From 80492e7d49bee0a280a84a39075a7857b92836b2 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Tue, 21 Apr 2009 16:08:39 +0200 Subject: rpcgss: remove redundant test on unsigned Signed-off-by: Roel Kluin Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 5839b22..6eb9181 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -464,16 +464,11 @@ static int secinfo_parse(char **mesg, char *buf, struct svc_export *exp) if (err) return err; /* - * Just a quick sanity check; we could also try to check - * whether this pseudoflavor is supported, but at worst - * an unsupported pseudoflavor on the export would just - * be a pseudoflavor that won't match the flavor of any - * authenticated request. The administrator will - * probably discover the problem when someone fails to - * authenticate. + * XXX: It would be nice to also check whether this + * pseudoflavor is supported, so we can discover the + * problem at export time instead of when a client fails + * to authenticate. */ - if (f->pseudoflavor < 0) - return -EINVAL; err = get_int(mesg, &f->flags); if (err) return err; -- cgit v0.10.2 From a9e61e25f9d2e7e43bf17625f5cb56c9e0a89b17 Mon Sep 17 00:00:00 2001 From: Felix Blyakher Date: Tue, 31 Mar 2009 15:12:56 -0500 Subject: lockd: call locks_release_private to cleanup per-filesystem state For every lock request lockd creates a new file_lock object in nlmsvc_setgrantargs() by copying the passed in file_lock with locks_copy_lock(). A filesystem can attach it's own lock_operations vector to the file_lock. It has to be cleaned up at the end of the file_lock's life. However, lockd doesn't do it today, yet it asserts in nlmclnt_release_lockargs() that the per-filesystem state is clean. This patch fixes it by exporting locks_release_private() and adding it to nlmsvc_freegrantargs(), to be symmetrical to creating a file_lock in nlmsvc_setgrantargs(). Signed-off-by: Felix Blyakher Signed-off-by: J. Bruce Fields diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index 83ee342..e577a78 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -326,6 +326,8 @@ static void nlmsvc_freegrantargs(struct nlm_rqst *call) { if (call->a_args.lock.oh.data != call->a_owner) kfree(call->a_args.lock.oh.data); + + locks_release_private(&call->a_args.lock.fl); } /* diff --git a/fs/locks.c b/fs/locks.c index ec3deea..b6440f5 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -151,7 +151,7 @@ static struct file_lock *locks_alloc_lock(void) return kmem_cache_alloc(filelock_cache, GFP_KERNEL); } -static void locks_release_private(struct file_lock *fl) +void locks_release_private(struct file_lock *fl) { if (fl->fl_ops) { if (fl->fl_ops->fl_release_private) @@ -165,6 +165,7 @@ static void locks_release_private(struct file_lock *fl) } } +EXPORT_SYMBOL_GPL(locks_release_private); /* Free a lock which is not in use. */ static void locks_free_lock(struct file_lock *fl) diff --git a/include/linux/fs.h b/include/linux/fs.h index 5bed436..5ba615e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1108,6 +1108,7 @@ extern void locks_copy_lock(struct file_lock *, struct file_lock *); extern void __locks_copy_lock(struct file_lock *, const struct file_lock *); extern void locks_remove_posix(struct file *, fl_owner_t); extern void locks_remove_flock(struct file *); +extern void locks_release_private(struct file_lock *); extern void posix_test_lock(struct file *, struct file_lock *); extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *); extern int posix_lock_file_wait(struct file *, struct file_lock *); -- cgit v0.10.2 From 78155ed75f470710f2aecb3e75e3d97107ba8374 Mon Sep 17 00:00:00 2001 From: Bian Naimeng Date: Wed, 22 Apr 2009 18:25:37 +0800 Subject: nfsd4: distinguish expired from stale stateids If we encode the time of client creation into the stateid instead of the time of server boot, then we can determine whether that stateid is from a previous instance of the a server, or from a client that has expired, and return an appropriate error to the client. Signed-off-by: Bian Naimeng Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index c65a27b..74e822e 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -206,7 +206,7 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f dp->dl_recall.cbr_dp = NULL; dp->dl_recall.cbr_ident = cb->cb_ident; dp->dl_recall.cbr_trunc = 0; - dp->dl_stateid.si_boot = boot_time; + dp->dl_stateid.si_boot = get_seconds(); dp->dl_stateid.si_stateownerid = current_delegid++; dp->dl_stateid.si_fileid = 0; dp->dl_stateid.si_generation = 0; @@ -1883,7 +1883,7 @@ init_stateid(struct nfs4_stateid *stp, struct nfs4_file *fp, struct nfsd4_open * stp->st_stateowner = sop; get_nfs4_file(fp); stp->st_file = fp; - stp->st_stateid.si_boot = boot_time; + stp->st_stateid.si_boot = get_seconds(); stp->st_stateid.si_stateownerid = sop->so_id; stp->st_stateid.si_fileid = fp->fi_id; stp->st_stateid.si_generation = 0; @@ -2739,12 +2739,42 @@ nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stateid *stp) static int STALE_STATEID(stateid_t *stateid) { - if (stateid->si_boot == boot_time) - return 0; - dprintk("NFSD: stale stateid (%08x/%08x/%08x/%08x)!\n", - stateid->si_boot, stateid->si_stateownerid, stateid->si_fileid, - stateid->si_generation); - return 1; + if (time_after((unsigned long)boot_time, + (unsigned long)stateid->si_boot)) { + dprintk("NFSD: stale stateid (%08x/%08x/%08x/%08x)!\n", + stateid->si_boot, stateid->si_stateownerid, + stateid->si_fileid, stateid->si_generation); + return 1; + } + return 0; +} + +static int +EXPIRED_STATEID(stateid_t *stateid) +{ + if (time_before((unsigned long)boot_time, + ((unsigned long)stateid->si_boot)) && + time_before((stateid->si_boot + lease_time), get_seconds())) { + dprintk("NFSD: expired stateid (%08x/%08x/%08x/%08x)!\n", + stateid->si_boot, stateid->si_stateownerid, + stateid->si_fileid, stateid->si_generation); + return 1; + } + return 0; +} + +static __be32 +stateid_error_map(stateid_t *stateid) +{ + if (STALE_STATEID(stateid)) + return nfserr_stale_stateid; + if (EXPIRED_STATEID(stateid)) + return nfserr_expired; + + dprintk("NFSD: bad stateid (%08x/%08x/%08x/%08x)!\n", + stateid->si_boot, stateid->si_stateownerid, + stateid->si_fileid, stateid->si_generation); + return nfserr_bad_stateid; } static inline int @@ -2868,8 +2898,10 @@ nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate, status = nfserr_bad_stateid; if (is_delegation_stateid(stateid)) { dp = find_delegation_stateid(ino, stateid); - if (!dp) + if (!dp) { + status = stateid_error_map(stateid); goto out; + } status = check_stateid_generation(stateid, &dp->dl_stateid, flags); if (status) @@ -2882,8 +2914,10 @@ nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate, *filpp = dp->dl_vfs_file; } else { /* open or lock stateid */ stp = find_stateid(stateid, flags); - if (!stp) + if (!stp) { + status = stateid_error_map(stateid); goto out; + } if (nfs4_check_fh(current_fh, stp)) goto out; if (!stp->st_stateowner->so_confirmed) @@ -2957,7 +2991,7 @@ nfs4_preprocess_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid, */ sop = search_close_lru(stateid->si_stateownerid, flags); if (sop == NULL) - return nfserr_bad_stateid; + return stateid_error_map(stateid); *sopp = sop; goto check_replay; } @@ -3228,8 +3262,10 @@ nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (!is_delegation_stateid(stateid)) goto out; dp = find_delegation_stateid(inode, stateid); - if (!dp) + if (!dp) { + status = stateid_error_map(stateid); goto out; + } status = check_stateid_generation(stateid, &dp->dl_stateid, flags); if (status) goto out; @@ -3456,7 +3492,7 @@ alloc_init_lock_stateid(struct nfs4_stateowner *sop, struct nfs4_file *fp, struc stp->st_stateowner = sop; get_nfs4_file(fp); stp->st_file = fp; - stp->st_stateid.si_boot = boot_time; + stp->st_stateid.si_boot = get_seconds(); stp->st_stateid.si_stateownerid = sop->so_id; stp->st_stateid.si_fileid = fp->fi_id; stp->st_stateid.si_generation = 0; -- cgit v0.10.2 From 4ba170c2bb77713e999ac6fb9ffe6ddd99d2f25a Mon Sep 17 00:00:00 2001 From: Benny Halevy Date: Tue, 7 Apr 2009 21:45:37 +0300 Subject: update Documentation/filesystems/00-INDEX with new nfsd related docs. Signed-off-by: Benny Halevy Cc: James Lentini Signed-off-by: J. Bruce Fields diff --git a/Documentation/filesystems/00-INDEX b/Documentation/filesystems/00-INDEX index 8dd6db7..f15621e 100644 --- a/Documentation/filesystems/00-INDEX +++ b/Documentation/filesystems/00-INDEX @@ -66,6 +66,10 @@ mandatory-locking.txt - info on the Linux implementation of Sys V mandatory file locking. ncpfs.txt - info on Novell Netware(tm) filesystem using NCP protocol. +nfs41-server.txt + - info on the Linux server implementation of NFSv4 minor version 1. +nfs-rdma.txt + - how to install and setup the Linux NFS/RDMA client and server software. nfsroot.txt - short guide on setting up a diskless box with NFS root filesystem. nilfs2.txt -- cgit v0.10.2 From dcf1a3573eae69937fb14462369c4d3e6f4a37f1 Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Wed, 22 Apr 2009 20:18:19 -0400 Subject: net/sunrpc/svc_xprt.c: fix sparse warnings Fix the following sparse warnings in net/sunrpc/svc_xprt.c. warning: symbol 'svc_recv' was not declared. Should it be static? warning: symbol 'svc_drop' was not declared. Should it be static? warning: symbol 'svc_send' was not declared. Should it be static? warning: symbol 'svc_close_all' was not declared. Should it be static? Signed-off-by: H Hartley Sweeten Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index c200d92..f200393 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -11,6 +11,7 @@ #include #include #include +#include #define RPCDBG_FACILITY RPCDBG_SVCXPRT -- cgit v0.10.2 From abc5c44d6284fab8fb21bcfc52c0f16f980637df Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:31:25 -0400 Subject: SUNRPC: Fix error return value of svc_addr_len() The svc_addr_len() helper function returns -EAFNOSUPPORT if it doesn't recognize the address family of the passed-in socket address. However, the return type of this function is size_t, which means -EAFNOSUPPORT is turned into a very large positive value in this case. The check in svc_udp_recvfrom() to see if the return value is less than zero therefore won't work at all. Additionally, handle_connect_req() passes this value directly to memset(). This could cause memset() to clobber a large chunk of memory if svc_addr_len() has returned an error. Currently the address family of these addresses, however, is known to be supported long before handle_connect_req() is called, so this isn't a real risk. Change the error return value of svc_addr_len() to zero, which fits in the range of size_t, and is safer to pass to memset() directly. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/include/linux/sunrpc/svc_xprt.h b/include/linux/sunrpc/svc_xprt.h index 0d9cb6e..d790c52 100644 --- a/include/linux/sunrpc/svc_xprt.h +++ b/include/linux/sunrpc/svc_xprt.h @@ -118,7 +118,7 @@ static inline unsigned short svc_addr_port(const struct sockaddr *sa) return 0; } -static inline size_t svc_addr_len(struct sockaddr *sa) +static inline size_t svc_addr_len(const struct sockaddr *sa) { switch (sa->sa_family) { case AF_INET: @@ -126,7 +126,8 @@ static inline size_t svc_addr_len(struct sockaddr *sa) case AF_INET6: return sizeof(struct sockaddr_in6); } - return -EAFNOSUPPORT; + + return 0; } static inline unsigned short svc_xprt_local_port(const struct svc_xprt *xprt) diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index af31988..8b08328 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -426,13 +426,14 @@ static int svc_udp_recvfrom(struct svc_rqst *rqstp) long all[SVC_PKTINFO_SPACE / sizeof(long)]; } buffer; struct cmsghdr *cmh = &buffer.hdr; - int err, len; struct msghdr msg = { .msg_name = svc_addr(rqstp), .msg_control = cmh, .msg_controllen = sizeof(buffer), .msg_flags = MSG_DONTWAIT, }; + size_t len; + int err; if (test_and_clear_bit(XPT_CHNGBUF, &svsk->sk_xprt.xpt_flags)) /* udp sockets need large rcvbuf as all pending @@ -464,8 +465,8 @@ static int svc_udp_recvfrom(struct svc_rqst *rqstp) return -EAGAIN; } len = svc_addr_len(svc_addr(rqstp)); - if (len < 0) - return len; + if (len == 0) + return -EAFNOSUPPORT; rqstp->rq_addrlen = len; if (skb->tstamp.tv64 == 0) { skb->tstamp = ktime_get_real(); -- cgit v0.10.2 From 4cd5dc751a5889b5b37aa88752e33a58d8f6b624 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:31:32 -0400 Subject: NFSD: Refactor transport removal out of __write_ports() Clean up: Refactor transport removal out of __write_ports() to make it easier to understand and maintain. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index af16849..2c1dce8 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -910,6 +910,31 @@ static ssize_t write_versions(struct file *file, char *buf, size_t size) return rv; } +/* + * A transport listener is removed by writing a "-", it's transport + * name, and it's port number. + */ +static ssize_t __write_ports_delxprt(char *buf) +{ + struct svc_xprt *xprt; + char transport[16]; + int port; + + if (sscanf(&buf[1], "%15s %4u", transport, &port) != 2) + return -EINVAL; + + if (port < 1 || port > USHORT_MAX || nfsd_serv == NULL) + return -EINVAL; + + xprt = svc_find_xprt(nfsd_serv, transport, AF_UNSPEC, port); + if (xprt == NULL) + return -ENOTCONN; + + svc_close_xprt(xprt); + svc_xprt_put(xprt); + return 0; +} + static ssize_t __write_ports(struct file *file, char *buf, size_t size) { if (size == 0) { @@ -984,30 +1009,10 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size) return err < 0 ? err : 0; } } - /* - * Remove a transport by writing it's transport name and port number - */ - if (buf[0] == '-' && isalpha(buf[1])) { - struct svc_xprt *xprt; - int err = -EINVAL; - char transport[16]; - int port; - if (sscanf(&buf[1], "%15s %4d", transport, &port) == 2) { - if (port < 1 || port > 65535) - return -EINVAL; - if (nfsd_serv) { - xprt = svc_find_xprt(nfsd_serv, transport, - AF_UNSPEC, port); - if (xprt) { - svc_close_xprt(xprt); - svc_xprt_put(xprt); - err = 0; - } else - err = -ENOTCONN; - } - return err < 0 ? err : 0; - } - } + + if (buf[0] == '-' && isalpha(buf[1])) + return __write_ports_delxprt(buf); + return -EINVAL; } -- cgit v0.10.2 From 4eb68c266cb1754ffa0040e882882680ece8cf34 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:31:40 -0400 Subject: NFSD: Refactor transport addition out of __write_ports() Clean up: Refactor transport addition out of __write_ports() to make it easier to understand and maintain. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 2c1dce8..748532b 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -911,6 +911,36 @@ static ssize_t write_versions(struct file *file, char *buf, size_t size) } /* + * A transport listener is added by writing it's transport name and + * a port number. + */ +static ssize_t __write_ports_addxprt(char *buf) +{ + char transport[16]; + int port, err; + + if (sscanf(buf, "%15s %4u", transport, &port) != 2) + return -EINVAL; + + if (port < 1 || port > USHORT_MAX) + return -EINVAL; + + err = nfsd_create_serv(); + if (err != 0) + return err; + + err = svc_create_xprt(nfsd_serv, transport, + PF_INET, port, SVC_SOCK_ANONYMOUS); + if (err < 0) { + /* Give a reasonable perror msg for bad transport string */ + if (err == -ENOENT) + err = -EPROTONOSUPPORT; + return err; + } + return 0; +} + +/* * A transport listener is removed by writing a "-", it's transport * name, and it's port number. */ @@ -986,29 +1016,9 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size) kfree(toclose); return len; } - /* - * Add a transport listener by writing it's transport name - */ - if (isalpha(buf[0])) { - int err; - char transport[16]; - int port; - if (sscanf(buf, "%15s %4d", transport, &port) == 2) { - if (port < 1 || port > 65535) - return -EINVAL; - err = nfsd_create_serv(); - if (!err) { - err = svc_create_xprt(nfsd_serv, - transport, PF_INET, port, - SVC_SOCK_ANONYMOUS); - if (err == -ENOENT) - /* Give a reasonable perror msg for - * bad transport string */ - err = -EPROTONOSUPPORT; - } - return err < 0 ? err : 0; - } - } + + if (isalpha(buf[0])) + return __write_ports_addxprt(buf); if (buf[0] == '-' && isalpha(buf[1])) return __write_ports_delxprt(buf); -- cgit v0.10.2 From 82d565919aa4138bc481a7473491b71385af4018 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:31:48 -0400 Subject: NFSD: Refactor portlist socket closing into a helper Clean up: Refactor the socket closing logic out of __write_ports() to make it easier to understand and maintain. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 748532b..fa268d1 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -911,6 +911,27 @@ static ssize_t write_versions(struct file *file, char *buf, size_t size) } /* + * A '-' followed by the 'name' of a socket means we close the socket. + */ +static ssize_t __write_ports_delfd(char *buf) +{ + char *toclose; + int len = 0; + + toclose = kstrdup(buf + 1, GFP_KERNEL); + if (toclose == NULL) + return -ENOMEM; + + if (nfsd_serv != NULL) + len = svc_sock_names(buf, nfsd_serv, toclose); + if (len >= 0) + lockd_down(); + + kfree(toclose); + return len; +} + +/* * A transport listener is added by writing it's transport name and * a port number. */ @@ -1004,18 +1025,9 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size) } return err < 0 ? err : 0; } - if (buf[0] == '-' && isdigit(buf[1])) { - char *toclose = kstrdup(buf+1, GFP_KERNEL); - int len = 0; - if (!toclose) - return -ENOMEM; - if (nfsd_serv) - len = svc_sock_names(buf, nfsd_serv, toclose); - if (len >= 0) - lockd_down(); - kfree(toclose); - return len; - } + + if (buf[0] == '-' && isdigit(buf[1])) + return __write_ports_delfd(buf); if (isalpha(buf[0])) return __write_ports_addxprt(buf); -- cgit v0.10.2 From 0b7c2f6fc7e06867885c7b0f256ff5bd494e0653 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:31:55 -0400 Subject: NFSD: Refactor socket creation out of __write_ports() Clean up: Refactor the socket creation logic out of __write_ports() to make it easier to understand and maintain. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index fa268d1..b6a847f 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -911,6 +911,37 @@ static ssize_t write_versions(struct file *file, char *buf, size_t size) } /* + * A single 'fd' number was written, in which case it must be for + * a socket of a supported family/protocol, and we use it as an + * nfsd listener. + */ +static ssize_t __write_ports_addfd(char *buf) +{ + char *mesg = buf; + int fd, err; + + err = get_int(&mesg, &fd); + if (err != 0 || fd < 0) + return -EINVAL; + + err = nfsd_create_serv(); + if (err != 0) + return err; + + err = svc_addsock(nfsd_serv, fd, buf); + if (err >= 0) { + err = lockd_up(); + if (err < 0) + svc_sock_names(buf + strlen(buf) + 1, nfsd_serv, buf); + + /* Decrease the count, but don't shut down the service */ + nfsd_serv->sv_nrthreads--; + } + + return err < 0 ? err : 0; +} + +/* * A '-' followed by the 'name' of a socket means we close the socket. */ static ssize_t __write_ports_delfd(char *buf) @@ -995,36 +1026,9 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size) len = svc_xprt_names(nfsd_serv, buf, 0); return len; } - /* Either a single 'fd' number is written, in which - * case it must be for a socket of a supported family/protocol, - * and we use it as an nfsd socket, or - * A '-' followed by the 'name' of a socket in which case - * we close the socket. - */ - if (isdigit(buf[0])) { - char *mesg = buf; - int fd; - int err; - err = get_int(&mesg, &fd); - if (err) - return -EINVAL; - if (fd < 0) - return -EINVAL; - err = nfsd_create_serv(); - if (!err) { - err = svc_addsock(nfsd_serv, fd, buf); - if (err >= 0) { - err = lockd_up(); - if (err < 0) - svc_sock_names(buf+strlen(buf)+1, nfsd_serv, buf); - } - /* Decrease the count, but don't shutdown the - * the service - */ - nfsd_serv->sv_nrthreads--; - } - return err < 0 ? err : 0; - } + + if (isdigit(buf[0])) + return __write_ports_addfd(buf); if (buf[0] == '-' && isdigit(buf[1])) return __write_ports_delfd(buf); -- cgit v0.10.2 From c71206a7b4e829601e44c49d8a0f4df0fdfe9302 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:03 -0400 Subject: NFSD: Note an additional requirement when passing TCP sockets to portlist User space must call listen(3) on SOCK_STREAM sockets passed into /proc/fs/nfsd/portlist, otherwise that listener is ignored. Document this. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index b6a847f..d491fa9 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -1061,7 +1061,9 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size) * buf: C string containing an unsigned * integer value representing a bound * but unconnected socket that is to be - * used as an NFSD listener + * used as an NFSD listener; listen(3) + * must be called for a SOCK_STREAM + * socket, otherwise it is ignored * size: non-zero length of C string in @buf * Output: * On success: NFS service is started; -- cgit v0.10.2 From 0a5372d8a13c3f670bf7c9c4a68ab04c3664fa28 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:10 -0400 Subject: NFSD: Finish refactoring __write_ports() Clean up: Refactor transport name listing out of __write_ports() to make it easier to understand and maintain. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index d491fa9..caf4fdc 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -911,6 +911,17 @@ static ssize_t write_versions(struct file *file, char *buf, size_t size) } /* + * Zero-length write. Return a list of NFSD's current listener + * transports. + */ +static ssize_t __write_ports_names(char *buf) +{ + if (nfsd_serv == NULL) + return 0; + return svc_xprt_names(nfsd_serv, buf, 0); +} + +/* * A single 'fd' number was written, in which case it must be for * a socket of a supported family/protocol, and we use it as an * nfsd listener. @@ -1019,13 +1030,8 @@ static ssize_t __write_ports_delxprt(char *buf) static ssize_t __write_ports(struct file *file, char *buf, size_t size) { - if (size == 0) { - int len = 0; - - if (nfsd_serv) - len = svc_xprt_names(nfsd_serv, buf, 0); - return len; - } + if (size == 0) + return __write_ports_names(buf); if (isdigit(buf[0])) return __write_ports_addfd(buf); -- cgit v0.10.2 From ea068bad27cefc71ab03230dbf01a8f8d98da5ba Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:18 -0400 Subject: NFSD: move lockd_up() before svc_addsock() Clean up. A couple of years ago, a series of commits, finishing with commit 5680c446, swapped the order of the lockd_up() and svc_addsock() calls in __write_ports(). At that time lockd_up() needed to know the transport protocol of the passed-in socket to start a listener on the same transport protocol. These days, lockd_up() doesn't take a protocol argument; it always starts both a UDP and TCP listener. It's now more straightforward to try the lockd_up() first, then do a lockd_down() if the svc_addsock() fails. Careful review of this code shows that the svc_sock_names() call is used only to close the just-opened socket in case lockd_up() fails. So it is no longer needed if lockd_up() is done first. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index caf4fdc..e051847 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -939,17 +939,18 @@ static ssize_t __write_ports_addfd(char *buf) if (err != 0) return err; - err = svc_addsock(nfsd_serv, fd, buf); - if (err >= 0) { - err = lockd_up(); - if (err < 0) - svc_sock_names(buf + strlen(buf) + 1, nfsd_serv, buf); + err = lockd_up(); + if (err != 0) + goto out; - /* Decrease the count, but don't shut down the service */ - nfsd_serv->sv_nrthreads--; - } + err = svc_addsock(nfsd_serv, fd, buf); + if (err < 0) + lockd_down(); - return err < 0 ? err : 0; +out: + /* Decrease the count, but don't shut down the service */ + nfsd_serv->sv_nrthreads--; + return err; } /* -- cgit v0.10.2 From 335c54bdc4d3bacdbd619ec95cd0b352435bd37f Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:25 -0400 Subject: NFSD: Prevent a buffer overflow in svc_xprt_names() The svc_xprt_names() function can overflow its buffer if it's so near the end of the passed in buffer that the "name too long" string still doesn't fit. Of course, it could never tell if it was near the end of the passed in buffer, since its only caller passes in zero as the buffer length. Let's make this API a little safer. Change svc_xprt_names() so it *always* checks for a buffer overflow, and change its only caller to pass in the correct buffer length. If svc_xprt_names() does overflow its buffer, it now fails with an ENAMETOOLONG errno, instead of trying to write a message at the end of the buffer. I don't like this much, but I can't figure out a clean way that's always safe to return some of the names, *and* an indication that the buffer was not long enough. The displayed error when doing a 'cat /proc/fs/nfsd/portlist' is "File name too long". Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index e051847..6a1cd90 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -918,7 +918,7 @@ static ssize_t __write_ports_names(char *buf) { if (nfsd_serv == NULL) return 0; - return svc_xprt_names(nfsd_serv, buf, 0); + return svc_xprt_names(nfsd_serv, buf, SIMPLE_TRANSACTION_LIMIT); } /* diff --git a/include/linux/sunrpc/svc_xprt.h b/include/linux/sunrpc/svc_xprt.h index d790c52..2223ae0 100644 --- a/include/linux/sunrpc/svc_xprt.h +++ b/include/linux/sunrpc/svc_xprt.h @@ -83,7 +83,7 @@ int svc_port_is_privileged(struct sockaddr *sin); int svc_print_xprts(char *buf, int maxlen); struct svc_xprt *svc_find_xprt(struct svc_serv *serv, const char *xcl_name, const sa_family_t af, const unsigned short port); -int svc_xprt_names(struct svc_serv *serv, char *buf, int buflen); +int svc_xprt_names(struct svc_serv *serv, char *buf, const int buflen); static inline void svc_xprt_get(struct svc_xprt *xprt) { diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index f200393..6f33d33 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -1098,36 +1098,58 @@ struct svc_xprt *svc_find_xprt(struct svc_serv *serv, const char *xcl_name, } EXPORT_SYMBOL_GPL(svc_find_xprt); -/* - * Format a buffer with a list of the active transports. A zero for - * the buflen parameter disables target buffer overflow checking. +static int svc_one_xprt_name(const struct svc_xprt *xprt, + char *pos, int remaining) +{ + int len; + + len = snprintf(pos, remaining, "%s %u\n", + xprt->xpt_class->xcl_name, + svc_xprt_local_port(xprt)); + if (len >= remaining) + return -ENAMETOOLONG; + return len; +} + +/** + * svc_xprt_names - format a buffer with a list of transport names + * @serv: pointer to an RPC service + * @buf: pointer to a buffer to be filled in + * @buflen: length of buffer to be filled in + * + * Fills in @buf with a string containing a list of transport names, + * each name terminated with '\n'. + * + * Returns positive length of the filled-in string on success; otherwise + * a negative errno value is returned if an error occurs. */ -int svc_xprt_names(struct svc_serv *serv, char *buf, int buflen) +int svc_xprt_names(struct svc_serv *serv, char *buf, const int buflen) { struct svc_xprt *xprt; - char xprt_str[64]; - int totlen = 0; - int len; + int len, totlen; + char *pos; /* Sanity check args */ if (!serv) return 0; spin_lock_bh(&serv->sv_lock); + + pos = buf; + totlen = 0; list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list) { - len = snprintf(xprt_str, sizeof(xprt_str), - "%s %d\n", xprt->xpt_class->xcl_name, - svc_xprt_local_port(xprt)); - /* If the string was truncated, replace with error string */ - if (len >= sizeof(xprt_str)) - strcpy(xprt_str, "name-too-long\n"); - /* Don't overflow buffer */ - len = strlen(xprt_str); - if (buflen && (len + totlen >= buflen)) + len = svc_one_xprt_name(xprt, pos, buflen - totlen); + if (len < 0) { + *buf = '\0'; + totlen = len; + } + if (len <= 0) break; - strcpy(buf+totlen, xprt_str); + + pos += len; totlen += len; } + spin_unlock_bh(&serv->sv_lock); return totlen; } -- cgit v0.10.2 From bfba9ab4c64f0e5c33930711e6c073c285e01fcf Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:33 -0400 Subject: SUNRPC: pass buffer size to svc_addsock() Adjust the synopsis of svc_addsock() to pass in the size of the output buffer. Add a documenting comment. This is a cosmetic change for now. A subsequent patch will make sure the buffer length is passed to one_sock_name(), where the length will actually be useful. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 6a1cd90..1f1c215 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -943,7 +943,7 @@ static ssize_t __write_ports_addfd(char *buf) if (err != 0) goto out; - err = svc_addsock(nfsd_serv, fd, buf); + err = svc_addsock(nfsd_serv, fd, buf, SIMPLE_TRANSACTION_LIMIT); if (err < 0) lockd_down(); diff --git a/include/linux/sunrpc/svcsock.h b/include/linux/sunrpc/svcsock.h index 483e103..e23241c 100644 --- a/include/linux/sunrpc/svcsock.h +++ b/include/linux/sunrpc/svcsock.h @@ -39,7 +39,8 @@ int svc_send(struct svc_rqst *); void svc_drop(struct svc_rqst *); void svc_sock_update_bufs(struct svc_serv *serv); int svc_sock_names(char *buf, struct svc_serv *serv, char *toclose); -int svc_addsock(struct svc_serv *serv, int fd, char *name_return); +int svc_addsock(struct svc_serv *serv, const int fd, + char *name_return, const size_t len); void svc_init_xprt_sock(void); void svc_cleanup_xprt_sock(void); diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 8b08328..6bec1e2 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -1128,9 +1128,19 @@ static struct svc_sock *svc_setup_socket(struct svc_serv *serv, return svsk; } -int svc_addsock(struct svc_serv *serv, - int fd, - char *name_return) +/** + * svc_addsock - add a listener socket to an RPC service + * @serv: pointer to RPC service to which to add a new listener + * @fd: file descriptor of the new listener + * @name_return: pointer to buffer to fill in with name of listener + * @len: size of the buffer + * + * Fills in socket name and returns positive length of name if successful. + * Name is terminated with '\n'. On error, returns a negative errno + * value. + */ +int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, + const size_t len) { int err = 0; struct socket *so = sockfd_lookup(fd, &err); -- cgit v0.10.2 From 8435d34dbbe75678c3cdad3d53b1e7996a79b3bf Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:40 -0400 Subject: SUNRPC: pass buffer size to svc_sock_names() Adjust the synopsis of svc_sock_names() to pass in the size of the output buffer. Add a documenting comment. This is a cosmetic change for now. A subsequent patch will make sure the buffer length is passed to one_sock_name(), where the length will actually be useful. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 1f1c215..b64a7fb 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -966,7 +966,8 @@ static ssize_t __write_ports_delfd(char *buf) return -ENOMEM; if (nfsd_serv != NULL) - len = svc_sock_names(buf, nfsd_serv, toclose); + len = svc_sock_names(nfsd_serv, buf, + SIMPLE_TRANSACTION_LIMIT, toclose); if (len >= 0) lockd_down(); diff --git a/include/linux/sunrpc/svcsock.h b/include/linux/sunrpc/svcsock.h index e23241c..8271631 100644 --- a/include/linux/sunrpc/svcsock.h +++ b/include/linux/sunrpc/svcsock.h @@ -38,7 +38,9 @@ int svc_recv(struct svc_rqst *, long); int svc_send(struct svc_rqst *); void svc_drop(struct svc_rqst *); void svc_sock_update_bufs(struct svc_serv *serv); -int svc_sock_names(char *buf, struct svc_serv *serv, char *toclose); +int svc_sock_names(struct svc_serv *serv, char *buf, + const size_t buflen, + const char *toclose); int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, const size_t len); void svc_init_xprt_sock(void); diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 6bec1e2..032b52e 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -259,8 +259,23 @@ static int one_sock_name(char *buf, struct svc_sock *svsk) return len; } -int -svc_sock_names(char *buf, struct svc_serv *serv, char *toclose) +/** + * svc_sock_names - construct a list of listener names in a string + * @serv: pointer to RPC service + * @buf: pointer to a buffer to fill in with socket names + * @buflen: size of the buffer to be filled + * @toclose: pointer to '\0'-terminated C string containing the name + * of a listener to be closed + * + * Fills in @buf with a '\n'-separated list of names of listener + * sockets. If @toclose is not NULL, the socket named by @toclose + * is closed, and is not included in the output list. + * + * Returns positive length of the socket name string, or a negative + * errno value on error. + */ +int svc_sock_names(struct svc_serv *serv, char *buf, const size_t buflen, + const char *toclose) { struct svc_sock *svsk, *closesk = NULL; int len = 0; -- cgit v0.10.2 From e7942b9f2586fb15e1b898acc7c198ffd60aee4e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:48 -0400 Subject: SUNRPC: Switch one_sock_name() to use snprintf() Use snprintf() in one_sock_name() to prevent overflowing the output buffer. If the name doesn't fit in the buffer, the buffer is filled in with an empty string, and -ENAMETOOLONG is returned. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 032b52e..61d4a32 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -240,22 +240,27 @@ out: /* * Report socket names for nfsdfs */ -static int one_sock_name(char *buf, struct svc_sock *svsk) +static int svc_one_sock_name(struct svc_sock *svsk, char *buf, int remaining) { int len; switch(svsk->sk_sk->sk_family) { - case AF_INET: - len = sprintf(buf, "ipv4 %s %pI4 %d\n", + case PF_INET: + len = snprintf(buf, remaining, "ipv4 %s %pI4 %d\n", svsk->sk_sk->sk_protocol == IPPROTO_UDP ? "udp" : "tcp", &inet_sk(svsk->sk_sk)->rcv_saddr, inet_sk(svsk->sk_sk)->num); break; default: - len = sprintf(buf, "*unknown-%d*\n", + len = snprintf(buf, remaining, "*unknown-%d*\n", svsk->sk_sk->sk_family); } + + if (len >= remaining) { + *buf = '\0'; + return -ENAMETOOLONG; + } return len; } @@ -282,15 +287,21 @@ int svc_sock_names(struct svc_serv *serv, char *buf, const size_t buflen, if (!serv) return 0; + spin_lock_bh(&serv->sv_lock); list_for_each_entry(svsk, &serv->sv_permsocks, sk_xprt.xpt_list) { - int onelen = one_sock_name(buf+len, svsk); - if (toclose && strcmp(toclose, buf+len) == 0) + int onelen = svc_one_sock_name(svsk, buf + len, buflen - len); + if (onelen < 0) { + len = onelen; + break; + } + if (toclose && strcmp(toclose, buf + len) == 0) closesk = svsk; else len += onelen; } spin_unlock_bh(&serv->sv_lock); + if (closesk) /* Should unregister with portmap, but you cannot * unregister just one protocol... @@ -1195,7 +1206,7 @@ int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, sockfd_put(so); return err; } - return one_sock_name(name_return, svsk); + return svc_one_sock_name(svsk, name_return, len); } EXPORT_SYMBOL_GPL(svc_addsock); -- cgit v0.10.2 From 58de2f86585dd8fc785aca6625adee32af84b8d7 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:32:55 -0400 Subject: SUNRPC: Support PF_INET6 in one_sock_name() Add an arm to the switch statement in svc_one_sock_name() so it can construct the name of PF_INET6 sockets properly. Signed-off-by: Chuck Lever Cc: Aime Le Rouzic Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 61d4a32..983bfa9 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -252,6 +252,13 @@ static int svc_one_sock_name(struct svc_sock *svsk, char *buf, int remaining) &inet_sk(svsk->sk_sk)->rcv_saddr, inet_sk(svsk->sk_sk)->num); break; + case PF_INET6: + len = snprintf(buf, remaining, "ipv6 %s %pI6 %d\n", + svsk->sk_sk->sk_protocol == IPPROTO_UDP ? + "udp" : "tcp", + &inet6_sk(svsk->sk_sk)->rcv_saddr, + inet_sk(svsk->sk_sk)->num); + break; default: len = snprintf(buf, remaining, "*unknown-%d*\n", svsk->sk_sk->sk_family); -- cgit v0.10.2 From 017cb47f46722f29d101a709a2094d151111ed6d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:33:03 -0400 Subject: SUNRPC: Clean up one_sock_name() Clean up svc_one_sock_name() by setting up automatic variables for frequently used expressions. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 983bfa9..4e6d406 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -242,26 +242,27 @@ out: */ static int svc_one_sock_name(struct svc_sock *svsk, char *buf, int remaining) { + const struct sock *sk = svsk->sk_sk; + const char *proto_name = sk->sk_protocol == IPPROTO_UDP ? + "udp" : "tcp"; int len; - switch(svsk->sk_sk->sk_family) { + switch (sk->sk_family) { case PF_INET: len = snprintf(buf, remaining, "ipv4 %s %pI4 %d\n", - svsk->sk_sk->sk_protocol == IPPROTO_UDP ? - "udp" : "tcp", - &inet_sk(svsk->sk_sk)->rcv_saddr, - inet_sk(svsk->sk_sk)->num); + proto_name, + &inet_sk(sk)->rcv_saddr, + inet_sk(sk)->num); break; case PF_INET6: len = snprintf(buf, remaining, "ipv6 %s %pI6 %d\n", - svsk->sk_sk->sk_protocol == IPPROTO_UDP ? - "udp" : "tcp", - &inet6_sk(svsk->sk_sk)->rcv_saddr, - inet_sk(svsk->sk_sk)->num); + proto_name, + &inet6_sk(sk)->rcv_saddr, + inet_sk(sk)->num); break; default: len = snprintf(buf, remaining, "*unknown-%d*\n", - svsk->sk_sk->sk_family); + sk->sk_family); } if (len >= remaining) { -- cgit v0.10.2 From 3d72ab8fdd44c872633b210dd1a4afd2910d0bbb Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:33:10 -0400 Subject: NFSD: Stricter buffer size checking in write_recoverydir() While it's not likely a pathname will be longer than SIMPLE_TRANSACTION_SIZE, we should be more careful about just plopping it into the output buffer without bounds checking. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index b64a7fb..c484346 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -1260,8 +1260,9 @@ static ssize_t __write_recoverydir(struct file *file, char *buf, size_t size) status = nfs4_reset_recoverydir(recdir); } - sprintf(buf, "%s\n", nfs4_recoverydir()); - return strlen(buf); + + return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%s\n", + nfs4_recoverydir()); } /** -- cgit v0.10.2 From 261758b5c3dfeac73ca364c47ed538f5ce4250ee Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:33:18 -0400 Subject: NFSD: Stricter buffer size checking in write_versions() While it's not likely today that there are enough NFS versions to overflow the output buffer in write_versions(), we should be more careful about detecting the end of the buffer. The number of NFS versions will only increase as NFSv4 minor versions are added. Note that this API doesn't behave the same as portlist. Here we attempt to display as many versions as will fit in the buffer, and do not provide any indication that an overflow would have occurred. I don't have any good rationale for that. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index c484346..a152694 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -793,7 +793,7 @@ static ssize_t __write_versions(struct file *file, char *buf, size_t size) { char *mesg = buf; char *vers, *minorp, sign; - int len, num; + int len, num, remaining; unsigned minor; ssize_t tlen = 0; char *sep; @@ -840,32 +840,50 @@ static ssize_t __write_versions(struct file *file, char *buf, size_t size) } next: vers += len + 1; - tlen += len; } while ((len = qword_get(&mesg, vers, size)) > 0); /* If all get turned off, turn them back on, as * having no versions is BAD */ nfsd_reset_versions(); } + /* Now write current state into reply buffer */ len = 0; sep = ""; + remaining = SIMPLE_TRANSACTION_LIMIT; for (num=2 ; num <= 4 ; num++) if (nfsd_vers(num, NFSD_AVAIL)) { - len += sprintf(buf+len, "%s%c%d", sep, + len = snprintf(buf, remaining, "%s%c%d", sep, nfsd_vers(num, NFSD_TEST)?'+':'-', num); sep = " "; + + if (len > remaining) + break; + remaining -= len; + buf += len; + tlen += len; } if (nfsd_vers(4, NFSD_AVAIL)) - for (minor = 1; minor <= NFSD_SUPPORTED_MINOR_VERSION; minor++) - len += sprintf(buf+len, " %c4.%u", + for (minor = 1; minor <= NFSD_SUPPORTED_MINOR_VERSION; + minor++) { + len = snprintf(buf, remaining, " %c4.%u", (nfsd_vers(4, NFSD_TEST) && nfsd_minorversion(minor, NFSD_TEST)) ? '+' : '-', minor); - len += sprintf(buf+len, "\n"); - return len; + + if (len > remaining) + break; + remaining -= len; + buf += len; + tlen += len; + } + + len = snprintf(buf, remaining, "\n"); + if (len > remaining) + return -EINVAL; + return tlen + len; } /** -- cgit v0.10.2 From e06b64050ebea1220bca3c291d0431f5206b5c95 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 23 Apr 2009 19:33:25 -0400 Subject: NFSD: Stricter buffer size checking in fs/nfsd/nfsctl.c Clean up: For consistency, handle output buffer size checking in a other nfsctl functions the same way it's done for write_versions(). Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index a152694..877e713 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -695,8 +695,9 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size) if (rv) return rv; } - sprintf(buf, "%d\n", nfsd_nrthreads()); - return strlen(buf); + + return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%d\n", + nfsd_nrthreads()); } /** @@ -1197,7 +1198,9 @@ static ssize_t write_maxblksize(struct file *file, char *buf, size_t size) nfsd_max_blksize = bsize; mutex_unlock(&nfsd_mutex); } - return sprintf(buf, "%d\n", nfsd_max_blksize); + + return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%d\n", + nfsd_max_blksize); } #ifdef CONFIG_NFSD_V4 @@ -1221,8 +1224,9 @@ static ssize_t __write_leasetime(struct file *file, char *buf, size_t size) return -EINVAL; nfs4_reset_lease(lease); } - sprintf(buf, "%ld\n", nfs4_lease_time()); - return strlen(buf); + + return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%ld\n", + nfs4_lease_time()); } /** -- cgit v0.10.2 From bc749ca4c405d507e6ec6e3f3e5475e9a09faf0a Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 7 Apr 2009 16:55:27 -0700 Subject: nfsd: eliminate ENCODE_HEAD macro This macro doesn't serve any useful purpose. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index b820c31..fe46ede 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -1503,8 +1503,6 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp) * task to translate them into Linux-specific versions which are more * consistent with the style used in NFSv2/v3... */ -#define ENCODE_HEAD __be32 *p - #define WRITE32(n) *p++ = htonl(n) #define WRITE64(n) do { \ *p++ = htonl((u32)((n) >> 32)); \ @@ -2334,7 +2332,7 @@ fail: static void nfsd4_encode_stateid(struct nfsd4_compoundres *resp, stateid_t *sid) { - ENCODE_HEAD; + __be32 *p; RESERVE_SPACE(sizeof(stateid_t)); WRITE32(sid->si_generation); @@ -2345,7 +2343,7 @@ nfsd4_encode_stateid(struct nfsd4_compoundres *resp, stateid_t *sid) static __be32 nfsd4_encode_access(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_access *access) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(8); @@ -2372,7 +2370,7 @@ nfsd4_encode_close(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_c static __be32 nfsd4_encode_commit(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_commit *commit) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(8); @@ -2385,7 +2383,7 @@ nfsd4_encode_commit(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_ static __be32 nfsd4_encode_create(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_create *create) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(32); @@ -2421,7 +2419,7 @@ nfsd4_encode_getfh(struct nfsd4_compoundres *resp, __be32 nfserr, struct svc_fh { struct svc_fh *fhp = *fhpp; unsigned int len; - ENCODE_HEAD; + __be32 *p; if (!nfserr) { len = fhp->fh_handle.fh_size; @@ -2440,7 +2438,7 @@ nfsd4_encode_getfh(struct nfsd4_compoundres *resp, __be32 nfserr, struct svc_fh static void nfsd4_encode_lock_denied(struct nfsd4_compoundres *resp, struct nfsd4_lock_denied *ld) { - ENCODE_HEAD; + __be32 *p; RESERVE_SPACE(32 + XDR_LEN(ld->ld_sop ? ld->ld_sop->so_owner.len : 0)); WRITE64(ld->ld_start); @@ -2496,7 +2494,7 @@ nfsd4_encode_locku(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_l static __be32 nfsd4_encode_link(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_link *link) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(20); @@ -2510,7 +2508,7 @@ nfsd4_encode_link(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_li static __be32 nfsd4_encode_open(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_open *open) { - ENCODE_HEAD; + __be32 *p; ENCODE_SEQID_OP_HEAD; if (nfserr) @@ -2605,7 +2603,7 @@ nfsd4_encode_read(struct nfsd4_compoundres *resp, __be32 nfserr, int v, pn; unsigned long maxcount; long len; - ENCODE_HEAD; + __be32 *p; if (nfserr) return nfserr; @@ -2667,7 +2665,7 @@ nfsd4_encode_readlink(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd { int maxcount; char *page; - ENCODE_HEAD; + __be32 *p; if (nfserr) return nfserr; @@ -2716,7 +2714,7 @@ nfsd4_encode_readdir(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4 int maxcount; loff_t offset; __be32 *page, *savep, *tailbase; - ENCODE_HEAD; + __be32 *p; if (nfserr) return nfserr; @@ -2792,7 +2790,7 @@ err_no_verf: static __be32 nfsd4_encode_remove(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_remove *remove) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(20); @@ -2805,7 +2803,7 @@ nfsd4_encode_remove(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_ static __be32 nfsd4_encode_rename(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_rename *rename) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(40); @@ -2825,7 +2823,7 @@ nfsd4_encode_secinfo(struct nfsd4_compoundres *resp, __be32 nfserr, u32 nflavs; struct exp_flavor_info *flavs; struct exp_flavor_info def_flavs[2]; - ENCODE_HEAD; + __be32 *p; if (nfserr) goto out; @@ -2890,7 +2888,7 @@ out: static __be32 nfsd4_encode_setattr(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_setattr *setattr) { - ENCODE_HEAD; + __be32 *p; RESERVE_SPACE(12); if (nfserr) { @@ -2910,7 +2908,7 @@ nfsd4_encode_setattr(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4 static __be32 nfsd4_encode_setclientid(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_setclientid *scd) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(8 + sizeof(nfs4_verifier)); @@ -2930,7 +2928,7 @@ nfsd4_encode_setclientid(struct nfsd4_compoundres *resp, __be32 nfserr, struct n static __be32 nfsd4_encode_write(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_write *write) { - ENCODE_HEAD; + __be32 *p; if (!nfserr) { RESERVE_SPACE(16); @@ -2946,7 +2944,7 @@ static __be32 nfsd4_encode_exchange_id(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_exchange_id *exid) { - ENCODE_HEAD; + __be32 *p; char *major_id; char *server_scope; int major_id_sz; @@ -3001,7 +2999,7 @@ static __be32 nfsd4_encode_create_session(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_create_session *sess) { - ENCODE_HEAD; + __be32 *p; if (nfserr) return nfserr; @@ -3057,7 +3055,7 @@ __be32 nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_sequence *seq) { - ENCODE_HEAD; + __be32 *p; if (nfserr) return nfserr; @@ -3205,7 +3203,7 @@ void nfsd4_encode_operation(struct nfsd4_compoundres *resp, struct nfsd4_op *op) { __be32 *statp; - ENCODE_HEAD; + __be32 *p; RESERVE_SPACE(8); WRITE32(op->opnum); @@ -3239,7 +3237,7 @@ status: void nfsd4_encode_replay(struct nfsd4_compoundres *resp, struct nfsd4_op *op) { - ENCODE_HEAD; + __be32 *p; struct nfs4_replay *rp = op->replay; BUG_ON(!rp); -- cgit v0.10.2 From 3352d2c2d0540955a7bbb3421a28330af7f9d79c Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 7 Apr 2009 17:03:19 -0700 Subject: nfsd4: delete obsolete xdr comments We don't need comments to tell us these macros are ugly. And we're long past trying to share any of this code with the BSD's. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index fe46ede..4a71fcd 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -83,16 +83,6 @@ check_filename(char *str, int len, __be32 err) return 0; } -/* - * START OF "GENERIC" DECODE ROUTINES. - * These may look a little ugly since they are imported from a "generic" - * set of XDR encode/decode routines which are intended to be shared by - * all of our NFSv4 implementations (OpenBSD, MacOS X...). - * - * If the pain of reading these is too great, it should be a straightforward - * task to translate them into Linux-specific versions which are more - * consistent with the style used in NFSv2/v3... - */ #define DECODE_HEAD \ __be32 *p; \ __be32 status @@ -1489,20 +1479,7 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp) DECODE_TAIL; } -/* - * END OF "GENERIC" DECODE ROUTINES. - */ -/* - * START OF "GENERIC" ENCODE ROUTINES. - * These may look a little ugly since they are imported from a "generic" - * set of XDR encode/decode routines which are intended to be shared by - * all of our NFSv4 implementations (OpenBSD, MacOS X...). - * - * If the pain of reading these is too great, it should be a straightforward - * task to translate them into Linux-specific versions which are more - * consistent with the style used in NFSv2/v3... - */ #define WRITE32(n) *p++ = htonl(n) #define WRITE64(n) do { \ *p++ = htonl((u32)((n) >> 32)); \ @@ -3252,10 +3229,6 @@ nfsd4_encode_replay(struct nfsd4_compoundres *resp, struct nfsd4_op *op) ADJUST_ARGS(); } -/* - * END OF "GENERIC" ENCODE ROUTINES. - */ - int nfs4svc_encode_voidres(struct svc_rqst *rqstp, __be32 *p, void *dummy) { -- cgit v0.10.2 From c654b8a9cba6002aad1c01919e4928a79a4a6dcf Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 16 Apr 2009 17:33:25 -0400 Subject: nfsd: support ext4 i_version ext4 supports a real NFSv4 change attribute, which is bumped whenever the ctime would be updated, including times when two updates arrive within a jiffy of each other. (Note that although ext4 has space for nanosecond-precision ctime, the real resolution is lower: it actually uses jiffies as the time-source.) This ensures clients will invalidate their caches when they need to. There is some fear that keeping the i_version up-to-date could have performance drawbacks, so for now it's turned on only by a mount option. We hope to do something better eventually. Signed-off-by: J. Bruce Fields Cc: Theodore Tso diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c index 17d0dd9..01d4ec1 100644 --- a/fs/nfsd/nfs3xdr.c +++ b/fs/nfsd/nfs3xdr.c @@ -272,6 +272,7 @@ void fill_post_wcc(struct svc_fh *fhp) err = vfs_getattr(fhp->fh_export->ex_path.mnt, fhp->fh_dentry, &fhp->fh_post_attr); + fhp->fh_post_change = fhp->fh_dentry->d_inode->i_version; if (err) fhp->fh_post_saved = 0; else diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 4a71fcd..12d36a7 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -1490,13 +1490,41 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp) memcpy(p, ptr, nbytes); \ p += XDR_QUADLEN(nbytes); \ }} while (0) -#define WRITECINFO(c) do { \ - *p++ = htonl(c.atomic); \ - *p++ = htonl(c.before_ctime_sec); \ - *p++ = htonl(c.before_ctime_nsec); \ - *p++ = htonl(c.after_ctime_sec); \ - *p++ = htonl(c.after_ctime_nsec); \ -} while (0) + +static void write32(__be32 **p, u32 n) +{ + *(*p)++ = n; +} + +static void write64(__be32 **p, u64 n) +{ + write32(p, (u32)(n >> 32)); + write32(p, (u32)n); +} + +static void write_change(__be32 **p, struct kstat *stat, struct inode *inode) +{ + if (IS_I_VERSION(inode)) { + write64(p, inode->i_version); + } else { + write32(p, stat->ctime.tv_sec); + write32(p, stat->ctime.tv_nsec); + } +} + +static void write_cinfo(__be32 **p, struct nfsd4_change_info *c) +{ + write32(p, c->atomic); + if (c->change_supported) { + write64(p, c->before_change); + write64(p, c->after_change); + } else { + write32(p, c->before_ctime_sec); + write32(p, c->before_ctime_nsec); + write32(p, c->after_ctime_sec); + write32(p, c->after_ctime_nsec); + } +} #define RESERVE_SPACE(nbytes) do { \ p = resp->p; \ @@ -1849,16 +1877,9 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp, WRITE32(NFS4_FH_PERSISTENT|NFS4_FH_VOL_RENAME); } if (bmval0 & FATTR4_WORD0_CHANGE) { - /* - * Note: This _must_ be consistent with the scheme for writing - * change_info, so any changes made here must be reflected there - * as well. (See xdr4.h:set_change_info() and the WRITECINFO() - * macro above.) - */ if ((buflen -= 8) < 0) goto out_resource; - WRITE32(stat.ctime.tv_sec); - WRITE32(stat.ctime.tv_nsec); + write_change(&p, &stat, dentry->d_inode); } if (bmval0 & FATTR4_WORD0_SIZE) { if ((buflen -= 8) < 0) @@ -2364,7 +2385,7 @@ nfsd4_encode_create(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_ if (!nfserr) { RESERVE_SPACE(32); - WRITECINFO(create->cr_cinfo); + write_cinfo(&p, &create->cr_cinfo); WRITE32(2); WRITE32(create->cr_bmval[0]); WRITE32(create->cr_bmval[1]); @@ -2475,7 +2496,7 @@ nfsd4_encode_link(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_li if (!nfserr) { RESERVE_SPACE(20); - WRITECINFO(link->li_cinfo); + write_cinfo(&p, &link->li_cinfo); ADJUST_ARGS(); } return nfserr; @@ -2493,7 +2514,7 @@ nfsd4_encode_open(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_op nfsd4_encode_stateid(resp, &open->op_stateid); RESERVE_SPACE(40); - WRITECINFO(open->op_cinfo); + write_cinfo(&p, &open->op_cinfo); WRITE32(open->op_rflags); WRITE32(2); WRITE32(open->op_bmval[0]); @@ -2771,7 +2792,7 @@ nfsd4_encode_remove(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_ if (!nfserr) { RESERVE_SPACE(20); - WRITECINFO(remove->rm_cinfo); + write_cinfo(&p, &remove->rm_cinfo); ADJUST_ARGS(); } return nfserr; @@ -2784,8 +2805,8 @@ nfsd4_encode_rename(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_ if (!nfserr) { RESERVE_SPACE(40); - WRITECINFO(rename->rn_sinfo); - WRITECINFO(rename->rn_tinfo); + write_cinfo(&p, &rename->rn_sinfo); + write_cinfo(&p, &rename->rn_tinfo); ADJUST_ARGS(); } return nfserr; diff --git a/include/linux/nfsd/nfsfh.h b/include/linux/nfsd/nfsfh.h index afa1901..8f641c9 100644 --- a/include/linux/nfsd/nfsfh.h +++ b/include/linux/nfsd/nfsfh.h @@ -151,9 +151,15 @@ typedef struct svc_fh { __u64 fh_pre_size; /* size before operation */ struct timespec fh_pre_mtime; /* mtime before oper */ struct timespec fh_pre_ctime; /* ctime before oper */ + /* + * pre-op nfsv4 change attr: note must check IS_I_VERSION(inode) + * to find out if it is valid. + */ + u64 fh_pre_change; /* Post-op attributes saved in fh_unlock */ struct kstat fh_post_attr; /* full attrs after operation */ + u64 fh_post_change; /* nfsv4 change; see above */ #endif /* CONFIG_NFSD_V3 */ } svc_fh; @@ -298,6 +304,7 @@ fill_pre_wcc(struct svc_fh *fhp) fhp->fh_pre_mtime = inode->i_mtime; fhp->fh_pre_ctime = inode->i_ctime; fhp->fh_pre_size = inode->i_size; + fhp->fh_pre_change = inode->i_version; fhp->fh_pre_saved = 1; } } diff --git a/include/linux/nfsd/xdr4.h b/include/linux/nfsd/xdr4.h index f80d601..d0f050f 100644 --- a/include/linux/nfsd/xdr4.h +++ b/include/linux/nfsd/xdr4.h @@ -64,10 +64,13 @@ static inline bool nfsd4_has_session(struct nfsd4_compound_state *cs) struct nfsd4_change_info { u32 atomic; + bool change_supported; u32 before_ctime_sec; u32 before_ctime_nsec; + u64 before_change; u32 after_ctime_sec; u32 after_ctime_nsec; + u64 after_change; }; struct nfsd4_access { @@ -503,10 +506,16 @@ set_change_info(struct nfsd4_change_info *cinfo, struct svc_fh *fhp) { BUG_ON(!fhp->fh_pre_saved || !fhp->fh_post_saved); cinfo->atomic = 1; - cinfo->before_ctime_sec = fhp->fh_pre_ctime.tv_sec; - cinfo->before_ctime_nsec = fhp->fh_pre_ctime.tv_nsec; - cinfo->after_ctime_sec = fhp->fh_post_attr.ctime.tv_sec; - cinfo->after_ctime_nsec = fhp->fh_post_attr.ctime.tv_nsec; + cinfo->change_supported = IS_I_VERSION(fhp->fh_dentry->d_inode); + if (cinfo->change_supported) { + cinfo->before_change = fhp->fh_pre_change; + cinfo->after_change = fhp->fh_post_change; + } else { + cinfo->before_ctime_sec = fhp->fh_pre_ctime.tv_sec; + cinfo->before_ctime_nsec = fhp->fh_pre_ctime.tv_nsec; + cinfo->after_ctime_sec = fhp->fh_post_attr.ctime.tv_sec; + cinfo->after_ctime_nsec = fhp->fh_post_attr.ctime.tv_nsec; + } } int nfs4svc_encode_voidres(struct svc_rqst *, __be32 *, void *); -- cgit v0.10.2 From b8fd47aefa5f13df1edacbc7e68d9874635109e5 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 29 Apr 2009 11:36:17 -0400 Subject: nfsd: quiet compile warning Stephen Rothwell said: "Today's linux-next build (powerpc ppc64_defconfig) produced this new warning: fs/nfsd/nfs4state.c: In function 'EXPIRED_STATEID': fs/nfsd/nfs4state.c:2757: warning: comparison of distinct pointer types lacks a cast Caused by commit 78155ed75f470710f2aecb3e75e3d97107ba8374 ("nfsd4: distinguish expired from stale stateids")." Signed-off-by: J. Bruce Fields Reported-by: Stephen Rothwell diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 74e822e..d24dd12 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2754,7 +2754,7 @@ EXPIRED_STATEID(stateid_t *stateid) { if (time_before((unsigned long)boot_time, ((unsigned long)stateid->si_boot)) && - time_before((stateid->si_boot + lease_time), get_seconds())) { + time_before((unsigned long)(stateid->si_boot + lease_time), get_seconds())) { dprintk("NFSD: expired stateid (%08x/%08x/%08x/%08x)!\n", stateid->si_boot, stateid->si_stateownerid, stateid->si_fileid, stateid->si_generation); -- cgit v0.10.2 From f64f79ea5f5e02ba8585f35a10b4a3bcab0cea52 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 29 Apr 2009 13:45:36 -0400 Subject: nfsd4: setclientid_confirm callback-change fixes This setclientid_confirm case should allow the client to change callbacks, but it currently has a dummy implementation that just turns off callbacks completely. That dummy implementation isn't completely correct either, though: - There's no need to remove any client recovery directory in this case. - New clientid confirm verifiers should be generated (and returned) in setclientid; there's no need to generate a new one here. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index d24dd12..7e1fcc3 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1688,8 +1688,6 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp, /* XXX: We just turn off callbacks until we can handle * change request correctly. */ atomic_set(&conf->cl_callback.cb_set, 0); - gen_confirm(conf); - nfsd4_remove_clid_dir(unconf); expire_client(unconf); status = nfs_ok; -- cgit v0.10.2 From 595947acaaef373445131471a78650003f5d8e7d Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 5 Mar 2009 17:18:10 -0500 Subject: nfsd4: set shorter timeout We tried to do something overly complicated with the callback rpc timeouts here. And they're wrong--the result is that by the time a single callback times out, it's already too late to tell the client (using the cb_path_down return to RENEW) that the callback is down. Use a much shorter, simpler timeout. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 290289b..049f052 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -358,6 +358,11 @@ static struct rpc_program cb_program = { .pipe_dir_name = "/nfsd4_cb", }; +static int max_cb_time(void) +{ + return max(NFSD_LEASE_TIME/10, (time_t)1) * HZ; +} + /* Reference counting, callback cleanup, etc., all look racy as heck. * And why is cb_set an atomic? */ @@ -366,10 +371,8 @@ static struct rpc_clnt *setup_callback_client(struct nfs4_client *clp) struct sockaddr_in addr; struct nfs4_callback *cb = &clp->cl_callback; struct rpc_timeout timeparms = { - .to_initval = (NFSD_LEASE_TIME/4) * HZ, - .to_retries = 5, - .to_maxval = (NFSD_LEASE_TIME/2) * HZ, - .to_exponential = 1, + .to_initval = max_cb_time(), + .to_retries = 0, }; struct rpc_create_args args = { .protocol = IPPROTO_TCP, -- cgit v0.10.2 From e1cab5a5896e142190cd66a8287099b52e5855a7 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 23 Feb 2009 10:45:27 -0800 Subject: nfsd4: set cb_client inside setup_callback_client This is just a minor code simplification. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 049f052..4788d09 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -366,7 +366,7 @@ static int max_cb_time(void) /* Reference counting, callback cleanup, etc., all look racy as heck. * And why is cb_set an atomic? */ -static struct rpc_clnt *setup_callback_client(struct nfs4_client *clp) +int setup_callback_client(struct nfs4_client *clp) { struct sockaddr_in addr; struct nfs4_callback *cb = &clp->cl_callback; @@ -389,7 +389,7 @@ static struct rpc_clnt *setup_callback_client(struct nfs4_client *clp) struct rpc_clnt *client; if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5)) - return ERR_PTR(-EINVAL); + return -EINVAL; /* Initialize address */ memset(&addr, 0, sizeof(addr)); @@ -399,10 +399,13 @@ static struct rpc_clnt *setup_callback_client(struct nfs4_client *clp) /* Create RPC client */ client = rpc_create(&args); - if (IS_ERR(client)) + if (IS_ERR(client)) { dprintk("NFSD: couldn't create callback client: %ld\n", PTR_ERR(client)); - return client; + return PTR_ERR(client); + } + cb->cb_client = client; + return 0; } @@ -414,28 +417,23 @@ static int do_probe_callback(void *data) .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], .rpc_argp = clp, }; - struct rpc_clnt *client; int status; - client = setup_callback_client(clp); - if (IS_ERR(client)) { - status = PTR_ERR(client); - dprintk("NFSD: couldn't create callback client: %d\n", - status); + status = setup_callback_client(clp); + if (status) goto out_err; - } - status = rpc_call_sync(client, &msg, RPC_TASK_SOFT); + status = rpc_call_sync(cb->cb_client, &msg, RPC_TASK_SOFT); if (status) goto out_release_client; - cb->cb_client = client; atomic_set(&cb->cb_set, 1); put_nfs4_client(clp); return 0; out_release_client: - rpc_shutdown_client(client); + rpc_shutdown_client(cb->cb_client); + cb->cb_client = NULL; out_err: dprintk("NFSD: warning: no callback path to client %.*s: error %d\n", (int)clp->cl_name.len, clp->cl_name.data, status); -- cgit v0.10.2 From ecdd03b7914c91ef849e49c4d466c87f4981b5cd Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 23 Feb 2009 19:35:22 -0800 Subject: nfsd4: create rpc callback client from server thread The code is a little simpler, and it should be easier to avoid races, if we just do all rpc client creation/destruction from nfsd or laundromat threads and do only the rpc calls themselves asynchronously. The rpc creation doesn't involve any significant waiting (it doesn't call the client, for example), so there's no reason not to do this. Also don't bother destroying the client on failure of the rpc null probe. We may want to retry the probe later anyway. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 4788d09..711c628 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -409,6 +409,12 @@ int setup_callback_client(struct nfs4_client *clp) } +static void warn_no_callback_path(struct nfs4_client *clp, int reason) +{ + dprintk("NFSD: warning: no callback path to client %.*s: error %d\n", + (int)clp->cl_name.len, clp->cl_name.data, reason); +} + static int do_probe_callback(void *data) { struct nfs4_client *clp = data; @@ -419,24 +425,12 @@ static int do_probe_callback(void *data) }; int status; - status = setup_callback_client(clp); - if (status) - goto out_err; - status = rpc_call_sync(cb->cb_client, &msg, RPC_TASK_SOFT); - if (status) - goto out_release_client; + warn_no_callback_path(clp, status); + else + atomic_set(&cb->cb_set, 1); - atomic_set(&cb->cb_set, 1); - put_nfs4_client(clp); - return 0; -out_release_client: - rpc_shutdown_client(cb->cb_client); - cb->cb_client = NULL; -out_err: - dprintk("NFSD: warning: no callback path to client %.*s: error %d\n", - (int)clp->cl_name.len, clp->cl_name.data, status); put_nfs4_client(clp); return 0; } @@ -448,9 +442,16 @@ void nfsd4_probe_callback(struct nfs4_client *clp) { struct task_struct *t; + int status; BUG_ON(atomic_read(&clp->cl_callback.cb_set)); + status = setup_callback_client(clp); + if (status) { + warn_no_callback_path(clp, status); + return; + } + /* the task holds a reference to the nfs4_client struct */ atomic_inc(&clp->cl_count); -- cgit v0.10.2 From 3cef9ab266a932899e756f7e1ea7a988a97bf3b2 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 23 Feb 2009 21:42:10 -0800 Subject: nfsd4: lookup up callback cred only once Lookup the callback cred once and then use it for all subsequent callbacks. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 711c628..cc10ed3 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -415,6 +415,22 @@ static void warn_no_callback_path(struct nfs4_client *clp, int reason) (int)clp->cl_name.len, clp->cl_name.data, reason); } +static struct rpc_cred *lookup_cb_cred(struct nfs4_callback *cb) +{ + struct auth_cred acred = { + .machine_cred = 1 + }; + + /* + * Note in the gss case this doesn't actually have to wait for a + * gss upcall (or any calls to the client); this just creates a + * non-uptodate cred which the rpc state machine will fill in with + * a refresh_upcall later. + */ + return rpcauth_lookup_credcache(cb->cb_client->cl_auth, &acred, + RPCAUTH_LOOKUP_NEW); +} + static int do_probe_callback(void *data) { struct nfs4_client *clp = data; @@ -423,9 +439,18 @@ static int do_probe_callback(void *data) .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], .rpc_argp = clp, }; + struct rpc_cred *cred; int status; + cred = lookup_cb_cred(cb); + if (IS_ERR(cred)) { + status = PTR_ERR(cred); + goto out; + } + cb->cb_cred = cred; + msg.rpc_cred = cb->cb_cred; status = rpc_call_sync(cb->cb_client, &msg, RPC_TASK_SOFT); +out: if (status) warn_no_callback_path(clp, status); else @@ -475,6 +500,7 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL], .rpc_argp = cbr, + .rpc_cred = clp->cl_callback.cb_cred }; int retries = 1; int status = 0; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 7e1fcc3..b205c7d 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -643,6 +643,10 @@ shutdown_callback_client(struct nfs4_client *clp) clp->cl_callback.cb_client = NULL; rpc_shutdown_client(clnt); } + if (clp->cl_callback.cb_cred) { + put_rpccred(clp->cl_callback.cb_cred); + clp->cl_callback.cb_cred = NULL; + } } static inline void diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 4d61c87..8d882a3 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -97,6 +97,7 @@ struct nfs4_callback { /* RPC client info */ atomic_t cb_set; /* successful CB_NULL call */ struct rpc_clnt * cb_client; + struct rpc_cred * cb_cred; }; /* Maximum number of slots per session. 128 is useful for long haul TCP */ -- cgit v0.10.2 From e300a63ce4ccec073d254d883a3638d5dca1d771 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 5 Mar 2009 15:01:11 -0500 Subject: nfsd4: replace callback thread by asynchronous rpc We don't really need a synchronous rpc, and moving to an asynchronous rpc allows us to do without this extra kthread. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index cc10ed3..0aaf68b 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -415,6 +415,21 @@ static void warn_no_callback_path(struct nfs4_client *clp, int reason) (int)clp->cl_name.len, clp->cl_name.data, reason); } +static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata) +{ + struct nfs4_client *clp = calldata; + + if (task->tk_status) + warn_no_callback_path(clp, task->tk_status); + else + atomic_set(&clp->cl_callback.cb_set, 1); + put_nfs4_client(clp); +} + +static const struct rpc_call_ops nfsd4_cb_probe_ops = { + .rpc_call_done = nfsd4_cb_probe_done, +}; + static struct rpc_cred *lookup_cb_cred(struct nfs4_callback *cb) { struct auth_cred acred = { @@ -431,9 +446,8 @@ static struct rpc_cred *lookup_cb_cred(struct nfs4_callback *cb) RPCAUTH_LOOKUP_NEW); } -static int do_probe_callback(void *data) +void do_probe_callback(struct nfs4_client *clp) { - struct nfs4_client *clp = data; struct nfs4_callback *cb = &clp->cl_callback; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], @@ -449,15 +463,13 @@ static int do_probe_callback(void *data) } cb->cb_cred = cred; msg.rpc_cred = cb->cb_cred; - status = rpc_call_sync(cb->cb_client, &msg, RPC_TASK_SOFT); + status = rpc_call_async(cb->cb_client, &msg, RPC_TASK_SOFT, + &nfsd4_cb_probe_ops, (void *)clp); out: - if (status) + if (status) { warn_no_callback_path(clp, status); - else - atomic_set(&cb->cb_set, 1); - - put_nfs4_client(clp); - return 0; + put_nfs4_client(clp); + } } /* @@ -466,7 +478,6 @@ out: void nfsd4_probe_callback(struct nfs4_client *clp) { - struct task_struct *t; int status; BUG_ON(atomic_read(&clp->cl_callback.cb_set)); @@ -480,12 +491,7 @@ nfsd4_probe_callback(struct nfs4_client *clp) /* the task holds a reference to the nfs4_client struct */ atomic_inc(&clp->cl_count); - t = kthread_run(do_probe_callback, clp, "nfs4_cb_probe"); - - if (IS_ERR(t)) - atomic_dec(&clp->cl_count); - - return; + do_probe_callback(clp); } /* -- cgit v0.10.2 From c237dc0303bcf6f4cc2e0efe4fe4e341c6f34dac Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 29 Apr 2009 19:09:19 -0400 Subject: nfsd4: rename callback struct to cb_conn I want to use the name for a struct that actually does represent a single callback. (Actually, I've never been sure it helps to a separate struct for the callback information. Some day maybe those fields could just be dumped into struct nfs4_client. I don't know.) Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 0aaf68b..ed860d7 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -369,7 +369,7 @@ static int max_cb_time(void) int setup_callback_client(struct nfs4_client *clp) { struct sockaddr_in addr; - struct nfs4_callback *cb = &clp->cl_callback; + struct nfs4_cb_conn *cb = &clp->cl_cb_conn; struct rpc_timeout timeparms = { .to_initval = max_cb_time(), .to_retries = 0, @@ -422,7 +422,7 @@ static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata) if (task->tk_status) warn_no_callback_path(clp, task->tk_status); else - atomic_set(&clp->cl_callback.cb_set, 1); + atomic_set(&clp->cl_cb_conn.cb_set, 1); put_nfs4_client(clp); } @@ -430,7 +430,7 @@ static const struct rpc_call_ops nfsd4_cb_probe_ops = { .rpc_call_done = nfsd4_cb_probe_done, }; -static struct rpc_cred *lookup_cb_cred(struct nfs4_callback *cb) +static struct rpc_cred *lookup_cb_cred(struct nfs4_cb_conn *cb) { struct auth_cred acred = { .machine_cred = 1 @@ -448,7 +448,7 @@ static struct rpc_cred *lookup_cb_cred(struct nfs4_callback *cb) void do_probe_callback(struct nfs4_client *clp) { - struct nfs4_callback *cb = &clp->cl_callback; + struct nfs4_cb_conn *cb = &clp->cl_cb_conn; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], .rpc_argp = clp, @@ -480,7 +480,7 @@ nfsd4_probe_callback(struct nfs4_client *clp) { int status; - BUG_ON(atomic_read(&clp->cl_callback.cb_set)); + BUG_ON(atomic_read(&clp->cl_cb_conn.cb_set)); status = setup_callback_client(clp); if (status) { @@ -501,12 +501,12 @@ void nfsd4_cb_recall(struct nfs4_delegation *dp) { struct nfs4_client *clp = dp->dl_client; - struct rpc_clnt *clnt = clp->cl_callback.cb_client; + struct rpc_clnt *clnt = clp->cl_cb_conn.cb_client; struct nfs4_cb_recall *cbr = &dp->dl_recall; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL], .rpc_argp = cbr, - .rpc_cred = clp->cl_callback.cb_cred + .rpc_cred = clp->cl_cb_conn.cb_cred }; int retries = 1; int status = 0; @@ -519,7 +519,7 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) switch (status) { case -EIO: /* Network partition? */ - atomic_set(&clp->cl_callback.cb_set, 0); + atomic_set(&clp->cl_cb_conn.cb_set, 0); case -EBADHANDLE: case -NFS4ERR_BAD_STATEID: /* Race: client probably got cb_recall diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index b205c7d..d7b5e6b 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -182,7 +182,7 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f { struct nfs4_delegation *dp; struct nfs4_file *fp = stp->st_file; - struct nfs4_callback *cb = &stp->st_stateowner->so_client->cl_callback; + struct nfs4_cb_conn *cb = &stp->st_stateowner->so_client->cl_cb_conn; dprintk("NFSD alloc_init_deleg\n"); if (fp->fi_had_conflict) @@ -633,19 +633,19 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name) static void shutdown_callback_client(struct nfs4_client *clp) { - struct rpc_clnt *clnt = clp->cl_callback.cb_client; + struct rpc_clnt *clnt = clp->cl_cb_conn.cb_client; if (clnt) { /* * Callback threads take a reference on the client, so there * should be no outstanding callbacks at this point. */ - clp->cl_callback.cb_client = NULL; + clp->cl_cb_conn.cb_client = NULL; rpc_shutdown_client(clnt); } - if (clp->cl_callback.cb_cred) { - put_rpccred(clp->cl_callback.cb_cred); - clp->cl_callback.cb_cred = NULL; + if (clp->cl_cb_conn.cb_cred) { + put_rpccred(clp->cl_cb_conn.cb_cred); + clp->cl_cb_conn.cb_cred = NULL; } } @@ -719,7 +719,7 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir) return NULL; memcpy(clp->cl_recdir, recdir, HEXDIR_LEN); atomic_set(&clp->cl_count, 1); - atomic_set(&clp->cl_callback.cb_set, 0); + atomic_set(&clp->cl_cb_conn.cb_set, 0); INIT_LIST_HEAD(&clp->cl_idhash); INIT_LIST_HEAD(&clp->cl_strhash); INIT_LIST_HEAD(&clp->cl_openowners); @@ -971,7 +971,7 @@ parse_ipv4(unsigned int addr_len, char *addr_val, unsigned int *cbaddrp, unsigne static void gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se) { - struct nfs4_callback *cb = &clp->cl_callback; + struct nfs4_cb_conn *cb = &clp->cl_cb_conn; /* Currently, we only support tcp for the callback channel */ if ((se->se_callback_netid_len != 3) || memcmp((char *)se->se_callback_netid_val, "tcp", 3)) @@ -1691,7 +1691,7 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp, else { /* XXX: We just turn off callbacks until we can handle * change request correctly. */ - atomic_set(&conf->cl_callback.cb_set, 0); + atomic_set(&conf->cl_cb_conn.cb_set, 0); expire_client(unconf); status = nfs_ok; @@ -2425,7 +2425,7 @@ nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_sta { struct nfs4_delegation *dp; struct nfs4_stateowner *sop = stp->st_stateowner; - struct nfs4_callback *cb = &sop->so_client->cl_callback; + struct nfs4_cb_conn *cb = &sop->so_client->cl_cb_conn; struct file_lock fl, *flp = &fl; int status, flag = 0; @@ -2617,7 +2617,7 @@ nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, renew_client(clp); status = nfserr_cb_path_down; if (!list_empty(&clp->cl_delegations) - && !atomic_read(&clp->cl_callback.cb_set)) + && !atomic_read(&clp->cl_cb_conn.cb_set)) goto out; status = nfs_ok; out: diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 8d882a3..563c367 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -88,7 +88,7 @@ struct nfs4_delegation { #define dl_fh dl_recall.cbr_fh /* client delegation callback info */ -struct nfs4_callback { +struct nfs4_cb_conn { /* SETCLIENTID info */ u32 cb_addr; unsigned short cb_port; @@ -186,7 +186,7 @@ struct nfs4_client { struct svc_cred cl_cred; /* setclientid principal */ clientid_t cl_clientid; /* generated by server */ nfs4_verifier cl_confirm; /* generated by server */ - struct nfs4_callback cl_callback; /* callback info */ + struct nfs4_cb_conn cl_cb_conn; /* callback info */ atomic_t cl_count; /* ref count */ u32 cl_firststate; /* recovery dir creation */ -- cgit v0.10.2 From b53d40c5070bffde1b2bcaf848412a50d8894794 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 1 May 2009 19:50:00 -0400 Subject: nfsd4: eliminate struct nfs4_cb_recall The nfs4_cb_recall struct is used only in nfs4_delegation, so its pointer to the containing delegation is unnecessary--we could just use container_of(). But there's no real reason to have this a separate struct at all--just move these fields to nfs4_delegation. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index ed860d7..2509305 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -215,18 +215,18 @@ encode_cb_compound_hdr(struct xdr_stream *xdr, struct nfs4_cb_compound_hdr *hdr) } static int -encode_cb_recall(struct xdr_stream *xdr, struct nfs4_cb_recall *cb_rec) +encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp) { __be32 *p; - int len = cb_rec->cbr_fh.fh_size; + int len = dp->dl_fh.fh_size; - RESERVE_SPACE(12+sizeof(cb_rec->cbr_stateid) + len); + RESERVE_SPACE(12+sizeof(dp->dl_stateid) + len); WRITE32(OP_CB_RECALL); - WRITE32(cb_rec->cbr_stateid.si_generation); - WRITEMEM(&cb_rec->cbr_stateid.si_opaque, sizeof(stateid_opaque_t)); - WRITE32(cb_rec->cbr_trunc); + WRITE32(dp->dl_stateid.si_generation); + WRITEMEM(&dp->dl_stateid.si_opaque, sizeof(stateid_opaque_t)); + WRITE32(dp->dl_trunc); WRITE32(len); - WRITEMEM(&cb_rec->cbr_fh.fh_base, len); + WRITEMEM(&dp->dl_fh.fh_base, len); return 0; } @@ -241,11 +241,11 @@ nfs4_xdr_enc_cb_null(struct rpc_rqst *req, __be32 *p) } static int -nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p, struct nfs4_cb_recall *args) +nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p, struct nfs4_delegation *args) { struct xdr_stream xdr; struct nfs4_cb_compound_hdr hdr = { - .ident = args->cbr_ident, + .ident = args->dl_ident, .nops = 1, }; @@ -502,17 +502,15 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) { struct nfs4_client *clp = dp->dl_client; struct rpc_clnt *clnt = clp->cl_cb_conn.cb_client; - struct nfs4_cb_recall *cbr = &dp->dl_recall; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL], - .rpc_argp = cbr, + .rpc_argp = dp, .rpc_cred = clp->cl_cb_conn.cb_cred }; int retries = 1; int status = 0; - cbr->cbr_trunc = 0; /* XXX need to implement truncate optimization */ - cbr->cbr_dp = dp; + dp->dl_trunc = 0; /* XXX need to implement truncate optimization */ status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT); while (retries--) { diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index d7b5e6b..3e5345e 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -203,9 +203,8 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f get_file(stp->st_vfs_file); dp->dl_vfs_file = stp->st_vfs_file; dp->dl_type = type; - dp->dl_recall.cbr_dp = NULL; - dp->dl_recall.cbr_ident = cb->cb_ident; - dp->dl_recall.cbr_trunc = 0; + dp->dl_ident = cb->cb_ident; + dp->dl_trunc = 0; dp->dl_stateid.si_boot = get_seconds(); dp->dl_stateid.si_stateownerid = current_delegid++; dp->dl_stateid.si_fileid = 0; diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 563c367..233b60d 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -61,15 +61,6 @@ typedef struct { #define si_stateownerid si_opaque.so_stateownerid #define si_fileid si_opaque.so_fileid - -struct nfs4_cb_recall { - u32 cbr_ident; - int cbr_trunc; - stateid_t cbr_stateid; - struct knfsd_fh cbr_fh; - struct nfs4_delegation *cbr_dp; -}; - struct nfs4_delegation { struct list_head dl_perfile; struct list_head dl_perclnt; @@ -81,12 +72,13 @@ struct nfs4_delegation { struct file *dl_vfs_file; u32 dl_type; time_t dl_time; - struct nfs4_cb_recall dl_recall; +/* For recall: */ + u32 dl_ident; + int dl_trunc; + stateid_t dl_stateid; + struct knfsd_fh dl_fh; }; -#define dl_stateid dl_recall.cbr_stateid -#define dl_fh dl_recall.cbr_fh - /* client delegation callback info */ struct nfs4_cb_conn { /* SETCLIENTID info */ -- cgit v0.10.2 From 6707bd3d420f53ae8f090dac871843f6f43c9980 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 1 May 2009 19:57:46 -0400 Subject: nfsd4: remove unused dl_trunc There's no point in keeping this field around--it's always zero. (Background: the protocol allows you to tell the client that the file is about to be truncated, as an optimization to save the client from writing back dirty pages that will just be discarded. We don't implement this hint. If we do some day, adding this field back in will be the least of the work involved.) Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 2509305..0420b5e 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -224,7 +224,7 @@ encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp) WRITE32(OP_CB_RECALL); WRITE32(dp->dl_stateid.si_generation); WRITEMEM(&dp->dl_stateid.si_opaque, sizeof(stateid_opaque_t)); - WRITE32(dp->dl_trunc); + WRITE32(0); /* truncate optimization not implemented */ WRITE32(len); WRITEMEM(&dp->dl_fh.fh_base, len); return 0; @@ -510,8 +510,6 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) int retries = 1; int status = 0; - dp->dl_trunc = 0; /* XXX need to implement truncate optimization */ - status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT); while (retries--) { switch (status) { diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 3e5345e..cbb16e1 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -204,7 +204,6 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f dp->dl_vfs_file = stp->st_vfs_file; dp->dl_type = type; dp->dl_ident = cb->cb_ident; - dp->dl_trunc = 0; dp->dl_stateid.si_boot = get_seconds(); dp->dl_stateid.si_stateownerid = current_delegid++; dp->dl_stateid.si_fileid = 0; diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 233b60d..346b603 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -74,7 +74,6 @@ struct nfs4_delegation { time_t dl_time; /* For recall: */ u32 dl_ident; - int dl_trunc; stateid_t dl_stateid; struct knfsd_fh dl_fh; }; -- cgit v0.10.2 From 3aea09dc9106407d8bc18e593fbffda9ad632844 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 1 May 2009 20:11:12 -0400 Subject: nfsd4: track recall retries in nfs4_delegation Move this out of a local variable into the nfs4_delegation object in preparation for making this an async rpc call (at which point we'll need any state like this in a common object that's preserved across function calls). Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 0420b5e..b88b207 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -507,11 +507,11 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) .rpc_argp = dp, .rpc_cred = clp->cl_cb_conn.cb_cred }; - int retries = 1; int status = 0; + dp->dl_retries = 1; status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT); - while (retries--) { + while (dp->dl_retries--) { switch (status) { case -EIO: /* Network partition? */ diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 346b603..c0c4921 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -76,6 +76,7 @@ struct nfs4_delegation { u32 dl_ident; stateid_t dl_stateid; struct knfsd_fh dl_fh; + int dl_retries; }; /* client delegation callback info */ -- cgit v0.10.2 From 63e4863fabc6e165a6ca813051305be58966da45 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 1 May 2009 22:36:55 -0400 Subject: nfsd4: make recall callback an asynchronous rpc As with the probe, this removes the need for another kthread. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index b88b207..f4fab69 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -494,6 +494,49 @@ nfsd4_probe_callback(struct nfs4_client *clp) do_probe_callback(clp); } +static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata) +{ + struct nfs4_delegation *dp = calldata; + struct nfs4_client *clp = dp->dl_client; + + switch (task->tk_status) { + case -EIO: + /* Network partition? */ + atomic_set(&clp->cl_cb_conn.cb_set, 0); + warn_no_callback_path(clp, task->tk_status); + case -EBADHANDLE: + case -NFS4ERR_BAD_STATEID: + /* Race: client probably got cb_recall + * before open reply granting delegation */ + break; + default: + /* success, or error we can't handle */ + return; + } + if (dp->dl_retries--) { + rpc_delay(task, 2*HZ); + task->tk_status = 0; + rpc_restart_call(task); + } else { + atomic_set(&clp->cl_cb_conn.cb_set, 0); + warn_no_callback_path(clp, task->tk_status); + } +} + +static void nfsd4_cb_recall_release(void *calldata) +{ + struct nfs4_delegation *dp = calldata; + struct nfs4_client *clp = dp->dl_client; + + nfs4_put_delegation(dp); + put_nfs4_client(clp); +} + +static const struct rpc_call_ops nfsd4_cb_recall_ops = { + .rpc_call_done = nfsd4_cb_recall_done, + .rpc_release = nfsd4_cb_recall_release, +}; + /* * called with dp->dl_count inc'ed. */ @@ -507,32 +550,13 @@ nfsd4_cb_recall(struct nfs4_delegation *dp) .rpc_argp = dp, .rpc_cred = clp->cl_cb_conn.cb_cred }; - int status = 0; + int status; dp->dl_retries = 1; - status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT); - while (dp->dl_retries--) { - switch (status) { - case -EIO: - /* Network partition? */ - atomic_set(&clp->cl_cb_conn.cb_set, 0); - case -EBADHANDLE: - case -NFS4ERR_BAD_STATEID: - /* Race: client probably got cb_recall - * before open reply granting delegation */ - break; - default: - goto out_put_cred; - } - ssleep(2); - status = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT); + status = rpc_call_async(clnt, &msg, RPC_TASK_SOFT, + &nfsd4_cb_recall_ops, dp); + if (status) { + put_nfs4_client(clp); + nfs4_put_delegation(dp); } -out_put_cred: - /* - * Success or failure, now we're either waiting for lease expiration - * or deleg_return. - */ - put_nfs4_client(clp); - nfs4_put_delegation(dp); - return; } diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index cbb16e1..a4bdf25 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2060,19 +2060,6 @@ nfs4_file_downgrade(struct file *filp, unsigned int share_access) } /* - * Recall a delegation - */ -static int -do_recall(void *__dp) -{ - struct nfs4_delegation *dp = __dp; - - dp->dl_file->fi_had_conflict = true; - nfsd4_cb_recall(dp); - return 0; -} - -/* * Spawn a thread to perform a recall on the delegation represented * by the lease (file_lock) * @@ -2083,8 +2070,7 @@ do_recall(void *__dp) static void nfsd_break_deleg_cb(struct file_lock *fl) { - struct nfs4_delegation *dp= (struct nfs4_delegation *)fl->fl_owner; - struct task_struct *t; + struct nfs4_delegation *dp = (struct nfs4_delegation *)fl->fl_owner; dprintk("NFSD nfsd_break_deleg_cb: dp %p fl %p\n",dp,fl); if (!dp) @@ -2112,16 +2098,8 @@ void nfsd_break_deleg_cb(struct file_lock *fl) */ fl->fl_break_time = 0; - t = kthread_run(do_recall, dp, "%s", "nfs4_cb_recall"); - if (IS_ERR(t)) { - struct nfs4_client *clp = dp->dl_client; - - printk(KERN_INFO "NFSD: Callback thread failed for " - "for client (clientid %08x/%08x)\n", - clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id); - put_nfs4_client(dp->dl_client); - nfs4_put_delegation(dp); - } + dp->dl_file->fi_had_conflict = true; + nfsd4_cb_recall(dp); } /* -- cgit v0.10.2 From 9064caae8f47bb7ed5d91712e81f01c1aba2fa3c Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 28 Apr 2009 16:48:25 -0700 Subject: nfsd: use C99 struct initializers Eliminate 56 sparse warnings like this one: fs/nfsd/nfs4xdr.c:1331:15: warning: obsolete array initializer, use C99 syntax Signed-off-by: Randy Dunlap Cc: Neil Brown Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 12d36a7..ab005fc 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -1318,64 +1318,64 @@ static nfsd4_dec nfsd4_dec_ops[] = { }; static nfsd4_dec nfsd41_dec_ops[] = { - [OP_ACCESS] (nfsd4_dec)nfsd4_decode_access, - [OP_CLOSE] (nfsd4_dec)nfsd4_decode_close, - [OP_COMMIT] (nfsd4_dec)nfsd4_decode_commit, - [OP_CREATE] (nfsd4_dec)nfsd4_decode_create, - [OP_DELEGPURGE] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_DELEGRETURN] (nfsd4_dec)nfsd4_decode_delegreturn, - [OP_GETATTR] (nfsd4_dec)nfsd4_decode_getattr, - [OP_GETFH] (nfsd4_dec)nfsd4_decode_noop, - [OP_LINK] (nfsd4_dec)nfsd4_decode_link, - [OP_LOCK] (nfsd4_dec)nfsd4_decode_lock, - [OP_LOCKT] (nfsd4_dec)nfsd4_decode_lockt, - [OP_LOCKU] (nfsd4_dec)nfsd4_decode_locku, - [OP_LOOKUP] (nfsd4_dec)nfsd4_decode_lookup, - [OP_LOOKUPP] (nfsd4_dec)nfsd4_decode_noop, - [OP_NVERIFY] (nfsd4_dec)nfsd4_decode_verify, - [OP_OPEN] (nfsd4_dec)nfsd4_decode_open, - [OP_OPENATTR] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_OPEN_CONFIRM] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_OPEN_DOWNGRADE] (nfsd4_dec)nfsd4_decode_open_downgrade, - [OP_PUTFH] (nfsd4_dec)nfsd4_decode_putfh, - [OP_PUTPUBFH] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_PUTROOTFH] (nfsd4_dec)nfsd4_decode_noop, - [OP_READ] (nfsd4_dec)nfsd4_decode_read, - [OP_READDIR] (nfsd4_dec)nfsd4_decode_readdir, - [OP_READLINK] (nfsd4_dec)nfsd4_decode_noop, - [OP_REMOVE] (nfsd4_dec)nfsd4_decode_remove, - [OP_RENAME] (nfsd4_dec)nfsd4_decode_rename, - [OP_RENEW] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_RESTOREFH] (nfsd4_dec)nfsd4_decode_noop, - [OP_SAVEFH] (nfsd4_dec)nfsd4_decode_noop, - [OP_SECINFO] (nfsd4_dec)nfsd4_decode_secinfo, - [OP_SETATTR] (nfsd4_dec)nfsd4_decode_setattr, - [OP_SETCLIENTID] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_SETCLIENTID_CONFIRM](nfsd4_dec)nfsd4_decode_notsupp, - [OP_VERIFY] (nfsd4_dec)nfsd4_decode_verify, - [OP_WRITE] (nfsd4_dec)nfsd4_decode_write, - [OP_RELEASE_LOCKOWNER] (nfsd4_dec)nfsd4_decode_notsupp, + [OP_ACCESS] = (nfsd4_dec)nfsd4_decode_access, + [OP_CLOSE] = (nfsd4_dec)nfsd4_decode_close, + [OP_COMMIT] = (nfsd4_dec)nfsd4_decode_commit, + [OP_CREATE] = (nfsd4_dec)nfsd4_decode_create, + [OP_DELEGPURGE] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_DELEGRETURN] = (nfsd4_dec)nfsd4_decode_delegreturn, + [OP_GETATTR] = (nfsd4_dec)nfsd4_decode_getattr, + [OP_GETFH] = (nfsd4_dec)nfsd4_decode_noop, + [OP_LINK] = (nfsd4_dec)nfsd4_decode_link, + [OP_LOCK] = (nfsd4_dec)nfsd4_decode_lock, + [OP_LOCKT] = (nfsd4_dec)nfsd4_decode_lockt, + [OP_LOCKU] = (nfsd4_dec)nfsd4_decode_locku, + [OP_LOOKUP] = (nfsd4_dec)nfsd4_decode_lookup, + [OP_LOOKUPP] = (nfsd4_dec)nfsd4_decode_noop, + [OP_NVERIFY] = (nfsd4_dec)nfsd4_decode_verify, + [OP_OPEN] = (nfsd4_dec)nfsd4_decode_open, + [OP_OPENATTR] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_OPEN_CONFIRM] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_OPEN_DOWNGRADE] = (nfsd4_dec)nfsd4_decode_open_downgrade, + [OP_PUTFH] = (nfsd4_dec)nfsd4_decode_putfh, + [OP_PUTPUBFH] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_PUTROOTFH] = (nfsd4_dec)nfsd4_decode_noop, + [OP_READ] = (nfsd4_dec)nfsd4_decode_read, + [OP_READDIR] = (nfsd4_dec)nfsd4_decode_readdir, + [OP_READLINK] = (nfsd4_dec)nfsd4_decode_noop, + [OP_REMOVE] = (nfsd4_dec)nfsd4_decode_remove, + [OP_RENAME] = (nfsd4_dec)nfsd4_decode_rename, + [OP_RENEW] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_RESTOREFH] = (nfsd4_dec)nfsd4_decode_noop, + [OP_SAVEFH] = (nfsd4_dec)nfsd4_decode_noop, + [OP_SECINFO] = (nfsd4_dec)nfsd4_decode_secinfo, + [OP_SETATTR] = (nfsd4_dec)nfsd4_decode_setattr, + [OP_SETCLIENTID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_SETCLIENTID_CONFIRM]= (nfsd4_dec)nfsd4_decode_notsupp, + [OP_VERIFY] = (nfsd4_dec)nfsd4_decode_verify, + [OP_WRITE] = (nfsd4_dec)nfsd4_decode_write, + [OP_RELEASE_LOCKOWNER] = (nfsd4_dec)nfsd4_decode_notsupp, /* new operations for NFSv4.1 */ - [OP_BACKCHANNEL_CTL] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_BIND_CONN_TO_SESSION](nfsd4_dec)nfsd4_decode_notsupp, - [OP_EXCHANGE_ID] (nfsd4_dec)nfsd4_decode_exchange_id, - [OP_CREATE_SESSION] (nfsd4_dec)nfsd4_decode_create_session, - [OP_DESTROY_SESSION] (nfsd4_dec)nfsd4_decode_destroy_session, - [OP_FREE_STATEID] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_GET_DIR_DELEGATION] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_GETDEVICEINFO] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_GETDEVICELIST] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_LAYOUTCOMMIT] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_LAYOUTGET] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_LAYOUTRETURN] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_SECINFO_NO_NAME] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_SEQUENCE] (nfsd4_dec)nfsd4_decode_sequence, - [OP_SET_SSV] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_TEST_STATEID] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_WANT_DELEGATION] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_DESTROY_CLIENTID] (nfsd4_dec)nfsd4_decode_notsupp, - [OP_RECLAIM_COMPLETE] (nfsd4_dec)nfsd4_decode_notsupp, + [OP_BACKCHANNEL_CTL] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_BIND_CONN_TO_SESSION]= (nfsd4_dec)nfsd4_decode_notsupp, + [OP_EXCHANGE_ID] = (nfsd4_dec)nfsd4_decode_exchange_id, + [OP_CREATE_SESSION] = (nfsd4_dec)nfsd4_decode_create_session, + [OP_DESTROY_SESSION] = (nfsd4_dec)nfsd4_decode_destroy_session, + [OP_FREE_STATEID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_GET_DIR_DELEGATION] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_GETDEVICEINFO] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_GETDEVICELIST] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_LAYOUTCOMMIT] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_LAYOUTGET] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_LAYOUTRETURN] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_SECINFO_NO_NAME] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_SEQUENCE] = (nfsd4_dec)nfsd4_decode_sequence, + [OP_SET_SSV] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_TEST_STATEID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_WANT_DELEGATION] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_DESTROY_CLIENTID] = (nfsd4_dec)nfsd4_decode_notsupp, + [OP_RECLAIM_COMPLETE] = (nfsd4_dec)nfsd4_decode_notsupp, }; struct nfsd4_minorversion_ops { -- cgit v0.10.2 From 02cb2858dbfe216bd4a91c4afb4e259ef3b9e151 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Fri, 24 Apr 2009 15:41:57 +0800 Subject: nfsd: nfs4_stat_init cleanup Save some loop time. Signed-off-by: Wang Chen Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index a4bdf25..b276624 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -4002,6 +4002,7 @@ nfs4_state_init(void) INIT_LIST_HEAD(&conf_str_hashtbl[i]); INIT_LIST_HEAD(&unconf_str_hashtbl[i]); INIT_LIST_HEAD(&unconf_id_hashtbl[i]); + INIT_LIST_HEAD(&reclaim_str_hashtbl[i]); } for (i = 0; i < SESSION_HASH_SIZE; i++) INIT_LIST_HEAD(&sessionid_hashtbl[i]); @@ -4024,8 +4025,6 @@ nfs4_state_init(void) INIT_LIST_HEAD(&close_lru); INIT_LIST_HEAD(&client_lru); INIT_LIST_HEAD(&del_recall_lru); - for (i = 0; i < CLIENT_HASH_SIZE; i++) - INIT_LIST_HEAD(&reclaim_str_hashtbl[i]); reclaim_str_hashtbl_size = 0; return 0; } -- cgit v0.10.2 From dd4dc82d4c129babca9b48b219cff0add5ff4132 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 12 May 2009 13:28:09 -0700 Subject: lockd: fix FILE_LOCKING=n build error lockd/svclock.c is missing a header file . is missing a definition of locks_release_private() for the config case of FILE_LOCKING=n, causing a build error: fs/lockd/svclock.c:330: error: implicit declaration of function 'locks_release_private' lockd without FILE_LOCKING doesn't make sense, so make LOCKD and LOCKD_V4 depend on FILE_LOCKING, and make NFS depend on FILE_LOCKING. Signed-off-by: Randy Dunlap Cc: Trond Myklebust Signed-off-by: Andrew Morton Signed-off-by: J. Bruce Fields diff --git a/fs/Kconfig b/fs/Kconfig index 9f7270f..44ab328 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -226,10 +226,12 @@ source "fs/nfsd/Kconfig" config LOCKD tristate + depends on FILE_LOCKING config LOCKD_V4 bool depends on NFSD_V3 || NFS_V3 + depends on FILE_LOCKING default y config EXPORTFS diff --git a/fs/nfs/Kconfig b/fs/nfs/Kconfig index e67f3ec..7dbb8f2 100644 --- a/fs/nfs/Kconfig +++ b/fs/nfs/Kconfig @@ -1,6 +1,6 @@ config NFS_FS tristate "NFS client support" - depends on INET + depends on INET && FILE_LOCKING select LOCKD select SUNRPC select NFS_ACL_SUPPORT if NFS_V3_ACL -- cgit v0.10.2 From fca4217c5bab31019b5247e977673c9fcc385f6b Mon Sep 17 00:00:00 2001 From: Greg Banks Date: Wed, 1 Apr 2009 07:28:13 +1100 Subject: knfsd: reply cache cleanups Make REQHASH() an inline function. Rename hash_list to cache_hash. Fix an obsolete comment. Signed-off-by: Greg Banks Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c index 5bfc2ac..6f0aa49 100644 --- a/fs/nfsd/nfscache.c +++ b/fs/nfsd/nfscache.c @@ -29,15 +29,24 @@ */ #define CACHESIZE 1024 #define HASHSIZE 64 -#define REQHASH(xid) (((((__force __u32)xid) >> 24) ^ ((__force __u32)xid)) & (HASHSIZE-1)) -static struct hlist_head * hash_list; +static struct hlist_head * cache_hash; static struct list_head lru_head; static int cache_disabled = 1; +/* + * Calculate the hash index from an XID. + */ +static inline u32 request_hash(u32 xid) +{ + u32 h = xid; + h ^= (xid >> 24); + return h & (HASHSIZE-1); +} + static int nfsd_cache_append(struct svc_rqst *rqstp, struct kvec *vec); -/* +/* * locking for the reply cache: * A cache entry is "single use" if c_state == RC_INPROG * Otherwise, it when accessing _prev or _next, the lock must be held. @@ -62,8 +71,8 @@ int nfsd_reply_cache_init(void) i--; } - hash_list = kcalloc (HASHSIZE, sizeof(struct hlist_head), GFP_KERNEL); - if (!hash_list) + cache_hash = kcalloc (HASHSIZE, sizeof(struct hlist_head), GFP_KERNEL); + if (!cache_hash) goto out_nomem; cache_disabled = 0; @@ -88,8 +97,8 @@ void nfsd_reply_cache_shutdown(void) cache_disabled = 1; - kfree (hash_list); - hash_list = NULL; + kfree (cache_hash); + cache_hash = NULL; } /* @@ -108,7 +117,7 @@ static void hash_refile(struct svc_cacherep *rp) { hlist_del_init(&rp->c_hash); - hlist_add_head(&rp->c_hash, hash_list + REQHASH(rp->c_xid)); + hlist_add_head(&rp->c_hash, cache_hash + request_hash(rp->c_xid)); } /* @@ -138,7 +147,7 @@ nfsd_cache_lookup(struct svc_rqst *rqstp, int type) spin_lock(&cache_lock); rtn = RC_DOIT; - rh = &hash_list[REQHASH(xid)]; + rh = &cache_hash[request_hash(xid)]; hlist_for_each_entry(rp, hn, rh, c_hash) { if (rp->c_state != RC_UNUSED && xid == rp->c_xid && proc == rp->c_proc && @@ -264,7 +273,7 @@ nfsd_cache_update(struct svc_rqst *rqstp, int cachetype, __be32 *statp) len = resv->iov_len - ((char*)statp - (char*)resv->iov_base); len >>= 2; - + /* Don't cache excessive amounts of data and XDR failures */ if (!statp || len > (256 >> 2)) { rp->c_state = RC_UNUSED; diff --git a/include/linux/nfsd/cache.h b/include/linux/nfsd/cache.h index 5bccaab..3a3f589 100644 --- a/include/linux/nfsd/cache.h +++ b/include/linux/nfsd/cache.h @@ -14,8 +14,7 @@ #include /* - * Representation of a reply cache entry. The first two members *must* - * be hash_next and hash_prev. + * Representation of a reply cache entry. */ struct svc_cacherep { struct hlist_node c_hash; -- cgit v0.10.2 From cf0a586cf41a1779edeee7562afb5d0ab46c7cf4 Mon Sep 17 00:00:00 2001 From: Greg Banks Date: Wed, 1 Apr 2009 07:28:15 +1100 Subject: knfsd: fix reply cache memory corruption Fix a regression in the reply cache introduced when the code was converted to use proper Linux lists. When a new entry needs to be inserted, the case where all the entries are currently being used by threads is not correctly detected. This can result in memory corruption and a crash. In the current code this is an extremely unlikely corner case; it would require the machine to have 1024 nfsd threads and all of them to be busy at the same time. However, upcoming reply cache changes make this more likely; a crash due to this problem was actually observed in field. Signed-off-by: Greg Banks Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c index 6f0aa49..4638635 100644 --- a/fs/nfsd/nfscache.c +++ b/fs/nfsd/nfscache.c @@ -174,8 +174,8 @@ nfsd_cache_lookup(struct svc_rqst *rqstp, int type) } } - /* This should not happen */ - if (rp == NULL) { + /* All entries on the LRU are in-progress. This should not happen */ + if (&rp->c_lru == &lru_head) { static int complaints; printk(KERN_WARNING "nfsd: all repcache entries locked!\n"); -- cgit v0.10.2 From 1dbd0d53f394cd9a86fc801dd68fdbcbcdb45718 Mon Sep 17 00:00:00 2001 From: Greg Banks Date: Wed, 1 Apr 2009 07:28:21 +1100 Subject: knfsd: remove unreported filehandle stats counters The file nfsfh.c contains two static variables nfsd_nr_verified and nfsd_nr_put. These are counters which are incremented as a side effect of the fh_verify() fh_compose() and fh_put() operations, i.e. at least twice per NFS call for any non-trivial workload. Needless to say this makes the cacheline that contains them (and any other innocent victims) a very hot contention point indeed under high call-rate workloads on multiprocessor NFS server. It also turns out that these counters are not used anywhere. They're not reported to userspace, they're not used in logic, they're not even exported from the object file (let alone the module). All they do is waste CPU time. So this patch removes them. Tests on a 16 CPU Altix A4700 with 2 10gige Myricom cards, configured separately (no bonding). Workload is 640 client threads doing directory traverals with random small reads, from server RAM. Before ====== Kernel profile: % cumulative self self total time samples samples calls 1/call 1/call name 6.05 2716.00 2716.00 30406 0.09 1.02 svc_process 4.44 4706.00 1990.00 1975 1.01 1.01 spin_unlock_irqrestore 3.72 6376.00 1670.00 1666 1.00 1.00 svc_export_put 3.41 7907.00 1531.00 1786 0.86 1.02 nfsd_ofcache_lookup 3.25 9363.00 1456.00 10965 0.13 1.01 nfsd_dispatch 3.10 10752.00 1389.00 1376 1.01 1.01 nfsd_cache_lookup 2.57 11907.00 1155.00 4517 0.26 1.03 svc_tcp_recvfrom ... 2.21 15352.00 1003.00 1081 0.93 1.00 nfsd_choose_ofc <---- ^^^^ Here the function nfsd_choose_ofc() reads a global variable which by accident happened to be located in the same cacheline as nfsd_nr_verified. Call rate: nullarbor:~ # pmdumptext nfs3.server.calls ... Thu Dec 13 00:15:27 184780.663 Thu Dec 13 00:15:28 184885.881 Thu Dec 13 00:15:29 184449.215 Thu Dec 13 00:15:30 184971.058 Thu Dec 13 00:15:31 185036.052 Thu Dec 13 00:15:32 185250.475 Thu Dec 13 00:15:33 184481.319 Thu Dec 13 00:15:34 185225.737 Thu Dec 13 00:15:35 185408.018 Thu Dec 13 00:15:36 185335.764 After ===== kernel profile: % cumulative self self total time samples samples calls 1/call 1/call name 6.33 2813.00 2813.00 29979 0.09 1.01 svc_process 4.66 4883.00 2070.00 2065 1.00 1.00 spin_unlock_irqrestore 4.06 6687.00 1804.00 2182 0.83 1.00 nfsd_ofcache_lookup 3.20 8110.00 1423.00 10932 0.13 1.00 nfsd_dispatch 3.03 9456.00 1346.00 1343 1.00 1.00 nfsd_cache_lookup 2.62 10622.00 1166.00 4645 0.25 1.01 svc_tcp_recvfrom [...] 0.10 42586.00 44.00 74 0.59 1.00 nfsd_choose_ofc <--- HA!! ^^^^ Call rate: nullarbor:~ # pmdumptext nfs3.server.calls ... Thu Dec 13 01:45:28 194677.118 Thu Dec 13 01:45:29 193932.692 Thu Dec 13 01:45:30 194294.364 Thu Dec 13 01:45:31 194971.276 Thu Dec 13 01:45:32 194111.207 Thu Dec 13 01:45:33 194999.635 Thu Dec 13 01:45:34 195312.594 Thu Dec 13 01:45:35 195707.293 Thu Dec 13 01:45:36 194610.353 Thu Dec 13 01:45:37 195913.662 Thu Dec 13 01:45:38 194808.675 i.e. about a 5.3% improvement in call rate. Signed-off-by: Greg Banks Reviewed-by: David Chinner Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 9f1ca17..8847f3f 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -27,9 +27,6 @@ #define NFSDDBG_FACILITY NFSDDBG_FH -static int nfsd_nr_verified; -static int nfsd_nr_put; - /* * our acceptability function. * if NOSUBTREECHECK, accept anything @@ -251,7 +248,6 @@ static __be32 nfsd_set_fh_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp) fhp->fh_dentry = dentry; fhp->fh_export = exp; - nfsd_nr_verified++; return 0; out: exp_put(exp); @@ -552,7 +548,6 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, return nfserr_opnotsupp; } - nfsd_nr_verified++; return 0; } @@ -609,7 +604,6 @@ fh_put(struct svc_fh *fhp) fhp->fh_pre_saved = 0; fhp->fh_post_saved = 0; #endif - nfsd_nr_put++; } if (exp) { cache_put(&exp->h, &svc_export_cache); -- cgit v0.10.2 From 3c8e03166ae234d16e7871f8009638e0946d303c Mon Sep 17 00:00:00 2001 From: Yu Zhiguo Date: Sat, 16 May 2009 16:22:31 +0800 Subject: NFSv4: do exact check about attribute specified Server should return NFS4ERR_ATTRNOTSUPP if an attribute specified is not supported in current environment. Operations CREATE, NVERIFY, OPEN, SETATTR and VERIFY should do this check. This bug is found when do newpynfs tests. The names of the tests that failed are following: CR12 NVF7a NVF7b NVF7c NVF7d NVF7f NVF7r NVF7s OPEN15 VF7a VF7b VF7c VF7d VF7f VF7r VF7s Add function do_check_fattr() to do exact check: 1, Check attribute specified is supported by the NFSv4 server or not. 2, Check FATTR4_WORD0_ACL & FATTR4_WORD0_FS_LOCATIONS are supported in current environment or not. 3, Check attribute specified is writable or not. step 1 and 3 are done in function nfsd4_decode_fattr() but removed to this function now. Signed-off-by: Yu Zhiguo Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index b2883e9..9272e1f 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -51,6 +51,78 @@ #define NFSDDBG_FACILITY NFSDDBG_PROC +static u32 nfsd_attrmask[] = { + NFSD_WRITEABLE_ATTRS_WORD0, + NFSD_WRITEABLE_ATTRS_WORD1, + NFSD_WRITEABLE_ATTRS_WORD2 +}; + +static u32 nfsd41_ex_attrmask[] = { + NFSD_SUPPATTR_EXCLCREAT_WORD0, + NFSD_SUPPATTR_EXCLCREAT_WORD1, + NFSD_SUPPATTR_EXCLCREAT_WORD2 +}; + +static __be32 +check_attr_support(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, + u32 *bmval, u32 *writable) +{ + struct dentry *dentry = cstate->current_fh.fh_dentry; + struct svc_export *exp = cstate->current_fh.fh_export; + + /* + * Check about attributes are supported by the NFSv4 server or not. + * According to spec, unsupported attributes return ERR_ATTRNOTSUPP. + */ + if ((bmval[0] & ~nfsd_suppattrs0(cstate->minorversion)) || + (bmval[1] & ~nfsd_suppattrs1(cstate->minorversion)) || + (bmval[2] & ~nfsd_suppattrs2(cstate->minorversion))) + return nfserr_attrnotsupp; + + /* + * Check FATTR4_WORD0_ACL & FATTR4_WORD0_FS_LOCATIONS can be supported + * in current environment or not. + */ + if (bmval[0] & FATTR4_WORD0_ACL) { + if (!IS_POSIXACL(dentry->d_inode)) + return nfserr_attrnotsupp; + } + if (bmval[0] & FATTR4_WORD0_FS_LOCATIONS) { + if (exp->ex_fslocs.locations == NULL) + return nfserr_attrnotsupp; + } + + /* + * According to spec, read-only attributes return ERR_INVAL. + */ + if (writable) { + if ((bmval[0] & ~writable[0]) || (bmval[1] & ~writable[1]) || + (bmval[2] & ~writable[2])) + return nfserr_inval; + } + + return nfs_ok; +} + +static __be32 +nfsd4_check_open_attributes(struct svc_rqst *rqstp, + struct nfsd4_compound_state *cstate, struct nfsd4_open *open) +{ + __be32 status = nfs_ok; + + if (open->op_create == NFS4_OPEN_CREATE) { + if (open->op_createmode == NFS4_CREATE_UNCHECKED + || open->op_createmode == NFS4_CREATE_GUARDED) + status = check_attr_support(rqstp, cstate, + open->op_bmval, nfsd_attrmask); + else if (open->op_createmode == NFS4_CREATE_EXCLUSIVE4_1) + status = check_attr_support(rqstp, cstate, + open->op_bmval, nfsd41_ex_attrmask); + } + + return status; +} + static inline void fh_dup2(struct svc_fh *dst, struct svc_fh *src) { @@ -225,6 +297,10 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (status) goto out; + status = nfsd4_check_open_attributes(rqstp, cstate, open); + if (status) + goto out; + /* Openowner is now set, so sequence id will get bumped. Now we need * these checks before we do any creates: */ status = nfserr_grace; @@ -395,6 +471,11 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (status) return status; + status = check_attr_support(rqstp, cstate, create->cr_bmval, + nfsd_attrmask); + if (status) + return status; + switch (create->cr_type) { case NF4LNK: /* ugh! we have to null-terminate the linktext, or @@ -689,6 +770,12 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (status) return status; status = nfs_ok; + + status = check_attr_support(rqstp, cstate, setattr->sa_bmval, + nfsd_attrmask); + if (status) + goto out; + if (setattr->sa_acl != NULL) status = nfsd4_set_nfs4_acl(rqstp, &cstate->current_fh, setattr->sa_acl); @@ -763,10 +850,10 @@ _nfsd4_verify(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (status) return status; - if ((verify->ve_bmval[0] & ~nfsd_suppattrs0(cstate->minorversion)) - || (verify->ve_bmval[1] & ~nfsd_suppattrs1(cstate->minorversion)) - || (verify->ve_bmval[2] & ~nfsd_suppattrs2(cstate->minorversion))) - return nfserr_attrnotsupp; + status = check_attr_support(rqstp, cstate, verify->ve_bmval, NULL); + if (status) + return status; + if ((verify->ve_bmval[0] & FATTR4_WORD0_RDATTR_ERROR) || (verify->ve_bmval[1] & NFSD_WRITEONLY_ATTRS_WORD1)) return nfserr_inval; diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index ab005fc..254e5b2 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -244,20 +244,8 @@ nfsd4_decode_bitmap(struct nfsd4_compoundargs *argp, u32 *bmval) DECODE_TAIL; } -static u32 nfsd_attrmask[] = { - NFSD_WRITEABLE_ATTRS_WORD0, - NFSD_WRITEABLE_ATTRS_WORD1, - NFSD_WRITEABLE_ATTRS_WORD2 -}; - -static u32 nfsd41_ex_attrmask[] = { - NFSD_SUPPATTR_EXCLCREAT_WORD0, - NFSD_SUPPATTR_EXCLCREAT_WORD1, - NFSD_SUPPATTR_EXCLCREAT_WORD2 -}; - static __be32 -nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, u32 *writable, +nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, struct iattr *iattr, struct nfs4_acl **acl) { int expected_len, len = 0; @@ -270,18 +258,6 @@ nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, u32 *writable, if ((status = nfsd4_decode_bitmap(argp, bmval))) return status; - /* - * According to spec, unsupported attributes return ERR_ATTRNOTSUPP; - * read-only attributes return ERR_INVAL. - */ - if ((bmval[0] & ~nfsd_suppattrs0(argp->minorversion)) || - (bmval[1] & ~nfsd_suppattrs1(argp->minorversion)) || - (bmval[2] & ~nfsd_suppattrs2(argp->minorversion))) - return nfserr_attrnotsupp; - if ((bmval[0] & ~writable[0]) || (bmval[1] & ~writable[1]) || - (bmval[2] & ~writable[2])) - return nfserr_inval; - READ_BUF(4); READ32(expected_len); @@ -414,8 +390,11 @@ nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, u32 *writable, goto xdr_error; } } - BUG_ON(bmval[2]); /* no such writeable attr supported yet */ - if (len != expected_len) + if (bmval[0] & ~NFSD_WRITEABLE_ATTRS_WORD0 + || bmval[1] & ~NFSD_WRITEABLE_ATTRS_WORD1 + || bmval[2] & ~NFSD_WRITEABLE_ATTRS_WORD2) + READ_BUF(expected_len - len); + else if (len != expected_len) goto xdr_error; DECODE_TAIL; @@ -508,8 +487,8 @@ nfsd4_decode_create(struct nfsd4_compoundargs *argp, struct nfsd4_create *create if ((status = check_filename(create->cr_name, create->cr_namelen, nfserr_inval))) return status; - status = nfsd4_decode_fattr(argp, create->cr_bmval, nfsd_attrmask, - &create->cr_iattr, &create->cr_acl); + status = nfsd4_decode_fattr(argp, create->cr_bmval, &create->cr_iattr, + &create->cr_acl); if (status) goto out; @@ -672,7 +651,7 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) case NFS4_CREATE_UNCHECKED: case NFS4_CREATE_GUARDED: status = nfsd4_decode_fattr(argp, open->op_bmval, - nfsd_attrmask, &open->op_iattr, &open->op_acl); + &open->op_iattr, &open->op_acl); if (status) goto out; break; @@ -686,8 +665,7 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) READ_BUF(8); COPYMEM(open->op_verf.data, 8); status = nfsd4_decode_fattr(argp, open->op_bmval, - nfsd41_ex_attrmask, &open->op_iattr, - &open->op_acl); + &open->op_iattr, &open->op_acl); if (status) goto out; break; @@ -883,8 +861,8 @@ nfsd4_decode_setattr(struct nfsd4_compoundargs *argp, struct nfsd4_setattr *seta status = nfsd4_decode_stateid(argp, &setattr->sa_stateid); if (status) return status; - return nfsd4_decode_fattr(argp, setattr->sa_bmval, nfsd_attrmask, - &setattr->sa_iattr, &setattr->sa_acl); + return nfsd4_decode_fattr(argp, setattr->sa_bmval, &setattr->sa_iattr, + &setattr->sa_acl); } static __be32 -- cgit v0.10.2 From 0a93a47f042c459f0f46942c3a920e3c81878031 Mon Sep 17 00:00:00 2001 From: Yu Zhiguo Date: Tue, 19 May 2009 14:09:54 +0800 Subject: NFSv4: kill off complicated macro 'PROC' J. Bruce Fields wrote: ... > (This is extremely confusing code to track down: note that > proc->pc_decode is set to nfs4svc_decode_compoundargs() by the PROC() > macro at the end of fs/nfsd/nfs4proc.c. Which means, for example, that > grepping for nfs4svc_decode_compoundargs() gets you nowhere. Patches to > kill off that macro would be welcomed....) the macro 'PROC' is complicated and obscure, it had better be killed off in order to make the code more clear. Signed-off-by: Yu Zhiguo Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 9272e1f..7c88017 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -1313,24 +1313,9 @@ static const char *nfsd4_op_name(unsigned opnum) return "unknown_operation"; } -#define nfs4svc_decode_voidargs NULL -#define nfs4svc_release_void NULL #define nfsd4_voidres nfsd4_voidargs -#define nfs4svc_release_compound NULL struct nfsd4_voidargs { int dummy; }; -#define PROC(name, argt, rest, relt, cache, respsize) \ - { (svc_procfunc) nfsd4_proc_##name, \ - (kxdrproc_t) nfs4svc_decode_##argt##args, \ - (kxdrproc_t) nfs4svc_encode_##rest##res, \ - (kxdrproc_t) nfs4svc_release_##relt, \ - sizeof(struct nfsd4_##argt##args), \ - sizeof(struct nfsd4_##rest##res), \ - 0, \ - cache, \ - respsize, \ - } - /* * TODO: At the present time, the NFSv4 server does not do XID caching * of requests. Implementing XID caching would not be a serious problem, @@ -1342,8 +1327,23 @@ struct nfsd4_voidargs { int dummy; }; * better XID's. */ static struct svc_procedure nfsd_procedures4[2] = { - PROC(null, void, void, void, RC_NOCACHE, 1), - PROC(compound, compound, compound, compound, RC_NOCACHE, NFSD_BUFSIZE/4) + [NFSPROC4_NULL] = { + .pc_func = (svc_procfunc) nfsd4_proc_null, + .pc_encode = (kxdrproc_t) nfs4svc_encode_voidres, + .pc_argsize = sizeof(struct nfsd4_voidargs), + .pc_ressize = sizeof(struct nfsd4_voidres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = 1, + }, + [NFSPROC4_COMPOUND] = { + .pc_func = (svc_procfunc) nfsd4_proc_compound, + .pc_decode = (kxdrproc_t) nfs4svc_decode_compoundargs, + .pc_encode = (kxdrproc_t) nfs4svc_encode_compoundres, + .pc_argsize = sizeof(struct nfsd4_compoundargs), + .pc_ressize = sizeof(struct nfsd4_compoundres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = NFSD_BUFSIZE/4, + }, }; struct svc_version nfsd_version4 = { -- cgit v0.10.2 From 48e03bc515cff75718de5460458680a230ae997e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 5 Jun 2009 12:35:15 -0400 Subject: nfsd: Use write gathering only with NFSv2 NFSv3 and above can use unstable writes whenever they are sending more than one write, rather than relying on the flaky write gathering heuristics. More often than not, write gathering is currently getting it wrong when the NFSv3 clients are sending a single write with FILE_SYNC for efficiency reasons. This patch turns off write gathering for NFSv3/v4, and ensures that it only applies to the one case that can actually benefit: namely NFSv2. Signed-off-by: Trond Myklebust Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index b660435..f30cc4e 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -975,6 +975,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, __be32 err = 0; int host_err; int stable = *stablep; + int use_wgather; #ifdef MSNFS err = nfserr_perm; @@ -993,9 +994,10 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, * - the sync export option has been set, or * - the client requested O_SYNC behavior (NFSv3 feature). * - The file system doesn't support fsync(). - * When gathered writes have been configured for this volume, + * When NFSv2 gathered writes have been configured for this volume, * flushing the data to disk is handled separately below. */ + use_wgather = (rqstp->rq_vers == 2) && EX_WGATHER(exp); if (!file->f_op->fsync) {/* COMMIT3 cannot work */ stable = 2; @@ -1004,7 +1006,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, if (!EX_ISSYNC(exp)) stable = 0; - if (stable && !EX_WGATHER(exp)) { + if (stable && !use_wgather) { spin_lock(&file->f_lock); file->f_flags |= O_SYNC; spin_unlock(&file->f_lock); @@ -1040,7 +1042,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, * nice and simple solution (IMHO), and it seems to * work:-) */ - if (EX_WGATHER(exp)) { + if (use_wgather) { if (atomic_read(&inode->i_writecount) > 1 || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) { dprintk("nfsd: write defer %d\n", task_pid_nr(current)); -- cgit v0.10.2 From 6aad89c8376e432231b78d1bdbcc10ef7708b011 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 10 Jun 2009 12:52:21 -0700 Subject: sunrpc: align cache_clean work's timer Align cache_clean work. Signed-off-by: Anton Blanchard Cc: Neil Brown Cc: Trond Myklebust Cc: "David S. Miller" Signed-off-by: Andrew Morton Acked-by: David S. Miller Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 20029a7..ff0c230 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -488,7 +488,7 @@ static void do_cache_clean(struct work_struct *work) { int delay = 5; if (cache_clean() == -1) - delay = 30*HZ; + delay = round_jiffies_relative(30*HZ); if (list_empty(&cache_list)) delay = 0; -- cgit v0.10.2 From 9d2a3f31d6d7832cd441eeda08bc2266cdd5d972 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 15 Jun 2009 18:47:56 -0700 Subject: nfsd: track last inode only in use_wgather case Updating last_ino and last_dev probably isn't useful in the !use_wgather case. Also remove some pointless ifdef'd-out code. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index f30cc4e..ebf56c6 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1026,7 +1026,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, if (host_err >= 0 && (inode->i_mode & (S_ISUID | S_ISGID))) kill_suid(dentry); - if (host_err >= 0 && stable) { + if (host_err >= 0 && stable && use_wgather) { static ino_t last_ino; static dev_t last_dev; @@ -1042,21 +1042,16 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, * nice and simple solution (IMHO), and it seems to * work:-) */ - if (use_wgather) { - if (atomic_read(&inode->i_writecount) > 1 - || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) { - dprintk("nfsd: write defer %d\n", task_pid_nr(current)); - msleep(10); - dprintk("nfsd: write resume %d\n", task_pid_nr(current)); - } + if (atomic_read(&inode->i_writecount) > 1 + || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) { + dprintk("nfsd: write defer %d\n", task_pid_nr(current)); + msleep(10); + dprintk("nfsd: write resume %d\n", task_pid_nr(current)); + } - if (inode->i_state & I_DIRTY) { - dprintk("nfsd: write sync %d\n", task_pid_nr(current)); - host_err=nfsd_sync(file); - } -#if 0 - wake_up(&inode->i_wait); -#endif + if (inode->i_state & I_DIRTY) { + dprintk("nfsd: write sync %d\n", task_pid_nr(current)); + host_err=nfsd_sync(file); } last_ino = inode->i_ino; last_dev = inode->i_sb->s_dev; -- cgit v0.10.2 From d911df7b8d44de41661363a4e29ee710180ba025 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 15 Jun 2009 16:03:53 -0700 Subject: nfsd: Pull write-gathering code out of nfsd_vfs_write This is a relatively self-contained piece of code that handles a special case--move it to its own function. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index ebf56c6..6ad76a4 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -963,6 +963,43 @@ static void kill_suid(struct dentry *dentry) mutex_unlock(&dentry->d_inode->i_mutex); } +/* + * Gathered writes: If another process is currently writing to the file, + * there's a high chance this is another nfsd (triggered by a bulk write + * from a client's biod). Rather than syncing the file with each write + * request, we sleep for 10 msec. + * + * I don't know if this roughly approximates C. Juszak's idea of + * gathered writes, but it's a nice and simple solution (IMHO), and it + * seems to work:-) + * + * Note: we do this only in the NFSv2 case, since v3 and higher have a + * better tool (separate unstable writes and commits) for solving this + * problem. + */ +static int wait_for_concurrent_writes(struct file *file) +{ + struct inode *inode = file->f_path.dentry->d_inode; + static ino_t last_ino; + static dev_t last_dev; + int err = 0; + + if (atomic_read(&inode->i_writecount) > 1 + || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) { + dprintk("nfsd: write defer %d\n", task_pid_nr(current)); + msleep(10); + dprintk("nfsd: write resume %d\n", task_pid_nr(current)); + } + + if (inode->i_state & I_DIRTY) { + dprintk("nfsd: write sync %d\n", task_pid_nr(current)); + err = nfsd_sync(file); + } + last_ino = inode->i_ino; + last_dev = inode->i_sb->s_dev; + return err; +} + static __be32 nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, loff_t offset, struct kvec *vec, int vlen, @@ -1026,36 +1063,8 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, if (host_err >= 0 && (inode->i_mode & (S_ISUID | S_ISGID))) kill_suid(dentry); - if (host_err >= 0 && stable && use_wgather) { - static ino_t last_ino; - static dev_t last_dev; - - /* - * Gathered writes: If another process is currently - * writing to the file, there's a high chance - * this is another nfsd (triggered by a bulk write - * from a client's biod). Rather than syncing the - * file with each write request, we sleep for 10 msec. - * - * I don't know if this roughly approximates - * C. Juszak's idea of gathered writes, but it's a - * nice and simple solution (IMHO), and it seems to - * work:-) - */ - if (atomic_read(&inode->i_writecount) > 1 - || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) { - dprintk("nfsd: write defer %d\n", task_pid_nr(current)); - msleep(10); - dprintk("nfsd: write resume %d\n", task_pid_nr(current)); - } - - if (inode->i_state & I_DIRTY) { - dprintk("nfsd: write sync %d\n", task_pid_nr(current)); - host_err=nfsd_sync(file); - } - last_ino = inode->i_ino; - last_dev = inode->i_sb->s_dev; - } + if (host_err >= 0 && stable && use_wgather) + host_err = wait_for_concurrent_writes(file); dprintk("nfsd: write complete host_err=%d\n", host_err); if (host_err >= 0) -- cgit v0.10.2 From e4636d535e32768c8c500641ddb144f56e3dc5c0 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 15 Jun 2009 19:07:13 -0700 Subject: nfsd: minor nfsd_vfs_write cleanup There's no need to check host_err >= 0 every time here when we could check host_err < 0 once, following the usual kernel style. Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 6ad76a4..1cf7061 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1053,19 +1053,20 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, oldfs = get_fs(); set_fs(KERNEL_DS); host_err = vfs_writev(file, (struct iovec __user *)vec, vlen, &offset); set_fs(oldfs); - if (host_err >= 0) { - *cnt = host_err; - nfsdstats.io_write += host_err; - fsnotify_modify(file->f_path.dentry); - } + if (host_err < 0) + goto out_nfserr; + *cnt = host_err; + nfsdstats.io_write += host_err; + fsnotify_modify(file->f_path.dentry); /* clear setuid/setgid flag after write */ - if (host_err >= 0 && (inode->i_mode & (S_ISUID | S_ISGID))) + if (inode->i_mode & (S_ISUID | S_ISGID)) kill_suid(dentry); - if (host_err >= 0 && stable && use_wgather) + if (stable && use_wgather) host_err = wait_for_concurrent_writes(file); +out_nfserr: dprintk("nfsd: write complete host_err=%d\n", host_err); if (host_err >= 0) err = 0; -- cgit v0.10.2 From 59fb30660be3f3ead54b3850c401d462449caaaa Mon Sep 17 00:00:00 2001 From: Christian Engelmayer Date: Sun, 14 Jun 2009 00:05:26 +0200 Subject: sunrpc: potential memory leak in function rdma_read_xdr In case the check on ch_count fails the cleanup path is skipped and the previously allocated memory 'rpl_map', 'chl_map' is not freed. Reported by Coverity. Signed-off-by: Christian Engelmayer Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c index 42a6f9f..9e88438 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c +++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c @@ -397,14 +397,14 @@ static int rdma_read_xdr(struct svcxprt_rdma *xprt, if (!ch) return 0; - /* Allocate temporary reply and chunk maps */ - rpl_map = svc_rdma_get_req_map(); - chl_map = svc_rdma_get_req_map(); - svc_rdma_rcl_chunk_counts(ch, &ch_count, &byte_count); if (ch_count > RPCSVC_MAXPAGES) return -EINVAL; + /* Allocate temporary reply and chunk maps */ + rpl_map = svc_rdma_get_req_map(); + chl_map = svc_rdma_get_req_map(); + if (!xprt->sc_frmr_pg_list_len) sge_count = map_read_chunks(xprt, rqstp, hdr_ctxt, rmsgp, rpl_map, chl_map, ch_count, -- cgit v0.10.2 From b9081d90f5b989cd927052084b16b4f950c8c8d7 Mon Sep 17 00:00:00 2001 From: Yu Zhiguo Date: Tue, 9 Jun 2009 17:33:34 +0800 Subject: NFS: kill off complicated macro 'PROC' kill off obscure macro 'PROC' of NFSv2&3 in order to make the code more clear. Among other things, this makes it simpler to grep for callers of these functions--something which has frequently caused confusion among nfs developers. Signed-off-by: Yu Zhiguo Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs3proc.c b/fs/nfsd/nfs3proc.c index 7c9fe83..a713c41 100644 --- a/fs/nfsd/nfs3proc.c +++ b/fs/nfsd/nfs3proc.c @@ -652,8 +652,6 @@ nfsd3_proc_commit(struct svc_rqst * rqstp, struct nfsd3_commitargs *argp, * NFSv3 Server procedures. * Only the results of non-idempotent operations are cached. */ -#define nfs3svc_decode_voidargs NULL -#define nfs3svc_release_void NULL #define nfs3svc_decode_fhandleargs nfs3svc_decode_fhandle #define nfs3svc_encode_attrstatres nfs3svc_encode_attrstat #define nfs3svc_encode_wccstatres nfs3svc_encode_wccstat @@ -686,28 +684,219 @@ struct nfsd3_voidargs { int dummy; }; #define WC (7+pAT) /* WCC attributes */ static struct svc_procedure nfsd_procedures3[22] = { - PROC(null, void, void, void, RC_NOCACHE, ST), - PROC(getattr, fhandle, attrstat, fhandle, RC_NOCACHE, ST+AT), - PROC(setattr, sattr, wccstat, fhandle, RC_REPLBUFF, ST+WC), - PROC(lookup, dirop, dirop, fhandle2, RC_NOCACHE, ST+FH+pAT+pAT), - PROC(access, access, access, fhandle, RC_NOCACHE, ST+pAT+1), - PROC(readlink, readlink, readlink, fhandle, RC_NOCACHE, ST+pAT+1+NFS3_MAXPATHLEN/4), - PROC(read, read, read, fhandle, RC_NOCACHE, ST+pAT+4+NFSSVC_MAXBLKSIZE/4), - PROC(write, write, write, fhandle, RC_REPLBUFF, ST+WC+4), - PROC(create, create, create, fhandle2, RC_REPLBUFF, ST+(1+FH+pAT)+WC), - PROC(mkdir, mkdir, create, fhandle2, RC_REPLBUFF, ST+(1+FH+pAT)+WC), - PROC(symlink, symlink, create, fhandle2, RC_REPLBUFF, ST+(1+FH+pAT)+WC), - PROC(mknod, mknod, create, fhandle2, RC_REPLBUFF, ST+(1+FH+pAT)+WC), - PROC(remove, dirop, wccstat, fhandle, RC_REPLBUFF, ST+WC), - PROC(rmdir, dirop, wccstat, fhandle, RC_REPLBUFF, ST+WC), - PROC(rename, rename, rename, fhandle2, RC_REPLBUFF, ST+WC+WC), - PROC(link, link, link, fhandle2, RC_REPLBUFF, ST+pAT+WC), - PROC(readdir, readdir, readdir, fhandle, RC_NOCACHE, 0), - PROC(readdirplus,readdirplus, readdir, fhandle, RC_NOCACHE, 0), - PROC(fsstat, fhandle, fsstat, void, RC_NOCACHE, ST+pAT+2*6+1), - PROC(fsinfo, fhandle, fsinfo, void, RC_NOCACHE, ST+pAT+12), - PROC(pathconf, fhandle, pathconf, void, RC_NOCACHE, ST+pAT+6), - PROC(commit, commit, commit, fhandle, RC_NOCACHE, ST+WC+2), + [NFS3PROC_NULL] = { + .pc_func = (svc_procfunc) nfsd3_proc_null, + .pc_encode = (kxdrproc_t) nfs3svc_encode_voidres, + .pc_argsize = sizeof(struct nfsd3_voidargs), + .pc_ressize = sizeof(struct nfsd3_voidres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST, + }, + [NFS3PROC_GETATTR] = { + .pc_func = (svc_procfunc) nfsd3_proc_getattr, + .pc_decode = (kxdrproc_t) nfs3svc_decode_fhandleargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_attrstatres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_fhandleargs), + .pc_ressize = sizeof(struct nfsd3_attrstatres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+AT, + }, + [NFS3PROC_SETATTR] = { + .pc_func = (svc_procfunc) nfsd3_proc_setattr, + .pc_decode = (kxdrproc_t) nfs3svc_decode_sattrargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_wccstatres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_sattrargs), + .pc_ressize = sizeof(struct nfsd3_wccstatres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+WC, + }, + [NFS3PROC_LOOKUP] = { + .pc_func = (svc_procfunc) nfsd3_proc_lookup, + .pc_decode = (kxdrproc_t) nfs3svc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_diropres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_diropargs), + .pc_ressize = sizeof(struct nfsd3_diropres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+FH+pAT+pAT, + }, + [NFS3PROC_ACCESS] = { + .pc_func = (svc_procfunc) nfsd3_proc_access, + .pc_decode = (kxdrproc_t) nfs3svc_decode_accessargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_accessres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_accessargs), + .pc_ressize = sizeof(struct nfsd3_accessres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+1, + }, + [NFS3PROC_READLINK] = { + .pc_func = (svc_procfunc) nfsd3_proc_readlink, + .pc_decode = (kxdrproc_t) nfs3svc_decode_readlinkargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_readlinkres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_readlinkargs), + .pc_ressize = sizeof(struct nfsd3_readlinkres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+1+NFS3_MAXPATHLEN/4, + }, + [NFS3PROC_READ] = { + .pc_func = (svc_procfunc) nfsd3_proc_read, + .pc_decode = (kxdrproc_t) nfs3svc_decode_readargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_readres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_readargs), + .pc_ressize = sizeof(struct nfsd3_readres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+4+NFSSVC_MAXBLKSIZE/4, + }, + [NFS3PROC_WRITE] = { + .pc_func = (svc_procfunc) nfsd3_proc_write, + .pc_decode = (kxdrproc_t) nfs3svc_decode_writeargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_writeres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_writeargs), + .pc_ressize = sizeof(struct nfsd3_writeres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+WC+4, + }, + [NFS3PROC_CREATE] = { + .pc_func = (svc_procfunc) nfsd3_proc_create, + .pc_decode = (kxdrproc_t) nfs3svc_decode_createargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_createres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_createargs), + .pc_ressize = sizeof(struct nfsd3_createres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+(1+FH+pAT)+WC, + }, + [NFS3PROC_MKDIR] = { + .pc_func = (svc_procfunc) nfsd3_proc_mkdir, + .pc_decode = (kxdrproc_t) nfs3svc_decode_mkdirargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_createres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_mkdirargs), + .pc_ressize = sizeof(struct nfsd3_createres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+(1+FH+pAT)+WC, + }, + [NFS3PROC_SYMLINK] = { + .pc_func = (svc_procfunc) nfsd3_proc_symlink, + .pc_decode = (kxdrproc_t) nfs3svc_decode_symlinkargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_createres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_symlinkargs), + .pc_ressize = sizeof(struct nfsd3_createres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+(1+FH+pAT)+WC, + }, + [NFS3PROC_MKNOD] = { + .pc_func = (svc_procfunc) nfsd3_proc_mknod, + .pc_decode = (kxdrproc_t) nfs3svc_decode_mknodargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_createres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_mknodargs), + .pc_ressize = sizeof(struct nfsd3_createres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+(1+FH+pAT)+WC, + }, + [NFS3PROC_REMOVE] = { + .pc_func = (svc_procfunc) nfsd3_proc_remove, + .pc_decode = (kxdrproc_t) nfs3svc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_wccstatres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_diropargs), + .pc_ressize = sizeof(struct nfsd3_wccstatres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+WC, + }, + [NFS3PROC_RMDIR] = { + .pc_func = (svc_procfunc) nfsd3_proc_rmdir, + .pc_decode = (kxdrproc_t) nfs3svc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_wccstatres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_diropargs), + .pc_ressize = sizeof(struct nfsd3_wccstatres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+WC, + }, + [NFS3PROC_RENAME] = { + .pc_func = (svc_procfunc) nfsd3_proc_rename, + .pc_decode = (kxdrproc_t) nfs3svc_decode_renameargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_renameres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_renameargs), + .pc_ressize = sizeof(struct nfsd3_renameres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+WC+WC, + }, + [NFS3PROC_LINK] = { + .pc_func = (svc_procfunc) nfsd3_proc_link, + .pc_decode = (kxdrproc_t) nfs3svc_decode_linkargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_linkres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle2, + .pc_argsize = sizeof(struct nfsd3_linkargs), + .pc_ressize = sizeof(struct nfsd3_linkres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+pAT+WC, + }, + [NFS3PROC_READDIR] = { + .pc_func = (svc_procfunc) nfsd3_proc_readdir, + .pc_decode = (kxdrproc_t) nfs3svc_decode_readdirargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_readdirres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_readdirargs), + .pc_ressize = sizeof(struct nfsd3_readdirres), + .pc_cachetype = RC_NOCACHE, + }, + [NFS3PROC_READDIRPLUS] = { + .pc_func = (svc_procfunc) nfsd3_proc_readdirplus, + .pc_decode = (kxdrproc_t) nfs3svc_decode_readdirplusargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_readdirres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_readdirplusargs), + .pc_ressize = sizeof(struct nfsd3_readdirres), + .pc_cachetype = RC_NOCACHE, + }, + [NFS3PROC_FSSTAT] = { + .pc_func = (svc_procfunc) nfsd3_proc_fsstat, + .pc_decode = (kxdrproc_t) nfs3svc_decode_fhandleargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_fsstatres, + .pc_argsize = sizeof(struct nfsd3_fhandleargs), + .pc_ressize = sizeof(struct nfsd3_fsstatres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+2*6+1, + }, + [NFS3PROC_FSINFO] = { + .pc_func = (svc_procfunc) nfsd3_proc_fsinfo, + .pc_decode = (kxdrproc_t) nfs3svc_decode_fhandleargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_fsinfores, + .pc_argsize = sizeof(struct nfsd3_fhandleargs), + .pc_ressize = sizeof(struct nfsd3_fsinfores), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+12, + }, + [NFS3PROC_PATHCONF] = { + .pc_func = (svc_procfunc) nfsd3_proc_pathconf, + .pc_decode = (kxdrproc_t) nfs3svc_decode_fhandleargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_pathconfres, + .pc_argsize = sizeof(struct nfsd3_fhandleargs), + .pc_ressize = sizeof(struct nfsd3_pathconfres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+pAT+6, + }, + [NFS3PROC_COMMIT] = { + .pc_func = (svc_procfunc) nfsd3_proc_commit, + .pc_decode = (kxdrproc_t) nfs3svc_decode_commitargs, + .pc_encode = (kxdrproc_t) nfs3svc_encode_commitres, + .pc_release = (kxdrproc_t) nfs3svc_release_fhandle, + .pc_argsize = sizeof(struct nfsd3_commitargs), + .pc_ressize = sizeof(struct nfsd3_commitres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+WC+2, + }, }; struct svc_version nfsd_version3 = { diff --git a/fs/nfsd/nfsproc.c b/fs/nfsd/nfsproc.c index e298e26..0eb9c82 100644 --- a/fs/nfsd/nfsproc.c +++ b/fs/nfsd/nfsproc.c @@ -533,45 +533,179 @@ nfsd_proc_statfs(struct svc_rqst * rqstp, struct nfsd_fhandle *argp, * NFSv2 Server procedures. * Only the results of non-idempotent operations are cached. */ -#define nfsd_proc_none NULL -#define nfssvc_release_none NULL struct nfsd_void { int dummy; }; -#define PROC(name, argt, rest, relt, cache, respsize) \ - { (svc_procfunc) nfsd_proc_##name, \ - (kxdrproc_t) nfssvc_decode_##argt, \ - (kxdrproc_t) nfssvc_encode_##rest, \ - (kxdrproc_t) nfssvc_release_##relt, \ - sizeof(struct nfsd_##argt), \ - sizeof(struct nfsd_##rest), \ - 0, \ - cache, \ - respsize, \ - } - #define ST 1 /* status */ #define FH 8 /* filehandle */ #define AT 18 /* attributes */ static struct svc_procedure nfsd_procedures2[18] = { - PROC(null, void, void, none, RC_NOCACHE, ST), - PROC(getattr, fhandle, attrstat, fhandle, RC_NOCACHE, ST+AT), - PROC(setattr, sattrargs, attrstat, fhandle, RC_REPLBUFF, ST+AT), - PROC(none, void, void, none, RC_NOCACHE, ST), - PROC(lookup, diropargs, diropres, fhandle, RC_NOCACHE, ST+FH+AT), - PROC(readlink, readlinkargs, readlinkres, none, RC_NOCACHE, ST+1+NFS_MAXPATHLEN/4), - PROC(read, readargs, readres, fhandle, RC_NOCACHE, ST+AT+1+NFSSVC_MAXBLKSIZE_V2/4), - PROC(none, void, void, none, RC_NOCACHE, ST), - PROC(write, writeargs, attrstat, fhandle, RC_REPLBUFF, ST+AT), - PROC(create, createargs, diropres, fhandle, RC_REPLBUFF, ST+FH+AT), - PROC(remove, diropargs, void, none, RC_REPLSTAT, ST), - PROC(rename, renameargs, void, none, RC_REPLSTAT, ST), - PROC(link, linkargs, void, none, RC_REPLSTAT, ST), - PROC(symlink, symlinkargs, void, none, RC_REPLSTAT, ST), - PROC(mkdir, createargs, diropres, fhandle, RC_REPLBUFF, ST+FH+AT), - PROC(rmdir, diropargs, void, none, RC_REPLSTAT, ST), - PROC(readdir, readdirargs, readdirres, none, RC_NOCACHE, 0), - PROC(statfs, fhandle, statfsres, none, RC_NOCACHE, ST+5), + [NFSPROC_NULL] = { + .pc_func = (svc_procfunc) nfsd_proc_null, + .pc_decode = (kxdrproc_t) nfssvc_decode_void, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_void), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST, + }, + [NFSPROC_GETATTR] = { + .pc_func = (svc_procfunc) nfsd_proc_getattr, + .pc_decode = (kxdrproc_t) nfssvc_decode_fhandle, + .pc_encode = (kxdrproc_t) nfssvc_encode_attrstat, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_fhandle), + .pc_ressize = sizeof(struct nfsd_attrstat), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+AT, + }, + [NFSPROC_SETATTR] = { + .pc_func = (svc_procfunc) nfsd_proc_setattr, + .pc_decode = (kxdrproc_t) nfssvc_decode_sattrargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_attrstat, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_sattrargs), + .pc_ressize = sizeof(struct nfsd_attrstat), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+AT, + }, + [NFSPROC_ROOT] = { + .pc_decode = (kxdrproc_t) nfssvc_decode_void, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_void), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST, + }, + [NFSPROC_LOOKUP] = { + .pc_func = (svc_procfunc) nfsd_proc_lookup, + .pc_decode = (kxdrproc_t) nfssvc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_diropres, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_diropargs), + .pc_ressize = sizeof(struct nfsd_diropres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+FH+AT, + }, + [NFSPROC_READLINK] = { + .pc_func = (svc_procfunc) nfsd_proc_readlink, + .pc_decode = (kxdrproc_t) nfssvc_decode_readlinkargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_readlinkres, + .pc_argsize = sizeof(struct nfsd_readlinkargs), + .pc_ressize = sizeof(struct nfsd_readlinkres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+1+NFS_MAXPATHLEN/4, + }, + [NFSPROC_READ] = { + .pc_func = (svc_procfunc) nfsd_proc_read, + .pc_decode = (kxdrproc_t) nfssvc_decode_readargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_readres, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_readargs), + .pc_ressize = sizeof(struct nfsd_readres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+AT+1+NFSSVC_MAXBLKSIZE_V2/4, + }, + [NFSPROC_WRITECACHE] = { + .pc_decode = (kxdrproc_t) nfssvc_decode_void, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_void), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST, + }, + [NFSPROC_WRITE] = { + .pc_func = (svc_procfunc) nfsd_proc_write, + .pc_decode = (kxdrproc_t) nfssvc_decode_writeargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_attrstat, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_writeargs), + .pc_ressize = sizeof(struct nfsd_attrstat), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+AT, + }, + [NFSPROC_CREATE] = { + .pc_func = (svc_procfunc) nfsd_proc_create, + .pc_decode = (kxdrproc_t) nfssvc_decode_createargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_diropres, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_createargs), + .pc_ressize = sizeof(struct nfsd_diropres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+FH+AT, + }, + [NFSPROC_REMOVE] = { + .pc_func = (svc_procfunc) nfsd_proc_remove, + .pc_decode = (kxdrproc_t) nfssvc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_diropargs), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_REPLSTAT, + .pc_xdrressize = ST, + }, + [NFSPROC_RENAME] = { + .pc_func = (svc_procfunc) nfsd_proc_rename, + .pc_decode = (kxdrproc_t) nfssvc_decode_renameargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_renameargs), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_REPLSTAT, + .pc_xdrressize = ST, + }, + [NFSPROC_LINK] = { + .pc_func = (svc_procfunc) nfsd_proc_link, + .pc_decode = (kxdrproc_t) nfssvc_decode_linkargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_linkargs), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_REPLSTAT, + .pc_xdrressize = ST, + }, + [NFSPROC_SYMLINK] = { + .pc_func = (svc_procfunc) nfsd_proc_symlink, + .pc_decode = (kxdrproc_t) nfssvc_decode_symlinkargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_symlinkargs), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_REPLSTAT, + .pc_xdrressize = ST, + }, + [NFSPROC_MKDIR] = { + .pc_func = (svc_procfunc) nfsd_proc_mkdir, + .pc_decode = (kxdrproc_t) nfssvc_decode_createargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_diropres, + .pc_release = (kxdrproc_t) nfssvc_release_fhandle, + .pc_argsize = sizeof(struct nfsd_createargs), + .pc_ressize = sizeof(struct nfsd_diropres), + .pc_cachetype = RC_REPLBUFF, + .pc_xdrressize = ST+FH+AT, + }, + [NFSPROC_RMDIR] = { + .pc_func = (svc_procfunc) nfsd_proc_rmdir, + .pc_decode = (kxdrproc_t) nfssvc_decode_diropargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_void, + .pc_argsize = sizeof(struct nfsd_diropargs), + .pc_ressize = sizeof(struct nfsd_void), + .pc_cachetype = RC_REPLSTAT, + .pc_xdrressize = ST, + }, + [NFSPROC_READDIR] = { + .pc_func = (svc_procfunc) nfsd_proc_readdir, + .pc_decode = (kxdrproc_t) nfssvc_decode_readdirargs, + .pc_encode = (kxdrproc_t) nfssvc_encode_readdirres, + .pc_argsize = sizeof(struct nfsd_readdirargs), + .pc_ressize = sizeof(struct nfsd_readdirres), + .pc_cachetype = RC_NOCACHE, + }, + [NFSPROC_STATFS] = { + .pc_func = (svc_procfunc) nfsd_proc_statfs, + .pc_decode = (kxdrproc_t) nfssvc_decode_fhandle, + .pc_encode = (kxdrproc_t) nfssvc_encode_statfsres, + .pc_argsize = sizeof(struct nfsd_fhandle), + .pc_ressize = sizeof(struct nfsd_statfsres), + .pc_cachetype = RC_NOCACHE, + .pc_xdrressize = ST+5, + }, }; -- cgit v0.10.2 From 6c18ba9f5e506b8115b89b1aa7bdc25178f40b0a Mon Sep 17 00:00:00 2001 From: Alexandros Batsakis Date: Tue, 16 Jun 2009 04:19:13 +0300 Subject: nfsd41: move channel attributes from nfsd4_session to a nfsd4_channel_attr struct the change is valid for both the forechannel and the backchannel (currently dummy) Signed-off-by: Alexandros Batsakis Signed-off-by: Benny Halevy Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 89d9ac5..d5caf2a 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -444,8 +444,8 @@ static int set_forechannel_maxreqs(struct nfsd4_channel_attrs *fchan) * fchan holds the client values on input, and the server values on output */ static int init_forechannel_attrs(struct svc_rqst *rqstp, - struct nfsd4_session *session, - struct nfsd4_channel_attrs *fchan) + struct nfsd4_channel_attrs *session_fchan, + struct nfsd4_channel_attrs *fchan) { int status = 0; __u32 maxcount = svc_max_payload(rqstp); @@ -455,21 +455,21 @@ static int init_forechannel_attrs(struct svc_rqst *rqstp, /* Use the client's max request and max response size if possible */ if (fchan->maxreq_sz > maxcount) fchan->maxreq_sz = maxcount; - session->se_fmaxreq_sz = fchan->maxreq_sz; + session_fchan->maxreq_sz = fchan->maxreq_sz; if (fchan->maxresp_sz > maxcount) fchan->maxresp_sz = maxcount; - session->se_fmaxresp_sz = fchan->maxresp_sz; + session_fchan->maxresp_sz = fchan->maxresp_sz; /* Set the max response cached size our default which is * a multiple of PAGE_SIZE and small */ - session->se_fmaxresp_cached = NFSD_PAGES_PER_SLOT * PAGE_SIZE; - fchan->maxresp_cached = session->se_fmaxresp_cached; + session_fchan->maxresp_cached = NFSD_PAGES_PER_SLOT * PAGE_SIZE; + fchan->maxresp_cached = session_fchan->maxresp_cached; /* Use the client's maxops if possible */ if (fchan->maxops > NFSD_MAX_OPS_PER_COMPOUND) fchan->maxops = NFSD_MAX_OPS_PER_COMPOUND; - session->se_fmaxops = fchan->maxops; + session_fchan->maxops = fchan->maxops; /* try to use the client requested number of slots */ if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION) @@ -481,7 +481,7 @@ static int init_forechannel_attrs(struct svc_rqst *rqstp, */ status = set_forechannel_maxreqs(fchan); - session->se_fnumslots = fchan->maxreqs; + session_fchan->maxreqs = fchan->maxreqs; return status; } @@ -495,12 +495,14 @@ alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp, memset(&tmp, 0, sizeof(tmp)); /* FIXME: For now, we just accept the client back channel attributes. */ - status = init_forechannel_attrs(rqstp, &tmp, &cses->fore_channel); + tmp.se_bchannel = cses->back_channel; + status = init_forechannel_attrs(rqstp, &tmp.se_fchannel, + &cses->fore_channel); if (status) goto out; /* allocate struct nfsd4_session and slot table in one piece */ - slotsize = tmp.se_fnumslots * sizeof(struct nfsd4_slot); + slotsize = tmp.se_fchannel.maxreqs * sizeof(struct nfsd4_slot); new = kzalloc(sizeof(*new) + slotsize, GFP_KERNEL); if (!new) goto out; @@ -574,7 +576,7 @@ free_session(struct kref *kref) int i; ses = container_of(kref, struct nfsd4_session, se_ref); - for (i = 0; i < ses->se_fnumslots; i++) { + for (i = 0; i < ses->se_fchannel.maxreqs; i++) { struct nfsd4_cache_entry *e = &ses->se_slots[i].sl_cache_entry; nfsd4_release_respages(e->ce_respages, e->ce_resused); } @@ -1130,7 +1132,7 @@ nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp, * is sent (lease renewal). */ if (seq && nfsd4_not_cached(resp)) { - seq->maxslots = resp->cstate.session->se_fnumslots; + seq->maxslots = resp->cstate.session->se_fchannel.maxreqs; return nfs_ok; } @@ -1473,7 +1475,7 @@ nfsd4_sequence(struct svc_rqst *rqstp, goto out; status = nfserr_badslot; - if (seq->slotid >= session->se_fnumslots) + if (seq->slotid >= session->se_fchannel.maxreqs) goto out; slot = &session->se_slots[seq->slotid]; diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index d07f704..2dcc7fe 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -3183,7 +3183,7 @@ static int nfsd4_check_drc_limit(struct nfsd4_compoundres *resp) dprintk("%s length %u, xb->page_len %u tlen %u pad %u\n", __func__, length, xb->page_len, tlen, pad); - if (length <= session->se_fmaxresp_cached) + if (length <= session->se_fchannel.maxresp_cached) return status; else return nfserr_rep_too_big_to_cache; diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index c0c4921..105cc10 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -115,6 +115,17 @@ struct nfsd4_slot { struct nfsd4_cache_entry sl_cache_entry; }; +struct nfsd4_channel_attrs { + u32 headerpadsz; + u32 maxreq_sz; + u32 maxresp_sz; + u32 maxresp_cached; + u32 maxops; + u32 maxreqs; + u32 nr_rdma_attrs; + u32 rdma_attrs; +}; + struct nfsd4_session { struct kref se_ref; struct list_head se_hash; /* hash by sessionid */ @@ -122,11 +133,8 @@ struct nfsd4_session { u32 se_flags; struct nfs4_client *se_client; /* for expire_client */ struct nfs4_sessionid se_sessionid; - u32 se_fmaxreq_sz; - u32 se_fmaxresp_sz; - u32 se_fmaxresp_cached; - u32 se_fmaxops; - u32 se_fnumslots; + struct nfsd4_channel_attrs se_fchannel; + struct nfsd4_channel_attrs se_bchannel; struct nfsd4_slot se_slots[]; /* forward channel slots */ }; diff --git a/include/linux/nfsd/xdr4.h b/include/linux/nfsd/xdr4.h index d0f050f..2bacf75 100644 --- a/include/linux/nfsd/xdr4.h +++ b/include/linux/nfsd/xdr4.h @@ -366,17 +366,6 @@ struct nfsd4_exchange_id { int spa_how; }; -struct nfsd4_channel_attrs { - u32 headerpadsz; - u32 maxreq_sz; - u32 maxresp_sz; - u32 maxresp_cached; - u32 maxops; - u32 maxreqs; - u32 nr_rdma_attrs; - u32 rdma_attrs; -}; - struct nfsd4_create_session { clientid_t clientid; struct nfs4_sessionid sessionid; -- cgit v0.10.2 From 5d77ddfbcb062f2617ea79d7a371b4bc78f28417 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Tue, 16 Jun 2009 04:19:38 +0300 Subject: nfsd41: sanity check client drc maxreqs Ensure the client requested maximum requests are between 1 and NFSD_MAX_SLOTS_PER_SESSION Signed-off-by: Andy Adamson Signed-off-by: Benny Halevy Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index d5caf2a..99570c4 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -425,6 +425,11 @@ static int set_forechannel_maxreqs(struct nfsd4_channel_attrs *fchan) { int status = 0, np = fchan->maxreqs * NFSD_PAGES_PER_SLOT; + if (fchan->maxreqs < 1) + return nfserr_inval; + else if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION) + fchan->maxreqs = NFSD_MAX_SLOTS_PER_SESSION; + spin_lock(&nfsd_serv->sv_lock); if (np + nfsd_serv->sv_drc_pages_used > nfsd_serv->sv_drc_max_pages) np = nfsd_serv->sv_drc_max_pages - nfsd_serv->sv_drc_pages_used; -- cgit v0.10.2 From 82e12fe9244ff653f703722a8937b595e10e71f4 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Tue, 16 Jun 2009 11:03:07 +1000 Subject: nfsd: don't take nfsd_mutex twice when setting number of threads. Currently when we write a number to 'threads' in nfsdfs, we take the nfsd_mutex, update the number of threads, then take the mutex again to read the number of threads. Mostly this isn't a big deal. However if we are write '0', and portmap happens to be dead, then we can get unpredictable behaviour. If the nfsd threads all got killed quickly and the last thread is waiting for portmap to respond, then the second time we take the mutex we will block waiting for the last thread. However if the nfsd threads didn't die quite that fast, then there will be no contention when we try to take the mutex again. Unpredictability isn't fun, and waiting for the last thread to exit is pointless, so avoid taking the lock twice. To achieve this, get nfsd_svc return a non-negative number of active threads when not returning a negative error. Signed-off-by: NeilBrown diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 877e713..1250fb9 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -207,10 +207,14 @@ static struct file_operations pool_stats_operations = { static ssize_t write_svc(struct file *file, char *buf, size_t size) { struct nfsctl_svc *data; + int err; if (size < sizeof(*data)) return -EINVAL; data = (struct nfsctl_svc*) buf; - return nfsd_svc(data->svc_port, data->svc_nthreads); + err = nfsd_svc(data->svc_port, data->svc_nthreads); + if (err < 0) + return err; + return 0; } /** @@ -692,12 +696,12 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size) if (newthreads < 0) return -EINVAL; rv = nfsd_svc(NFS_PORT, newthreads); - if (rv) + if (rv < 0) return rv; - } + } else + rv = nfsd_nrthreads(); - return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%d\n", - nfsd_nrthreads()); + return scnprintf(buf, SIMPLE_TRANSACTION_LIMIT, "%d\n", rv); } /** diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c index cbba4a9..209eaa0 100644 --- a/fs/nfsd/nfssvc.c +++ b/fs/nfsd/nfssvc.c @@ -413,6 +413,12 @@ nfsd_svc(unsigned short port, int nrservs) goto failure; error = svc_set_num_threads(nfsd_serv, NULL, nrservs); + if (error == 0) + /* We are holding a reference to nfsd_serv which + * we don't want to count in the return value, + * so subtract 1 + */ + error = nfsd_serv->sv_nrthreads - 1; failure: svc_destroy(nfsd_serv); /* Release server */ out: -- cgit v0.10.2 From 671e1fcf63fd115eabcb693b06cbc2e4a3d1a3a3 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Tue, 16 Jun 2009 11:03:20 +1000 Subject: nfsd: optimise the starting of zero threads when none are running. Currently, if we ask to set then number of nfsd threads to zero when there are none running, we set up all the sockets and register the service, and then tear it all down again. This is pointless. So detect that case and exit promptly. (also remove an assignment to 'error' which was never used. Signed-off-by: NeilBrown Acked-by: Jeff Layton diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c index 209eaa0..d4c9884 100644 --- a/fs/nfsd/nfssvc.c +++ b/fs/nfsd/nfssvc.c @@ -390,12 +390,14 @@ nfsd_svc(unsigned short port, int nrservs) mutex_lock(&nfsd_mutex); dprintk("nfsd: creating service\n"); - error = -EINVAL; if (nrservs <= 0) nrservs = 0; if (nrservs > NFSD_MAXSERVS) nrservs = NFSD_MAXSERVS; - + error = 0; + if (nrservs == 0 && nfsd_serv == NULL) + goto out; + /* Readahead param cache - will no-op if it already exists */ error = nfsd_racache_init(2*nrservs); if (error<0) -- cgit v0.10.2 From 6ddbbbfe52f35301ef5a1b595f912d8d2b3ec143 Mon Sep 17 00:00:00 2001 From: Mike Sager Date: Tue, 16 Jun 2009 04:20:47 +0300 Subject: nfsd41: Remove ip address collision detection case Verified that cthon and pynfs exchange id tests pass (except for the two expected fails: EID8 and EID50) Signed-off-by: Mike Sager Signed-off-by: Benny Halevy Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 99570c4..ef6944b 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1247,12 +1247,6 @@ nfsd4_exchange_id(struct svc_rqst *rqstp, expire_client(conf); goto out_new; } - if (ip_addr != conf->cl_addr && - !(exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A)) { - /* Client collision. 18.35.4 case 3 */ - status = nfserr_clid_inuse; - goto out; - } /* * Set bit when the owner id and verifier map to an already * confirmed client id (18.35.3). @@ -1266,12 +1260,12 @@ nfsd4_exchange_id(struct svc_rqst *rqstp, copy_verf(conf, &verf); new = conf; goto out_copy; - } else { - /* 18.35.4 case 7 */ - if (exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A) { - status = nfserr_noent; - goto out; - } + } + + /* 18.35.4 case 7 */ + if (exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A) { + status = nfserr_noent; + goto out; } unconf = find_unconfirmed_client_by_str(dname, strhashval, true); -- cgit v0.10.2 From ef52bff8409bba78b042f1bcf33a0f49debc9774 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Tue, 16 Jun 2009 04:20:50 +0300 Subject: nfsd41: Backchannel: cleanup nfs4.0 callback encode routines Mimic the client and prepare to share the back channel xdr with NFSv4.1. Bump the number of operations in each encode routine, then backfill the number of operations. Signed-off-by: Andy Adamson Signed-off-by: Benny Halevy Signed-off-by: Ricardo Labiaga Signed-off-by: Benny Halevy Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index f4fab69..353eb4a 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -140,8 +140,9 @@ struct nfs4_cb_compound_hdr { int status; u32 ident; u32 nops; + __be32 *nops_p; u32 taglen; - char * tag; + char *tag; }; static struct { @@ -201,7 +202,7 @@ nfs_cb_stat_to_errno(int stat) * XDR encode */ -static int +static void encode_cb_compound_hdr(struct xdr_stream *xdr, struct nfs4_cb_compound_hdr *hdr) { __be32 * p; @@ -210,12 +211,18 @@ encode_cb_compound_hdr(struct xdr_stream *xdr, struct nfs4_cb_compound_hdr *hdr) WRITE32(0); /* tag length is always 0 */ WRITE32(NFS4_MINOR_VERSION); WRITE32(hdr->ident); + hdr->nops_p = p; WRITE32(hdr->nops); - return 0; } -static int -encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp) +static void encode_cb_nops(struct nfs4_cb_compound_hdr *hdr) +{ + *hdr->nops_p = htonl(hdr->nops); +} + +static void +encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp, + struct nfs4_cb_compound_hdr *hdr) { __be32 *p; int len = dp->dl_fh.fh_size; @@ -227,7 +234,7 @@ encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp) WRITE32(0); /* truncate optimization not implemented */ WRITE32(len); WRITEMEM(&dp->dl_fh.fh_base, len); - return 0; + hdr->nops++; } static int @@ -246,12 +253,13 @@ nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p, struct nfs4_delegation * struct xdr_stream xdr; struct nfs4_cb_compound_hdr hdr = { .ident = args->dl_ident, - .nops = 1, }; xdr_init_encode(&xdr, &req->rq_snd_buf, p); encode_cb_compound_hdr(&xdr, &hdr); - return (encode_cb_recall(&xdr, args)); + encode_cb_recall(&xdr, args, &hdr); + encode_cb_nops(&hdr); + return 0; } -- cgit v0.10.2 From ab52ae6db035fa425f90146327ab7d2c5d3e5654 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Tue, 16 Jun 2009 04:20:53 +0300 Subject: nfsd41: Backchannel: minorversion support for the back channel Prepare to share backchannel code with NFSv4.1. Signed-off-by: Andy Adamson Signed-off-by: Benny Halevy Signed-off-by: Ricardo Labiaga [nfsd41: use nfsd4_cb_sequence for callback minorversion] Signed-off-by: Benny Halevy Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 353eb4a..3fd23f7 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -141,6 +141,7 @@ struct nfs4_cb_compound_hdr { u32 ident; u32 nops; __be32 *nops_p; + u32 minorversion; u32 taglen; char *tag; }; @@ -209,7 +210,7 @@ encode_cb_compound_hdr(struct xdr_stream *xdr, struct nfs4_cb_compound_hdr *hdr) RESERVE_SPACE(16); WRITE32(0); /* tag length is always 0 */ - WRITE32(NFS4_MINOR_VERSION); + WRITE32(hdr->minorversion); WRITE32(hdr->ident); hdr->nops_p = p; WRITE32(hdr->nops); diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index ef6944b..980a216 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -984,6 +984,7 @@ gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se) if ( !(parse_ipv4(se->se_callback_addr_len, se->se_callback_addr_val, &cb->cb_addr, &cb->cb_port))) goto out_err; + cb->cb_minorversion = 0; cb->cb_prog = se->se_callback_prog; cb->cb_ident = se->se_callback_ident; return; diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h index 105cc10..f5a95fd 100644 --- a/include/linux/nfsd/state.h +++ b/include/linux/nfsd/state.h @@ -85,7 +85,8 @@ struct nfs4_cb_conn { u32 cb_addr; unsigned short cb_port; u32 cb_prog; - u32 cb_ident; + u32 cb_minorversion; + u32 cb_ident; /* minorversion 0 only */ /* RPC client info */ atomic_t cb_set; /* successful CB_NULL call */ struct rpc_clnt * cb_client; -- cgit v0.10.2 From 47fcb03fefee2501e79176932a4184fc24d6f8ec Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 18 May 2009 17:47:56 -0400 Subject: SUNRPC: Fix the TCP server's send buffer accounting Currently, the sunrpc server is refusing to allow us to process new RPC calls if the TCP send buffer is 2/3 full, even if we do actually have enough free space to guarantee that we can send another request. The following patch fixes svc_tcp_has_wspace() so that we only stop processing requests if we know that the socket buffer cannot possibly fit another reply. It also fixes the tcp write_space() callback so that we only clear the SOCK_NOSPACE flag when the TCP send buffer is less than 2/3 full. This should ensure that the send window will grow as per the standard TCP socket code. Signed-off-by: Trond Myklebust Signed-off-by: J. Bruce Fields diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 004a2f9..b09c80c 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -380,6 +380,7 @@ static void svc_sock_setbufsize(struct socket *sock, unsigned int snd, sock->sk->sk_sndbuf = snd * 2; sock->sk->sk_rcvbuf = rcv * 2; sock->sk->sk_userlocks |= SOCK_SNDBUF_LOCK|SOCK_RCVBUF_LOCK; + sock->sk->sk_write_space(sock->sk); release_sock(sock->sk); #endif } @@ -421,6 +422,15 @@ static void svc_write_space(struct sock *sk) } } +static void svc_tcp_write_space(struct sock *sk) +{ + struct socket *sock = sk->sk_socket; + + if (sk_stream_wspace(sk) >= sk_stream_min_wspace(sk) && sock) + clear_bit(SOCK_NOSPACE, &sock->flags); + svc_write_space(sk); +} + /* * Copy the UDP datagram's destination address to the rqstp structure. * The 'destination' address in this case is the address to which the @@ -1015,25 +1025,16 @@ static void svc_tcp_prep_reply_hdr(struct svc_rqst *rqstp) static int svc_tcp_has_wspace(struct svc_xprt *xprt) { struct svc_sock *svsk = container_of(xprt, struct svc_sock, sk_xprt); - struct svc_serv *serv = svsk->sk_xprt.xpt_server; + struct svc_serv *serv = svsk->sk_xprt.xpt_server; int required; - int wspace; - /* - * Set the SOCK_NOSPACE flag before checking the available - * sock space. - */ + if (test_bit(XPT_LISTENER, &xprt->xpt_flags)) + return 1; + required = atomic_read(&xprt->xpt_reserved) + serv->sv_max_mesg; + if (sk_stream_wspace(svsk->sk_sk) >= required) + return 1; set_bit(SOCK_NOSPACE, &svsk->sk_sock->flags); - required = atomic_read(&svsk->sk_xprt.xpt_reserved) + serv->sv_max_mesg; - wspace = sk_stream_wspace(svsk->sk_sk); - - if (wspace < sk_stream_min_wspace(svsk->sk_sk)) - return 0; - if (required * 2 > wspace) - return 0; - - clear_bit(SOCK_NOSPACE, &svsk->sk_sock->flags); - return 1; + return 0; } static struct svc_xprt *svc_tcp_create(struct svc_serv *serv, @@ -1089,7 +1090,7 @@ static void svc_tcp_init(struct svc_sock *svsk, struct svc_serv *serv) dprintk("setting up TCP socket for reading\n"); sk->sk_state_change = svc_tcp_state_change; sk->sk_data_ready = svc_tcp_data_ready; - sk->sk_write_space = svc_write_space; + sk->sk_write_space = svc_tcp_write_space; svsk->sk_reclen = 0; svsk->sk_tcplen = 0; -- cgit v0.10.2