUG();
diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c
index bf137fec33ff..2961016097ac 100644
--- a/fs/nfsd/nfs4idmap.c
+++ b/fs/nfsd/nfs4idmap.c
@@ -634,7 +634,7 @@ nfsd_map_name_to_uid(struct svc_rqst *rqstp, const char *name, size_t namelen,
 		return nfserr_inval;
 
 	status = do_name_to_id(rqstp, IDMAP_TYPE_USER, name, namelen, &id);
-	*uid = make_kuid(&init_user_ns, id);
+	*uid = make_kuid(nfsd_user_namespace(rqstp), id);
 	if (!uid_valid(*uid))
 		status = nfserr_badowner;
 	return status;
@@ -651,7 +651,7 @@ nfsd_map_name_to_gid(struct svc_rqst *rqstp, const char *name, size_t namelen,
 		return nfserr_inval;
 
 	status = do_name_to_id(rqstp, IDMAP_TYPE_GROUP, name, namelen, &id);
-	*gid = make_kgid(&init_user_ns, id);
+	*gid = make_kgid(nfsd_user_namespace(rqstp), id);
 	if (!gid_valid(*gid))
 		status = nfserr_badowner;
 	return status;
@@ -660,13 +660,13 @@ nfsd_map_name_to_gid(struct svc_rqst *rqstp, const char *name, size_t namelen,
 __be32 nfsd4_encode_user(struct xdr_stream *xdr, struct svc_rqst *rqstp,
 			 kuid_t uid)
 {
-	u32 id = from_kuid(&init_user_ns, uid);
+	u32 id = from_kuid_munged(nfsd_user_namespace(rqstp), uid);
 	return encode_name_from_id(xdr, rqstp, IDMAP_TYPE_USER, id);
 }
 
 __be32 nfsd4_encode_group(struct xdr_stream *xdr, struct svc_rqst *rqstp,
 			  kgid_t gid)
 {
-	u32 id = from_kgid(&init_user_ns, gid);
+	u32 id = from_kgid_munged(nfsd_user_namespace(rqstp), gid);
 	return encode_name_from_id(xdr, rqstp, IDMAP_TYPE_GROUP, id);
 }
diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c
index 44517fb5c0de..a79e24b79095 100644
--- a/fs/nfsd/nfs4layouts.c
+++ b/fs/nfsd/nfs4layouts.c
@@ -693,7 +693,7 @@ nfsd4_cb_layout_done(struct nfsd4_callback *cb, struct rpc_task *task)
 			ops->fence_client(ls);
 		else
 			nfsd4_cb_layout_fail(ls);
-		return -1;
+		return 1;
 	case -NFS4ERR_NOMATCHING_LAYOUT:
 		trace_nfsd_layout_recall_done(&ls->ls_stid.sc_stateid);
 		task->tk_status = 0;
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 4680ad3bf55b..8beda999e134 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -1927,6 +1927,7 @@ nfsd4_proc_compound(struct svc_rqst *rqstp)
 	struct nfsd4_compound_state *cstate = &resp->cstate;
 	struct svc_fh *current_fh = &cstate->current_fh;
 	struct svc_fh *save_fh = &cstate->save_fh;
+	struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id);
 	__be32		status;
 
 	svcxdr_init_encode(rqstp, resp);
@@ -1949,7 +1950,7 @@ nfsd4_proc_compound(struct svc_rqst *rqstp)
 	 * According to RFC3010, this takes precedence over all other errors.
 	 */
 	status = nfserr_minor_vers_mismatch;
-	if (nfsd_minorversion(args->minorversion, NFSD_TEST) <= 0)
+	if (nfsd_minorversion(nn, args->minorversion, NFSD_TEST) <= 0)
 		goto out;
 	status = nfserr_resource;
 	if (args->opcnt > NFSD_MAX_OPS_PER_COMPOUND)
diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c
index 8c8563441208..87679557d0d6 100644
--- a/fs/nfsd/nfs4recover.c
+++ b/fs/nfsd/nfs4recover.c
@@ -169,12 +169,33 @@ legacy_recdir_name_error(struct nfs4_client *clp, int error)
 }
 
 static void
+__nfsd4_create_reclaim_record_grace(struct nfs4_client *clp,
+		const char *dname, int len, struct nfsd_net *nn)
+{
+	struct xdr_netobj name;
+	struct nfs4_client_reclaim *crp;
+
+	name.data = kmemdup(dname, len, GFP_KERNEL);
+	if (!name.data) {
+		dprintk("%s: failed to allocate memory for name.data!\n",
+			__func__);
+		return;
+	}
+	name.len = len;
+	crp = nfs4_client_to_reclaim(name, nn);
+	if (!crp) {
+		kfree(name.data);
+		return;
+	}
+	crp->cr_clp = clp;
+}
+
+static void
 nfsd4_create_clid_dir(struct nfs4_client *clp)
 {
 	const struct cred *original_cred;
 	char dname[HEXDIR_LEN];
 	struct dentry *dir, *dentry;
-	struct nfs4_client_reclaim *crp;
 	int status;
 	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
 
@@ -220,11 +241,9 @@ out_put:
 out_unlock:
 	inode_unlock(d_inode(dir));
 	if (status == 0) {
-		if (nn->in_grace) {
-			crp = nfs4_client_to_reclaim(dname, nn);
-			if (crp)
-				crp->cr_clp = clp;
-		}
+		if (nn->in_grace)
+			__nfsd4_create_reclaim_record_grace(clp, dname,
+					HEXDIR_LEN, nn);
 		vfs_fsync(nn->rec_file, 0);
 	} else {
 		printk(KERN_ERR "NFSD: failed to write recovery record"
@@ -345,10 +364,29 @@ out_unlock:
 }
 
 static void
+__nfsd4_remove_reclaim_record_grace(const char *dname, int len,
+		struct nfsd_net *nn)
+{
+	struct xdr_netobj name;
+	struct nfs4_client_reclaim *crp;
+
+	name.data = kmemdup(dname, len, GFP_KERNEL);
+	if (!name.data) {
+		dprintk("%s: failed to allocate memory for name.data!\n",
+			__func__);
+		return;
+	}
+	name.len = len;
+	crp = nfsd4_find_reclaim_client(name, nn);
+	kfree(name.data);
+	if (crp)
+		nfs4_remove_reclaim_record(crp, nn);
+}
+
+static void
 nfsd4_remove_clid_dir(struct nfs4_client *clp)
 {
 	const struct cred *original_cred;
-	struct nfs4_client_reclaim *crp;
 	char dname[HEXDIR_LEN];
 	int status;
 	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
@@ -373,12 +411,9 @@ nfsd4_remove_clid_dir(struct nfs4_client *clp)
 	nfs4_reset_creds(original_cred);
 	if (status == 0) {
 		vfs_fsync(nn->rec_file, 0);
-		if (nn->in_grace) {
-			/* remove reclaim record */
-			crp = nfsd4_find_reclaim_client(dname, nn);
-			if (crp)
-				nfs4_remove_reclaim_record(crp, nn);
-		}
+		if (nn->in_grace)
+			__nfsd4_remove_reclaim_record_grace(dname,
+					HEXDIR_LEN, nn);
 	}
 out_drop_write:
 	mnt_drop_write_file(nn->rec_file);
@@ -392,14 +427,31 @@ static int
 purge_old(struct dentry *parent, struct dentry *child, struct nfsd_net *nn)
 {
 	int status;
+	struct xdr_netobj name;
 
-	if (nfs4_has_reclaimed_state(child->d_name.name, nn))
+	if (child->d_name.len != HEXDIR_LEN - 1) {
+		printk("%s: illegal name %pd in recovery directory\n",
+				__func__, child);
+		/* Keep trying; maybe the others are OK: */
 		return 0;
+	}
+	name.data = kmemdup_nul(child->d_name.name, child->d_name.len, GFP_KERNEL);
+	if (!name.data) {
+		dprintk("%s: failed to allocate memory for name.data!\n",
+			__func__);
+		goto out;
+	}
+	name.len = HEXDIR_LEN;
+	if (nfs4_has_reclaimed_state(name, nn))
+		goto out_free;
 
 	status = vfs_rmdir(d_inode(parent), child);
 	if (status)
 		printk("failed to remove client recovery directory %pd\n",
 				child);
+out_free:
+	kfree(name.data);
+out:
 	/* Keep trying, success or failure: */
 	return 0;
 }
@@ -429,13 +481,24 @@ out:
 static int
 load_recdir(struct dentry *parent, struct dentry *child, struct nfsd_net *nn)
 {
+	struct xdr_netobj name;
+
 	if (child->d_name.len != HEXDIR_LEN - 1) {
-		printk("nfsd4: illegal name %pd in recovery directory\n",
-				child);
+		printk("%s: illegal name %pd in recovery directory\n",
+				__func__, child);
 		/* Keep trying; maybe the others are OK: */
 		return 0;
 	}
-	nfs4_client_to_reclaim(child->d_name.name, nn);
+	name.data = kmemdup_nul(child->d_name.name, child->d_name.len, GFP_KERNEL);
+	if (!name.data) {
+		dprintk("%s: failed to allocate memory for name.data!\n",
+			__func__);
+		goto out;
+	}
+	name.len = HEXDIR_LEN;
+	if (!nfs4_client_to_reclaim(name, nn))
+		kfree(name.data);
+out:
 	return 0;
 }
 
@@ -564,6 +627,7 @@ nfsd4_legacy_tracking_init(struct net *net)
 	status = nfsd4_load_reboot_recovery_data(net);
 	if (status)
 		goto err;
+	printk("NFSD: Using legacy client tracking operations.\n");
 	return 0;
 
 err:
@@ -615,6 +679,7 @@ nfsd4_check_legacy_client(struct nfs4_client *clp)
 	char dname[HEXDIR_LEN];
 	struct nfs4_client_reclaim *crp;
 	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
+	struct xdr_netobj name;
 
 	/* did we already find that this client is stable? */
 	if (test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags))
@@ -627,13 +692,22 @@ nfsd4_check_legacy_client(struct nfs4_client *clp)
 	}
 
 	/* look for it in the reclaim hashtable otherwise */
-	crp = nfsd4_find_reclaim_client(dname, nn);
+	name.data = kmemdup(dname, HEXDIR_LEN, GFP_KERNEL);
+	if (!name.data) {
+		dprintk("%s: failed to allocate memory for name.data!\n",
+			__func__);
+		goto out_enoent;
+	}
+	name.len = HEXDIR_LEN;
+	crp = nfsd4_find_reclaim_client(name, nn);
+	kfree(name.data);
 	if (crp) {
 		set_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags);
 		crp->cr_clp = clp;
 		return 0;
 	}
 
+out_enoent:
 	return -ENOENT;
 }
 
@@ -656,6 +730,7 @@ struct cld_net {
 	spinlock_t		 cn_lock;
 	struct list_head	 cn_list;
 	unsigned int		 cn_xid;
+	bool			 cn_has_legacy;
 };
 
 struct cld_upcall {
@@ -706,6 +781,40 @@ cld_pipe_upcall(struct rpc_pipe *pipe, struct cld_msg *cmsg)
 }
 
 static ssize_t
+__cld_pipe_inprogress_downcall(const struct cld_msg __user *cmsg,
+		struct nfsd_net *nn)
+{
+	uint8_t cmd;
+	struct xdr_netobj name;
+	uint16_t namelen;
+	struct cld_net *cn = nn->cld_net;
+
+	if (get_user(cmd, &cmsg->cm_cmd)) {
+		dprintk("%s: error when copying cmd from userspace", __func__);
+		return -EFAULT;
+	}
+	if (cmd == Cld_GraceStart) {
+		if (get_user(namelen, &cmsg->cm_u.cm_name.cn_len))
+			return -EFAULT;
+		name.data = memdup_user(&cmsg->cm_u.cm_name.cn_id, namelen);
+		if (IS_ERR_OR_NULL(name.data))
+			return -EFAULT;
+		name.len = namelen;
+		if (name.len > 5 && memcmp(name.data, "hash:", 5) == 0) {
+			name.len = name.len - 5;
+			memmove(name.data, name.data + 5, name.len);
+			cn->cn_has_legacy = true;
+		}
+		if (!nfs4_client_to_reclaim(name, nn)) {
+			kfree(name.data);
+			return -EFAULT;
+		}
+		return sizeof(*cmsg);
+	}
+	return -EFAULT;
+}
+
+static ssize_t
 cld_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 {
 	struct cld_upcall *tmp, *cup;
@@ -714,6 +823,7 @@ cld_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 	struct nfsd_net *nn = net_generic(file_inode(filp)->i_sb->s_fs_info,
 						nfsd_net_id);
 	struct cld_net *cn = nn->cld_net;
+	int16_t status;
 
 	if (mlen != sizeof(*cmsg)) {
 		dprintk("%s: got %zu bytes, expected %zu\n", __func__, mlen,
@@ -727,13 +837,24 @@ cld_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 		return -EFAULT;
 	}
 
+	/*
+	 * copy the status so we know whether to remove the upcall from the
+	 * list (for -EINPROGRESS, we just want to make sure the xid is
+	 * valid, not remove the upcall from the list)
+	 */
+	if (get_user(status, &cmsg->cm_status)) {
+		dprintk("%s: error when copying status from userspace", __func__);
+		return -EFAULT;
+	}
+
 	/* walk the list and find corresponding xid */
 	cup = NULL;
 	spin_lock(&cn->cn_lock);
 	list_for_each_entry(tmp, &cn->cn_list, cu_list) {
 		if (get_unaligned(&tmp->cu_msg.cm_xid) == xid) {
 			cup = tmp;
-			list_del_init(&cup->cu_list);
+			if (status != -EINPROGRESS)
+				list_del_init(&cup->cu_list);
 			break;
 		}
 	}
@@ -745,6 +866,9 @@ cld_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 		return -EINVAL;
 	}
 
+	if (status == -EINPROGRESS)
+		return __cld_pipe_inprogress_downcall(cmsg, nn);
+
 	if (copy_from_user(&cup->cu_msg, src, mlen) != 0)
 		return -EFAULT;
 
@@ -820,7 +944,7 @@ nfsd4_cld_unregister_net(struct net *net, struct rpc_pipe *pipe)
 
 /* Initialize rpc_pipefs pipe for communication with client tracking daemon */
 static int
-nfsd4_init_cld_pipe(struct net *net)
+__nfsd4_init_cld_pipe(struct net *net)
 {
 	int ret;
 	struct dentry *dentry;
@@ -851,6 +975,7 @@ nfsd4_init_cld_pipe(struct net *net)
 	}
 
 	cn->cn_pipe->dentry = dentry;
+	cn->cn_has_legacy = false;
 	nn->cld_net = cn;
 	return 0;
 
@@ -863,6 +988,17 @@ err:
 	return ret;
 }
 
+static int
+nfsd4_init_cld_pipe(struct net *net)
+{
+	int status;
+
+	status = __nfsd4_init_cld_pipe(net);
+	if (!status)
+		printk("NFSD: Using old nfsdcld client tracking operations.\n");
+	return status;
+}
+
 static void
 nfsd4_remove_cld_pipe(struct net *net)
 {
@@ -991,9 +1127,14 @@ out_err:
 				"record from stable storage: %d\n", ret);
 }
 
-/* Check for presence of a record, and update its timestamp */
+/*
+ * For older nfsdcld's that do not allow us to "slurp" the clients
+ * from the tracking database during startup.
+ *
+ * Check for presence of a record, and update its timestamp
+ */
 static int
-nfsd4_cld_check(struct nfs4_client *clp)
+nfsd4_cld_check_v0(struct nfs4_client *clp)
 {
 	int ret;
 	struct cld_upcall *cup;
@@ -1026,8 +1167,84 @@ nfsd4_cld_check(struct nfs4_client *clp)
 	return ret;
 }
 
+/*
+ * For newer nfsdcld's that allow us to "slurp" the clients
+ * from the tracking database during startup.
+ *
+ * Check for presence of a record in the reclaim_str_hashtbl
+ */
+static int
+nfsd4_cld_check(struct nfs4_client *clp)
+{
+	struct nfs4_client_reclaim *crp;
+	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
+	struct cld_net *cn = nn->cld_net;
+	int status;
+	char dname[HEXDIR_LEN];
+	struct xdr_netobj name;
+
+	/* did we already find that this client is stable? */
+	if (test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags))
+		return 0;
+
+	/* look for it in the reclaim hashtable otherwise */
+	crp = nfsd4_find_reclaim_client(clp->cl_name, nn);
+	if (crp)
+		goto found;
+
+	if (cn->cn_has_legacy) {
+		status = nfs4_make_rec_clidname(dname, &clp->cl_name);
+		if (status)
+			return -ENOENT;
+
+		name.data = kmemdup(dname, HEXDIR_LEN, GFP_KERNEL);
+		if (!name.data) {
+			dprintk("%s: failed to allocate memory for name.data!\n",
+				__func__);
+			return -ENOENT;
+		}
+		name.len = HEXDIR_LEN;
+		crp = nfsd4_find_reclaim_client(name, nn);
+		kfree(name.data);
+		if (crp)
+			goto found;
+
+	}
+	return -ENOENT;
+found:
+	crp->cr_clp = clp;
+	return 0;
+}
+
+static int
+nfsd4_cld_grace_start(struct nfsd_net *nn)
+{
+	int ret;
+	struct cld_upcall *cup;
+	struct cld_net *cn = nn->cld_net;
+
+	cup = alloc_cld_upcall(cn);
+	if (!cup) {
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
+	cup->cu_msg.cm_cmd = Cld_GraceStart;
+	ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg);
+	if (!ret)
+		ret = cup->cu_msg.cm_status;
+
+	free_cld_upcall(cup);
+out_err:
+	if (ret)
+		dprintk("%s: Unable to get clients from userspace: %d\n",
+			__func__, ret);
+	return ret;
+}
+
+/* For older nfsdcld's that need cm_gracetime */
 static void
-nfsd4_cld_grace_done(struct nfsd_net *nn)
+nfsd4_cld_grace_done_v0(struct nfsd_net *nn)
 {
 	int ret;
 	struct cld_upcall *cup;
@@ -1051,11 +1268,149 @@ out_err:
 		printk(KERN_ERR "NFSD: Unable to end grace period: %d\n", ret);
 }
 
-static const struct nfsd4_client_tracking_ops nfsd4_cld_tracking_ops = {
+/*
+ * For newer nfsdcld's that do not need cm_gracetime.  We also need to call
+ * nfs4_release_reclaim() to clear out the reclaim_str_hashtbl.
+ */
+static void
+nfsd4_cld_grace_done(struct nfsd_net *nn)
+{
+	int ret;
+	struct cld_upcall *cup;
+	struct cld_net *cn = nn->cld_net;
+
+	cup = alloc_cld_upcall(cn);
+	if (!cup) {
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
+	cup->cu_msg.cm_cmd = Cld_GraceDone;
+	ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg);
+	if (!ret)
+		ret = cup->cu_msg.cm_status;
+
+	free_cld_upcall(cup);
+out_err:
+	nfs4_release_reclaim(nn);
+	if (ret)
+		printk(KERN_ERR "NFSD: Unable to end grace period: %d\n", ret);
+}
+
+static int
+nfs4_cld_state_init(struct net *net)
+{
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	int i;
+
+	nn->reclaim_str_hashtbl = kmalloc_array(CLIENT_HASH_SIZE,
+						sizeof(struct list_head),
+						GFP_KERNEL);
+	if (!nn->reclaim_str_hashtbl)
+		return -ENOMEM;
+
+	for (i = 0; i < CLIENT_HASH_SIZE; i++)
+		INIT_LIST_HEAD(&nn->reclaim_str_hashtbl[i]);
+	nn->reclaim_str_hashtbl_size = 0;
+	nn->track_reclaim_completes = true;
+	atomic_set(&nn->nr_reclaim_complete, 0);
+
+	return 0;
+}
+
+static void
+nfs4_cld_state_shutdown(struct net *net)
+{
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+
+	nn->track_reclaim_completes = false;
+	kfree(nn->reclaim_str_hashtbl);
+}
+
+static bool
+cld_running(struct nfsd_net *nn)
+{
+	struct cld_net *cn = nn->cld_net;
+	struct rpc_pipe *pipe = cn->cn_pipe;
+
+	return pipe->nreaders || pipe->nwriters;
+}
+
+static int
+nfsd4_cld_tracking_init(struct net *net)
+{
+	int status;
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+	bool running;
+	int retries = 10;
+
+	status = nfs4_cld_state_init(net);
+	if (status)
+		return status;
+
+	status = __nfsd4_init_cld_pipe(net);
+	if (status)
+		goto err_shutdown;
+
+	/*
+	 * rpc pipe upcalls take 30 seconds to time out, so we don't want to
+	 * queue an upcall unless we know that nfsdcld is running (because we
+	 * want this to fail fast so that nfsd4_client_tracking_init() can try
+	 * the next client tracking method).  nfsdcld should already be running
+	 * before nfsd is started, so the wait here is for nfsdcld to open the
+	 * pipefs file we just created.
+	 */
+	while (!(running = cld_running(nn)) && retries--)
+		msleep(100);
+
+	if (!running) {
+		status = -ETIMEDOUT;
+		goto err_remove;
+	}
+
+	status = nfsd4_cld_grace_start(nn);
+	if (status) {
+		if (status == -EOPNOTSUPP)
+			printk(KERN_WARNING "NFSD: Please upgrade nfsdcld.\n");
+		nfs4_release_reclaim(nn);
+		goto err_remove;
+	} else
+		printk("NFSD: Using nfsdcld client tracking operations.\n");
+	return 0;
+
+err_remove:
+	nfsd4_remove_cld_pipe(net);
+err_shutdown:
+	nfs4_cld_state_shutdown(net);
+	return status;
+}
+
+static void
+nfsd4_cld_tracking_exit(struct net *net)
+{
+	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
+
+	nfs4_release_reclaim(nn);
+	nfsd4_remove_cld_pipe(net);
+	nfs4_cld_state_shutdown(net);
+}
+
+/* For older nfsdcld's */
+static const struct nfsd4_client_tracking_ops nfsd4_cld_tracking_ops_v0 = {
 	.init		= nfsd4_init_cld_pipe,
 	.exit		= nfsd4_remove_cld_pipe,
 	.create		= nfsd4_cld_create,
 	.remove		= nfsd4_cld_remove,
+	.check		= nfsd4_cld_check_v0,
+	.grace_done	= nfsd4_cld_grace_done_v0,
+};
+
+/* For newer nfsdcld's */
+static const struct nfsd4_client_tracking_ops nfsd4_cld_tracking_ops = {
+	.init		= nfsd4_cld_tracking_init,
+	.exit		= nfsd4_cld_tracking_exit,
+	.create		= nfsd4_cld_create,
+	.remove		= nfsd4_cld_remove,
 	.check		= nfsd4_cld_check,
 	.grace_done	= nfsd4_cld_grace_done,
 };
@@ -1267,6 +1622,8 @@ nfsd4_umh_cltrack_init(struct net *net)
 
 	ret = nfsd4_umh_cltrack_upcall("init", NULL, grace_start, NULL);
 	kfree(grace_start);
+	if (!ret)
+		printk("NFSD: Using UMH upcall client tracking operations.\n");
 	return ret;
 }
 
@@ -1416,9 +1773,20 @@ nfsd4_client_tracking_init(struct net *net)
 	if (nn->client_tracking_ops)
 		goto do_init;
 
+	/* First, try to use nfsdcld */
+	nn->client_tracking_ops = &nfsd4_cld_tracking_ops;
+	status = nn->client_tracking_ops->init(net);
+	if (!status)
+		return status;
+	if (status != -ETIMEDOUT) {
+		nn->client_tracking_ops = &nfsd4_cld_tracking_ops_v0;
+		status = nn->client_tracking_ops->init(net);
+		if (!status)
+			return status;
+	}
+
 	/*
-	 * First, try a UMH upcall. It should succeed or fail quickly, so
-	 * there's little harm in trying that first.
+	 * Next, try the UMH upcall.
 	 */
 	nn->client_tracking_ops = &nfsd4_umh_tracking_ops;
 	status = nn->client_tracking_ops->init(net);
@@ -1426,25 +1794,23 @@ nfsd4_client_tracking_init(struct net *net)
 		return status;
 
 	/*
-	 * See if the recoverydir exists and is a directory. If it is,
-	 * then use the legacy ops.
+	 * Finally, See if the recoverydir exists and is a directory.
+	 * If it is, then use the legacy ops.
 	 */
 	nn->client_tracking_ops = &nfsd4_legacy_tracking_ops;
 	status = kern_path(nfs4_recoverydir(), LOOKUP_FOLLOW, &path);
 	if (!status) {
 		status = d_is_dir(path.dentry);
 		path_put(&path);
-		if (status)
-			goto do_init;
+		if (!status) {
+			status = -EINVAL;
+			goto out;
+		}
 	}
 
-	/* Finally, try to use nfsdcld */
-	nn->client_tracking_ops = &nfsd4_cld_tracking_ops;
-	printk(KERN_WARNING "NFSD: the nfsdcld client tracking upcall will be "
-			"removed in 3.10. Please transition to using "
-			"nfsdcltrack.\n");
 do_init:
 	status = nn->client_tracking_ops->init(net);
+out:
 	if (status) {
 		printk(KERN_WARNING "NFSD: Unable to initialize client "
 				    "recovery tracking! (%d)\n", status);
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index eca4a23f93c8..618e66078ee5 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -77,6 +77,7 @@ static u64 current_sessionid = 1;
 /* forward declarations */
 static bool check_for_locks(struct nfs4_file *fp, struct nfs4_lockowner *lowner);
 static void nfs4_free_ol_stateid(struct nfs4_stid *stid);
+void nfsd4_end_grace(struct nfsd_net *nn);
 
 /* Locking: */
 
@@ -1067,9 +1068,9 @@ static unsigned int clientid_hashval(u32 id)
 	return id & CLIENT_HASH_MASK;
 }
 
-static unsigned int clientstr_hashval(const char *name)
+static unsigned int clientstr_hashval(struct xdr_netobj name)
 {
-	return opaque_hashval(name, 8) & CLIENT_HASH_MASK;
+	return opaque_hashval(name.data, 8) & CLIENT_HASH_MASK;
 }
 
 /*
@@ -1997,6 +1998,22 @@ destroy_client(struct nfs4_client *clp)
 	__destroy_client(clp);
 }
 
+static void inc_reclaim_complete(struct nfs4_client *clp)
+{
+	struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id);
+
+	if (!nn->track_reclaim_completes)
+		return;
+	if (!nfsd4_find_reclaim_client(clp->cl_name, nn))
+		return;
+	if (atomic_inc_return(&nn->nr_reclaim_complete) ==
+			nn->reclaim_str_hashtbl_size) {
+		printk(KERN_INFO "NFSD: all clients done reclaiming, ending NFSv4 grace period (net %x)\n",
+				clp->net->ns.inum);
+		nfsd4_end_grace(nn);
+	}
+}
+