From 88a9fe8cae3bb52e82489447f45e8d7ba1409ca8 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 23 Dec 2008 15:21:31 -0500 Subject: SUNRPC: Remove the last remnant of the BKL... Somehow, this escaped the previous purge. There should be no need to keep any extra locks in the XDR callbacks. The NFS client XDR code only writes into private objects, whereas all reads of shared objects are confined to fields that do not change, such as filehandles... Ditto for lockd, the NFSv2/v3 client mount code, and rpcbind. The nfsd XDR code may require the BKL, but since it does a synchronous RPC call from a thread that already holds the lock, that issue is moot. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xdr.h | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h index e4057d729f03..49e1eb454465 100644 --- a/include/linux/sunrpc/xdr.h +++ b/include/linux/sunrpc/xdr.h @@ -36,21 +36,6 @@ struct xdr_netobj { */ typedef int (*kxdrproc_t)(void *rqstp, __be32 *data, void *obj); -/* - * We're still requiring the BKL in the xdr code until it's been - * more carefully audited, at which point this wrapper will become - * unnecessary. - */ -static inline int rpc_call_xdrproc(kxdrproc_t xdrproc, void *rqstp, __be32 *data, void *obj) -{ - int ret; - - lock_kernel(); - ret = xdrproc(rqstp, data, obj); - unlock_kernel(); - return ret; -} - /* * Basic structure for transmission/reception of a client XDR message. * Features a header (for a linear buffer containing RPC headers -- cgit v1.2.3-70-g09d2 From 146ec944bbd31d241a44a00518b054fb01921d22 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 23 Dec 2008 15:21:34 -0500 Subject: NFS: Move declaration of nfs_mount() to fs/nfs/internal.h Clean up: The nfs_mount() function is not to be used outside of the NFS client. Move its public declaration to fs/nfs/internal.h. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 4 ++++ fs/nfs/nfsroot.c | 2 ++ include/linux/nfs_fs.h | 6 ------ 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index d212ee41caf2..7a38cc7b4137 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -63,6 +63,10 @@ struct nfs_parsed_mount_data { struct security_mnt_opts lsm_opts; }; +/* mount_clnt.c */ +extern int nfs_mount(struct sockaddr *, size_t, char *, char *, + int, int, struct nfs_fh *); + /* client.c */ extern struct rpc_program nfs_program; diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c index f1c0a066439f..a96e5fdb38af 100644 --- a/fs/nfs/nfsroot.c +++ b/fs/nfs/nfsroot.c @@ -86,6 +86,8 @@ #include #include +#include "internal.h" + /* Define this to allow debugging output */ #undef NFSROOT_DEBUG #define NFSDBG_FACILITY NFSDBG_ROOT diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 4eaa8347a0d9..f11077285a6c 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -532,12 +532,6 @@ static inline void nfs3_forget_cached_acls(struct inode *inode) } #endif /* CONFIG_NFS_V3_ACL */ -/* - * linux/fs/mount_clnt.c - */ -extern int nfs_mount(struct sockaddr *, size_t, char *, char *, - int, int, struct nfs_fh *); - /* * inline functions */ -- cgit v1.2.3-70-g09d2 From d740351bf0960e89ce1aef45cfe00167cb0f9e5b Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 23 Dec 2008 15:21:37 -0500 Subject: NFS: add "[no]resvport" mount option The standard default security setting for NFS is AUTH_SYS. An NFS client connects to NFS servers via a privileged source port and a fixed standard destination port (2049). The client sends raw uid and gid numbers to identify users making NFS requests, and the server assumes an appropriate authority on the client has vetted these values because the source port is privileged. On Linux, by default in-kernel RPC services use a privileged port in the range between 650 and 1023 to avoid using source ports of well- known IP services. Using such a small range limits the number of NFS mount points and the number of unique NFS servers to which a client can connect concurrently. An NFS client can use unprivileged source ports to expand the range of source port numbers, allowing more concurrent server connections and more NFS mount points. Servers must explicitly allow NFS connections from unprivileged ports for this to work. In the past, bumping the value of the sunrpc.max_resvport sysctl on the client would permit the NFS client to use unprivileged ports. Bumping this setting also changes the maximum port number used by other in-kernel RPC services, some of which still required a port number less than 1023. This is exacerbated by the way source port numbers are chosen by the Linux RPC client, which starts at the top of the range and works downwards. It means that bumping the maximum means all RPC services requesting a source port will likely get an unprivileged port instead of a privileged one. Changing this setting effects all NFS mount points on a client. A sysadmin could not selectively choose which mount points would use non-privileged ports and which could not. Lastly, this mechanism of expanding the limit on the number of NFS mount points was entirely undocumented. To address the need for the NFS client to use a large range of source ports without interfering with the activity of other in-kernel RPC services, we introduce a new NFS mount option. This option explicitly tells only the NFS client to use a non-privileged source port when communicating with the NFS server for one specific mount point. This new mount option is called "resvport," like the similar NFS mount option on FreeBSD and Mac OS X. A sister patch for nfs-utils will be submitted that documents this new option in nfs(5). The default setting for this new mount option requires the NFS client to use a privileged port, as before. Explicitly specifying the "noresvport" mount option allows the NFS client to use an unprivileged source port for this mount point when connecting to the NFS server port. This mount option is supported only for text-based NFS mounts. [ Sidebar: it is widely known that security mechanisms based on the use of privileged source ports are ineffective. However, the NFS client can combine the use of unprivileged ports with the use of secure authentication mechanisms, such as Kerberos. This allows a large number of connections and mount points while ensuring a useful level of security. Eventually we may change the default setting for this option depending on the security flavor used for the mount. For example, if the mount is using only AUTH_SYS, then the default setting will be "resvport;" if the mount is using a strong security flavor such as krb5, the default setting will be "noresvport." ] Signed-off-by: Chuck Lever [Trond.Myklebust@netapp.com: Fixed a bug whereby nfs4_init_client() was being called with incorrect arguments.] Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 11 +++++++---- fs/nfs/super.c | 12 +++++++++++- include/linux/nfs_mount.h | 3 ++- 3 files changed, 20 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 27190337fc13..3a69cacc4fa4 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -627,7 +627,8 @@ static int nfs_init_client(struct nfs_client *clp, * Create a client RPC handle for doing FSSTAT with UNIX auth only * - RFC 2623, sec 2.3.2 */ - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX, 0, 0); + error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX, + 0, data->flags & NFS_MOUNT_NORESVPORT); if (error < 0) goto error; nfs_mark_client_ready(clp, NFS_CS_READY); @@ -969,7 +970,8 @@ error: static int nfs4_init_client(struct nfs_client *clp, const struct rpc_timeout *timeparms, const char *ip_addr, - rpc_authflavor_t authflavour) + rpc_authflavor_t authflavour, + int flags) { int error; @@ -983,7 +985,7 @@ static int nfs4_init_client(struct nfs_client *clp, clp->rpc_ops = &nfs_v4_clientops; error = nfs_create_rpc_client(clp, timeparms, authflavour, - 1, 0); + 1, flags & NFS_MOUNT_NORESVPORT); if (error < 0) goto error; memcpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr)); @@ -1034,7 +1036,8 @@ static int nfs4_set_client(struct nfs_server *server, error = PTR_ERR(clp); goto error; } - error = nfs4_init_client(clp, timeparms, ip_addr, authflavour); + error = nfs4_init_client(clp, timeparms, ip_addr, authflavour, + server->flags); if (error < 0) goto error_put; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 2b0c8e132b54..e05a77be3068 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -75,6 +75,7 @@ enum { Opt_acl, Opt_noacl, Opt_rdirplus, Opt_nordirplus, Opt_sharecache, Opt_nosharecache, + Opt_resvport, Opt_noresvport, /* Mount options that take integer arguments */ Opt_port, @@ -129,6 +130,8 @@ static const match_table_t nfs_mount_option_tokens = { { Opt_nordirplus, "nordirplus" }, { Opt_sharecache, "sharecache" }, { Opt_nosharecache, "nosharecache" }, + { Opt_resvport, "resvport" }, + { Opt_noresvport, "noresvport" }, { Opt_port, "port=%u" }, { Opt_rsize, "rsize=%u" }, @@ -514,7 +517,8 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, { NFS_MOUNT_NONLM, ",nolock", "" }, { NFS_MOUNT_NOACL, ",noacl", "" }, { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" }, - { NFS_MOUNT_UNSHARED, ",nosharecache", ""}, + { NFS_MOUNT_UNSHARED, ",nosharecache", "" }, + { NFS_MOUNT_NORESVPORT, ",noresvport", "" }, { 0, NULL, NULL } }; const struct proc_nfs_info *nfs_infop; @@ -1035,6 +1039,12 @@ static int nfs_parse_mount_options(char *raw, case Opt_nosharecache: mnt->flags |= NFS_MOUNT_UNSHARED; break; + case Opt_resvport: + mnt->flags &= ~NFS_MOUNT_NORESVPORT; + break; + case Opt_noresvport: + mnt->flags |= NFS_MOUNT_NORESVPORT; + break; /* * options that take numeric values diff --git a/include/linux/nfs_mount.h b/include/linux/nfs_mount.h index 6549a06ac16e..4499016e6d0d 100644 --- a/include/linux/nfs_mount.h +++ b/include/linux/nfs_mount.h @@ -45,7 +45,7 @@ struct nfs_mount_data { char context[NFS_MAX_CONTEXT_LEN + 1]; /* 6 */ }; -/* bits in the flags field */ +/* bits in the flags field visible to user space */ #define NFS_MOUNT_SOFT 0x0001 /* 1 */ #define NFS_MOUNT_INTR 0x0002 /* 1 */ /* now unused, but ABI */ @@ -68,5 +68,6 @@ struct nfs_mount_data { /* The following are for internal use only */ #define NFS_MOUNT_LOOKUP_CACHE_NONEG 0x10000 #define NFS_MOUNT_LOOKUP_CACHE_NONE 0x20000 +#define NFS_MOUNT_NORESVPORT 0x40000 #endif -- cgit v1.2.3-70-g09d2 From 0cb2659b818eca99235e17c04291cfa9985c14f7 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 23 Dec 2008 15:21:38 -0500 Subject: NLM: allow lockd requests from an unprivileged port If the admin has specified the "noresvport" option for an NFS mount point, the kernel's NFS client uses an unprivileged source port for the main NFS transport. The kernel's lockd client should use an unprivileged port in this case as well. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/clntlock.c | 2 +- fs/lockd/host.c | 10 +++++++++- fs/nfs/client.c | 2 ++ include/linux/lockd/bind.h | 1 + include/linux/lockd/lockd.h | 4 +++- 5 files changed, 16 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c index 94d42cc4e393..1f3b0fc0d351 100644 --- a/fs/lockd/clntlock.c +++ b/fs/lockd/clntlock.c @@ -61,7 +61,7 @@ struct nlm_host *nlmclnt_init(const struct nlmclnt_initdata *nlm_init) host = nlmclnt_lookup_host(nlm_init->address, nlm_init->addrlen, nlm_init->protocol, nlm_version, - nlm_init->hostname); + nlm_init->hostname, nlm_init->noresvport); if (host == NULL) { lockd_down(); return ERR_PTR(-ENOLCK); diff --git a/fs/lockd/host.c b/fs/lockd/host.c index 70fc63a1727b..acc2aa5021d1 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -48,6 +48,7 @@ struct nlm_lookup_host_info { const size_t hostname_len; /* it's length */ const struct sockaddr *src_sap; /* our address (optional) */ const size_t src_len; /* it's length */ + const int noresvport; /* use non-priv port */ }; /* @@ -222,6 +223,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni) host->h_nsmstate = 0; /* real NSM state */ host->h_nsmhandle = nsm; host->h_server = ni->server; + host->h_noresvport = ni->noresvport; hlist_add_head(&host->h_hash, chain); INIT_LIST_HEAD(&host->h_lockowners); spin_lock_init(&host->h_lock); @@ -272,6 +274,7 @@ nlm_destroy_host(struct nlm_host *host) * @protocol: transport protocol to use * @version: NLM protocol version * @hostname: '\0'-terminated hostname of server + * @noresvport: 1 if non-privileged port should be used * * Returns an nlm_host structure that matches the passed-in * [server address, transport protocol, NLM version, server hostname]. @@ -281,7 +284,9 @@ nlm_destroy_host(struct nlm_host *host) struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap, const size_t salen, const unsigned short protocol, - const u32 version, const char *hostname) + const u32 version, + const char *hostname, + int noresvport) { const struct sockaddr source = { .sa_family = AF_UNSPEC, @@ -296,6 +301,7 @@ struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap, .hostname_len = strlen(hostname), .src_sap = &source, .src_len = sizeof(source), + .noresvport = noresvport, }; dprintk("lockd: %s(host='%s', vers=%u, proto=%s)\n", __func__, @@ -417,6 +423,8 @@ nlm_bind_host(struct nlm_host *host) */ if (!host->h_server) args.flags |= RPC_CLNT_CREATE_HARDRTRY; + if (host->h_noresvport) + args.flags |= RPC_CLNT_CREATE_NONPRIVPORT; clnt = rpc_create(&args); if (!IS_ERR(clnt)) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 3a69cacc4fa4..70b6d9e8517d 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -526,6 +526,8 @@ static int nfs_start_lockd(struct nfs_server *server) .protocol = server->flags & NFS_MOUNT_TCP ? IPPROTO_TCP : IPPROTO_UDP, .nfs_version = clp->rpc_ops->version, + .noresvport = server->flags & NFS_MOUNT_NORESVPORT ? + 1 : 0, }; if (nlm_init.nfs_version > 3) diff --git a/include/linux/lockd/bind.h b/include/linux/lockd/bind.h index e5872dc994c0..fbc48f898521 100644 --- a/include/linux/lockd/bind.h +++ b/include/linux/lockd/bind.h @@ -41,6 +41,7 @@ struct nlmclnt_initdata { size_t addrlen; unsigned short protocol; u32 nfs_version; + int noresvport; }; /* diff --git a/include/linux/lockd/lockd.h b/include/linux/lockd/lockd.h index b56d5aa9b194..23da3fa69efa 100644 --- a/include/linux/lockd/lockd.h +++ b/include/linux/lockd/lockd.h @@ -49,6 +49,7 @@ struct nlm_host { unsigned short h_proto; /* transport proto */ unsigned short h_reclaiming : 1, h_server : 1, /* server side, not client side */ + h_noresvport : 1, h_inuse : 1; wait_queue_head_t h_gracewait; /* wait while reclaiming */ struct rw_semaphore h_rwsem; /* Reboot recovery lock */ @@ -220,7 +221,8 @@ struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap, const size_t salen, const unsigned short protocol, const u32 version, - const char *hostname); + const char *hostname, + int noresvport); struct nlm_host *nlmsvc_lookup_host(const struct svc_rqst *rqstp, const char *hostname, const size_t hostname_len); -- cgit v1.2.3-70-g09d2 From 95d35cb4c473c754824967c0b069bbeb7efa4847 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 23 Dec 2008 15:21:45 -0500 Subject: NFSv4: Remove nfs_client->cl_sem Now that we're using the flags to indicate state that needs to be recovered, as well as having implemented proper refcounting and spinlocking on the state and open_owners, we can get rid of nfs_client->cl_sem. The only remaining case that was dubious was the file locking, and that case is now covered by the nfsi->rwsem. Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 1 - fs/nfs/delegation.c | 5 ----- fs/nfs/nfs4proc.c | 20 +------------------- fs/nfs/nfs4renewd.c | 3 --- fs/nfs/nfs4state.c | 17 ----------------- include/linux/nfs_fs_sb.h | 6 ------ 6 files changed, 1 insertion(+), 51 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 70b6d9e8517d..b643f0ec5c21 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -143,7 +143,6 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_ clp->cl_proto = cl_init->proto; #ifdef CONFIG_NFS_V4 - init_rwsem(&clp->cl_sem); INIT_LIST_HEAD(&clp->cl_delegations); spin_lock_init(&clp->cl_lock); INIT_DELAYED_WORK(&clp->cl_renewd, nfs4_renew_state); diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 646ba3e75a1e..ebc06f2da31a 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -243,16 +243,13 @@ static void nfs_msync_inode(struct inode *inode) */ static int __nfs_inode_return_delegation(struct inode *inode, struct nfs_delegation *delegation) { - struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); nfs_msync_inode(inode); - down_read(&clp->cl_sem); /* Guard against new delegated open calls */ down_write(&nfsi->rwsem); nfs_delegation_claim_opens(inode, &delegation->stateid); up_write(&nfsi->rwsem); - up_read(&clp->cl_sem); nfs_msync_inode(inode); return nfs_do_return_delegation(inode, delegation, 1); @@ -425,7 +422,6 @@ static int recall_thread(void *data) daemonize("nfsv4-delegreturn"); nfs_msync_inode(inode); - down_read(&clp->cl_sem); down_write(&nfsi->rwsem); spin_lock(&clp->cl_lock); delegation = nfs_detach_delegation_locked(nfsi, args->stateid); @@ -437,7 +433,6 @@ static int recall_thread(void *data) complete(&args->started); nfs_delegation_claim_opens(inode, args->stateid); up_write(&nfsi->rwsem); - up_read(&clp->cl_sem); nfs_msync_inode(inode); if (delegation != NULL) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index aec4e47c462d..26dcd77abb07 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -207,12 +207,8 @@ static int nfs4_wait_clnt_recover(struct nfs_client *clp) might_sleep(); - rwsem_acquire(&clp->cl_sem.dep_map, 0, 0, _RET_IP_); - res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER, nfs4_wait_bit_killable, TASK_KILLABLE); - - rwsem_release(&clp->cl_sem.dep_map, 1, _RET_IP_); return res; } @@ -1135,7 +1131,6 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct struct nfs4_state_owner *sp; struct nfs4_state *state = NULL; struct nfs_server *server = NFS_SERVER(dir); - struct nfs_client *clp = server->nfs_client; struct nfs4_opendata *opendata; int status; @@ -1150,11 +1145,10 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct goto err_put_state_owner; if (path->dentry->d_inode != NULL) nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE)); - down_read(&clp->cl_sem); status = -ENOMEM; opendata = nfs4_opendata_alloc(path, sp, flags, sattr); if (opendata == NULL) - goto err_release_rwsem; + goto err_put_state_owner; if (path->dentry->d_inode != NULL) opendata->state = nfs4_get_open_state(path->dentry->d_inode, sp); @@ -1172,13 +1166,10 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct goto err_opendata_put; nfs4_opendata_put(opendata); nfs4_put_state_owner(sp); - up_read(&clp->cl_sem); *res = state; return 0; err_opendata_put: nfs4_opendata_put(opendata); -err_release_rwsem: - up_read(&clp->cl_sem); err_put_state_owner: nfs4_put_state_owner(sp); out_err: @@ -3099,7 +3090,6 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock struct nfs4_lock_state *lsp; int status; - down_read(&clp->cl_sem); arg.lock_owner.clientid = clp->cl_clientid; status = nfs4_set_lock_state(state, request); if (status != 0) @@ -3116,7 +3106,6 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock } request->fl_ops->fl_release_private(request); out: - up_read(&clp->cl_sem); return status; } @@ -3273,7 +3262,6 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl, static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request) { - struct nfs_client *clp = state->owner->so_client; struct nfs_inode *nfsi = NFS_I(state->inode); struct nfs_seqid *seqid; struct nfs4_lock_state *lsp; @@ -3284,15 +3272,12 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock * status = nfs4_set_lock_state(state, request); /* Unlock _before_ we do the RPC call */ request->fl_flags |= FL_EXISTS; - down_read(&clp->cl_sem); down_read(&nfsi->rwsem); if (do_vfs_lock(request->fl_file, request) == -ENOENT) { up_read(&nfsi->rwsem); - up_read(&clp->cl_sem); goto out; } up_read(&nfsi->rwsem); - up_read(&clp->cl_sem); if (status != 0) goto out; /* Is this a delegated lock? */ @@ -3518,7 +3503,6 @@ static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request) { - struct nfs_client *clp = state->owner->so_client; struct nfs_inode *nfsi = NFS_I(state->inode); unsigned char fl_flags = request->fl_flags; int status; @@ -3531,7 +3515,6 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock status = do_vfs_lock(request->fl_file, request); if (status < 0) goto out; - down_read(&clp->cl_sem); down_read(&nfsi->rwsem); if (test_bit(NFS_DELEGATED_STATE, &state->flags)) { /* Yes: cache locks! */ @@ -3549,7 +3532,6 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __func__); out_unlock: up_read(&nfsi->rwsem); - up_read(&clp->cl_sem); out: request->fl_flags = fl_flags; return status; diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index 9fe8640a88eb..6101f955f231 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -65,7 +65,6 @@ nfs4_renew_state(struct work_struct *work) long lease, timeout; unsigned long last, now; - down_read(&clp->cl_sem); dprintk("%s: start\n", __func__); /* Are there any active superblocks? */ if (list_empty(&clp->cl_superblocks)) @@ -101,11 +100,9 @@ nfs4_renew_state(struct work_struct *work) schedule_delayed_work(&clp->cl_renewd, timeout); spin_unlock(&clp->cl_lock); out: - up_read(&clp->cl_sem); dprintk("%s: done\n", __func__); } -/* Must be called with clp->cl_sem locked for writes */ void nfs4_schedule_state_renewal(struct nfs_client *clp) { diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 16c9fbdf97b4..ec0cbe00a804 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -305,10 +305,6 @@ nfs4_drop_state_owner(struct nfs4_state_owner *sp) } } -/* - * Note: must be called with clp->cl_sem held in order to prevent races - * with reboot recovery! - */ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct rpc_cred *cred) { struct nfs_client *clp = server->nfs_client; @@ -337,10 +333,6 @@ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct return sp; } -/* - * Must be called with clp->cl_sem held in order to avoid races - * with state recovery... - */ void nfs4_put_state_owner(struct nfs4_state_owner *sp) { struct nfs_client *clp = sp->so_client; @@ -442,10 +434,6 @@ out: return state; } -/* - * Beware! Caller must be holding exactly one - * reference to clp->cl_sem! - */ void nfs4_put_open_state(struct nfs4_state *state) { struct inode *inode = state->inode; @@ -578,7 +566,6 @@ static void nfs4_free_lock_state(struct nfs4_lock_state *lsp) * Return a compatible lock_state. If no initialized lock_state structure * exists, return an uninitialized one. * - * The caller must be holding clp->cl_sem */ static struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t owner) { @@ -1127,7 +1114,6 @@ static int reclaimer(void *ptr) allow_signal(SIGKILL); /* Ensure exclusive access to NFSv4 state */ - down_write(&clp->cl_sem); while (!list_empty(&clp->cl_superblocks)) { if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) { /* We're going to have to re-establish a clientid */ @@ -1149,7 +1135,6 @@ static int reclaimer(void *ptr) /* First recover reboot state... */ if (test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state)) { - /* Note: list is protected by exclusive lock on cl->cl_sem */ status = nfs4_do_reclaim(clp, &nfs4_reboot_recovery_ops); if (status == -NFS4ERR_STALE_CLIENTID) continue; @@ -1159,7 +1144,6 @@ static int reclaimer(void *ptr) /* Now recover expired state... */ if (test_and_clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) { - /* Note: list is protected by exclusive lock on cl->cl_sem */ status = nfs4_do_reclaim(clp, &nfs4_nograce_recovery_ops); if (status < 0) { set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state); @@ -1175,7 +1159,6 @@ static int reclaimer(void *ptr) break; } out: - up_write(&clp->cl_sem); if (test_and_clear_bit(NFS4CLNT_CB_PATH_DOWN, &clp->cl_state)) nfs_handle_cb_pathdown(clp); nfs4_clear_recover_bit(clp); diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 4e477ae58699..9bb81aec91cf 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -42,12 +42,6 @@ struct nfs_client { struct rb_root cl_openowner_id; struct rb_root cl_lockowner_id; - /* - * The following rwsem ensures exclusive access to the server - * while we recover the state following a lease expiration. - */ - struct rw_semaphore cl_sem; - struct list_head cl_delegations; struct rb_root cl_state_owners; spinlock_t cl_lock; -- cgit v1.2.3-70-g09d2 From bd7bf9d540c001055fba796ebf146d90e4dd2eb2 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 23 Dec 2008 15:21:53 -0500 Subject: NFSv4: Convert delegation->type field to fmode_t Signed-off-by: Trond Myklebust --- fs/nfs/delegation.c | 2 +- fs/nfs/delegation.h | 6 +++--- fs/nfs/nfs4xdr.c | 4 ++-- include/linux/nfs_fs.h | 2 +- include/linux/nfs_xdr.h | 4 ++-- 5 files changed, 9 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index e75f2f8c5245..968225a88015 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -48,7 +48,7 @@ void nfs_mark_delegation_referenced(struct nfs_delegation *delegation) set_bit(NFS_DELEGATION_REFERENCED, &delegation->flags); } -int nfs_have_delegation(struct inode *inode, int flags) +int nfs_have_delegation(struct inode *inode, fmode_t flags) { struct nfs_delegation *delegation; int ret = 0; diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h index 2a74882e5d50..09f383795174 100644 --- a/fs/nfs/delegation.h +++ b/fs/nfs/delegation.h @@ -17,7 +17,7 @@ struct nfs_delegation { struct rpc_cred *cred; struct inode *inode; nfs4_stateid stateid; - int type; + fmode_t type; loff_t maxsize; __u64 change_attr; unsigned long flags; @@ -54,10 +54,10 @@ int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl); int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode); void nfs_mark_delegation_referenced(struct nfs_delegation *delegation); -int nfs_have_delegation(struct inode *inode, int flags); +int nfs_have_delegation(struct inode *inode, fmode_t flags); #else -static inline int nfs_have_delegation(struct inode *inode, int flags) +static inline int nfs_have_delegation(struct inode *inode, fmode_t flags) { return 0; } diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index b916297d2334..879911c99030 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1024,7 +1024,7 @@ static void encode_opentype(struct xdr_stream *xdr, const struct nfs_openargs *a } } -static inline void encode_delegation_type(struct xdr_stream *xdr, int delegation_type) +static inline void encode_delegation_type(struct xdr_stream *xdr, fmode_t delegation_type) { __be32 *p; @@ -1053,7 +1053,7 @@ static inline void encode_claim_null(struct xdr_stream *xdr, const struct qstr * encode_string(xdr, name->len, name->name); } -static inline void encode_claim_previous(struct xdr_stream *xdr, int type) +static inline void encode_claim_previous(struct xdr_stream *xdr, fmode_t type) { __be32 *p; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index f11077285a6c..8d71d7b7c78a 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -180,7 +180,7 @@ struct nfs_inode { /* NFSv4 state */ struct list_head open_states; struct nfs_delegation *delegation; - int delegation_state; + fmode_t delegation_state; struct rw_semaphore rwsem; #endif /* CONFIG_NFS_V4*/ struct inode vfs_inode; diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index c1c31acb8a2b..32c1a0ecdbff 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -126,7 +126,7 @@ struct nfs_openargs { struct iattr * attrs; /* UNCHECKED, GUARDED */ nfs4_verifier verifier; /* EXCLUSIVE */ nfs4_stateid delegation; /* CLAIM_DELEGATE_CUR */ - int delegation_type; /* CLAIM_PREVIOUS */ + fmode_t delegation_type; /* CLAIM_PREVIOUS */ } u; const struct qstr * name; const struct nfs_server *server; /* Needed for ID mapping */ @@ -143,7 +143,7 @@ struct nfs_openres { struct nfs_fattr * dir_attr; struct nfs_seqid * seqid; const struct nfs_server *server; - int delegation_type; + fmode_t delegation_type; nfs4_stateid delegation; __u32 do_recall; __u64 maxsize; -- cgit v1.2.3-70-g09d2 From dc0b027dfadfcb8a5504f7d8052754bf8d501ab9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 23 Dec 2008 15:21:56 -0500 Subject: NFSv4: Convert the open and close ops to use fmode Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- fs/nfs/nfs4_fs.h | 8 +-- fs/nfs/nfs4proc.c | 126 ++++++++++++++++++++++++++---------------------- fs/nfs/nfs4state.c | 24 ++++----- fs/nfs/nfs4xdr.c | 10 ++-- include/linux/nfs_fs.h | 4 +- include/linux/nfs_xdr.h | 3 +- 7 files changed, 94 insertions(+), 83 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index d22eb383e1cf..1cf39202c3ea 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -592,7 +592,7 @@ static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context /* * Given an inode, search for an open context with the desired characteristics */ -struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode) +struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, fmode_t mode) { struct nfs_inode *nfsi = NFS_I(inode); struct nfs_open_context *pos, *ctx = NULL; diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index ee5f51ef696d..d5f5efaf2b26 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -161,7 +161,7 @@ struct nfs4_state { unsigned int n_rdonly; /* Number of read-only references */ unsigned int n_wronly; /* Number of write-only references */ unsigned int n_rdwr; /* Number of read/write references */ - int state; /* State on the server (R,W, or RW) */ + fmode_t state; /* State on the server (R,W, or RW) */ atomic_t count; }; @@ -223,9 +223,9 @@ extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struc extern void nfs4_put_state_owner(struct nfs4_state_owner *); extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state_owner *); extern void nfs4_put_open_state(struct nfs4_state *); -extern void nfs4_close_state(struct path *, struct nfs4_state *, mode_t); -extern void nfs4_close_sync(struct path *, struct nfs4_state *, mode_t); -extern void nfs4_state_set_mode_locked(struct nfs4_state *, mode_t); +extern void nfs4_close_state(struct path *, struct nfs4_state *, fmode_t); +extern void nfs4_close_sync(struct path *, struct nfs4_state *, fmode_t); +extern void nfs4_state_set_mode_locked(struct nfs4_state *, fmode_t); extern void nfs4_schedule_state_recovery(struct nfs_client *); extern void nfs4_schedule_state_manager(struct nfs_client *); extern int nfs4_state_mark_reclaim_nograce(struct nfs_client *clp, struct nfs4_state *state); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index fc0c9d255cf7..e8df52dc2bc5 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -323,7 +323,7 @@ static void nfs4_init_opendata_res(struct nfs4_opendata *p) } static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path, - struct nfs4_state_owner *sp, int flags, + struct nfs4_state_owner *sp, fmode_t fmode, int flags, const struct iattr *attrs) { struct dentry *parent = dget_parent(path->dentry); @@ -343,7 +343,8 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path, p->owner = sp; atomic_inc(&sp->so_count); p->o_arg.fh = NFS_FH(dir); - p->o_arg.open_flags = flags, + p->o_arg.open_flags = flags; + p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE); p->o_arg.clientid = server->nfs_client->cl_clientid; p->o_arg.id = sp->so_owner_id.id; p->o_arg.name = &p->path.dentry->d_name; @@ -399,10 +400,13 @@ static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task) return ret; } -static int can_open_cached(struct nfs4_state *state, int mode) +static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode) { int ret = 0; - switch (mode & (FMODE_READ|FMODE_WRITE|O_EXCL)) { + + if (open_mode & O_EXCL) + goto out; + switch (mode & (FMODE_READ|FMODE_WRITE)) { case FMODE_READ: ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0; break; @@ -412,12 +416,13 @@ static int can_open_cached(struct nfs4_state *state, int mode) case FMODE_READ|FMODE_WRITE: ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0; } +out: return ret; } -static int can_open_delegated(struct nfs_delegation *delegation, mode_t open_flags) +static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode) { - if ((delegation->type & open_flags) != open_flags) + if ((delegation->type & fmode) != fmode) return 0; if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags)) return 0; @@ -425,9 +430,9 @@ static int can_open_delegated(struct nfs_delegation *delegation, mode_t open_fla return 1; } -static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags) +static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode) { - switch (open_flags) { + switch (fmode) { case FMODE_WRITE: state->n_wronly++; break; @@ -437,15 +442,15 @@ static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags) case FMODE_READ|FMODE_WRITE: state->n_rdwr++; } - nfs4_state_set_mode_locked(state, state->state | open_flags); + nfs4_state_set_mode_locked(state, state->state | fmode); } -static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags) +static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode) { if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0) memcpy(state->stateid.data, stateid->data, sizeof(state->stateid.data)); memcpy(state->open_stateid.data, stateid->data, sizeof(state->open_stateid.data)); - switch (open_flags) { + switch (fmode) { case FMODE_READ: set_bit(NFS_O_RDONLY_STATE, &state->flags); break; @@ -457,14 +462,14 @@ static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid * } } -static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags) +static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode) { write_seqlock(&state->seqlock); - nfs_set_open_stateid_locked(state, stateid, open_flags); + nfs_set_open_stateid_locked(state, stateid, fmode); write_sequnlock(&state->seqlock); } -static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, const nfs4_stateid *deleg_stateid, int open_flags) +static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, const nfs4_stateid *deleg_stateid, fmode_t fmode) { /* * Protect the call to nfs4_state_set_mode_locked and @@ -476,20 +481,20 @@ static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_s set_bit(NFS_DELEGATED_STATE, &state->flags); } if (open_stateid != NULL) - nfs_set_open_stateid_locked(state, open_stateid, open_flags); + nfs_set_open_stateid_locked(state, open_stateid, fmode); write_sequnlock(&state->seqlock); spin_lock(&state->owner->so_lock); - update_open_stateflags(state, open_flags); + update_open_stateflags(state, fmode); spin_unlock(&state->owner->so_lock); } -static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, int open_flags) +static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode) { struct nfs_inode *nfsi = NFS_I(state->inode); struct nfs_delegation *deleg_cur; int ret = 0; - open_flags &= (FMODE_READ|FMODE_WRITE); + fmode &= (FMODE_READ|FMODE_WRITE); rcu_read_lock(); deleg_cur = rcu_dereference(nfsi->delegation); @@ -498,7 +503,7 @@ static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stat spin_lock(&deleg_cur->lock); if (nfsi->delegation != deleg_cur || - (deleg_cur->type & open_flags) != open_flags) + (deleg_cur->type & fmode) != fmode) goto no_delegation_unlock; if (delegation == NULL) @@ -507,7 +512,7 @@ static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stat goto no_delegation_unlock; nfs_mark_delegation_referenced(deleg_cur); - __update_open_stateid(state, open_stateid, &deleg_cur->stateid, open_flags); + __update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode); ret = 1; no_delegation_unlock: spin_unlock(&deleg_cur->lock); @@ -515,7 +520,7 @@ no_delegation: rcu_read_unlock(); if (!ret && open_stateid != NULL) { - __update_open_stateid(state, open_stateid, NULL, open_flags); + __update_open_stateid(state, open_stateid, NULL, fmode); ret = 1; } @@ -523,13 +528,13 @@ no_delegation: } -static void nfs4_return_incompatible_delegation(struct inode *inode, mode_t open_flags) +static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode) { struct nfs_delegation *delegation; rcu_read_lock(); delegation = rcu_dereference(NFS_I(inode)->delegation); - if (delegation == NULL || (delegation->type & open_flags) == open_flags) { + if (delegation == NULL || (delegation->type & fmode) == fmode) { rcu_read_unlock(); return; } @@ -542,15 +547,16 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata) struct nfs4_state *state = opendata->state; struct nfs_inode *nfsi = NFS_I(state->inode); struct nfs_delegation *delegation; - int open_mode = opendata->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL); + int open_mode = opendata->o_arg.open_flags & O_EXCL; + fmode_t fmode = opendata->o_arg.fmode; nfs4_stateid stateid; int ret = -EAGAIN; for (;;) { - if (can_open_cached(state, open_mode)) { + if (can_open_cached(state, fmode, open_mode)) { spin_lock(&state->owner->so_lock); - if (can_open_cached(state, open_mode)) { - update_open_stateflags(state, open_mode); + if (can_open_cached(state, fmode, open_mode)) { + update_open_stateflags(state, fmode); spin_unlock(&state->owner->so_lock); goto out_return_state; } @@ -559,7 +565,7 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata) rcu_read_lock(); delegation = rcu_dereference(nfsi->delegation); if (delegation == NULL || - !can_open_delegated(delegation, open_mode)) { + !can_open_delegated(delegation, fmode)) { rcu_read_unlock(); break; } @@ -572,7 +578,7 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata) ret = -EAGAIN; /* Try to update the stateid using the delegation */ - if (update_open_stateid(state, NULL, &stateid, open_mode)) + if (update_open_stateid(state, NULL, &stateid, fmode)) goto out_return_state; } out: @@ -624,7 +630,7 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data } update_open_stateid(state, &data->o_res.stateid, NULL, - data->o_arg.open_flags); + data->o_arg.fmode); iput(inode); out: return state; @@ -655,7 +661,7 @@ static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context { struct nfs4_opendata *opendata; - opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, NULL); + opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, 0, NULL); if (opendata == NULL) return ERR_PTR(-ENOMEM); opendata->state = state; @@ -663,12 +669,13 @@ static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context return opendata; } -static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, struct nfs4_state **res) +static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res) { struct nfs4_state *newstate; int ret; - opendata->o_arg.open_flags = openflags; + opendata->o_arg.open_flags = 0; + opendata->o_arg.fmode = fmode; memset(&opendata->o_res, 0, sizeof(opendata->o_res)); memset(&opendata->c_res, 0, sizeof(opendata->c_res)); nfs4_init_opendata_res(opendata); @@ -678,7 +685,7 @@ static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openf newstate = nfs4_opendata_to_nfs4_state(opendata); if (IS_ERR(newstate)) return PTR_ERR(newstate); - nfs4_close_state(&opendata->path, newstate, openflags); + nfs4_close_state(&opendata->path, newstate, fmode); *res = newstate; return 0; } @@ -734,7 +741,7 @@ static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state { struct nfs_delegation *delegation; struct nfs4_opendata *opendata; - int delegation_type = 0; + fmode_t delegation_type = 0; int status; opendata = nfs4_open_recoverdata_alloc(ctx, state); @@ -847,7 +854,7 @@ static void nfs4_open_confirm_release(void *calldata) goto out_free; state = nfs4_opendata_to_nfs4_state(data); if (!IS_ERR(state)) - nfs4_close_state(&data->path, state, data->o_arg.open_flags); + nfs4_close_state(&data->path, state, data->o_arg.fmode); out_free: nfs4_opendata_put(data); } @@ -911,7 +918,7 @@ static void nfs4_open_prepare(struct rpc_task *task, void *calldata) if (data->state != NULL) { struct nfs_delegation *delegation; - if (can_open_cached(data->state, data->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL))) + if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags)) goto out_no_action; rcu_read_lock(); delegation = rcu_dereference(NFS_I(data->state->inode)->delegation); @@ -980,7 +987,7 @@ static void nfs4_open_release(void *calldata) goto out_free; state = nfs4_opendata_to_nfs4_state(data); if (!IS_ERR(state)) - nfs4_close_state(&data->path, state, data->o_arg.open_flags); + nfs4_close_state(&data->path, state, data->o_arg.fmode); out_free: nfs4_opendata_put(data); } @@ -1135,7 +1142,7 @@ static inline void nfs4_exclusive_attrset(struct nfs4_opendata *opendata, struct /* * Returns a referenced nfs4_state */ -static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res) +static int _nfs4_do_open(struct inode *dir, struct path *path, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res) { struct nfs4_state_owner *sp; struct nfs4_state *state = NULL; @@ -1153,9 +1160,9 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct if (status != 0) goto err_put_state_owner; if (path->dentry->d_inode != NULL) - nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE)); + nfs4_return_incompatible_delegation(path->dentry->d_inode, fmode); status = -ENOMEM; - opendata = nfs4_opendata_alloc(path, sp, flags, sattr); + opendata = nfs4_opendata_alloc(path, sp, fmode, flags, sattr); if (opendata == NULL) goto err_put_state_owner; @@ -1187,14 +1194,14 @@ out_err: } -static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred) +static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred) { struct nfs4_exception exception = { }; struct nfs4_state *res; int status; do { - status = _nfs4_do_open(dir, path, flags, sattr, cred, &res); + status = _nfs4_do_open(dir, path, fmode, flags, sattr, cred, &res); if (status == 0) break; /* NOTE: BAD_SEQID means the server and client disagree about the @@ -1332,7 +1339,7 @@ static void nfs4_close_done(struct rpc_task *task, void *data) case -NFS4ERR_OLD_STATEID: case -NFS4ERR_BAD_STATEID: case -NFS4ERR_EXPIRED: - if (calldata->arg.open_flags == 0) + if (calldata->arg.fmode == 0) break; default: if (nfs4_async_handle_error(task, server, state) == -EAGAIN) { @@ -1374,10 +1381,10 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data) nfs_fattr_init(calldata->res.fattr); if (test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0) { task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE]; - calldata->arg.open_flags = FMODE_READ; + calldata->arg.fmode = FMODE_READ; } else if (test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0) { task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE]; - calldata->arg.open_flags = FMODE_WRITE; + calldata->arg.fmode = FMODE_WRITE; } calldata->timestamp = jiffies; rpc_call_start(task); @@ -1430,7 +1437,7 @@ int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait) calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid); if (calldata->arg.seqid == NULL) goto out_free_calldata; - calldata->arg.open_flags = 0; + calldata->arg.fmode = 0; calldata->arg.bitmask = server->attr_bitmask; calldata->res.fattr = &calldata->fattr; calldata->res.seqid = calldata->arg.seqid; @@ -1457,13 +1464,13 @@ out: return status; } -static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state) +static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state, fmode_t fmode) { struct file *filp; int ret; /* If the open_intent is for execute, we have an extra check to make */ - if (nd->intent.open.flags & FMODE_EXEC) { + if (fmode & FMODE_EXEC) { ret = nfs_may_open(state->inode, state->owner->so_cred, nd->intent.open.flags); @@ -1479,7 +1486,7 @@ static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct } ret = PTR_ERR(filp); out_close: - nfs4_close_sync(path, state, nd->intent.open.flags); + nfs4_close_sync(path, state, fmode & (FMODE_READ|FMODE_WRITE)); return ret; } @@ -1495,6 +1502,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd) struct rpc_cred *cred; struct nfs4_state *state; struct dentry *res; + fmode_t fmode = nd->intent.open.flags & (FMODE_READ | FMODE_WRITE | FMODE_EXEC); if (nd->flags & LOOKUP_CREATE) { attr.ia_mode = nd->intent.open.create_mode; @@ -1512,7 +1520,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd) parent = dentry->d_parent; /* Protect against concurrent sillydeletes */ nfs_block_sillyrename(parent); - state = nfs4_do_open(dir, &path, nd->intent.open.flags, &attr, cred); + state = nfs4_do_open(dir, &path, fmode, nd->intent.open.flags, &attr, cred); put_rpccred(cred); if (IS_ERR(state)) { if (PTR_ERR(state) == -ENOENT) { @@ -1527,7 +1535,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd) path.dentry = res; nfs_set_verifier(path.dentry, nfs_save_change_attribute(dir)); nfs_unblock_sillyrename(parent); - nfs4_intent_set_file(nd, &path, state); + nfs4_intent_set_file(nd, &path, state, fmode); return res; } @@ -1540,11 +1548,12 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st }; struct rpc_cred *cred; struct nfs4_state *state; + fmode_t fmode = openflags & (FMODE_READ | FMODE_WRITE); cred = rpc_lookup_cred(); if (IS_ERR(cred)) return PTR_ERR(cred); - state = nfs4_do_open(dir, &path, openflags, NULL, cred); + state = nfs4_do_open(dir, &path, fmode, openflags, NULL, cred); put_rpccred(cred); if (IS_ERR(state)) { switch (PTR_ERR(state)) { @@ -1561,10 +1570,10 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st } if (state->inode == dentry->d_inode) { nfs_set_verifier(dentry, nfs_save_change_attribute(dir)); - nfs4_intent_set_file(nd, &path, state); + nfs4_intent_set_file(nd, &path, state, fmode); return 1; } - nfs4_close_sync(&path, state, openflags); + nfs4_close_sync(&path, state, fmode); out_drop: d_drop(dentry); return 0; @@ -1990,6 +1999,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, }; struct nfs4_state *state; struct rpc_cred *cred; + fmode_t fmode = flags & (FMODE_READ | FMODE_WRITE); int status = 0; cred = rpc_lookup_cred(); @@ -1997,7 +2007,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, status = PTR_ERR(cred); goto out; } - state = nfs4_do_open(dir, &path, flags, sattr, cred); + state = nfs4_do_open(dir, &path, fmode, flags, sattr, cred); d_drop(dentry); if (IS_ERR(state)) { status = PTR_ERR(state); @@ -2013,9 +2023,9 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, nfs_post_op_update_inode(state->inode, &fattr); } if (status == 0 && (nd->flags & LOOKUP_OPEN) != 0) - status = nfs4_intent_set_file(nd, &path, state); + status = nfs4_intent_set_file(nd, &path, state, fmode); else - nfs4_close_sync(&path, state, flags); + nfs4_close_sync(&path, state, fmode); out_putcred: put_rpccred(cred); out: diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index cd16b301f13c..2022fe47966f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -363,18 +363,18 @@ nfs4_alloc_open_state(void) } void -nfs4_state_set_mode_locked(struct nfs4_state *state, mode_t mode) +nfs4_state_set_mode_locked(struct nfs4_state *state, fmode_t fmode) { - if (state->state == mode) + if (state->state == fmode) return; /* NB! List reordering - see the reclaim code for why. */ - if ((mode & FMODE_WRITE) != (state->state & FMODE_WRITE)) { - if (mode & FMODE_WRITE) + if ((fmode & FMODE_WRITE) != (state->state & FMODE_WRITE)) { + if (fmode & FMODE_WRITE) list_move(&state->open_states, &state->owner->so_states); else list_move_tail(&state->open_states, &state->owner->so_states); } - state->state = mode; + state->state = fmode; } static struct nfs4_state * @@ -454,16 +454,16 @@ void nfs4_put_open_state(struct nfs4_state *state) /* * Close the current file. */ -static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mode, int wait) +static void __nfs4_close(struct path *path, struct nfs4_state *state, fmode_t fmode, int wait) { struct nfs4_state_owner *owner = state->owner; int call_close = 0; - int newstate; + fmode_t newstate; atomic_inc(&owner->so_count); /* Protect against nfs4_find_state() */ spin_lock(&owner->so_lock); - switch (mode & (FMODE_READ | FMODE_WRITE)) { + switch (fmode & (FMODE_READ | FMODE_WRITE)) { case FMODE_READ: state->n_rdonly--; break; @@ -498,14 +498,14 @@ static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mod nfs4_do_close(path, state, wait); } -void nfs4_close_state(struct path *path, struct nfs4_state *state, mode_t mode) +void nfs4_close_state(struct path *path, struct nfs4_state *state, fmode_t fmode) { - __nfs4_close(path, state, mode, 0); + __nfs4_close(path, state, fmode, 0); } -void nfs4_close_sync(struct path *path, struct nfs4_state *state, mode_t mode) +void nfs4_close_sync(struct path *path, struct nfs4_state *state, fmode_t fmode) { - __nfs4_close(path, state, mode, 1); + __nfs4_close(path, state, fmode, 1); } /* diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 879911c99030..d8ddfc5467d6 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -953,12 +953,12 @@ static int encode_lookup(struct xdr_stream *xdr, const struct qstr *name) return 0; } -static void encode_share_access(struct xdr_stream *xdr, int open_flags) +static void encode_share_access(struct xdr_stream *xdr, fmode_t fmode) { __be32 *p; RESERVE_SPACE(8); - switch (open_flags & (FMODE_READ|FMODE_WRITE)) { + switch (fmode & (FMODE_READ|FMODE_WRITE)) { case FMODE_READ: WRITE32(NFS4_SHARE_ACCESS_READ); break; @@ -969,7 +969,7 @@ static void encode_share_access(struct xdr_stream *xdr, int open_flags) WRITE32(NFS4_SHARE_ACCESS_BOTH); break; default: - BUG(); + WRITE32(0); } WRITE32(0); /* for linux, share_deny = 0 always */ } @@ -984,7 +984,7 @@ static inline void encode_openhdr(struct xdr_stream *xdr, const struct nfs_opena RESERVE_SPACE(8); WRITE32(OP_OPEN); WRITE32(arg->seqid->sequence->counter); - encode_share_access(xdr, arg->open_flags); + encode_share_access(xdr, arg->fmode); RESERVE_SPACE(28); WRITE64(arg->clientid); WRITE32(16); @@ -1112,7 +1112,7 @@ static int encode_open_downgrade(struct xdr_stream *xdr, const struct nfs_closea WRITE32(OP_OPEN_DOWNGRADE); WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE); WRITE32(arg->seqid->sequence->counter); - encode_share_access(xdr, arg->open_flags); + encode_share_access(xdr, arg->fmode); return 0; } diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 8d71d7b7c78a..b8d9c6dd4f63 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -83,7 +83,7 @@ struct nfs_open_context { struct rpc_cred *cred; struct nfs4_state *state; fl_owner_t lockowner; - int mode; + fmode_t mode; unsigned long flags; #define NFS_CONTEXT_ERROR_WRITE (0) @@ -342,7 +342,7 @@ extern int nfs_setattr(struct dentry *, struct iattr *); extern void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr); extern struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx); extern void put_nfs_open_context(struct nfs_open_context *ctx); -extern struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode); +extern struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, fmode_t mode); extern u64 nfs_compat_user_ino64(u64 fileid); extern void nfs_fattr_init(struct nfs_fattr *fattr); diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 32c1a0ecdbff..a550b528319f 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -120,6 +120,7 @@ struct nfs_openargs { const struct nfs_fh * fh; struct nfs_seqid * seqid; int open_flags; + fmode_t fmode; __u64 clientid; __u64 id; union { @@ -171,7 +172,7 @@ struct nfs_closeargs { struct nfs_fh * fh; nfs4_stateid * stateid; struct nfs_seqid * seqid; - int open_flags; + fmode_t fmode; const u32 * bitmask; }; -- cgit v1.2.3-70-g09d2 From 64672d55d93c26fb4035fd1a84a803cbc09cb058 Mon Sep 17 00:00:00 2001 From: Peter Staubach Date: Tue, 23 Dec 2008 15:21:56 -0500 Subject: optimize attribute timeouts for "noac" and "actimeo=0" Hi. I've been looking at a bugzilla which describes a problem where a customer was advised to use either the "noac" or "actimeo=0" mount options to solve a consistency problem that they were seeing in the file attributes. It turned out that this solution did not work reliably for them because sometimes, the local attribute cache was believed to be valid and not timed out. (With an attribute cache timeout of 0, the cache should always appear to be timed out.) In looking at this situation, it appears to me that the problem is that the attribute cache timeout code has an off-by-one error in it. It is assuming that the cache is valid in the region, [read_cache_jiffies, read_cache_jiffies + attrtimeo]. The cache should be considered valid only in the region, [read_cache_jiffies, read_cache_jiffies + attrtimeo). With this change, the options, "noac" and "actimeo=0", work as originally expected. This problem was previously addressed by special casing the attrtimeo == 0 case. However, since the problem is only an off- by-one error, the cleaner solution is address the off-by-one error and thus, not require the special case. Thanx... ps Signed-off-by: Peter Staubach Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 2 +- fs/nfs/inode.c | 11 ++--------- include/linux/jiffies.h | 10 ++++++++++ include/linux/nfs_fs.h | 5 ++++- net/sunrpc/auth.c | 2 +- 5 files changed, 18 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index ed7024c34885..e35c8199f82f 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1798,7 +1798,7 @@ static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, str if (cache == NULL) goto out; if (!nfs_have_delegation(inode, FMODE_READ) && - !time_in_range(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) + !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) goto out_stale; res->jiffies = cache->jiffies; res->cred = cache->cred; diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 1cf39202c3ea..0c381686171e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -712,14 +712,7 @@ int nfs_attribute_timeout(struct inode *inode) if (nfs_have_delegation(inode, FMODE_READ)) return 0; - /* - * Special case: if the attribute timeout is set to 0, then always - * treat the cache as having expired (unless holding - * a delegation). - */ - if (nfsi->attrtimeo == 0) - return 1; - return !time_in_range(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo); + return !time_in_range_open(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo); } /** @@ -1182,7 +1175,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) nfsi->attrtimeo_timestamp = now; nfsi->attr_gencount = nfs_inc_attr_generation_counter(); } else { - if (!time_in_range(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) { + if (!time_in_range_open(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) { if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode)) nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode); nfsi->attrtimeo_timestamp = now; diff --git a/include/linux/jiffies.h b/include/linux/jiffies.h index abb6ac639e8e..1a9cf78bfce5 100644 --- a/include/linux/jiffies.h +++ b/include/linux/jiffies.h @@ -115,10 +115,20 @@ static inline u64 get_jiffies_64(void) ((long)(a) - (long)(b) >= 0)) #define time_before_eq(a,b) time_after_eq(b,a) +/* + * Calculate whether a is in the range of [b, c]. + */ #define time_in_range(a,b,c) \ (time_after_eq(a,b) && \ time_before_eq(a,c)) +/* + * Calculate whether a is in the range of [b, c). + */ +#define time_in_range_open(a,b,c) \ + (time_after_eq(a,b) && \ + time_before(a,c)) + /* Same as above, but does so with platform independent 64bit types. * These must be used when utilizing jiffies_64 (i.e. return value of * get_jiffies_64() */ diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index b8d9c6dd4f63..db867b04ac3c 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -130,7 +130,10 @@ struct nfs_inode { * * We need to revalidate the cached attrs for this inode if * - * jiffies - read_cache_jiffies > attrtimeo + * jiffies - read_cache_jiffies >= attrtimeo + * + * Please note the comparison is greater than or equal + * so that zero timeout values can be specified. */ unsigned long read_cache_jiffies; unsigned long attrtimeo; diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c index 6e28744b1709..050e4e84d9e3 100644 --- a/net/sunrpc/auth.c +++ b/net/sunrpc/auth.c @@ -234,7 +234,7 @@ rpcauth_prune_expired(struct list_head *free, int nr_to_scan) list_for_each_entry_safe(cred, next, &cred_unused, cr_lru) { /* Enforce a 60 second garbage collection moratorium */ - if (time_in_range(cred->cr_expire, expired, jiffies) && + if (time_in_range_open(cred->cr_expire, expired, jiffies) && test_bit(RPCAUTH_CRED_HASHED, &cred->cr_flags) != 0) continue; -- cgit v1.2.3-70-g09d2 From c977a2ef40a38c45537ad03823d0a004f06373f0 Mon Sep 17 00:00:00 2001 From: Benny Halevy Date: Tue, 23 Dec 2008 16:06:13 -0500 Subject: sunrpc: get rid of rpc_rqst.rq_bufsize rq_bufsize is not used. Signed-off-by: Mike Sager Signed-off-by: Benny Halevy Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 4d80a118d538..11fc71d50c1e 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -76,8 +76,7 @@ struct rpc_rqst { struct list_head rq_list; __u32 * rq_buffer; /* XDR encode buffer */ - size_t rq_bufsize, - rq_callsize, + size_t rq_callsize, rq_rcvsize; struct xdr_buf rq_private_buf; /* The receive buffer -- cgit v1.2.3-70-g09d2 From c381060869317b3c84430d4f54965d409cbfe65f Mon Sep 17 00:00:00 2001 From: "\\\"J. Bruce Fields\\" Date: Tue, 23 Dec 2008 16:08:32 -0500 Subject: rpc: add an rpc_pipe_open method We want to transition to a new gssd upcall which is text-based and more easily extensible. To simplify upgrades, as well as testing and debugging, it will help if we can upgrade gssd (to a version which understands the new upcall) without having to choose at boot (or module-load) time whether we want the new or the old upcall. We will do this by providing two different pipes: one named, as currently, after the mechanism (normally "krb5"), and supporting the old upcall. One named "gssd" and supporting the new upcall version. We allow gssd to indicate which version it supports by its choice of which pipe to open. As we have no interest in supporting *simultaneous* use of both versions, we'll forbid opening both pipes at the same time. So, add a new pipe_open callback to the rpc_pipefs api, which the gss code can use to track which pipes have been open, and to refuse opens of incompatible pipes. We only need this to be called on the first open of a given pipe. Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust --- include/linux/sunrpc/rpc_pipe_fs.h | 1 + net/sunrpc/rpc_pipe.c | 22 +++++++++++++++------- 2 files changed, 16 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sunrpc/rpc_pipe_fs.h b/include/linux/sunrpc/rpc_pipe_fs.h index 51b977a4ca20..cea764c2359f 100644 --- a/include/linux/sunrpc/rpc_pipe_fs.h +++ b/include/linux/sunrpc/rpc_pipe_fs.h @@ -15,6 +15,7 @@ struct rpc_pipe_ops { ssize_t (*upcall)(struct file *, struct rpc_pipe_msg *, char __user *, size_t); ssize_t (*downcall)(struct file *, const char __user *, size_t); void (*release_pipe)(struct inode *); + int (*open_pipe)(struct inode *); void (*destroy_msg)(struct rpc_pipe_msg *); }; diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index 55b2049834c4..c9b57f47108c 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -169,16 +169,24 @@ static int rpc_pipe_open(struct inode *inode, struct file *filp) { struct rpc_inode *rpci = RPC_I(inode); + int first_open; int res = -ENXIO; mutex_lock(&inode->i_mutex); - if (rpci->ops != NULL) { - if (filp->f_mode & FMODE_READ) - rpci->nreaders ++; - if (filp->f_mode & FMODE_WRITE) - rpci->nwriters ++; - res = 0; + if (rpci->ops == NULL) + goto out; + first_open = rpci->nreaders == 0 && rpci->nwriters == 0; + if (first_open && rpci->ops->open_pipe) { + res = rpci->ops->open_pipe(inode); + if (res) + goto out; } + if (filp->f_mode & FMODE_READ) + rpci->nreaders++; + if (filp->f_mode & FMODE_WRITE) + rpci->nwriters++; + res = 0; +out: mutex_unlock(&inode->i_mutex); return res; } @@ -748,7 +756,7 @@ rpc_rmdir(struct dentry *dentry) * @name: name of pipe * @private: private data to associate with the pipe, for the caller's use * @ops: operations defining the behavior of the pipe: upcall, downcall, - * release_pipe, and destroy_msg. + * release_pipe, open_pipe, and destroy_msg. * @flags: rpc_inode flags * * Data is made available for userspace to read by calls to -- cgit v1.2.3-70-g09d2 From 68e76ad0baf8f5d5060377c2423ee6eed5c63057 Mon Sep 17 00:00:00 2001 From: Olga Kornievskaia Date: Tue, 23 Dec 2008 16:17:15 -0500 Subject: nfsd: pass client principal name in rsc downcall Two principals are involved in krb5 authentication: the target, who we authenticate *to* (normally the name of the server, like nfs/server.citi.umich.edu@CITI.UMICH.EDU), and the source, we we authenticate *as* (normally a user, like bfields@UMICH.EDU) In the case of NFSv4 callbacks, the target of the callback should be the source of the client's setclientid call, and the source should be the nfs server's own principal. Therefore we allow svcgssd to pass down the name of the principal that just authenticated, so that on setclientid we can store that principal name with the new client, to be used later on callbacks. Signed-off-by: Olga Kornievskaia Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust --- fs/nfsd/nfs4state.c | 11 +++++++++++ include/linux/nfsd/state.h | 1 + include/linux/sunrpc/svcauth_gss.h | 1 + net/sunrpc/auth_gss/svcauth_gss.c | 23 +++++++++++++++++++++++ 4 files changed, 36 insertions(+) (limited to 'include/linux') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 1a052ac2bde9..f3b9a8d064f3 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -54,6 +54,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_PROC @@ -377,6 +378,7 @@ free_client(struct nfs4_client *clp) shutdown_callback_client(clp); if (clp->cl_cred.cr_group_info) put_group_info(clp->cl_cred.cr_group_info); + kfree(clp->cl_principal); kfree(clp->cl_name.data); kfree(clp); } @@ -696,6 +698,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, unsigned int strhashval; struct nfs4_client *conf, *unconf, *new; __be32 status; + char *princ; char dname[HEXDIR_LEN]; if (!check_name(clname)) @@ -783,6 +786,14 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, } copy_verf(new, &clverifier); new->cl_addr = sin->sin_addr.s_addr; + princ = svc_gss_principal(rqstp); + if (princ) { + new->cl_principal = kstrdup(princ, GFP_KERNEL); + if (new->cl_principal == NULL) { + free_client(new); + goto out; + } + } copy_cred(&new->cl_cred, &rqstp->rq_cred); gen_confirm(new); gen_callback(new, setclid); diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index d0fe2e378452..ce7cbf4b7c93 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -124,6 +124,7 @@ struct nfs4_client { nfs4_verifier cl_verifier; /* generated by client */ time_t cl_time; /* time of last lease renewal */ __be32 cl_addr; /* client ipaddress */ + char *cl_principal; /* setclientid principal name */ struct svc_cred cl_cred; /* setclientid principal */ clientid_t cl_clientid; /* generated by server */ nfs4_verifier cl_confirm; /* generated by server */ diff --git a/include/linux/sunrpc/svcauth_gss.h b/include/linux/sunrpc/svcauth_gss.h index c9165d9771a8..ca7d725861fc 100644 --- a/include/linux/sunrpc/svcauth_gss.h +++ b/include/linux/sunrpc/svcauth_gss.h @@ -20,6 +20,7 @@ int gss_svc_init(void); void gss_svc_shutdown(void); int svcauth_gss_register_pseudoflavor(u32 pseudoflavor, char * name); u32 svcauth_gss_flavor(struct auth_domain *dom); +char *svc_gss_principal(struct svc_rqst *); #endif /* __KERNEL__ */ #endif /* _LINUX_SUNRPC_SVCAUTH_GSS_H */ diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c index 12803da95dc4..e9baa6ebb1dd 100644 --- a/net/sunrpc/auth_gss/svcauth_gss.c +++ b/net/sunrpc/auth_gss/svcauth_gss.c @@ -332,6 +332,7 @@ struct rsc { struct svc_cred cred; struct gss_svc_seq_data seqdata; struct gss_ctx *mechctx; + char *client_name; }; static struct cache_head *rsc_table[RSC_HASHMAX]; @@ -346,6 +347,7 @@ static void rsc_free(struct rsc *rsci) gss_delete_sec_context(&rsci->mechctx); if (rsci->cred.cr_group_info) put_group_info(rsci->cred.cr_group_info); + kfree(rsci->client_name); } static void rsc_put(struct kref *ref) @@ -383,6 +385,7 @@ rsc_init(struct cache_head *cnew, struct cache_head *ctmp) tmp->handle.data = NULL; new->mechctx = NULL; new->cred.cr_group_info = NULL; + new->client_name = NULL; } static void @@ -397,6 +400,8 @@ update_rsc(struct cache_head *cnew, struct cache_head *ctmp) spin_lock_init(&new->seqdata.sd_lock); new->cred = tmp->cred; tmp->cred.cr_group_info = NULL; + new->client_name = tmp->client_name; + tmp->client_name = NULL; } static struct cache_head * @@ -486,6 +491,15 @@ static int rsc_parse(struct cache_detail *cd, status = gss_import_sec_context(buf, len, gm, &rsci.mechctx); if (status) goto out; + + /* get client name */ + len = qword_get(&mesg, buf, mlen); + if (len > 0) { + rsci.client_name = kstrdup(buf, GFP_KERNEL); + if (!rsci.client_name) + goto out; + } + } rsci.h.expiry_time = expiry; rscp = rsc_update(&rsci, rscp); @@ -913,6 +927,15 @@ struct gss_svc_data { struct rsc *rsci; }; +char *svc_gss_principal(struct svc_rqst *rqstp) +{ + struct gss_svc_data *gd = (struct gss_svc_data *)rqstp->rq_auth_data; + + if (gd && gd->rsci) + return gd->rsci->client_name; + return NULL; +} + static int svcauth_gss_set_client(struct svc_rqst *rqstp) { -- cgit v1.2.3-70-g09d2 From 608207e8884e083ad8b8d33eda868da70f0d63e8 Mon Sep 17 00:00:00 2001 From: Olga Kornievskaia Date: Tue, 23 Dec 2008 16:17:40 -0500 Subject: rpc: pass target name down to rpc level on callbacks The rpc client needs to know the principal that the setclientid was done as, so it can tell gssd who to authenticate to. Signed-off-by: Olga Kornievskaia Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust --- fs/nfsd/nfs4callback.c | 6 ++++++ include/linux/sunrpc/clnt.h | 2 ++ net/sunrpc/clnt.c | 16 ++++++++++++++++ 3 files changed, 24 insertions(+) (limited to 'include/linux') diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 094747a1227c..3ca141782145 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -384,6 +384,7 @@ static int do_probe_callback(void *data) .version = nfs_cb_version[1]->number, .authflavor = RPC_AUTH_UNIX, /* XXX: need AUTH_GSS... */ .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET), + .client_name = clp->cl_principal, }; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], @@ -392,6 +393,11 @@ static int do_probe_callback(void *data) struct rpc_clnt *client; int status; + if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5)) { + status = nfserr_cb_path_down; + goto out_err; + } + /* Initialize address */ memset(&addr, 0, sizeof(addr)); addr.sin_family = AF_INET; diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 6f0ee1b84a4f..c39a21040dcb 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -58,6 +58,7 @@ struct rpc_clnt { struct rpc_timeout cl_timeout_default; struct rpc_program * cl_program; char cl_inline_name[32]; + char *cl_principal; /* target to authenticate to */ }; /* @@ -108,6 +109,7 @@ struct rpc_create_args { u32 version; rpc_authflavor_t authflavor; unsigned long flags; + char *client_name; }; /* Values for "flags" field */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 4895c341e46d..347f2a25abb6 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -197,6 +197,12 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, stru clnt->cl_rtt = &clnt->cl_rtt_default; rpc_init_rtt(&clnt->cl_rtt_default, clnt->cl_timeout->to_initval); + clnt->cl_principal = NULL; + if (args->client_name) { + clnt->cl_principal = kstrdup(args->client_name, GFP_KERNEL); + if (!clnt->cl_principal) + goto out_no_principal; + } kref_init(&clnt->cl_kref); @@ -226,6 +232,8 @@ out_no_auth: rpc_put_mount(); } out_no_path: + kfree(clnt->cl_principal); +out_no_principal: rpc_free_iostats(clnt->cl_metrics); out_no_stats: if (clnt->cl_server != clnt->cl_inline_name) @@ -354,6 +362,11 @@ rpc_clone_client(struct rpc_clnt *clnt) new->cl_metrics = rpc_alloc_iostats(clnt); if (new->cl_metrics == NULL) goto out_no_stats; + if (clnt->cl_principal) { + new->cl_principal = kstrdup(clnt->cl_principal, GFP_KERNEL); + if (new->cl_principal == NULL) + goto out_no_principal; + } kref_init(&new->cl_kref); err = rpc_setup_pipedir(new, clnt->cl_program->pipe_dir_name); if (err != 0) @@ -366,6 +379,8 @@ rpc_clone_client(struct rpc_clnt *clnt) rpciod_up(); return new; out_no_path: + kfree(new->cl_principal); +out_no_principal: rpc_free_iostats(new->cl_metrics); out_no_stats: kfree(new); @@ -417,6 +432,7 @@ rpc_free_client(struct kref *kref) out_free: rpc_unregister_client(clnt); rpc_free_iostats(clnt->cl_metrics); + kfree(clnt->cl_principal); clnt->cl_metrics = NULL; xprt_put(clnt->cl_xprt); rpciod_down(); -- cgit v1.2.3-70-g09d2 From 61054b14d545e257b9415d5ca0cd5f43762b4d0c Mon Sep 17 00:00:00 2001 From: Olga Kornievskaia Date: Tue, 23 Dec 2008 16:19:00 -0500 Subject: nfsd: support callbacks with gss flavors This patch adds server-side support for callbacks other than AUTH_SYS. Signed-off-by: Olga Kornievskaia Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust --- fs/nfsd/nfs4callback.c | 3 ++- fs/nfsd/nfs4state.c | 1 + include/linux/nfsd/state.h | 1 + net/sunrpc/rpc_pipe.c | 5 +++++ 4 files changed, 9 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 3ca141782145..6d7d8c02c197 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -358,6 +358,7 @@ static struct rpc_program cb_program = { .nrvers = ARRAY_SIZE(nfs_cb_version), .version = nfs_cb_version, .stats = &cb_stats, + .pipe_dir_name = "/nfsd4_cb", }; /* Reference counting, callback cleanup, etc., all look racy as heck. @@ -382,7 +383,7 @@ static int do_probe_callback(void *data) .program = &cb_program, .prognumber = cb->cb_prog, .version = nfs_cb_version[1]->number, - .authflavor = RPC_AUTH_UNIX, /* XXX: need AUTH_GSS... */ + .authflavor = clp->cl_flavor, .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET), .client_name = clp->cl_principal, }; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index f3b9a8d064f3..07db31568ac9 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -786,6 +786,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, } copy_verf(new, &clverifier); new->cl_addr = sin->sin_addr.s_addr; + new->cl_flavor = rqstp->rq_flavor; princ = svc_gss_principal(rqstp); if (princ) { new->cl_principal = kstrdup(princ, GFP_KERNEL); diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index ce7cbf4b7c93..128298c0362d 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -124,6 +124,7 @@ struct nfs4_client { nfs4_verifier cl_verifier; /* generated by client */ time_t cl_time; /* time of last lease renewal */ __be32 cl_addr; /* client ipaddress */ + u32 cl_flavor; /* setclientid pseudoflavor */ char *cl_principal; /* setclientid principal name */ struct svc_cred cl_cred; /* setclientid principal */ clientid_t cl_clientid; /* generated by server */ diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index 3105efbb182d..192453248870 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -407,6 +407,7 @@ enum { RPCAUTH_nfs, RPCAUTH_portmap, RPCAUTH_statd, + RPCAUTH_nfsd4_cb, RPCAUTH_RootEOF }; @@ -440,6 +441,10 @@ static struct rpc_filelist files[] = { .name = "statd", .mode = S_IFDIR | S_IRUGO | S_IXUGO, }, + [RPCAUTH_nfsd4_cb] = { + .name = "nfsd4_cb", + .mode = S_IFDIR | S_IRUGO | S_IXUGO, + }, }; enum { -- cgit v1.2.3-70-g09d2