From f3f4f4ed26b116f621596f74d42d2b736171e968 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:29 -0400 Subject: [PATCH 01/24] NFS: Fix up new minorversion= option The new minorversion= mount option (commit 3fd5be9e) was merged at the same time as the recent sloppy parser fixes (commit a5a16bae), so minorversion= still uses the old value parsing logic. If the minorversion= option specifies a bogus value, it should fail with "bad value" not "bad option." Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 0b4cbdc60ab..83a31070bc1 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -158,7 +158,7 @@ static const match_table_t nfs_mount_option_tokens = { { Opt_mountvers, "mountvers=%s" }, { Opt_nfsvers, "nfsvers=%s" }, { Opt_nfsvers, "vers=%s" }, - { Opt_minorversion, "minorversion=%u" }, + { Opt_minorversion, "minorversion=%s" }, { Opt_sec, "sec=%s" }, { Opt_proto, "proto=%s" }, @@ -1001,7 +1001,6 @@ static int nfs_parse_mount_options(char *raw, while ((p = strsep(&raw, ",")) != NULL) { substring_t args[MAX_OPT_ARGS]; unsigned long option; - int int_option; int token; if (!*p) @@ -1273,11 +1272,16 @@ static int nfs_parse_mount_options(char *raw, } break; case Opt_minorversion: - if (match_int(args, &int_option)) - return 0; - if (int_option < 0 || int_option > NFS4_MAX_MINOR_VERSION) - return 0; - mnt->minorversion = int_option; + string = match_strdup(args); + if (string == NULL) + goto out_nomem; + rc = strict_strtoul(string, 10, &option); + kfree(string); + if (rc != 0) + goto out_invalid_value; + if (option > NFS4_MAX_MINOR_VERSION) + goto out_invalid_value; + mnt->minorversion = option; break; /* From 0b524123c93893391ec9e6c9b04998a45235f9c8 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:30 -0400 Subject: [PATCH 02/24] NFS: Add ability to send MOUNTPROC_UMNT to the kernel's mountd client After certain failure modes of an NFS mount, an NFS client should send a MOUNTPROC_UMNT request to remove the just-added mount entry from the server's mount table. While no-one should rely on the accuracy of the server's mount table, sending a UMNT is simply being a good internet neighbor. Since NFS mount processing is handled in the kernel now, we will need a function in the kernel's mountd client that can post a MOUNTRPC_UMNT request, in order to handle these failure modes. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 1 + fs/nfs/mount_clnt.c | 79 +++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 80 insertions(+) diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 7dd90a6769d..8d2b71d57d2 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -102,6 +102,7 @@ struct nfs_mount_request { }; extern int nfs_mount(struct nfs_mount_request *info); +extern void nfs_umount(const struct nfs_mount_request *info); /* client.c */ extern struct rpc_program nfs_program; diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c index 38ef9eaec40..72dd8b6ccaf 100644 --- a/fs/nfs/mount_clnt.c +++ b/fs/nfs/mount_clnt.c @@ -209,6 +209,71 @@ out_mnt_err: goto out; } +/** + * nfs_umount - Notify a server that we have unmounted this export + * @info: pointer to umount request arguments + * + * MOUNTPROC_UMNT is advisory, so we set a short timeout, and always + * use UDP. + */ +void nfs_umount(const struct nfs_mount_request *info) +{ + static const struct rpc_timeout nfs_umnt_timeout = { + .to_initval = 1 * HZ, + .to_maxval = 3 * HZ, + .to_retries = 2, + }; + struct rpc_create_args args = { + .protocol = IPPROTO_UDP, + .address = info->sap, + .addrsize = info->salen, + .timeout = &nfs_umnt_timeout, + .servername = info->hostname, + .program = &mnt_program, + .version = info->version, + .authflavor = RPC_AUTH_UNIX, + .flags = RPC_CLNT_CREATE_NOPING, + }; + struct mountres result; + struct rpc_message msg = { + .rpc_argp = info->dirpath, + .rpc_resp = &result, + }; + struct rpc_clnt *clnt; + int status; + + if (info->noresvport) + args.flags |= RPC_CLNT_CREATE_NONPRIVPORT; + + clnt = rpc_create(&args); + if (unlikely(IS_ERR(clnt))) + goto out_clnt_err; + + dprintk("NFS: sending UMNT request for %s:%s\n", + (info->hostname ? info->hostname : "server"), info->dirpath); + + if (info->version == NFS_MNT3_VERSION) + msg.rpc_proc = &clnt->cl_procinfo[MOUNTPROC3_UMNT]; + else + msg.rpc_proc = &clnt->cl_procinfo[MOUNTPROC_UMNT]; + + status = rpc_call_sync(clnt, &msg, 0); + rpc_shutdown_client(clnt); + + if (unlikely(status < 0)) + goto out_call_err; + + return; + +out_clnt_err: + dprintk("NFS: failed to create UMNT RPC client, status=%ld\n", + PTR_ERR(clnt)); + return; + +out_call_err: + dprintk("NFS: UMNT request failed, status=%d\n", status); +} + /* * XDR encode/decode functions for MOUNT */ @@ -407,6 +472,13 @@ static struct rpc_procinfo mnt_procedures[] = { .p_statidx = MOUNTPROC_MNT, .p_name = "MOUNT", }, + [MOUNTPROC_UMNT] = { + .p_proc = MOUNTPROC_UMNT, + .p_encode = (kxdrproc_t)mnt_enc_dirpath, + .p_arglen = MNT_enc_dirpath_sz, + .p_statidx = MOUNTPROC_UMNT, + .p_name = "UMOUNT", + }, }; static struct rpc_procinfo mnt3_procedures[] = { @@ -419,6 +491,13 @@ static struct rpc_procinfo mnt3_procedures[] = { .p_statidx = MOUNTPROC3_MNT, .p_name = "MOUNT", }, + [MOUNTPROC3_UMNT] = { + .p_proc = MOUNTPROC3_UMNT, + .p_encode = (kxdrproc_t)mnt_enc_dirpath, + .p_arglen = MNT_enc_dirpath_sz, + .p_statidx = MOUNTPROC3_UMNT, + .p_name = "UMOUNT", + }, }; From 059f90b323c0f5d34656ab7e0548d7d033c2a51a Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:31 -0400 Subject: [PATCH 03/24] NFS: Fix auth flavor len accounting Previous logic in the NFS mount parsing code path assumed auth_flavor_len was set to zero for simple authentication flavors (like AUTH_UNIX), and 1 for compound flavors (like AUTH_GSS). At some earlier point (maybe even before the option parsers were merged?) specific checks for auth_flavor_len being zero were removed from the functions that validate the mount option that sets the mount point's authentication flavor. Since we are populating an array for authentication flavors, the auth_flavor_len should always be set to the number of flavors. Let's eliminate some cleverness here, and prepare for new logic that needs to know the number of flavors in the auth_flavors[] array. (auth_flavors[] is an array because at some point we want to allow a list of acceptable authentication flavors to be specified via the sec= mount option. For now it remains a single element array). Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 17 +++-------------- 1 file changed, 3 insertions(+), 14 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 83a31070bc1..a33e608713e 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -904,8 +904,6 @@ static void nfs_set_mount_transport_protocol(struct nfs_parsed_mount_data *mnt) /* * Parse the value of the 'sec=' option. - * - * The flavor_len setting is for v4 mounts. */ static int nfs_parse_security_flavors(char *value, struct nfs_parsed_mount_data *mnt) @@ -916,53 +914,43 @@ static int nfs_parse_security_flavors(char *value, switch (match_token(value, nfs_secflavor_tokens, args)) { case Opt_sec_none: - mnt->auth_flavor_len = 0; mnt->auth_flavors[0] = RPC_AUTH_NULL; break; case Opt_sec_sys: - mnt->auth_flavor_len = 0; mnt->auth_flavors[0] = RPC_AUTH_UNIX; break; case Opt_sec_krb5: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5; break; case Opt_sec_krb5i: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5I; break; case Opt_sec_krb5p: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5P; break; case Opt_sec_lkey: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEY; break; case Opt_sec_lkeyi: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEYI; break; case Opt_sec_lkeyp: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEYP; break; case Opt_sec_spkm: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKM; break; case Opt_sec_spkmi: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKMI; break; case Opt_sec_spkmp: - mnt->auth_flavor_len = 1; mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKMP; break; default: return 0; } + mnt->auth_flavor_len = 1; return 1; } @@ -1680,6 +1668,7 @@ static int nfs_validate_mount_data(void *options, args->nfs_server.port = 0; /* autobind unless user sets port */ args->nfs_server.protocol = XPRT_TRANSPORT_TCP; args->auth_flavors[0] = RPC_AUTH_UNIX; + args->auth_flavor_len = 1; switch (data->version) { case 1: @@ -2343,7 +2332,7 @@ static int nfs4_validate_mount_data(void *options, args->acdirmax = NFS_DEF_ACDIRMAX; args->nfs_server.port = NFS_PORT; /* 2049 unless user set port= */ args->auth_flavors[0] = RPC_AUTH_UNIX; - args->auth_flavor_len = 0; + args->auth_flavor_len = 1; args->minorversion = 0; switch (data->version) { From ec88f28d1eb77346f19ca324ceec76e645cdd9da Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:32 -0400 Subject: [PATCH 04/24] NFS: Use the authentication flavor list returned by mountd Commit a14017db added support in the kernel's NFS mount client to decode the authentication flavor list returned by mountd. The NFS client can now use this list to determine whether the authentication flavor requested by the user is actually supported by the server. Note we don't actually negotiate the security flavor if none was specified by the user. Instead, we try to use AUTH_SYS, and fail if the server does not support it. This prevents us from negotiating an inappropriate security flavor (some servers list AUTH_NULL first). If the server does not support AUTH_SYS, the user must provide an appropriate security flavor by specifying the "sec=" mount option. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 58 ++++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 51 insertions(+), 7 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index a33e608713e..8526008eba7 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1436,6 +1436,42 @@ out_security_failure: return 0; } +/* + * Match the requested auth flavors with the list returned by + * the server. Returns zero and sets the mount's authentication + * flavor on success; returns -EACCES if server does not support + * the requested flavor. + */ +static int nfs_walk_authlist(struct nfs_parsed_mount_data *args, + struct nfs_mount_request *request) +{ + unsigned int i, j, server_authlist_len = *(request->auth_flav_len); + + /* + * We avoid sophisticated negotiating here, as there are + * plenty of cases where we can get it wrong, providing + * either too little or too much security. + * + * RFC 2623, section 2.7 suggests we SHOULD prefer the + * flavor listed first. However, some servers list + * AUTH_NULL first. Our caller plants AUTH_SYS, the + * preferred default, in args->auth_flavors[0] if user + * didn't specify sec= mount option. + */ + for (i = 0; i < args->auth_flavor_len; i++) + for (j = 0; j < server_authlist_len; j++) + if (args->auth_flavors[i] == request->auth_flavs[j]) { + dfprintk(MOUNT, "NFS: using auth flavor %d\n", + request->auth_flavs[j]); + args->auth_flavors[0] = request->auth_flavs[j]; + return 0; + } + + dfprintk(MOUNT, "NFS: server does not support requested auth flavor\n"); + nfs_umount(request); + return -EACCES; +} + /* * Use the remote server's MOUNT service to request the NFS file handle * corresponding to the provided path. @@ -1443,7 +1479,8 @@ out_security_failure: static int nfs_try_mount(struct nfs_parsed_mount_data *args, struct nfs_fh *root_fh) { - unsigned int auth_flavor_len = 0; + rpc_authflavor_t server_authlist[NFS_MAX_SECFLAVORS]; + unsigned int server_authlist_len = ARRAY_SIZE(server_authlist); struct nfs_mount_request request = { .sap = (struct sockaddr *) &args->mount_server.address, @@ -1451,7 +1488,8 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args, .protocol = args->mount_server.protocol, .fh = root_fh, .noresvport = args->flags & NFS_MOUNT_NORESVPORT, - .auth_flav_len = &auth_flavor_len, + .auth_flav_len = &server_authlist_len, + .auth_flavs = server_authlist, }; int status; @@ -1488,12 +1526,18 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args, * to a file handle. */ status = nfs_mount(&request); - if (status == 0) - return 0; + if (status != 0) { + dfprintk(MOUNT, "NFS: unable to mount server %s, error %d\n", + request.hostname, status); + return status; + } - dfprintk(MOUNT, "NFS: unable to mount server %s, error %d\n", - request.hostname, status); - return status; + /* + * MNTv1 (NFSv2) does not support auth flavor negotiation. + */ + if (args->mount_server.version != NFS_MNT3_VERSION) + return 0; + return nfs_walk_authlist(args, &request); } static int nfs_parse_simple_hostname(const char *dev_name, From 169026a61e6f436dfc12c9d10d95455c4e9f945b Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:32 -0400 Subject: [PATCH 05/24] SUNRPC: Clean up RPCBIND_MAXUADDRLEN definitions Clean up: Replace the single-integer definition of RPCBIND_MAXUADDRLEN with a definition that is based on previously defined address string sizes, and document the way this maximum is calculated. Also provide a separate macro for the size of the port number extension. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/msg_prot.h | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/include/linux/sunrpc/msg_prot.h b/include/linux/sunrpc/msg_prot.h index 70df4f1d884..77e62488339 100644 --- a/include/linux/sunrpc/msg_prot.h +++ b/include/linux/sunrpc/msg_prot.h @@ -189,7 +189,22 @@ typedef __be32 rpc_fraghdr; * Additionally, the two alternative forms specified in Section 2.2 of * [RFC2373] are also acceptable. */ -#define RPCBIND_MAXUADDRLEN (56u) + +#include + +/* Maximum size of the port number part of a universal address */ +#define RPCBIND_MAXUADDRPLEN sizeof(".255.255") + +/* Maximum size of an IPv4 universal address */ +#define RPCBIND_MAXUADDR4LEN \ + (INET_ADDRSTRLEN + RPCBIND_MAXUADDRPLEN) + +/* Maximum size of an IPv6 universal address */ +#define RPCBIND_MAXUADDR6LEN \ + (INET6_ADDRSTRLEN + RPCBIND_MAXUADDRPLEN) + +/* Assume INET6_ADDRSTRLEN will always be larger than INET_ADDRSTRLEN... */ +#define RPCBIND_MAXUADDRLEN RPCBIND_MAXUADDR6LEN #endif /* __KERNEL__ */ #endif /* _LINUX_SUNRPC_MSGPROT_H_ */ From 0b10bf5e14d856d1d27a2117d07af2bebee81b75 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:33 -0400 Subject: [PATCH 06/24] SUNRPC: Move XDR data type size macros Clean up: To make subsequent patches cleaner, move the XDR data type size macros to the top of the file (similar to nfs4xdr.c) first. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 56 +++++++++++++++++++++++------------------- 1 file changed, 31 insertions(+), 25 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index beee6da3303..ad1d7315c49 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -75,6 +75,37 @@ enum { #define RPCB_OWNER_STRING "0" #define RPCB_MAXOWNERLEN sizeof(RPCB_OWNER_STRING) +/* + * XDR data type sizes + */ +#define RPCB_program_sz (1) +#define RPCB_version_sz (1) +#define RPCB_protocol_sz (1) +#define RPCB_port_sz (1) +#define RPCB_boolean_sz (1) + +#define RPCB_netid_sz (1 + XDR_QUADLEN(RPCBIND_MAXNETIDLEN)) +#define RPCB_addr_sz (1 + XDR_QUADLEN(RPCBIND_MAXUADDRLEN)) +#define RPCB_ownerstring_sz (1 + XDR_QUADLEN(RPCB_MAXOWNERLEN)) + +/* + * XDR argument and result sizes + */ +#define RPCB_mappingargs_sz (RPCB_program_sz + RPCB_version_sz + \ + RPCB_protocol_sz + RPCB_port_sz) +#define RPCB_getaddrargs_sz (RPCB_program_sz + RPCB_version_sz + \ + RPCB_netid_sz + RPCB_addr_sz + \ + RPCB_ownerstring_sz) + +#define RPCB_getportres_sz RPCB_port_sz +#define RPCB_setres_sz RPCB_boolean_sz + +/* + * Note that RFC 1833 does not put any size restrictions on the + * address string returned by the remote rpcbind database. + */ +#define RPCB_getaddrres_sz RPCB_addr_sz + static void rpcb_getport_done(struct rpc_task *, void *); static void rpcb_map_release(void *data); static struct rpc_program rpcb_program; @@ -768,31 +799,6 @@ out_err: return -EIO; } -#define RPCB_program_sz (1u) -#define RPCB_version_sz (1u) -#define RPCB_protocol_sz (1u) -#define RPCB_port_sz (1u) -#define RPCB_boolean_sz (1u) - -#define RPCB_netid_sz (1+XDR_QUADLEN(RPCBIND_MAXNETIDLEN)) -#define RPCB_addr_sz (1+XDR_QUADLEN(RPCBIND_MAXUADDRLEN)) -#define RPCB_ownerstring_sz (1+XDR_QUADLEN(RPCB_MAXOWNERLEN)) - -#define RPCB_mappingargs_sz RPCB_program_sz+RPCB_version_sz+ \ - RPCB_protocol_sz+RPCB_port_sz -#define RPCB_getaddrargs_sz RPCB_program_sz+RPCB_version_sz+ \ - RPCB_netid_sz+RPCB_addr_sz+ \ - RPCB_ownerstring_sz - -#define RPCB_setres_sz RPCB_boolean_sz -#define RPCB_getportres_sz RPCB_port_sz - -/* - * Note that RFC 1833 does not put any size restrictions on the - * address string returned by the remote rpcbind database. - */ -#define RPCB_getaddrres_sz RPCB_addr_sz - #define PROC(proc, argtype, restype) \ [RPCBPROC_##proc] = { \ .p_proc = RPCBPROC_##proc, \ From a02d692611348f11ee1bc37431a883c3ff2de23e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:34 -0400 Subject: [PATCH 07/24] SUNRPC: Provide functions for managing universal addresses Introduce a set of functions in the kernel's RPC implementation for converting between a socket address and either a standard presentation address string or an RPC universal address. The universal address functions will be used to encode and decode RPCB_FOO and NFSv4 SETCLIENTID arguments. The other functions are part of a previous promise to deliver shared functions that can be used by upper-layer protocols to display and manipulate IP addresses. The kernel's current address printf formatters were designed specifically for kernel to user-space APIs that require a particular string format for socket addresses, thus are somewhat limited for the purposes of sunrpc.ko. The formatter for IPv6 addresses, %pI6, does not support short-handing or scope IDs. Also, these printf formatters are unique per address family, so a separate formatter string is required for printing AF_INET and AF_INET6 addresses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 2 - include/linux/sunrpc/clnt.h | 38 ++++ net/sunrpc/Makefile | 2 +- net/sunrpc/addr.c | 364 ++++++++++++++++++++++++++++++++++++ 4 files changed, 403 insertions(+), 3 deletions(-) create mode 100644 net/sunrpc/addr.c diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 8d2b71d57d2..ff68397f9b1 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -370,8 +370,6 @@ unsigned int nfs_page_array_len(unsigned int base, size_t len) PAGE_SIZE - 1) >> PAGE_SHIFT; } -#define IPV6_SCOPE_DELIMITER '%' - /* * Set the port number in an address. Be agnostic about the address * family. diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 37881f1a0bd..2636e8ad551 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -9,6 +9,10 @@ #ifndef _LINUX_SUNRPC_CLNT_H #define _LINUX_SUNRPC_CLNT_H +#include +#include +#include + #include #include #include @@ -151,5 +155,39 @@ void rpc_force_rebind(struct rpc_clnt *); size_t rpc_peeraddr(struct rpc_clnt *, struct sockaddr *, size_t); const char *rpc_peeraddr2str(struct rpc_clnt *, enum rpc_display_format_t); +size_t rpc_ntop(const struct sockaddr *, char *, const size_t); +size_t rpc_pton(const char *, const size_t, + struct sockaddr *, const size_t); +char * rpc_sockaddr2uaddr(const struct sockaddr *); +size_t rpc_uaddr2sockaddr(const char *, const size_t, + struct sockaddr *, const size_t); + +static inline unsigned short rpc_get_port(const struct sockaddr *sap) +{ + switch (sap->sa_family) { + case AF_INET: + return ntohs(((struct sockaddr_in *)sap)->sin_port); + case AF_INET6: + return ntohs(((struct sockaddr_in6 *)sap)->sin6_port); + } + return 0; +} + +static inline void rpc_set_port(struct sockaddr *sap, + const unsigned short port) +{ + switch (sap->sa_family) { + case AF_INET: + ((struct sockaddr_in *)sap)->sin_port = htons(port); + break; + case AF_INET6: + ((struct sockaddr_in6 *)sap)->sin6_port = htons(port); + break; + } +} + +#define IPV6_SCOPE_DELIMITER '%' +#define IPV6_SCOPE_ID_LEN sizeof("%nnnnnnnnnn") + #endif /* __KERNEL__ */ #endif /* _LINUX_SUNRPC_CLNT_H */ diff --git a/net/sunrpc/Makefile b/net/sunrpc/Makefile index db73fd2a3f0..9d2fca5ad14 100644 --- a/net/sunrpc/Makefile +++ b/net/sunrpc/Makefile @@ -10,7 +10,7 @@ obj-$(CONFIG_SUNRPC_XPRT_RDMA) += xprtrdma/ sunrpc-y := clnt.o xprt.o socklib.o xprtsock.o sched.o \ auth.o auth_null.o auth_unix.o auth_generic.o \ svc.o svcsock.o svcauth.o svcauth_unix.o \ - rpcb_clnt.o timer.o xdr.o \ + addr.o rpcb_clnt.o timer.o xdr.o \ sunrpc_syms.o cache.o rpc_pipe.o \ svc_xprt.o sunrpc-$(CONFIG_NFS_V4_1) += backchannel_rqst.o bc_svc.o diff --git a/net/sunrpc/addr.c b/net/sunrpc/addr.c new file mode 100644 index 00000000000..22e8fd89477 --- /dev/null +++ b/net/sunrpc/addr.c @@ -0,0 +1,364 @@ +/* + * Copyright 2009, Oracle. All rights reserved. + * + * Convert socket addresses to presentation addresses and universal + * addresses, and vice versa. + * + * Universal addresses are introduced by RFC 1833 and further refined by + * recent RFCs describing NFSv4. The universal address format is part + * of the external (network) interface provided by rpcbind version 3 + * and 4, and by NFSv4. Such an address is a string containing a + * presentation format IP address followed by a port number in + * "hibyte.lobyte" format. + * + * IPv6 addresses can also include a scope ID, typically denoted by + * a '%' followed by a device name or a non-negative integer. Refer to + * RFC 4291, Section 2.2 for details on IPv6 presentation formats. + */ + +#include +#include + +#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) + +static size_t rpc_ntop6_noscopeid(const struct sockaddr *sap, + char *buf, const int buflen) +{ + const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; + const struct in6_addr *addr = &sin6->sin6_addr; + + /* + * RFC 4291, Section 2.2.2 + * + * Shorthanded ANY address + */ + if (ipv6_addr_any(addr)) + return snprintf(buf, buflen, "::"); + + /* + * RFC 4291, Section 2.2.2 + * + * Shorthanded loopback address + */ + if (ipv6_addr_loopback(addr)) + return snprintf(buf, buflen, "::1"); + + /* + * RFC 4291, Section 2.2.3 + * + * Special presentation address format for mapped v4 + * addresses. + */ + if (ipv6_addr_v4mapped(addr)) + return snprintf(buf, buflen, "::ffff:%pI4", + &addr->s6_addr32[3]); + + /* + * RFC 4291, Section 2.2.1 + * + * To keep the result as short as possible, especially + * since we don't shorthand, we don't want leading zeros + * in each halfword, so avoid %pI6. + */ + return snprintf(buf, buflen, "%x:%x:%x:%x:%x:%x:%x:%x", + ntohs(addr->s6_addr16[0]), ntohs(addr->s6_addr16[1]), + ntohs(addr->s6_addr16[2]), ntohs(addr->s6_addr16[3]), + ntohs(addr->s6_addr16[4]), ntohs(addr->s6_addr16[5]), + ntohs(addr->s6_addr16[6]), ntohs(addr->s6_addr16[7])); +} + +static size_t rpc_ntop6(const struct sockaddr *sap, + char *buf, const size_t buflen) +{ + const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; + char scopebuf[IPV6_SCOPE_ID_LEN]; + size_t len; + int rc; + + len = rpc_ntop6_noscopeid(sap, buf, buflen); + if (unlikely(len == 0)) + return len; + + if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL) && + !(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_SITELOCAL)) + return len; + + rc = snprintf(scopebuf, sizeof(scopebuf), "%c%u", + IPV6_SCOPE_DELIMITER, sin6->sin6_scope_id); + if (unlikely((size_t)rc > sizeof(scopebuf))) + return 0; + + len += rc; + if (unlikely(len > buflen)) + return 0; + + strcat(buf, scopebuf); + return len; +} + +#else /* !(defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)) */ + +static size_t rpc_ntop6_noscopeid(const struct sockaddr *sap, + char *buf, const int buflen) +{ + return 0; +} + +static size_t rpc_ntop6(const struct sockaddr *sap, + char *buf, const size_t buflen) +{ + return 0; +} + +#endif /* !(defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)) */ + +static int rpc_ntop4(const struct sockaddr *sap, + char *buf, const size_t buflen) +{ + const struct sockaddr_in *sin = (struct sockaddr_in *)sap; + + return snprintf(buf, buflen, "%pI4", &sin->sin_addr); +} + +/** + * rpc_ntop - construct a presentation address in @buf + * @sap: socket address + * @buf: construction area + * @buflen: size of @buf, in bytes + * + * Plants a %NUL-terminated string in @buf and returns the length + * of the string, excluding the %NUL. Otherwise zero is returned. + */ +size_t rpc_ntop(const struct sockaddr *sap, char *buf, const size_t buflen) +{ + switch (sap->sa_family) { + case AF_INET: + return rpc_ntop4(sap, buf, buflen); + case AF_INET6: + return rpc_ntop6(sap, buf, buflen); + } + + return 0; +} +EXPORT_SYMBOL_GPL(rpc_ntop); + +static size_t rpc_pton4(const char *buf, const size_t buflen, + struct sockaddr *sap, const size_t salen) +{ + struct sockaddr_in *sin = (struct sockaddr_in *)sap; + u8 *addr = (u8 *)&sin->sin_addr.s_addr; + + if (buflen > INET_ADDRSTRLEN || salen < sizeof(struct sockaddr_in)) + return 0; + + memset(sap, 0, sizeof(struct sockaddr_in)); + + if (in4_pton(buf, buflen, addr, '\0', NULL) == 0) + return 0; + + sin->sin_family = AF_INET; + return sizeof(struct sockaddr_in);; +} + +#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) +static int rpc_parse_scope_id(const char *buf, const size_t buflen, + const char *delim, struct sockaddr_in6 *sin6) +{ + char *p; + size_t len; + + if ((buf + buflen) == delim) + return 1; + + if (*delim != IPV6_SCOPE_DELIMITER) + return 0; + + if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL) && + !(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_SITELOCAL)) + return 0; + + len = (buf + buflen) - delim - 1; + p = kstrndup(delim + 1, len, GFP_KERNEL); + if (p) { + unsigned long scope_id = 0; + struct net_device *dev; + + dev = dev_get_by_name(&init_net, p); + if (dev != NULL) { + scope_id = dev->ifindex; + dev_put(dev); + } else { + if (strict_strtoul(p, 10, &scope_id) == 0) { + kfree(p); + return 0; + } + } + + kfree(p); + + sin6->sin6_scope_id = scope_id; + return 1; + } + + return 0; +} + +static size_t rpc_pton6(const char *buf, const size_t buflen, + struct sockaddr *sap, const size_t salen) +{ + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; + u8 *addr = (u8 *)&sin6->sin6_addr.in6_u; + const char *delim; + + if (buflen > (INET6_ADDRSTRLEN + IPV6_SCOPE_ID_LEN) || + salen < sizeof(struct sockaddr_in6)) + return 0; + + memset(sap, 0, sizeof(struct sockaddr_in6)); + + if (in6_pton(buf, buflen, addr, IPV6_SCOPE_DELIMITER, &delim) == 0) + return 0; + + if (!rpc_parse_scope_id(buf, buflen, delim, sin6)) + return 0; + + sin6->sin6_family = AF_INET6; + return sizeof(struct sockaddr_in6); +} +#else +static size_t rpc_pton6(const char *buf, const size_t buflen, + struct sockaddr *sap, const size_t salen) +{ + return 0; +} +#endif + +/** + * rpc_pton - Construct a sockaddr in @sap + * @buf: C string containing presentation format IP address + * @buflen: length of presentation address in bytes + * @sap: buffer into which to plant socket address + * @salen: size of buffer in bytes + * + * Returns the size of the socket address if successful; otherwise + * zero is returned. + * + * Plants a socket address in @sap and returns the size of the + * socket address, if successful. Returns zero if an error + * occurred. + */ +size_t rpc_pton(const char *buf, const size_t buflen, + struct sockaddr *sap, const size_t salen) +{ + unsigned int i; + + for (i = 0; i < buflen; i++) + if (buf[i] == ':') + return rpc_pton6(buf, buflen, sap, salen); + return rpc_pton4(buf, buflen, sap, salen); +} +EXPORT_SYMBOL_GPL(rpc_pton); + +/** + * rpc_sockaddr2uaddr - Construct a universal address string from @sap. + * @sap: socket address + * + * Returns a %NUL-terminated string in dynamically allocated memory; + * otherwise NULL is returned if an error occurred. Caller must + * free the returned string. + */ +char *rpc_sockaddr2uaddr(const struct sockaddr *sap) +{ + char portbuf[RPCBIND_MAXUADDRPLEN]; + char addrbuf[RPCBIND_MAXUADDRLEN]; + unsigned short port; + + switch (sap->sa_family) { + case AF_INET: + if (rpc_ntop4(sap, addrbuf, sizeof(addrbuf)) == 0) + return NULL; + port = ntohs(((struct sockaddr_in *)sap)->sin_port); + break; + case AF_INET6: + if (rpc_ntop6_noscopeid(sap, addrbuf, sizeof(addrbuf)) == 0) + return NULL; + port = ntohs(((struct sockaddr_in6 *)sap)->sin6_port); + break; + default: + return NULL; + } + + if (snprintf(portbuf, sizeof(portbuf), + ".%u.%u", port >> 8, port & 0xff) > (int)sizeof(portbuf)) + return NULL; + + if (strlcat(addrbuf, portbuf, sizeof(addrbuf)) > sizeof(addrbuf)) + return NULL; + + return kstrdup(addrbuf, GFP_KERNEL); +} +EXPORT_SYMBOL_GPL(rpc_sockaddr2uaddr); + +/** + * rpc_uaddr2sockaddr - convert a universal address to a socket address. + * @uaddr: C string containing universal address to convert + * @uaddr_len: length of universal address string + * @sap: buffer into which to plant socket address + * @salen: size of buffer + * + * Returns the size of the socket address if successful; otherwise + * zero is returned. + */ +size_t rpc_uaddr2sockaddr(const char *uaddr, const size_t uaddr_len, + struct sockaddr *sap, const size_t salen) +{ + char *c, buf[RPCBIND_MAXUADDRLEN]; + unsigned long portlo, porthi; + unsigned short port; + + if (uaddr_len > sizeof(buf)) + return 0; + + memcpy(buf, uaddr, uaddr_len); + + buf[uaddr_len] = '\n'; + buf[uaddr_len + 1] = '\0'; + + c = strrchr(buf, '.'); + if (unlikely(c == NULL)) + return 0; + if (unlikely(strict_strtoul(c + 1, 10, &portlo) != 0)) + return 0; + if (unlikely(portlo > 255)) + return 0; + + c[0] = '\n'; + c[1] = '\0'; + + c = strrchr(buf, '.'); + if (unlikely(c == NULL)) + return 0; + if (unlikely(strict_strtoul(c + 1, 10, &porthi) != 0)) + return 0; + if (unlikely(porthi > 255)) + return 0; + + port = (unsigned short)((porthi << 8) | portlo); + + c[0] = '\0'; + + if (rpc_pton(buf, strlen(buf), sap, salen) == 0) + return 0; + + switch (sap->sa_family) { + case AF_INET: + ((struct sockaddr_in *)sap)->sin_port = htons(port); + return sizeof(struct sockaddr_in); + case AF_INET6: + ((struct sockaddr_in6 *)sap)->sin6_port = htons(port); + return sizeof(struct sockaddr_in6); + } + + return 0; +} +EXPORT_SYMBOL_GPL(rpc_uaddr2sockaddr); From ba809130bc260fce04141aca01ef9e068d32af2a Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:35 -0400 Subject: [PATCH 08/24] SUNRPC: Remove duplicate universal address generation RPC universal address generation is currently done in several places: rpcb_clnt.c, nfs4proc.c xprtsock.c, and xprtrdma.c. Remove the redundant cases that convert a socket address to a universal address. The nfs4proc.c case takes a pre-formatted presentation address string, not a socket address, so we'll leave that one. Because the new uaddr constructor uses the recently introduced rpc_ntop(), it now supports proper "::" shorthanding for IPv6 addresses. This allows the kernel to register properly formed universal addresses with the local rpcbind service, in _all_ cases. The kernel can now also send properly formed universal addresses in RPCB_GETADDR requests, and support link-local properly when encoding and decoding IPv6 addresses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 1 - net/sunrpc/rpcb_clnt.c | 48 ++++++++++++++++++++------------- net/sunrpc/xprtrdma/transport.c | 8 ------ net/sunrpc/xprtsock.c | 18 ------------- 4 files changed, 29 insertions(+), 46 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 1175d58efc2..65fad9534d9 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -41,7 +41,6 @@ enum rpc_display_format_t { RPC_DISPLAY_ALL, RPC_DISPLAY_HEX_ADDR, RPC_DISPLAY_HEX_PORT, - RPC_DISPLAY_UNIVERSAL_ADDR, RPC_DISPLAY_NETID, RPC_DISPLAY_MAX, }; diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index ad1d7315c49..1fb1c070c19 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -153,6 +153,7 @@ static void rpcb_map_release(void *data) rpcb_wake_rpcbind_waiters(map->r_xprt, map->r_status); xprt_put(map->r_xprt); + kfree(map->r_addr); kfree(map); } @@ -299,12 +300,9 @@ static int rpcb_register_inet4(const struct sockaddr *sap, const struct sockaddr_in *sin = (const struct sockaddr_in *)sap; struct rpcbind_args *map = msg->rpc_argp; unsigned short port = ntohs(sin->sin_port); - char buf[32]; + int result; - /* Construct AF_INET universal address */ - snprintf(buf, sizeof(buf), "%pI4.%u.%u", - &sin->sin_addr.s_addr, port >> 8, port & 0xff); - map->r_addr = buf; + map->r_addr = rpc_sockaddr2uaddr(sap); dprintk("RPC: %sregistering [%u, %u, %s, '%s'] with " "local rpcbind\n", (port ? "" : "un"), @@ -315,7 +313,9 @@ static int rpcb_register_inet4(const struct sockaddr *sap, if (port) msg->rpc_proc = &rpcb_procedures4[RPCBPROC_SET]; - return rpcb_register_call(RPCBVERS_4, msg); + result = rpcb_register_call(RPCBVERS_4, msg); + kfree(map->r_addr); + return result; } /* @@ -327,16 +327,9 @@ static int rpcb_register_inet6(const struct sockaddr *sap, const struct sockaddr_in6 *sin6 = (const struct sockaddr_in6 *)sap; struct rpcbind_args *map = msg->rpc_argp; unsigned short port = ntohs(sin6->sin6_port); - char buf[64]; + int result; - /* Construct AF_INET6 universal address */ - if (ipv6_addr_any(&sin6->sin6_addr)) - snprintf(buf, sizeof(buf), "::.%u.%u", - port >> 8, port & 0xff); - else - snprintf(buf, sizeof(buf), "%pI6.%u.%u", - &sin6->sin6_addr, port >> 8, port & 0xff); - map->r_addr = buf; + map->r_addr = rpc_sockaddr2uaddr(sap); dprintk("RPC: %sregistering [%u, %u, %s, '%s'] with " "local rpcbind\n", (port ? "" : "un"), @@ -347,7 +340,9 @@ static int rpcb_register_inet6(const struct sockaddr *sap, if (port) msg->rpc_proc = &rpcb_procedures4[RPCBPROC_SET]; - return rpcb_register_call(RPCBVERS_4, msg); + result = rpcb_register_call(RPCBVERS_4, msg); + kfree(map->r_addr); + return result; } static int rpcb_unregister_all_protofamilies(struct rpc_message *msg) @@ -570,6 +565,7 @@ void rpcb_getport_async(struct rpc_task *task) goto bailout_nofree; } + /* Parent transport's destination address */ salen = rpc_peeraddr(clnt, sap, sizeof(addr)); /* Don't ever use rpcbind v2 for AF_INET6 requests */ @@ -620,11 +616,22 @@ void rpcb_getport_async(struct rpc_task *task) map->r_prot = xprt->prot; map->r_port = 0; map->r_xprt = xprt_get(xprt); - map->r_netid = rpc_peeraddr2str(clnt, RPC_DISPLAY_NETID); - map->r_addr = rpc_peeraddr2str(rpcb_clnt, RPC_DISPLAY_UNIVERSAL_ADDR); - map->r_owner = ""; map->r_status = -EIO; + switch (bind_version) { + case RPCBVERS_4: + case RPCBVERS_3: + map->r_netid = rpc_peeraddr2str(clnt, RPC_DISPLAY_NETID); + map->r_addr = rpc_sockaddr2uaddr(sap); + map->r_owner = ""; + break; + case RPCBVERS_2: + map->r_addr = NULL; + break; + default: + BUG(); + } + child = rpcb_call_async(rpcb_clnt, map, proc); rpc_release_client(rpcb_clnt); if (IS_ERR(child)) { @@ -722,6 +729,9 @@ static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p, static int rpcb_encode_getaddr(struct rpc_rqst *req, __be32 *p, struct rpcbind_args *rpcb) { + if (rpcb->r_addr == NULL) + return -EIO; + dprintk("RPC: encoding rpcb request (%u, %u, %s)\n", rpcb->r_prog, rpcb->r_vers, rpcb->r_addr); *p++ = htonl(rpcb->r_prog); diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 1dd6123070e..537c210a8b9 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -202,14 +202,6 @@ xprt_rdma_format_addresses(struct rpc_xprt *xprt) snprintf(buf, 8, "%4hx", ntohs(addr->sin_port)); xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; - buf = kzalloc(30, GFP_KERNEL); - if (buf) - snprintf(buf, 30, "%pI4.%u.%u", - &addr->sin_addr.s_addr, - ntohs(addr->sin_port) >> 8, - ntohs(addr->sin_port) & 0xff); - xprt->address_strings[RPC_DISPLAY_UNIVERSAL_ADDR] = buf; - /* netid */ xprt->address_strings[RPC_DISPLAY_NETID] = "rdma"; } diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 83c73c4d017..a42c2adda59 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -341,15 +341,6 @@ static void xs_format_ipv4_peer_addresses(struct rpc_xprt *xprt, } xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; - buf = kzalloc(30, GFP_KERNEL); - if (buf) { - snprintf(buf, 30, "%pI4.%u.%u", - &addr->sin_addr.s_addr, - ntohs(addr->sin_port) >> 8, - ntohs(addr->sin_port) & 0xff); - } - xprt->address_strings[RPC_DISPLAY_UNIVERSAL_ADDR] = buf; - xprt->address_strings[RPC_DISPLAY_NETID] = netid; } @@ -397,15 +388,6 @@ static void xs_format_ipv6_peer_addresses(struct rpc_xprt *xprt, } xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; - buf = kzalloc(50, GFP_KERNEL); - if (buf) { - snprintf(buf, 50, "%pI6.%u.%u", - &addr->sin6_addr, - ntohs(addr->sin6_port) >> 8, - ntohs(addr->sin6_port) & 0xff); - } - xprt->address_strings[RPC_DISPLAY_UNIVERSAL_ADDR] = buf; - xprt->address_strings[RPC_DISPLAY_NETID] = netid; } From c877b849d302d1275452af80b7221a2555dc02e1 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:36 -0400 Subject: [PATCH 09/24] SUNRPC: Use rpc_ntop() for constructing transport address strings Clean up: In addition to using the new generic rpc_ntop() and rpc_get_port() functions, have the RPC client compute the presentation address buffer sizes dynamically using kstrdup(). Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/xprtrdma/transport.c | 41 +++++------- net/sunrpc/xprtsock.c | 113 ++++++++++---------------------- 2 files changed, 52 insertions(+), 102 deletions(-) diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 537c210a8b9..5f9b8676b6b 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -168,39 +168,30 @@ static struct rpc_xprt_ops xprt_rdma_procs; /* forward reference */ static void xprt_rdma_format_addresses(struct rpc_xprt *xprt) { - struct sockaddr_in *addr = (struct sockaddr_in *) + struct sockaddr *sap = (struct sockaddr *) &rpcx_to_rdmad(xprt).addr; - char *buf; + struct sockaddr_in *sin = (struct sockaddr_in *)sap; + char buf[64]; - buf = kzalloc(20, GFP_KERNEL); - if (buf) - snprintf(buf, 20, "%pI4", &addr->sin_addr.s_addr); - xprt->address_strings[RPC_DISPLAY_ADDR] = buf; + (void)rpc_ntop(sap, buf, sizeof(buf)); + xprt->address_strings[RPC_DISPLAY_ADDR] = kstrdup(buf, GFP_KERNEL); - buf = kzalloc(8, GFP_KERNEL); - if (buf) - snprintf(buf, 8, "%u", ntohs(addr->sin_port)); - xprt->address_strings[RPC_DISPLAY_PORT] = buf; + (void)snprintf(buf, sizeof(buf), "%u", rpc_get_port(sap)); + xprt->address_strings[RPC_DISPLAY_PORT] = kstrdup(buf, GFP_KERNEL); xprt->address_strings[RPC_DISPLAY_PROTO] = "rdma"; - buf = kzalloc(48, GFP_KERNEL); - if (buf) - snprintf(buf, 48, "addr=%pI4 port=%u proto=%s", - &addr->sin_addr.s_addr, - ntohs(addr->sin_port), "rdma"); - xprt->address_strings[RPC_DISPLAY_ALL] = buf; + (void)snprintf(buf, sizeof(buf), "addr=%s port=%s proto=rdma", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT]); + xprt->address_strings[RPC_DISPLAY_ALL] = kstrdup(buf, GFP_KERNEL); - buf = kzalloc(10, GFP_KERNEL); - if (buf) - snprintf(buf, 10, "%02x%02x%02x%02x", - NIPQUAD(addr->sin_addr.s_addr)); - xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = buf; + (void)snprintf(buf, sizeof(buf), "%02x%02x%02x%02x", + NIPQUAD(sin->sin_addr.s_addr)); + xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); - buf = kzalloc(8, GFP_KERNEL); - if (buf) - snprintf(buf, 8, "%4hx", ntohs(addr->sin_port)); - xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; + (void)snprintf(buf, sizeof(buf), "%4hx", rpc_get_port(sap)); + xprt->address_strings[RPC_DISPLAY_HEX_PORT] = kstrdup(buf, GFP_KERNEL); /* netid */ xprt->address_strings[RPC_DISPLAY_NETID] = "rdma"; diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index a42c2adda59..302a4090492 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -296,99 +296,58 @@ static inline struct sockaddr_in6 *xs_addr_in6(struct rpc_xprt *xprt) return (struct sockaddr_in6 *) &xprt->addr; } +static void xs_format_common_peer_addresses(struct rpc_xprt *xprt) +{ + struct sockaddr *sap = xs_addr(xprt); + char buf[128]; + + (void)rpc_ntop(sap, buf, sizeof(buf)); + xprt->address_strings[RPC_DISPLAY_ADDR] = kstrdup(buf, GFP_KERNEL); + + (void)snprintf(buf, sizeof(buf), "%u", rpc_get_port(sap)); + xprt->address_strings[RPC_DISPLAY_PORT] = kstrdup(buf, GFP_KERNEL); + + (void)snprintf(buf, sizeof(buf), "addr=%s port=%s proto=%s", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT], + xprt->address_strings[RPC_DISPLAY_PROTO]); + xprt->address_strings[RPC_DISPLAY_ALL] = kstrdup(buf, GFP_KERNEL); + + (void)snprintf(buf, sizeof(buf), "%4hx", rpc_get_port(sap)); + xprt->address_strings[RPC_DISPLAY_HEX_PORT] = kstrdup(buf, GFP_KERNEL); +} + static void xs_format_ipv4_peer_addresses(struct rpc_xprt *xprt, const char *protocol, const char *netid) { - struct sockaddr_in *addr = xs_addr_in(xprt); - char *buf; - - buf = kzalloc(20, GFP_KERNEL); - if (buf) { - snprintf(buf, 20, "%pI4", &addr->sin_addr.s_addr); - } - xprt->address_strings[RPC_DISPLAY_ADDR] = buf; - - buf = kzalloc(8, GFP_KERNEL); - if (buf) { - snprintf(buf, 8, "%u", - ntohs(addr->sin_port)); - } - xprt->address_strings[RPC_DISPLAY_PORT] = buf; + struct sockaddr_in *sin = xs_addr_in(xprt); + char buf[16]; xprt->address_strings[RPC_DISPLAY_PROTO] = protocol; - - buf = kzalloc(48, GFP_KERNEL); - if (buf) { - snprintf(buf, 48, "addr=%pI4 port=%u proto=%s", - &addr->sin_addr.s_addr, - ntohs(addr->sin_port), - protocol); - } - xprt->address_strings[RPC_DISPLAY_ALL] = buf; - - buf = kzalloc(10, GFP_KERNEL); - if (buf) { - snprintf(buf, 10, "%02x%02x%02x%02x", - NIPQUAD(addr->sin_addr.s_addr)); - } - xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = buf; - - buf = kzalloc(8, GFP_KERNEL); - if (buf) { - snprintf(buf, 8, "%4hx", - ntohs(addr->sin_port)); - } - xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; - xprt->address_strings[RPC_DISPLAY_NETID] = netid; + + (void)snprintf(buf, sizeof(buf), "%02x%02x%02x%02x", + NIPQUAD(sin->sin_addr.s_addr)); + xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); + + xs_format_common_peer_addresses(xprt); } static void xs_format_ipv6_peer_addresses(struct rpc_xprt *xprt, const char *protocol, const char *netid) { - struct sockaddr_in6 *addr = xs_addr_in6(xprt); - char *buf; - - buf = kzalloc(40, GFP_KERNEL); - if (buf) { - snprintf(buf, 40, "%pI6",&addr->sin6_addr); - } - xprt->address_strings[RPC_DISPLAY_ADDR] = buf; - - buf = kzalloc(8, GFP_KERNEL); - if (buf) { - snprintf(buf, 8, "%u", - ntohs(addr->sin6_port)); - } - xprt->address_strings[RPC_DISPLAY_PORT] = buf; + struct sockaddr_in6 *sin6 = xs_addr_in6(xprt); + char buf[48]; xprt->address_strings[RPC_DISPLAY_PROTO] = protocol; - - buf = kzalloc(64, GFP_KERNEL); - if (buf) { - snprintf(buf, 64, "addr=%pI6 port=%u proto=%s", - &addr->sin6_addr, - ntohs(addr->sin6_port), - protocol); - } - xprt->address_strings[RPC_DISPLAY_ALL] = buf; - - buf = kzalloc(36, GFP_KERNEL); - if (buf) - snprintf(buf, 36, "%pi6", &addr->sin6_addr); - - xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = buf; - - buf = kzalloc(8, GFP_KERNEL); - if (buf) { - snprintf(buf, 8, "%4hx", - ntohs(addr->sin6_port)); - } - xprt->address_strings[RPC_DISPLAY_HEX_PORT] = buf; - xprt->address_strings[RPC_DISPLAY_NETID] = netid; + + (void)snprintf(buf, sizeof(buf), "%pi6", &sin6->sin6_addr); + xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); + + xs_format_common_peer_addresses(xprt); } static void xs_free_peer_addresses(struct rpc_xprt *xprt) From 53a0b9c4c99ab0085a06421f71592722e5b3fd5f Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:36 -0400 Subject: [PATCH 10/24] NFS: Replace nfs_parse_ip_address() with rpc_pton() Clean up: Use the common routine now provided in sunrpc.ko for parsing mount addresses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 1 - fs/nfs/nfs4namespace.c | 6 +- fs/nfs/super.c | 148 ++++++----------------------------------- 3 files changed, 22 insertions(+), 133 deletions(-) diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index ff68397f9b1..cf1da3e2200 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -214,7 +214,6 @@ void nfs_zap_acl_cache(struct inode *inode); extern int nfs_wait_bit_killable(void *word); /* super.c */ -void nfs_parse_ip_address(char *, size_t, struct sockaddr *, size_t *); extern struct file_system_type nfs_xdev_fs_type; #ifdef CONFIG_NFS_V4 extern struct file_system_type nfs4_xdev_fs_type; diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c index 2a2a0a7143a..43c86b7556e 100644 --- a/fs/nfs/nfs4namespace.c +++ b/fs/nfs/nfs4namespace.c @@ -121,9 +121,9 @@ static struct vfsmount *try_location(struct nfs_clone_mount *mountdata, if (memchr(buf->data, IPV6_SCOPE_DELIMITER, buf->len)) continue; - nfs_parse_ip_address(buf->data, buf->len, - mountdata->addr, &mountdata->addrlen); - if (mountdata->addr->sa_family == AF_UNSPEC) + mountdata->addrlen = rpc_pton(buf->data, buf->len, + mountdata->addr, mountdata->addrlen); + if (mountdata->addrlen == 0) continue; nfs_set_port(mountdata->addr, NFS_PORT); diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 8526008eba7..1eeba8e5380 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -742,129 +742,10 @@ static int nfs_verify_server_address(struct sockaddr *addr) } } + dfprintk(MOUNT, "NFS: Invalid IP address specified\n"); return 0; } -static void nfs_parse_ipv4_address(char *string, size_t str_len, - struct sockaddr *sap, size_t *addr_len) -{ - struct sockaddr_in *sin = (struct sockaddr_in *)sap; - u8 *addr = (u8 *)&sin->sin_addr.s_addr; - - if (str_len <= INET_ADDRSTRLEN) { - dfprintk(MOUNT, "NFS: parsing IPv4 address %*s\n", - (int)str_len, string); - - sin->sin_family = AF_INET; - *addr_len = sizeof(*sin); - if (in4_pton(string, str_len, addr, '\0', NULL)) - return; - } - - sap->sa_family = AF_UNSPEC; - *addr_len = 0; -} - -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) -static int nfs_parse_ipv6_scope_id(const char *string, const size_t str_len, - const char *delim, - struct sockaddr_in6 *sin6) -{ - char *p; - size_t len; - - if ((string + str_len) == delim) - return 1; - - if (*delim != IPV6_SCOPE_DELIMITER) - return 0; - - if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL)) - return 0; - - len = (string + str_len) - delim - 1; - p = kstrndup(delim + 1, len, GFP_KERNEL); - if (p) { - unsigned long scope_id = 0; - struct net_device *dev; - - dev = dev_get_by_name(&init_net, p); - if (dev != NULL) { - scope_id = dev->ifindex; - dev_put(dev); - } else { - if (strict_strtoul(p, 10, &scope_id) == 0) { - kfree(p); - return 0; - } - } - - kfree(p); - - sin6->sin6_scope_id = scope_id; - dfprintk(MOUNT, "NFS: IPv6 scope ID = %lu\n", scope_id); - return 1; - } - - return 0; -} - -static void nfs_parse_ipv6_address(char *string, size_t str_len, - struct sockaddr *sap, size_t *addr_len) -{ - struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; - u8 *addr = (u8 *)&sin6->sin6_addr.in6_u; - const char *delim; - - if (str_len <= INET6_ADDRSTRLEN) { - dfprintk(MOUNT, "NFS: parsing IPv6 address %*s\n", - (int)str_len, string); - - sin6->sin6_family = AF_INET6; - *addr_len = sizeof(*sin6); - if (in6_pton(string, str_len, addr, - IPV6_SCOPE_DELIMITER, &delim) != 0) { - if (nfs_parse_ipv6_scope_id(string, str_len, - delim, sin6) != 0) - return; - } - } - - sap->sa_family = AF_UNSPEC; - *addr_len = 0; -} -#else -static void nfs_parse_ipv6_address(char *string, size_t str_len, - struct sockaddr *sap, size_t *addr_len) -{ - sap->sa_family = AF_UNSPEC; - *addr_len = 0; -} -#endif - -/* - * Construct a sockaddr based on the contents of a string that contains - * an IP address in presentation format. - * - * If there is a problem constructing the new sockaddr, set the address - * family to AF_UNSPEC. - */ -void nfs_parse_ip_address(char *string, size_t str_len, - struct sockaddr *sap, size_t *addr_len) -{ - unsigned int i, colons; - - colons = 0; - for (i = 0; i < str_len; i++) - if (string[i] == ':') - colons++; - - if (colons >= 2) - nfs_parse_ipv6_address(string, str_len, sap, addr_len); - else - nfs_parse_ipv4_address(string, str_len, sap, addr_len); -} - /* * Sanity check the NFS transport protocol. * @@ -1344,11 +1225,14 @@ static int nfs_parse_mount_options(char *raw, string = match_strdup(args); if (string == NULL) goto out_nomem; - nfs_parse_ip_address(string, strlen(string), - (struct sockaddr *) - &mnt->nfs_server.address, - &mnt->nfs_server.addrlen); + mnt->nfs_server.addrlen = + rpc_pton(string, strlen(string), + (struct sockaddr *) + &mnt->nfs_server.address, + sizeof(mnt->nfs_server.address)); kfree(string); + if (mnt->nfs_server.addrlen == 0) + goto out_invalid_address; break; case Opt_clientaddr: string = match_strdup(args); @@ -1368,11 +1252,14 @@ static int nfs_parse_mount_options(char *raw, string = match_strdup(args); if (string == NULL) goto out_nomem; - nfs_parse_ip_address(string, strlen(string), - (struct sockaddr *) - &mnt->mount_server.address, - &mnt->mount_server.addrlen); + mnt->mount_server.addrlen = + rpc_pton(string, strlen(string), + (struct sockaddr *) + &mnt->mount_server.address, + sizeof(mnt->mount_server.address)); kfree(string); + if (mnt->mount_server.addrlen == 0) + goto out_invalid_address; break; case Opt_lookupcache: string = match_strdup(args); @@ -1424,8 +1311,11 @@ static int nfs_parse_mount_options(char *raw, return 1; +out_invalid_address: + printk(KERN_INFO "NFS: bad IP address specified: %s\n", p); + return 0; out_invalid_value: - printk(KERN_INFO "NFS: bad mount option value specified: %s \n", p); + printk(KERN_INFO "NFS: bad mount option value specified: %s\n", p); return 0; out_nomem: printk(KERN_INFO "NFS: not enough memory to parse option\n"); From ec6ee61250acfccbc5578dd4014735fb2cbe53b5 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:37 -0400 Subject: [PATCH 11/24] NFS: Replace nfs_set_port() with rpc_set_port() Clean up. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 19 ------------------- fs/nfs/nfs4namespace.c | 2 +- fs/nfs/super.c | 6 +++--- 3 files changed, 4 insertions(+), 23 deletions(-) diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index cf1da3e2200..c2f171a3d70 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -368,22 +368,3 @@ unsigned int nfs_page_array_len(unsigned int base, size_t len) return ((unsigned long)len + (unsigned long)base + PAGE_SIZE - 1) >> PAGE_SHIFT; } - -/* - * Set the port number in an address. Be agnostic about the address - * family. - */ -static inline void nfs_set_port(struct sockaddr *sap, unsigned short port) -{ - struct sockaddr_in *ap = (struct sockaddr_in *)sap; - struct sockaddr_in6 *ap6 = (struct sockaddr_in6 *)sap; - - switch (sap->sa_family) { - case AF_INET: - ap->sin_port = htons(port); - break; - case AF_INET6: - ap6->sin6_port = htons(port); - break; - } -} diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c index 43c86b7556e..ef22ee89aa7 100644 --- a/fs/nfs/nfs4namespace.c +++ b/fs/nfs/nfs4namespace.c @@ -125,7 +125,7 @@ static struct vfsmount *try_location(struct nfs_clone_mount *mountdata, mountdata->addr, mountdata->addrlen); if (mountdata->addrlen == 0) continue; - nfs_set_port(mountdata->addr, NFS_PORT); + rpc_set_port(mountdata->addr, NFS_PORT); memcpy(page2, buf->data, buf->len); page2[buf->len] = '\0'; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 1eeba8e5380..9c85cdb353a 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1409,7 +1409,7 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args, /* * autobind will be used if mount_server.port == 0 */ - nfs_set_port(request.sap, args->mount_server.port); + rpc_set_port(request.sap, args->mount_server.port); /* * Now ask the mount server to map our export path @@ -1703,7 +1703,7 @@ static int nfs_validate_mount_data(void *options, &args->nfs_server.address)) goto out_no_address; - nfs_set_port((struct sockaddr *)&args->nfs_server.address, + rpc_set_port((struct sockaddr *)&args->nfs_server.address, args->nfs_server.port); nfs_set_mount_transport_protocol(args); @@ -2336,7 +2336,7 @@ static int nfs4_validate_mount_data(void *options, &args->nfs_server.address)) return -EINVAL; - nfs_set_port((struct sockaddr *)&args->nfs_server.address, + rpc_set_port((struct sockaddr *)&args->nfs_server.address, args->nfs_server.port); nfs_validate_transport_protocol(args); From b97a56747ea3f6c1a27dd0719bf1424959f1ebae Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:38 -0400 Subject: [PATCH 12/24] lockd: Replace nlm_clear_port() Clean up: Use shared rpc_set_port() function instead of nlm_clear_port(). Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/host.c | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) diff --git a/fs/lockd/host.c b/fs/lockd/host.c index 99d737bd432..7cb076ac6b4 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -87,18 +87,6 @@ static unsigned int nlm_hash_address(const struct sockaddr *sap) return hash & (NLM_HOST_NRHASH - 1); } -static void nlm_clear_port(struct sockaddr *sap) -{ - switch (sap->sa_family) { - case AF_INET: - ((struct sockaddr_in *)sap)->sin_port = 0; - break; - case AF_INET6: - ((struct sockaddr_in6 *)sap)->sin6_port = 0; - break; - } -} - /* * Common host lookup routine for server & client */ @@ -177,7 +165,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni) host->h_addrbuf = nsm->sm_addrbuf; memcpy(nlm_addr(host), ni->sap, ni->salen); host->h_addrlen = ni->salen; - nlm_clear_port(nlm_addr(host)); + rpc_set_port(nlm_addr(host), 0); memcpy(nlm_srcaddr(host), ni->src_sap, ni->src_len); host->h_version = ni->version; host->h_proto = ni->protocol; From c15128c5e428af1e8ada1476484a74c970665edd Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:39 -0400 Subject: [PATCH 13/24] lockd: Replace nsm_display_address() with rpc_ntop() Clean up. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-) diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 7fce1b52584..30c933188dd 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -61,43 +61,6 @@ static inline struct sockaddr *nsm_addr(const struct nsm_handle *nsm) return (struct sockaddr *)&nsm->sm_addr; } -static void nsm_display_ipv4_address(const struct sockaddr *sap, char *buf, - const size_t len) -{ - const struct sockaddr_in *sin = (struct sockaddr_in *)sap; - snprintf(buf, len, "%pI4", &sin->sin_addr.s_addr); -} - -static void nsm_display_ipv6_address(const struct sockaddr *sap, char *buf, - const size_t len) -{ - const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; - - if (ipv6_addr_v4mapped(&sin6->sin6_addr)) - snprintf(buf, len, "%pI4", &sin6->sin6_addr.s6_addr32[3]); - else if (sin6->sin6_scope_id != 0) - snprintf(buf, len, "%pI6%%%u", &sin6->sin6_addr, - sin6->sin6_scope_id); - else - snprintf(buf, len, "%pI6", &sin6->sin6_addr); -} - -static void nsm_display_address(const struct sockaddr *sap, - char *buf, const size_t len) -{ - switch (sap->sa_family) { - case AF_INET: - nsm_display_ipv4_address(sap, buf, len); - break; - case AF_INET6: - nsm_display_ipv6_address(sap, buf, len); - break; - default: - snprintf(buf, len, "unsupported address family"); - break; - } -} - static struct rpc_clnt *nsm_create(void) { struct sockaddr_in sin = { @@ -307,8 +270,11 @@ static struct nsm_handle *nsm_create_handle(const struct sockaddr *sap, memcpy(nsm_addr(new), sap, salen); new->sm_addrlen = salen; nsm_init_private(new); - nsm_display_address((const struct sockaddr *)&new->sm_addr, - new->sm_addrbuf, sizeof(new->sm_addrbuf)); + + if (rpc_ntop(nsm_addr(new), new->sm_addrbuf, + sizeof(new->sm_addrbuf)) == 0) + (void)snprintf(new->sm_addrbuf, sizeof(new->sm_addrbuf), + "unsupported address family"); memcpy(new->sm_name, hostname, hostname_len); new->sm_name[hostname_len] = '\0'; From 4116092b92f859e5e9a90c99d740933e651ee8c0 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:40 -0400 Subject: [PATCH 14/24] NFSD: Support IPv6 addresses in write_failover_ip() In write_failover_ip(), replace the sscanf() with a call to the common sunrpc.ko presentation address parser. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfsd/nfsctl.c | 21 +++++++-------------- 1 file changed, 7 insertions(+), 14 deletions(-) diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 6d0847562d8..7e906c5b767 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -490,22 +491,18 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size) * * Input: * buf: '\n'-terminated C string containing a - * presentation format IPv4 address + * presentation format IP address * size: length of C string in @buf * Output: * On success: returns zero if all specified locks were released; * returns one if one or more locks were not released * On error: return code is negative errno value - * - * Note: Only AF_INET client addresses are passed in */ static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size) { - struct sockaddr_in sin = { - .sin_family = AF_INET, - }; - int b1, b2, b3, b4; - char c; + struct sockaddr_storage address; + struct sockaddr *sap = (struct sockaddr *)&address; + size_t salen = sizeof(address); char *fo_path; /* sanity check */ @@ -519,14 +516,10 @@ static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size) if (qword_get(&buf, fo_path, size) < 0) return -EINVAL; - /* get ipv4 address */ - if (sscanf(fo_path, "%u.%u.%u.%u%c", &b1, &b2, &b3, &b4, &c) != 4) + if (rpc_pton(fo_path, size, sap, salen) == 0) return -EINVAL; - if (b1 > 255 || b2 > 255 || b3 > 255 || b4 > 255) - return -EINVAL; - sin.sin_addr.s_addr = htonl((b1 << 24) | (b2 << 16) | (b3 << 8) | b4); - return nlmsvc_unlock_all_by_ip((struct sockaddr *)&sin); + return nlmsvc_unlock_all_by_ip(sap); } /** From 6f2c2db7a46243bd86e3d7ff5f9ff982f26a9fe8 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:40 -0400 Subject: [PATCH 15/24] SUNRPC: Introduce new xdr_stream-based encoders to rpcb_clnt.c Replace the open-coded encode logic for rpcbind arguments with an xdr_stream-based implementation, similar to what NFSv4 uses, to better protect against buffer overflows. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 76 +++++++++++++++++++++++++++++++++++++++++- 1 file changed, 75 insertions(+), 1 deletion(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 1fb1c070c19..823d20dbad0 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -708,6 +708,30 @@ static int rpcb_encode_mapping(struct rpc_rqst *req, __be32 *p, return 0; } +static int rpcb_enc_mapping(struct rpc_rqst *req, __be32 *p, + const struct rpcbind_args *rpcb) +{ + struct rpc_task *task = req->rq_task; + struct xdr_stream xdr; + + dprintk("RPC: %5u encoding PMAP_%s call (%u, %u, %d, %u)\n", + task->tk_pid, task->tk_msg.rpc_proc->p_name, + rpcb->r_prog, rpcb->r_vers, rpcb->r_prot, rpcb->r_port); + + xdr_init_encode(&xdr, &req->rq_snd_buf, p); + + p = xdr_reserve_space(&xdr, sizeof(__be32) * RPCB_mappingargs_sz); + if (unlikely(p == NULL)) + return -EIO; + + *p++ = htonl(rpcb->r_prog); + *p++ = htonl(rpcb->r_vers); + *p++ = htonl(rpcb->r_prot); + *p = htonl(rpcb->r_port); + + return 0; +} + static int rpcb_decode_getport(struct rpc_rqst *req, __be32 *p, unsigned short *portp) { @@ -746,6 +770,56 @@ static int rpcb_encode_getaddr(struct rpc_rqst *req, __be32 *p, return 0; } +static int encode_rpcb_string(struct xdr_stream *xdr, const char *string, + const u32 maxstrlen) +{ + u32 len; + __be32 *p; + + if (unlikely(string == NULL)) + return -EIO; + len = strlen(string); + if (unlikely(len > maxstrlen)) + return -EIO; + + p = xdr_reserve_space(xdr, sizeof(__be32) + len); + if (unlikely(p == NULL)) + return -EIO; + xdr_encode_opaque(p, string, len); + + return 0; +} + +static int rpcb_enc_getaddr(struct rpc_rqst *req, __be32 *p, + const struct rpcbind_args *rpcb) +{ + struct rpc_task *task = req->rq_task; + struct xdr_stream xdr; + + dprintk("RPC: %5u encoding RPCB_%s call (%u, %u, '%s', '%s')\n", + task->tk_pid, task->tk_msg.rpc_proc->p_name, + rpcb->r_prog, rpcb->r_vers, + rpcb->r_netid, rpcb->r_addr); + + xdr_init_encode(&xdr, &req->rq_snd_buf, p); + + p = xdr_reserve_space(&xdr, + sizeof(__be32) * (RPCB_program_sz + RPCB_version_sz)); + if (unlikely(p == NULL)) + return -EIO; + *p++ = htonl(rpcb->r_prog); + *p = htonl(rpcb->r_vers); + + if (encode_rpcb_string(&xdr, rpcb->r_netid, RPCBIND_MAXNETIDLEN)) + return -EIO; + if (encode_rpcb_string(&xdr, rpcb->r_addr, RPCBIND_MAXUADDRLEN)) + return -EIO; + if (encode_rpcb_string(&xdr, rpcb->r_owner, RPCB_MAXOWNERLEN)) + return -EIO; + + return 0; +} + static int rpcb_decode_getaddr(struct rpc_rqst *req, __be32 *p, unsigned short *portp) { @@ -812,7 +886,7 @@ out_err: #define PROC(proc, argtype, restype) \ [RPCBPROC_##proc] = { \ .p_proc = RPCBPROC_##proc, \ - .p_encode = (kxdrproc_t) rpcb_encode_##argtype, \ + .p_encode = (kxdrproc_t) rpcb_enc_##argtype, \ .p_decode = (kxdrproc_t) rpcb_decode_##restype, \ .p_arglen = RPCB_##argtype##args_sz, \ .p_replen = RPCB_##restype##res_sz, \ From 0d36c4f7574d5a33bedd8f0e3c793490d45d83c6 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:41 -0400 Subject: [PATCH 16/24] SUNRPC: Clean up: Remove unused XDR encoder functions from rpcb_clnt.c Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 34 ---------------------------------- 1 file changed, 34 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 823d20dbad0..e0d7b31e54a 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -694,20 +694,6 @@ static void rpcb_getport_done(struct rpc_task *child, void *data) * XDR functions for rpcbind */ -static int rpcb_encode_mapping(struct rpc_rqst *req, __be32 *p, - struct rpcbind_args *rpcb) -{ - dprintk("RPC: encoding rpcb request (%u, %u, %d, %u)\n", - rpcb->r_prog, rpcb->r_vers, rpcb->r_prot, rpcb->r_port); - *p++ = htonl(rpcb->r_prog); - *p++ = htonl(rpcb->r_vers); - *p++ = htonl(rpcb->r_prot); - *p++ = htonl(rpcb->r_port); - - req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); - return 0; -} - static int rpcb_enc_mapping(struct rpc_rqst *req, __be32 *p, const struct rpcbind_args *rpcb) { @@ -750,26 +736,6 @@ static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p, return 0; } -static int rpcb_encode_getaddr(struct rpc_rqst *req, __be32 *p, - struct rpcbind_args *rpcb) -{ - if (rpcb->r_addr == NULL) - return -EIO; - - dprintk("RPC: encoding rpcb request (%u, %u, %s)\n", - rpcb->r_prog, rpcb->r_vers, rpcb->r_addr); - *p++ = htonl(rpcb->r_prog); - *p++ = htonl(rpcb->r_vers); - - p = xdr_encode_string(p, rpcb->r_netid); - p = xdr_encode_string(p, rpcb->r_addr); - p = xdr_encode_string(p, rpcb->r_owner); - - req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); - - return 0; -} - static int encode_rpcb_string(struct xdr_stream *xdr, const char *string, const u32 maxstrlen) { From 7ed0ff983c8ad30bf4e2b9fdbb299a3e3ec08d08 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:42 -0400 Subject: [PATCH 17/24] SUNRPC: Introduce xdr_stream-based decoders for RPCB_UNSET Replace the open-coded decode logic for rpcbind UNSET results with an xdr_stream-based implementation, similar to what NFSv4 uses, to protect against buffer overflows. The new function is unused for the moment. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index e0d7b31e54a..fe183af5cc7 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -736,6 +736,28 @@ static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p, return 0; } +static int rpcb_dec_set(struct rpc_rqst *req, __be32 *p, + unsigned int *boolp) +{ + struct rpc_task *task = req->rq_task; + struct xdr_stream xdr; + + xdr_init_decode(&xdr, &req->rq_rcv_buf, p); + + p = xdr_inline_decode(&xdr, sizeof(__be32)); + if (unlikely(p == NULL)) + return -EIO; + + *boolp = 0; + if (*p) + *boolp = 1; + + dprintk("RPC: %5u RPCB_%s call %s\n", + task->tk_pid, task->tk_msg.rpc_proc->p_name, + (*boolp ? "succeeded" : "failed")); + return 0; +} + static int encode_rpcb_string(struct xdr_stream *xdr, const char *string, const u32 maxstrlen) { From c0c077df009f2f329875051ac5283df235288689 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:43 -0400 Subject: [PATCH 18/24] SUNRPC: Introduce new xdr_stream-based decoders to rpcb_clnt.c Replace the open-coded decode logic for PMAP_GETPORT/RPCB_GETADDR with an xdr_stream-based implementation, similar to what NFSv4 uses, to protect against buffer overflows. The new implementation also checks that the incoming port number is reasonable. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 80 ++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 77 insertions(+), 3 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index fe183af5cc7..88de754e3b0 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -454,7 +454,7 @@ int rpcb_getport_sync(struct sockaddr_in *sin, u32 prog, u32 vers, int prot) struct rpc_message msg = { .rpc_proc = &rpcb_procedures2[RPCBPROC_GETPORT], .rpc_argp = &map, - .rpc_resp = &map.r_port, + .rpc_resp = &map, }; struct rpc_clnt *rpcb_clnt; int status; @@ -484,7 +484,7 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi struct rpc_message msg = { .rpc_proc = proc, .rpc_argp = map, - .rpc_resp = &map->r_port, + .rpc_resp = map, }; struct rpc_task_setup task_setup_data = { .rpc_client = rpcb_clnt, @@ -727,6 +727,31 @@ static int rpcb_decode_getport(struct rpc_rqst *req, __be32 *p, return 0; } +static int rpcb_dec_getport(struct rpc_rqst *req, __be32 *p, + struct rpcbind_args *rpcb) +{ + struct rpc_task *task = req->rq_task; + struct xdr_stream xdr; + unsigned long port; + + xdr_init_decode(&xdr, &req->rq_rcv_buf, p); + + rpcb->r_port = 0; + + p = xdr_inline_decode(&xdr, sizeof(__be32)); + if (unlikely(p == NULL)) + return -EIO; + + port = ntohl(*p); + dprintk("RPC: %5u PMAP_%s result: %lu\n", task->tk_pid, + task->tk_msg.rpc_proc->p_name, port); + if (unlikely(port > USHORT_MAX)) + return -EIO; + + rpcb->r_port = port; + return 0; +} + static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p, unsigned int *boolp) { @@ -871,11 +896,60 @@ out_err: return -EIO; } +static int rpcb_dec_getaddr(struct rpc_rqst *req, __be32 *p, + struct rpcbind_args *rpcb) +{ + struct sockaddr_storage address; + struct sockaddr *sap = (struct sockaddr *)&address; + struct rpc_task *task = req->rq_task; + struct xdr_stream xdr; + u32 len; + + rpcb->r_port = 0; + + xdr_init_decode(&xdr, &req->rq_rcv_buf, p); + + p = xdr_inline_decode(&xdr, sizeof(__be32)); + if (unlikely(p == NULL)) + goto out_fail; + len = ntohl(*p); + + /* + * If the returned universal address is a null string, + * the requested RPC service was not registered. + */ + if (len == 0) { + dprintk("RPC: %5u RPCB reply: program not registered\n", + task->tk_pid); + return 0; + } + + if (unlikely(len > RPCBIND_MAXUADDRLEN)) + goto out_fail; + + p = xdr_inline_decode(&xdr, len); + if (unlikely(p == NULL)) + goto out_fail; + dprintk("RPC: %5u RPCB_%s reply: %s\n", task->tk_pid, + task->tk_msg.rpc_proc->p_name, (char *)p); + + if (rpc_uaddr2sockaddr((char *)p, len, sap, sizeof(address)) == 0) + goto out_fail; + rpcb->r_port = rpc_get_port(sap); + + return 0; + +out_fail: + dprintk("RPC: %5u malformed RPCB_%s reply\n", + task->tk_pid, task->tk_msg.rpc_proc->p_name); + return -EIO; +} + #define PROC(proc, argtype, restype) \ [RPCBPROC_##proc] = { \ .p_proc = RPCBPROC_##proc, \ .p_encode = (kxdrproc_t) rpcb_enc_##argtype, \ - .p_decode = (kxdrproc_t) rpcb_decode_##restype, \ + .p_decode = (kxdrproc_t) rpcb_dec_##restype, \ .p_arglen = RPCB_##argtype##args_sz, \ .p_replen = RPCB_##restype##res_sz, \ .p_statidx = RPCBPROC_##proc, \ From 0e47f0d665efcf1b45f293422d21d8383f07d122 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:43 -0400 Subject: [PATCH 19/24] SUNRPC: Clean up: Remove unused XDR decoder functions from rpcb_clnt.c Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 81 ------------------------------------------ 1 file changed, 81 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 88de754e3b0..c4b716cacf2 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -718,15 +718,6 @@ static int rpcb_enc_mapping(struct rpc_rqst *req, __be32 *p, return 0; } -static int rpcb_decode_getport(struct rpc_rqst *req, __be32 *p, - unsigned short *portp) -{ - *portp = (unsigned short) ntohl(*p++); - dprintk("RPC: rpcb getport result: %u\n", - *portp); - return 0; -} - static int rpcb_dec_getport(struct rpc_rqst *req, __be32 *p, struct rpcbind_args *rpcb) { @@ -752,15 +743,6 @@ static int rpcb_dec_getport(struct rpc_rqst *req, __be32 *p, return 0; } -static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p, - unsigned int *boolp) -{ - *boolp = (unsigned int) ntohl(*p++); - dprintk("RPC: rpcb set/unset call %s\n", - (*boolp ? "succeeded" : "failed")); - return 0; -} - static int rpcb_dec_set(struct rpc_rqst *req, __be32 *p, unsigned int *boolp) { @@ -833,69 +815,6 @@ static int rpcb_enc_getaddr(struct rpc_rqst *req, __be32 *p, return 0; } -static int rpcb_decode_getaddr(struct rpc_rqst *req, __be32 *p, - unsigned short *portp) -{ - char *addr; - u32 addr_len; - int c, i, f, first, val; - - *portp = 0; - addr_len = ntohl(*p++); - - if (addr_len == 0) { - dprintk("RPC: rpcb_decode_getaddr: " - "service is not registered\n"); - return 0; - } - - /* - * Simple sanity check. - */ - if (addr_len > RPCBIND_MAXUADDRLEN) - goto out_err; - - /* - * Start at the end and walk backwards until the first dot - * is encountered. When the second dot is found, we have - * both parts of the port number. - */ - addr = (char *)p; - val = 0; - first = 1; - f = 1; - for (i = addr_len - 1; i > 0; i--) { - c = addr[i]; - if (c >= '0' && c <= '9') { - val += (c - '0') * f; - f *= 10; - } else if (c == '.') { - if (first) { - *portp = val; - val = first = 0; - f = 1; - } else { - *portp |= (val << 8); - break; - } - } - } - - /* - * Simple sanity check. If we never saw a dot in the reply, - * then this was probably just garbage. - */ - if (first) - goto out_err; - - dprintk("RPC: rpcb_decode_getaddr port=%u\n", *portp); - return 0; - -out_err: - dprintk("RPC: rpcbind server returned malformed reply\n"); - return -EIO; -} - static int rpcb_dec_getaddr(struct rpc_rqst *req, __be32 *p, struct rpcbind_args *rpcb) { From f8b761eff1e49fde43dd940ac32b38cf7cee2a95 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:44 -0400 Subject: [PATCH 20/24] SUNRPC: Eliminate PROC macro from rpcb_clnt Clean up: Replace PROC macro with open coded C99 structure initializers to improve readability. The rpcbind v4 GETVERSADDR procedure is never sent by the current implementation, so it is not copied to the new structures. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 113 +++++++++++++++++++++++++++++++++-------- 1 file changed, 91 insertions(+), 22 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index c4b716cacf2..830faf4d999 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -864,39 +864,108 @@ out_fail: return -EIO; } -#define PROC(proc, argtype, restype) \ - [RPCBPROC_##proc] = { \ - .p_proc = RPCBPROC_##proc, \ - .p_encode = (kxdrproc_t) rpcb_enc_##argtype, \ - .p_decode = (kxdrproc_t) rpcb_dec_##restype, \ - .p_arglen = RPCB_##argtype##args_sz, \ - .p_replen = RPCB_##restype##res_sz, \ - .p_statidx = RPCBPROC_##proc, \ - .p_timer = 0, \ - .p_name = #proc, \ - } - /* * Not all rpcbind procedures described in RFC 1833 are implemented * since the Linux kernel RPC code requires only these. */ + static struct rpc_procinfo rpcb_procedures2[] = { - PROC(SET, mapping, set), - PROC(UNSET, mapping, set), - PROC(GETPORT, mapping, getport), + [RPCBPROC_SET] = { + .p_proc = RPCBPROC_SET, + .p_encode = (kxdrproc_t)rpcb_enc_mapping, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_mappingargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_SET, + .p_timer = 0, + .p_name = "SET", + }, + [RPCBPROC_UNSET] = { + .p_proc = RPCBPROC_UNSET, + .p_encode = (kxdrproc_t)rpcb_enc_mapping, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_mappingargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_UNSET, + .p_timer = 0, + .p_name = "UNSET", + }, + [RPCBPROC_GETPORT] = { + .p_proc = RPCBPROC_GETPORT, + .p_encode = (kxdrproc_t)rpcb_enc_mapping, + .p_decode = (kxdrproc_t)rpcb_dec_getport, + .p_arglen = RPCB_mappingargs_sz, + .p_replen = RPCB_getportres_sz, + .p_statidx = RPCBPROC_GETPORT, + .p_timer = 0, + .p_name = "GETPORT", + }, }; static struct rpc_procinfo rpcb_procedures3[] = { - PROC(SET, getaddr, set), - PROC(UNSET, getaddr, set), - PROC(GETADDR, getaddr, getaddr), + [RPCBPROC_SET] = { + .p_proc = RPCBPROC_SET, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_SET, + .p_timer = 0, + .p_name = "SET", + }, + [RPCBPROC_UNSET] = { + .p_proc = RPCBPROC_UNSET, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_UNSET, + .p_timer = 0, + .p_name = "UNSET", + }, + [RPCBPROC_GETADDR] = { + .p_proc = RPCBPROC_GETADDR, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_getaddr, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_getaddrres_sz, + .p_statidx = RPCBPROC_GETADDR, + .p_timer = 0, + .p_name = "GETADDR", + }, }; static struct rpc_procinfo rpcb_procedures4[] = { - PROC(SET, getaddr, set), - PROC(UNSET, getaddr, set), - PROC(GETADDR, getaddr, getaddr), - PROC(GETVERSADDR, getaddr, getaddr), + [RPCBPROC_SET] = { + .p_proc = RPCBPROC_SET, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_SET, + .p_timer = 0, + .p_name = "SET", + }, + [RPCBPROC_UNSET] = { + .p_proc = RPCBPROC_UNSET, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_set, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_setres_sz, + .p_statidx = RPCBPROC_UNSET, + .p_timer = 0, + .p_name = "UNSET", + }, + [RPCBPROC_GETADDR] = { + .p_proc = RPCBPROC_GETADDR, + .p_encode = (kxdrproc_t)rpcb_enc_getaddr, + .p_decode = (kxdrproc_t)rpcb_dec_getaddr, + .p_arglen = RPCB_getaddrargs_sz, + .p_replen = RPCB_getaddrres_sz, + .p_statidx = RPCBPROC_GETADDR, + .p_timer = 0, + .p_name = "GETADDR", + }, }; static struct rpcb_info rpcb_next_version[] = { From fbfffbd5e74c5fa8c9165e110cb5899ec21e6364 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:46 -0400 Subject: [PATCH 21/24] SUNRPC: Rename sock_xprt.addr as sock_xprt.srcaddr Clean up: Give the "addr" and "port" field less ambiguous names. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/xprtsock.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 302a4090492..7bc3c178ccb 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -248,8 +248,8 @@ struct sock_xprt { * Connection of transports */ struct delayed_work connect_worker; - struct sockaddr_storage addr; - unsigned short port; + struct sockaddr_storage srcaddr; + unsigned short srcport; /* * UDP socket buffer size parameters @@ -1546,7 +1546,7 @@ static void xs_set_port(struct rpc_xprt *xprt, unsigned short port) static unsigned short xs_get_srcport(struct sock_xprt *transport, struct socket *sock) { - unsigned short port = transport->port; + unsigned short port = transport->srcport; if (port == 0 && transport->xprt.resvport) port = xs_get_random_port(); @@ -1555,8 +1555,8 @@ static unsigned short xs_get_srcport(struct sock_xprt *transport, struct socket static unsigned short xs_next_srcport(struct sock_xprt *transport, struct socket *sock, unsigned short port) { - if (transport->port != 0) - transport->port = 0; + if (transport->srcport != 0) + transport->srcport = 0; if (!transport->xprt.resvport) return 0; if (port <= xprt_min_resvport || port > xprt_max_resvport) @@ -1574,7 +1574,7 @@ static int xs_bind4(struct sock_xprt *transport, struct socket *sock) unsigned short port = xs_get_srcport(transport, sock); unsigned short last; - sa = (struct sockaddr_in *)&transport->addr; + sa = (struct sockaddr_in *)&transport->srcaddr; myaddr.sin_addr = sa->sin_addr; do { myaddr.sin_port = htons(port); @@ -1583,7 +1583,7 @@ static int xs_bind4(struct sock_xprt *transport, struct socket *sock) if (port == 0) break; if (err == 0) { - transport->port = port; + transport->srcport = port; break; } last = port; @@ -1607,7 +1607,7 @@ static int xs_bind6(struct sock_xprt *transport, struct socket *sock) unsigned short port = xs_get_srcport(transport, sock); unsigned short last; - sa = (struct sockaddr_in6 *)&transport->addr; + sa = (struct sockaddr_in6 *)&transport->srcaddr; myaddr.sin6_addr = sa->sin6_addr; do { myaddr.sin6_port = htons(port); @@ -1616,7 +1616,7 @@ static int xs_bind6(struct sock_xprt *transport, struct socket *sock) if (port == 0) break; if (err == 0) { - transport->port = port; + transport->srcport = port; break; } last = port; @@ -2061,7 +2061,7 @@ static void xs_udp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq) struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt); seq_printf(seq, "\txprt:\tudp %u %lu %lu %lu %lu %Lu %Lu\n", - transport->port, + transport->srcport, xprt->stat.bind_count, xprt->stat.sends, xprt->stat.recvs, @@ -2085,7 +2085,7 @@ static void xs_tcp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq) idle_time = (long)(jiffies - xprt->last_used) / HZ; seq_printf(seq, "\txprt:\ttcp %u %lu %lu %lu %ld %lu %lu %lu %Lu %Lu\n", - transport->port, + transport->srcport, xprt->stat.bind_count, xprt->stat.connect_count, xprt->stat.connect_time, @@ -2164,7 +2164,7 @@ static struct rpc_xprt *xs_setup_xprt(struct xprt_create *args, memcpy(&xprt->addr, args->dstaddr, args->addrlen); xprt->addrlen = args->addrlen; if (args->srcaddr) - memcpy(&new->addr, args->srcaddr, args->addrlen); + memcpy(&new->srcaddr, args->srcaddr, args->addrlen); return xprt; } From c740eff84bcfd63c0497ef880e80171931cb8222 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:46 -0400 Subject: [PATCH 22/24] SUNRPC: Kill RPC_DISPLAY_ALL At some point, I recall that rpc_pipe_fs used RPC_DISPLAY_ALL. Currently there are no uses of RPC_DISPLAY_ALL outside the transport modules themselves, so we can safely get rid of it. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 1 - net/sunrpc/xprtrdma/transport.c | 5 ---- net/sunrpc/xprtsock.c | 50 ++++++++++++++++++++++----------- 3 files changed, 34 insertions(+), 22 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 65fad9534d9..c090df44257 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -38,7 +38,6 @@ enum rpc_display_format_t { RPC_DISPLAY_ADDR = 0, RPC_DISPLAY_PORT, RPC_DISPLAY_PROTO, - RPC_DISPLAY_ALL, RPC_DISPLAY_HEX_ADDR, RPC_DISPLAY_HEX_PORT, RPC_DISPLAY_NETID, diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 5f9b8676b6b..9a63f669ece 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -181,11 +181,6 @@ xprt_rdma_format_addresses(struct rpc_xprt *xprt) xprt->address_strings[RPC_DISPLAY_PROTO] = "rdma"; - (void)snprintf(buf, sizeof(buf), "addr=%s port=%s proto=rdma", - xprt->address_strings[RPC_DISPLAY_ADDR], - xprt->address_strings[RPC_DISPLAY_PORT]); - xprt->address_strings[RPC_DISPLAY_ALL] = kstrdup(buf, GFP_KERNEL); - (void)snprintf(buf, sizeof(buf), "%02x%02x%02x%02x", NIPQUAD(sin->sin_addr.s_addr)); xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 7bc3c178ccb..eee5ac96e17 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -307,12 +307,6 @@ static void xs_format_common_peer_addresses(struct rpc_xprt *xprt) (void)snprintf(buf, sizeof(buf), "%u", rpc_get_port(sap)); xprt->address_strings[RPC_DISPLAY_PORT] = kstrdup(buf, GFP_KERNEL); - (void)snprintf(buf, sizeof(buf), "addr=%s port=%s proto=%s", - xprt->address_strings[RPC_DISPLAY_ADDR], - xprt->address_strings[RPC_DISPLAY_PORT], - xprt->address_strings[RPC_DISPLAY_PROTO]); - xprt->address_strings[RPC_DISPLAY_ALL] = kstrdup(buf, GFP_KERNEL); - (void)snprintf(buf, sizeof(buf), "%4hx", rpc_get_port(sap)); xprt->address_strings[RPC_DISPLAY_HEX_PORT] = kstrdup(buf, GFP_KERNEL); } @@ -1721,8 +1715,11 @@ static void xs_udp_connect_worker4(struct work_struct *work) goto out; } - dprintk("RPC: worker connecting xprt %p to address: %s\n", - xprt, xprt->address_strings[RPC_DISPLAY_ALL]); + dprintk("RPC: worker connecting xprt %p via %s to " + "%s (port %s)\n", xprt, + xprt->address_strings[RPC_DISPLAY_PROTO], + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT]); xs_udp_finish_connecting(xprt, sock); status = 0; @@ -1763,8 +1760,11 @@ static void xs_udp_connect_worker6(struct work_struct *work) goto out; } - dprintk("RPC: worker connecting xprt %p to address: %s\n", - xprt, xprt->address_strings[RPC_DISPLAY_ALL]); + dprintk("RPC: worker connecting xprt %p via %s to " + "%s (port %s)\n", xprt, + xprt->address_strings[RPC_DISPLAY_PROTO], + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT]); xs_udp_finish_connecting(xprt, sock); status = 0; @@ -1889,8 +1889,11 @@ static void xs_tcp_setup_socket(struct rpc_xprt *xprt, goto out_eagain; } - dprintk("RPC: worker connecting xprt %p to address: %s\n", - xprt, xprt->address_strings[RPC_DISPLAY_ALL]); + dprintk("RPC: worker connecting xprt %p via %s to " + "%s (port %s)\n", xprt, + xprt->address_strings[RPC_DISPLAY_PROTO], + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT]); status = xs_tcp_finish_connecting(xprt, sock); dprintk("RPC: %p connect status %d connected %d sock state %d\n", @@ -2228,8 +2231,15 @@ static struct rpc_xprt *xs_setup_udp(struct xprt_create *args) return ERR_PTR(-EAFNOSUPPORT); } - dprintk("RPC: set up transport to address %s\n", - xprt->address_strings[RPC_DISPLAY_ALL]); + if (xprt_bound(xprt)) + dprintk("RPC: set up xprt to %s (port %s) via %s\n", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT], + xprt->address_strings[RPC_DISPLAY_PROTO]); + else + dprintk("RPC: set up xprt to %s (autobind) via %s\n", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PROTO]); if (try_module_get(THIS_MODULE)) return xprt; @@ -2293,8 +2303,16 @@ static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args) return ERR_PTR(-EAFNOSUPPORT); } - dprintk("RPC: set up transport to address %s\n", - xprt->address_strings[RPC_DISPLAY_ALL]); + if (xprt_bound(xprt)) + dprintk("RPC: set up xprt to %s (port %s) via %s\n", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PORT], + xprt->address_strings[RPC_DISPLAY_PROTO]); + else + dprintk("RPC: set up xprt to %s (autobind) via %s\n", + xprt->address_strings[RPC_DISPLAY_ADDR], + xprt->address_strings[RPC_DISPLAY_PROTO]); + if (try_module_get(THIS_MODULE)) return xprt; From 9dc3b095b78347bfb02c324b5ee2e558f7267396 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:46 -0400 Subject: [PATCH 23/24] SUNRPC: Update xprt address strings after an rpcbind completes After a bind completes, update the transport instance's address strings so debugging messages display the current port the transport is connected to. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/xprtsock.c | 82 +++++++++++++++++++++---------------------- 1 file changed, 41 insertions(+), 41 deletions(-) diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index eee5ac96e17..8aaf9003fe6 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -299,11 +299,34 @@ static inline struct sockaddr_in6 *xs_addr_in6(struct rpc_xprt *xprt) static void xs_format_common_peer_addresses(struct rpc_xprt *xprt) { struct sockaddr *sap = xs_addr(xprt); + struct sockaddr_in6 *sin6; + struct sockaddr_in *sin; char buf[128]; (void)rpc_ntop(sap, buf, sizeof(buf)); xprt->address_strings[RPC_DISPLAY_ADDR] = kstrdup(buf, GFP_KERNEL); + switch (sap->sa_family) { + case AF_INET: + sin = xs_addr_in(xprt); + (void)snprintf(buf, sizeof(buf), "%02x%02x%02x%02x", + NIPQUAD(sin->sin_addr.s_addr)); + break; + case AF_INET6: + sin6 = xs_addr_in6(xprt); + (void)snprintf(buf, sizeof(buf), "%pi6", &sin6->sin6_addr); + break; + default: + BUG(); + } + xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); +} + +static void xs_format_common_peer_ports(struct rpc_xprt *xprt) +{ + struct sockaddr *sap = xs_addr(xprt); + char buf[128]; + (void)snprintf(buf, sizeof(buf), "%u", rpc_get_port(sap)); xprt->address_strings[RPC_DISPLAY_PORT] = kstrdup(buf, GFP_KERNEL); @@ -311,37 +334,22 @@ static void xs_format_common_peer_addresses(struct rpc_xprt *xprt) xprt->address_strings[RPC_DISPLAY_HEX_PORT] = kstrdup(buf, GFP_KERNEL); } -static void xs_format_ipv4_peer_addresses(struct rpc_xprt *xprt, - const char *protocol, - const char *netid) +static void xs_format_peer_addresses(struct rpc_xprt *xprt, + const char *protocol, + const char *netid) { - struct sockaddr_in *sin = xs_addr_in(xprt); - char buf[16]; - xprt->address_strings[RPC_DISPLAY_PROTO] = protocol; xprt->address_strings[RPC_DISPLAY_NETID] = netid; - - (void)snprintf(buf, sizeof(buf), "%02x%02x%02x%02x", - NIPQUAD(sin->sin_addr.s_addr)); - xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); - xs_format_common_peer_addresses(xprt); + xs_format_common_peer_ports(xprt); } -static void xs_format_ipv6_peer_addresses(struct rpc_xprt *xprt, - const char *protocol, - const char *netid) +static void xs_update_peer_port(struct rpc_xprt *xprt) { - struct sockaddr_in6 *sin6 = xs_addr_in6(xprt); - char buf[48]; + kfree(xprt->address_strings[RPC_DISPLAY_HEX_PORT]); + kfree(xprt->address_strings[RPC_DISPLAY_PORT]); - xprt->address_strings[RPC_DISPLAY_PROTO] = protocol; - xprt->address_strings[RPC_DISPLAY_NETID] = netid; - - (void)snprintf(buf, sizeof(buf), "%pi6", &sin6->sin6_addr); - xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL); - - xs_format_common_peer_addresses(xprt); + xs_format_common_peer_ports(xprt); } static void xs_free_peer_addresses(struct rpc_xprt *xprt) @@ -1522,20 +1530,10 @@ static unsigned short xs_get_random_port(void) */ static void xs_set_port(struct rpc_xprt *xprt, unsigned short port) { - struct sockaddr *addr = xs_addr(xprt); - dprintk("RPC: setting port for xprt %p to %u\n", xprt, port); - switch (addr->sa_family) { - case AF_INET: - ((struct sockaddr_in *)addr)->sin_port = htons(port); - break; - case AF_INET6: - ((struct sockaddr_in6 *)addr)->sin6_port = htons(port); - break; - default: - BUG(); - } + rpc_set_port(xs_addr(xprt), port); + xs_update_peer_port(xprt); } static unsigned short xs_get_srcport(struct sock_xprt *transport, struct socket *sock) @@ -2216,7 +2214,7 @@ static struct rpc_xprt *xs_setup_udp(struct xprt_create *args) INIT_DELAYED_WORK(&transport->connect_worker, xs_udp_connect_worker4); - xs_format_ipv4_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP); + xs_format_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP); break; case AF_INET6: if (((struct sockaddr_in6 *)addr)->sin6_port != htons(0)) @@ -2224,7 +2222,7 @@ static struct rpc_xprt *xs_setup_udp(struct xprt_create *args) INIT_DELAYED_WORK(&transport->connect_worker, xs_udp_connect_worker6); - xs_format_ipv6_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP6); + xs_format_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP6); break; default: kfree(xprt); @@ -2288,15 +2286,17 @@ static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args) if (((struct sockaddr_in *)addr)->sin_port != htons(0)) xprt_set_bound(xprt); - INIT_DELAYED_WORK(&transport->connect_worker, xs_tcp_connect_worker4); - xs_format_ipv4_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP); + INIT_DELAYED_WORK(&transport->connect_worker, + xs_tcp_connect_worker4); + xs_format_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP); break; case AF_INET6: if (((struct sockaddr_in6 *)addr)->sin6_port != htons(0)) xprt_set_bound(xprt); - INIT_DELAYED_WORK(&transport->connect_worker, xs_tcp_connect_worker6); - xs_format_ipv6_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP6); + INIT_DELAYED_WORK(&transport->connect_worker, + xs_tcp_connect_worker6); + xs_format_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP6); break; default: kfree(xprt); From c05988cdb06237738d361ef82fbf4df1020aa3db Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 9 Aug 2009 15:09:47 -0400 Subject: [PATCH 24/24] SUNRPC: Add documenting comments in net/sunrpc/timer.c Clean up: provide documenting comments for the functions in net/sunrpc/timer.c. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/timer.c | 45 +++++++++++++++++++++++++++++---------------- 1 file changed, 29 insertions(+), 16 deletions(-) diff --git a/net/sunrpc/timer.c b/net/sunrpc/timer.c index 31becbf0926..dd824341c34 100644 --- a/net/sunrpc/timer.c +++ b/net/sunrpc/timer.c @@ -25,8 +25,13 @@ #define RPC_RTO_INIT (HZ/5) #define RPC_RTO_MIN (HZ/10) -void -rpc_init_rtt(struct rpc_rtt *rt, unsigned long timeo) +/** + * rpc_init_rtt - Initialize an RPC RTT estimator context + * @rt: context to initialize + * @timeo: initial timeout value, in jiffies + * + */ +void rpc_init_rtt(struct rpc_rtt *rt, unsigned long timeo) { unsigned long init = 0; unsigned i; @@ -43,12 +48,16 @@ rpc_init_rtt(struct rpc_rtt *rt, unsigned long timeo) } EXPORT_SYMBOL_GPL(rpc_init_rtt); -/* +/** + * rpc_update_rtt - Update an RPC RTT estimator context + * @rt: context to update + * @timer: timer array index (request type) + * @m: recent actual RTT, in jiffies + * * NB: When computing the smoothed RTT and standard deviation, * be careful not to produce negative intermediate results. */ -void -rpc_update_rtt(struct rpc_rtt *rt, unsigned timer, long m) +void rpc_update_rtt(struct rpc_rtt *rt, unsigned timer, long m) { long *srtt, *sdrtt; @@ -79,21 +88,25 @@ rpc_update_rtt(struct rpc_rtt *rt, unsigned timer, long m) } EXPORT_SYMBOL_GPL(rpc_update_rtt); -/* - * Estimate rto for an nfs rpc sent via. an unreliable datagram. - * Use the mean and mean deviation of rtt for the appropriate type of rpc - * for the frequent rpcs and a default for the others. - * The justification for doing "other" this way is that these rpcs - * happen so infrequently that timer est. would probably be stale. - * Also, since many of these rpcs are - * non-idempotent, a conservative timeout is desired. +/** + * rpc_calc_rto - Provide an estimated timeout value + * @rt: context to use for calculation + * @timer: timer array index (request type) + * + * Estimate RTO for an NFS RPC sent via an unreliable datagram. Use + * the mean and mean deviation of RTT for the appropriate type of RPC + * for frequently issued RPCs, and a fixed default for the others. + * + * The justification for doing "other" this way is that these RPCs + * happen so infrequently that timer estimation would probably be + * stale. Also, since many of these RPCs are non-idempotent, a + * conservative timeout is desired. + * * getattr, lookup, * read, write, commit - A+4D * other - timeo */ - -unsigned long -rpc_calc_rto(struct rpc_rtt *rt, unsigned timer) +unsigned long rpc_calc_rto(struct rpc_rtt *rt, unsigned timer) { unsigned long res;