/* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ /* * Copyright (c) 1983,1984,1985,1986,1987,1988,1989 AT&T. * All rights reserved. * Use is subject to license terms. */ #pragma ident "%Z%%M% %I% %E% SMI" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define MAXHOST 32 const char *kinet_ntop6(uchar_t *, char *, size_t); /* * Module linkage information. */ static struct modlmisc modlmisc = { &mod_miscops, "NFS server module" }; static struct modlinkage modlinkage = { MODREV_1, (void *)&modlmisc, NULL }; char _depends_on[] = "misc/klmmod"; int _init(void) { int status; if ((status = nfs_srvinit()) != 0) { cmn_err(CE_WARN, "_init: nfs_srvinit failed"); return (status); } status = mod_install((struct modlinkage *)&modlinkage); if (status != 0) { /* * Could not load module, cleanup previous * initialization work. */ nfs_srvfini(); } /* * Initialise some placeholders for nfssys() calls. These have * to be declared by the nfs module, since that handles nfssys() * calls - also used by NFS clients - but are provided by this * nfssrv module. These also then serve as confirmation to the * relevant code in nfs that nfssrv has been loaded, as they're * initially NULL. */ nfs_srv_quiesce_func = nfs_srv_quiesce_all; nfs_srv_dss_func = rfs4_dss_setpaths; /* setup DSS paths here; must be done before initial server startup */ rfs4_dss_paths = rfs4_dss_oldpaths = NULL; return (status); } int _fini() { return (EBUSY); } int _info(struct modinfo *modinfop) { return (mod_info(&modlinkage, modinfop)); } /* * PUBLICFH_CHECK() checks if the dispatch routine supports * RPC_PUBLICFH_OK, if the filesystem is exported public, and if the * incoming request is using the public filehandle. The check duplicates * the exportmatch() call done in checkexport(), and we should consider * modifying those routines to avoid the duplication. For now, we optimize * by calling exportmatch() only after checking that the dispatch routine * supports RPC_PUBLICFH_OK, and if the filesystem is explicitly exported * public (i.e., not the placeholder). */ #define PUBLICFH_CHECK(disp, exi, fsid, xfid) \ ((disp->dis_flags & RPC_PUBLICFH_OK) && \ ((exi->exi_export.ex_flags & EX_PUBLIC) || \ (exi == exi_public && exportmatch(exi_root, \ fsid, xfid)))) static void nfs_srv_shutdown_all(int); static void rfs4_server_start(int); static void nullfree(void); static void rfs_dispatch(struct svc_req *, SVCXPRT *); static void acl_dispatch(struct svc_req *, SVCXPRT *); static void common_dispatch(struct svc_req *, SVCXPRT *, rpcvers_t, rpcvers_t, char *, struct rpc_disptable *); static void hanfsv4_failover(void); static int checkauth(struct exportinfo *, struct svc_req *, cred_t *, int, bool_t); static char *client_name(struct svc_req *req); static char *client_addr(struct svc_req *req, char *buf); extern int sec_svc_getcred(struct svc_req *, cred_t *cr, char **, int *); extern bool_t sec_svc_inrootlist(int, caddr_t, int, caddr_t *); #define NFSLOG_COPY_NETBUF(exi, xprt, nb) { \ (nb)->maxlen = (xprt)->xp_rtaddr.maxlen; \ (nb)->len = (xprt)->xp_rtaddr.len; \ (nb)->buf = kmem_alloc((nb)->len, KM_SLEEP); \ bcopy((xprt)->xp_rtaddr.buf, (nb)->buf, (nb)->len); \ } /* * Public Filehandle common nfs routines */ static int MCLpath(char **); static void URLparse(char *); /* * NFS callout table. * This table is used by svc_getreq() to dispatch a request with * a given prog/vers pair to an appropriate service provider * dispatch routine. * * NOTE: ordering is relied upon below when resetting the version min/max * for NFS_PROGRAM. Careful, if this is ever changed. */ static SVC_CALLOUT __nfs_sc_clts[] = { { NFS_PROGRAM, NFS_VERSMIN, NFS_VERSMAX, rfs_dispatch }, { NFS_ACL_PROGRAM, NFS_ACL_VERSMIN, NFS_ACL_VERSMAX, acl_dispatch } }; static SVC_CALLOUT_TABLE nfs_sct_clts = { sizeof (__nfs_sc_clts) / sizeof (__nfs_sc_clts[0]), FALSE, __nfs_sc_clts }; static SVC_CALLOUT __nfs_sc_cots[] = { { NFS_PROGRAM, NFS_VERSMIN, NFS_VERSMAX, rfs_dispatch }, { NFS_ACL_PROGRAM, NFS_ACL_VERSMIN, NFS_ACL_VERSMAX, acl_dispatch } }; static SVC_CALLOUT_TABLE nfs_sct_cots = { sizeof (__nfs_sc_cots) / sizeof (__nfs_sc_cots[0]), FALSE, __nfs_sc_cots }; static SVC_CALLOUT __nfs_sc_rdma[] = { { NFS_PROGRAM, NFS_VERSMIN, NFS_VERSMAX, rfs_dispatch }, { NFS_ACL_PROGRAM, NFS_ACL_VERSMIN, NFS_ACL_VERSMAX, acl_dispatch } }; static SVC_CALLOUT_TABLE nfs_sct_rdma = { sizeof (__nfs_sc_rdma) / sizeof (__nfs_sc_rdma[0]), FALSE, __nfs_sc_rdma }; rpcvers_t nfs_versmin = NFS_VERSMIN_DEFAULT; rpcvers_t nfs_versmax = NFS_VERSMAX_DEFAULT; /* * Used to track the state of the server so that initialization * can be done properly. */ typedef enum { NFS_SERVER_STOPPED, /* server state destroyed */ NFS_SERVER_STOPPING, /* server state being destroyed */ NFS_SERVER_RUNNING, NFS_SERVER_QUIESCED, /* server state preserved */ NFS_SERVER_OFFLINE /* server pool offline */ } nfs_server_running_t; static nfs_server_running_t nfs_server_upordown; static kmutex_t nfs_server_upordown_lock; static kcondvar_t nfs_server_upordown_cv; /* * DSS: distributed stable storage * lists of all DSS paths: current, and before last warmstart */ nvlist_t *rfs4_dss_paths, *rfs4_dss_oldpaths; int rfs4_dispatch(struct rpcdisp *, struct svc_req *, SVCXPRT *, char *); bool_t rfs4_minorvers_mismatch(struct svc_req *, SVCXPRT *, void *); /* * RDMA wait variables. */ static kcondvar_t rdma_wait_cv; static kmutex_t rdma_wait_mutex; /* * Will be called at the point the server pool is being unregistered * from the pool list. From that point onwards, the pool is waiting * to be drained and as such the server state is stale and pertains * to the old instantiation of the NFS server pool. */ void nfs_srv_offline(void) { mutex_enter(&nfs_server_upordown_lock); if (nfs_server_upordown == NFS_SERVER_RUNNING) { nfs_server_upordown = NFS_SERVER_OFFLINE; } mutex_exit(&nfs_server_upordown_lock); } /* * Will be called at the point the server pool is being destroyed so * all transports have been closed and no service threads are in * existence. * * If we quiesce the server, we're shutting it down without destroying the * server state. This allows it to warm start subsequently. */ void nfs_srv_stop_all(void) { int quiesce = 0; nfs_srv_shutdown_all(quiesce); } /* * This alternative shutdown routine can be requested via nfssys() */ void nfs_srv_quiesce_all(void) { int quiesce = 1; nfs_srv_shutdown_all(quiesce); } static void nfs_srv_shutdown_all(int quiesce) { mutex_enter(&nfs_server_upordown_lock); if (quiesce) { if (nfs_server_upordown == NFS_SERVER_RUNNING || nfs_server_upordown == NFS_SERVER_OFFLINE) { nfs_server_upordown = NFS_SERVER_QUIESCED; cv_signal(&nfs_server_upordown_cv); /* reset DSS state, for subsequent warm restart */ rfs4_dss_numnewpaths = 0; rfs4_dss_newpaths = NULL; cmn_err(CE_NOTE, "nfs_server: server is now quiesced; " "NFSv4 state has been preserved"); } } else { if (nfs_server_upordown == NFS_SERVER_OFFLINE) { nfs_server_upordown = NFS_SERVER_STOPPING; mutex_exit(&nfs_server_upordown_lock); rfs4_state_fini(); rfs4_fini_drc(nfs4_drc); mutex_enter(&nfs_server_upordown_lock); nfs_server_upordown = NFS_SERVER_STOPPED; cv_signal(&nfs_server_upordown_cv); } } mutex_exit(&nfs_server_upordown_lock); } static int nfs_srv_set_sc_versions(struct file *fp, SVC_CALLOUT_TABLE **sctpp, rpcvers_t versmin, rpcvers_t versmax) { struct strioctl strioc; struct T_info_ack tinfo; int error, retval; /* * Find out what type of transport this is. */ strioc.ic_cmd = TI_GETINFO; strioc.ic_timout = -1; strioc.ic_len = sizeof (tinfo); strioc.ic_dp = (char *)&tinfo; tinfo.PRIM_type = T_INFO_REQ; error = strioctl(fp->f_vnode, I_STR, (intptr_t)&strioc, 0, K_TO_K, CRED(), &retval); if (error || retval) return (error); /* * Based on our query of the transport type... * * Reset the min/max versions based on the caller's request * NOTE: This assumes that NFS_PROGRAM is first in the array!! * And the second entry is the NFS_ACL_PROGRAM. */ switch (tinfo.SERV_type) { case T_CLTS: if (versmax == NFS_V4) return (EINVAL); __nfs_sc_clts[0].sc_versmin = versmin; __nfs_sc_clts[0].sc_versmax = versmax; __nfs_sc_clts[1].sc_versmin = versmin; __nfs_sc_clts[1].sc_versmax = versmax; *sctpp = &nfs_sct_clts; break; case T_COTS: case T_COTS_ORD: __nfs_sc_cots[0].sc_versmin = versmin; __nfs_sc_cots[0].sc_versmax = versmax; /* For the NFS_ACL program, check the max version */ if (versmax > NFS_ACL_VERSMAX) versmax = NFS_ACL_VERSMAX; __nfs_sc_cots[1].sc_versmin = versmin; __nfs_sc_cots[1].sc_versmax = versmax; *sctpp = &nfs_sct_cots; break; default: error = EINVAL; } return (error); } /* * NFS Server system call. * Does all of the work of running a NFS server. * uap->fd is the fd of an open transport provider */ int nfs_svc(struct nfs_svc_args *arg, model_t model) { file_t *fp; SVCMASTERXPRT *xprt; int error; int readsize; char buf[KNC_STRSIZE]; size_t len; STRUCT_HANDLE(nfs_svc_args, uap); struct netbuf addrmask; SVC_CALLOUT_TABLE *sctp = NULL; #ifdef lint model = model; /* STRUCT macros don't always refer to it */ #endif STRUCT_SET_HANDLE(uap, model, arg); /* Check privileges in nfssys() */ if ((fp = getf(STRUCT_FGET(uap, fd))) == NULL) return (EBADF); /* * Set read buffer size to rsize * and add room for RPC headers. */ readsize = nfs3tsize() + (RPC_MAXDATASIZE - NFS_MAXDATA); if (readsize < RPC_MAXDATASIZE) readsize = RPC_MAXDATASIZE; error = copyinstr((const char *)STRUCT_FGETP(uap, netid), buf, KNC_STRSIZE, &len); if (error) { releasef(STRUCT_FGET(uap, fd)); return (error); } addrmask.len = STRUCT_FGET(uap, addrmask.len); addrmask.maxlen = STRUCT_FGET(uap, addrmask.maxlen); addrmask.buf = kmem_alloc(addrmask.maxlen, KM_SLEEP); error = copyin(STRUCT_FGETP(uap, addrmask.buf), addrmask.buf, addrmask.len); if (error) { releasef(STRUCT_FGET(uap, fd)); kmem_free(addrmask.buf, addrmask.maxlen); return (error); } nfs_versmin = STRUCT_FGET(uap, versmin); nfs_versmax = STRUCT_FGET(uap, versmax); /* Double check the vers min/max ranges */ if ((nfs_versmin > nfs_versmax) || (nfs_versmin < NFS_VERSMIN) || (nfs_versmax > NFS_VERSMAX)) { nfs_versmin = NFS_VERSMIN_DEFAULT; nfs_versmax = NFS_VERSMAX_DEFAULT; } if (error = nfs_srv_set_sc_versions(fp, &sctp, nfs_versmin, nfs_versmax)) { releasef(STRUCT_FGET(uap, fd)); kmem_free(addrmask.buf, addrmask.maxlen); return (error); } /* Initialize nfsv4 server */ if (nfs_versmax == (rpcvers_t)NFS_V4) rfs4_server_start(STRUCT_FGET(uap, delegation)); /* Create a transport handle. */ error = svc_tli_kcreate(fp, readsize, buf, &addrmask, &xprt, sctp, NULL, NFS_SVCPOOL_ID, TRUE); if (error) kmem_free(addrmask.buf, addrmask.maxlen); releasef(STRUCT_FGET(uap, fd)); /* HA-NFSv4: save the cluster nodeid */ if (cluster_bootflags & CLUSTER_BOOTED) lm_global_nlmid = clconf_get_nodeid(); return (error); } static void rfs4_server_start(int nfs4_srv_delegation) { /* * Determine if the server has previously been "started" and * if not, do the per instance initialization */ mutex_enter(&nfs_server_upordown_lock); if (nfs_server_upordown != NFS_SERVER_RUNNING) { /* Do we need to stop and wait on the previous server? */ while (nfs_server_upordown == NFS_SERVER_STOPPING || nfs_server_upordown == NFS_SERVER_OFFLINE) cv_wait(&nfs_server_upordown_cv, &nfs_server_upordown_lock); if (nfs_server_upordown != NFS_SERVER_RUNNING) { (void) svc_pool_control(NFS_SVCPOOL_ID, SVCPSET_UNREGISTER_PROC, (void *)&nfs_srv_offline); (void) svc_pool_control(NFS_SVCPOOL_ID, SVCPSET_SHUTDOWN_PROC, (void *)&nfs_srv_stop_all); /* is this an nfsd warm start? */ if (nfs_server_upordown == NFS_SERVER_QUIESCED) { cmn_err(CE_NOTE, "nfs_server: " "server was previously quiesced; " "existing NFSv4 state will be re-used"); /* * HA-NFSv4: this is also the signal * that a Resource Group failover has * occurred. */ if (cluster_bootflags & CLUSTER_BOOTED) hanfsv4_failover(); } else { /* cold start */ rfs4_state_init(); nfs4_drc = rfs4_init_drc(nfs4_drc_max, nfs4_drc_hash); } /* * Check to see if delegation is to be * enabled at the server */ if (nfs4_srv_delegation != FALSE) rfs4_set_deleg_policy(SRV_NORMAL_DELEGATE); nfs_server_upordown = NFS_SERVER_RUNNING; } cv_signal(&nfs_server_upordown_cv); } mutex_exit(&nfs_server_upordown_lock); } /* * If RDMA device available, * start RDMA listener. */ int rdma_start(struct rdma_svc_args *rsa) { int error; rdma_xprt_group_t started_rdma_xprts; /* Double check the vers min/max ranges */ if ((rsa->nfs_versmin > rsa->nfs_versmax) || (rsa->nfs_versmin < NFS_VERSMIN) || (rsa->nfs_versmax > NFS_VERSMAX)) { rsa->nfs_versmin = NFS_VERSMIN_DEFAULT; rsa->nfs_versmax = NFS_VERSMAX_DEFAULT; } nfs_versmin = rsa->nfs_versmin; nfs_versmax = rsa->nfs_versmax; /* Set the versions in the callout table */ __nfs_sc_rdma[0].sc_versmin = rsa->nfs_versmin; __nfs_sc_rdma[0].sc_versmax = rsa->nfs_versmax; /* For the NFS_ACL program, check the max version */ __nfs_sc_rdma[1].sc_versmin = rsa->nfs_versmin; if (rsa->nfs_versmax > NFS_ACL_VERSMAX) __nfs_sc_rdma[1].sc_versmax = NFS_ACL_VERSMAX; else __nfs_sc_rdma[1].sc_versmax = rsa->nfs_versmax; /* Initialize nfsv4 server */ if (rsa->nfs_versmax == (rpcvers_t)NFS_V4) rfs4_server_start(rsa->delegation); started_rdma_xprts.rtg_count = 0; started_rdma_xprts.rtg_listhead = NULL; started_rdma_xprts.rtg_poolid = rsa->poolid; error = svc_rdma_kcreate(rsa->netid, &nfs_sct_rdma, rsa->poolid, &started_rdma_xprts); if (error == 0) { mutex_enter(&rdma_wait_mutex); if (!cv_wait_sig(&rdma_wait_cv, &rdma_wait_mutex)) { rdma_stop(started_rdma_xprts); } mutex_exit(&rdma_wait_mutex); } return (error); } /* ARGSUSED */ void rpc_null(caddr_t *argp, caddr_t *resp) { } /* ARGSUSED */ void rpc_null_v3(caddr_t *argp, caddr_t *resp, struct exportinfo *exi, struct svc_req *req, cred_t *cr) { DTRACE_NFSV3_3(op__null__start, struct svc_req *, req, cred_t *, cr, vnode_t *, NULL); DTRACE_NFSV3_3(op__null__done, struct svc_req *, req, cred_t *, cr, vnode_t *, NULL); } /* ARGSUSED */ static void rfs_error(caddr_t *argp, caddr_t *resp) { /* return (EOPNOTSUPP); */ } static void nullfree(void) { } static char *rfscallnames_v2[] = { "RFS2_NULL", "RFS2_GETATTR", "RFS2_SETATTR", "RFS2_ROOT", "RFS2_LOOKUP", "RFS2_READLINK", "RFS2_READ", "RFS2_WRITECACHE", "RFS2_WRITE", "RFS2_CREATE", "RFS2_REMOVE", "RFS2_RENAME", "RFS2_LINK", "RFS2_SYMLINK", "RFS2_MKDIR", "RFS2_RMDIR", "RFS2_READDIR", "RFS2_STATFS" }; static struct rpcdisp rfsdisptab_v2[] = { /* * NFS VERSION 2 */ /* RFS_NULL = 0 */ {rpc_null, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* RFS_GETATTR = 1 */ {rfs_getattr, xdr_fhandle, xdr_fastfhandle, sizeof (fhandle_t), xdr_attrstat, xdr_fastattrstat, sizeof (struct nfsattrstat), nullfree, RPC_IDEMPOTENT|RPC_ALLOWANON|RPC_MAPRESP, rfs_getattr_getfh}, /* RFS_SETATTR = 2 */ {rfs_setattr, xdr_saargs, NULL_xdrproc_t, sizeof (struct nfssaargs), xdr_attrstat, xdr_fastattrstat, sizeof (struct nfsattrstat), nullfree, RPC_MAPRESP, rfs_setattr_getfh}, /* RFS_ROOT = 3 *** NO LONGER SUPPORTED *** */ {rfs_error, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* RFS_LOOKUP = 4 */ {rfs_lookup, xdr_diropargs, NULL_xdrproc_t, sizeof (struct nfsdiropargs), xdr_diropres, xdr_fastdiropres, sizeof (struct nfsdiropres), nullfree, RPC_IDEMPOTENT|RPC_MAPRESP|RPC_PUBLICFH_OK, rfs_lookup_getfh}, /* RFS_READLINK = 5 */ {rfs_readlink, xdr_fhandle, xdr_fastfhandle, sizeof (fhandle_t), xdr_rdlnres, NULL_xdrproc_t, sizeof (struct nfsrdlnres), rfs_rlfree, RPC_IDEMPOTENT, rfs_readlink_getfh}, /* RFS_READ = 6 */ {rfs_read, xdr_readargs, NULL_xdrproc_t, sizeof (struct nfsreadargs), xdr_rdresult, NULL_xdrproc_t, sizeof (struct nfsrdresult), rfs_rdfree, RPC_IDEMPOTENT, rfs_read_getfh}, /* RFS_WRITECACHE = 7 *** NO LONGER SUPPORTED *** */ {rfs_error, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* RFS_WRITE = 8 */ {rfs_write, xdr_writeargs, NULL_xdrproc_t, sizeof (struct nfswriteargs), xdr_attrstat, xdr_fastattrstat, sizeof (struct nfsattrstat), nullfree, RPC_MAPRESP, rfs_write_getfh}, /* RFS_CREATE = 9 */ {rfs_create, xdr_creatargs, NULL_xdrproc_t, sizeof (struct nfscreatargs), xdr_diropres, xdr_fastdiropres, sizeof (struct nfsdiropres), nullfree, RPC_MAPRESP, rfs_create_getfh}, /* RFS_REMOVE = 10 */ {rfs_remove, xdr_diropargs, NULL_xdrproc_t, sizeof (struct nfsdiropargs), #ifdef _LITTLE_ENDIAN xdr_enum, xdr_fastenum, sizeof (enum nfsstat), #else xdr_enum, NULL_xdrproc_t, sizeof (enum nfsstat), #endif nullfree, RPC_MAPRESP, rfs_remove_getfh}, /* RFS_RENAME = 11 */ {rfs_rename, xdr_rnmargs, NULL_xdrproc_t, sizeof (struct nfsrnmargs), #ifdef _LITTLE_ENDIAN xdr_enum, xdr_fastenum, sizeof (enum nfsstat), #else xdr_enum, NULL_xdrproc_t, sizeof (enum nfsstat), #endif nullfree, RPC_MAPRESP, rfs_rename_getfh}, /* RFS_LINK = 12 */ {rfs_link, xdr_linkargs, NULL_xdrproc_t, sizeof (struct nfslinkargs), #ifdef _LITTLE_ENDIAN xdr_enum, xdr_fastenum, sizeof (enum nfsstat), #else xdr_enum, NULL_xdrproc_t, sizeof (enum nfsstat), #endif nullfree, RPC_MAPRESP, rfs_link_getfh}, /* RFS_SYMLINK = 13 */ {rfs_symlink, xdr_slargs, NULL_xdrproc_t, sizeof (struct nfsslargs), #ifdef _LITTLE_ENDIAN xdr_enum, xdr_fastenum, sizeof (enum nfsstat), #else xdr_enum, NULL_xdrproc_t, sizeof (enum nfsstat), #endif nullfree, RPC_MAPRESP, rfs_symlink_getfh}, /* RFS_MKDIR = 14 */ {rfs_mkdir, xdr_creatargs, NULL_xdrproc_t, sizeof (struct nfscreatargs), xdr_diropres, xdr_fastdiropres, sizeof (struct nfsdiropres), nullfree, RPC_MAPRESP, rfs_mkdir_getfh}, /* RFS_RMDIR = 15 */ {rfs_rmdir, xdr_diropargs, NULL_xdrproc_t, sizeof (struct nfsdiropargs), #ifdef _LITTLE_ENDIAN xdr_enum, xdr_fastenum, sizeof (enum nfsstat), #else xdr_enum, NULL_xdrproc_t, sizeof (enum nfsstat), #endif nullfree, RPC_MAPRESP, rfs_rmdir_getfh}, /* RFS_READDIR = 16 */ {rfs_readdir, xdr_rddirargs, NULL_xdrproc_t, sizeof (struct nfsrddirargs), xdr_putrddirres, NULL_xdrproc_t, sizeof (struct nfsrddirres), rfs_rddirfree, RPC_IDEMPOTENT, rfs_readdir_getfh}, /* RFS_STATFS = 17 */ {rfs_statfs, xdr_fhandle, xdr_fastfhandle, sizeof (fhandle_t), xdr_statfs, xdr_faststatfs, sizeof (struct nfsstatfs), nullfree, RPC_IDEMPOTENT|RPC_ALLOWANON|RPC_MAPRESP, rfs_statfs_getfh}, }; static char *rfscallnames_v3[] = { "RFS3_NULL", "RFS3_GETATTR", "RFS3_SETATTR", "RFS3_LOOKUP", "RFS3_ACCESS", "RFS3_READLINK", "RFS3_READ", "RFS3_WRITE", "RFS3_CREATE", "RFS3_MKDIR", "RFS3_SYMLINK", "RFS3_MKNOD", "RFS3_REMOVE", "RFS3_RMDIR", "RFS3_RENAME", "RFS3_LINK", "RFS3_READDIR", "RFS3_READDIRPLUS", "RFS3_FSSTAT", "RFS3_FSINFO", "RFS3_PATHCONF", "RFS3_COMMIT" }; static struct rpcdisp rfsdisptab_v3[] = { /* * NFS VERSION 3 */ /* RFS_NULL = 0 */ {rpc_null_v3, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* RFS3_GETATTR = 1 */ {rfs3_getattr, xdr_nfs_fh3_server, NULL_xdrproc_t, sizeof (GETATTR3args), xdr_GETATTR3res, NULL_xdrproc_t, sizeof (GETATTR3res), nullfree, (RPC_IDEMPOTENT | RPC_ALLOWANON), rfs3_getattr_getfh}, /* RFS3_SETATTR = 2 */ {rfs3_setattr, xdr_SETATTR3args, NULL_xdrproc_t, sizeof (SETATTR3args), xdr_SETATTR3res, NULL_xdrproc_t, sizeof (SETATTR3res), nullfree, 0, rfs3_setattr_getfh}, /* RFS3_LOOKUP = 3 */ {rfs3_lookup, xdr_diropargs3, NULL_xdrproc_t, sizeof (LOOKUP3args), xdr_LOOKUP3res, NULL_xdrproc_t, sizeof (LOOKUP3res), nullfree, (RPC_IDEMPOTENT | RPC_PUBLICFH_OK), rfs3_lookup_getfh}, /* RFS3_ACCESS = 4 */ {rfs3_access, xdr_ACCESS3args, NULL_xdrproc_t, sizeof (ACCESS3args), xdr_ACCESS3res, NULL_xdrproc_t, sizeof (ACCESS3res), nullfree, RPC_IDEMPOTENT, rfs3_access_getfh}, /* RFS3_READLINK = 5 */ {rfs3_readlink, xdr_nfs_fh3_server, NULL_xdrproc_t, sizeof (READLINK3args), xdr_READLINK3res, NULL_xdrproc_t, sizeof (READLINK3res), rfs3_readlink_free, RPC_IDEMPOTENT, rfs3_readlink_getfh}, /* RFS3_READ = 6 */ {rfs3_read, xdr_READ3args, NULL_xdrproc_t, sizeof (READ3args), xdr_READ3res, NULL_xdrproc_t, sizeof (READ3res), rfs3_read_free, RPC_IDEMPOTENT, rfs3_read_getfh}, /* RFS3_WRITE = 7 */ {rfs3_write, xdr_WRITE3args, NULL_xdrproc_t, sizeof (WRITE3args), xdr_WRITE3res, NULL_xdrproc_t, sizeof (WRITE3res), nullfree, 0, rfs3_write_getfh}, /* RFS3_CREATE = 8 */ {rfs3_create, xdr_CREATE3args, NULL_xdrproc_t, sizeof (CREATE3args), xdr_CREATE3res, NULL_xdrproc_t, sizeof (CREATE3res), nullfree, 0, rfs3_create_getfh}, /* RFS3_MKDIR = 9 */ {rfs3_mkdir, xdr_MKDIR3args, NULL_xdrproc_t, sizeof (MKDIR3args), xdr_MKDIR3res, NULL_xdrproc_t, sizeof (MKDIR3res), nullfree, 0, rfs3_mkdir_getfh}, /* RFS3_SYMLINK = 10 */ {rfs3_symlink, xdr_SYMLINK3args, NULL_xdrproc_t, sizeof (SYMLINK3args), xdr_SYMLINK3res, NULL_xdrproc_t, sizeof (SYMLINK3res), nullfree, 0, rfs3_symlink_getfh}, /* RFS3_MKNOD = 11 */ {rfs3_mknod, xdr_MKNOD3args, NULL_xdrproc_t, sizeof (MKNOD3args), xdr_MKNOD3res, NULL_xdrproc_t, sizeof (MKNOD3res), nullfree, 0, rfs3_mknod_getfh}, /* RFS3_REMOVE = 12 */ {rfs3_remove, xdr_diropargs3, NULL_xdrproc_t, sizeof (REMOVE3args), xdr_REMOVE3res, NULL_xdrproc_t, sizeof (REMOVE3res), nullfree, 0, rfs3_remove_getfh}, /* RFS3_RMDIR = 13 */ {rfs3_rmdir, xdr_diropargs3, NULL_xdrproc_t, sizeof (RMDIR3args), xdr_RMDIR3res, NULL_xdrproc_t, sizeof (RMDIR3res), nullfree, 0, rfs3_rmdir_getfh}, /* RFS3_RENAME = 14 */ {rfs3_rename, xdr_RENAME3args, NULL_xdrproc_t, sizeof (RENAME3args), xdr_RENAME3res, NULL_xdrproc_t, sizeof (RENAME3res), nullfree, 0, rfs3_rename_getfh}, /* RFS3_LINK = 15 */ {rfs3_link, xdr_LINK3args, NULL_xdrproc_t, sizeof (LINK3args), xdr_LINK3res, NULL_xdrproc_t, sizeof (LINK3res), nullfree, 0, rfs3_link_getfh}, /* RFS3_READDIR = 16 */ {rfs3_readdir, xdr_READDIR3args, NULL_xdrproc_t, sizeof (READDIR3args), xdr_READDIR3res, NULL_xdrproc_t, sizeof (READDIR3res), rfs3_readdir_free, RPC_IDEMPOTENT, rfs3_readdir_getfh}, /* RFS3_READDIRPLUS = 17 */ {rfs3_readdirplus, xdr_READDIRPLUS3args, NULL_xdrproc_t, sizeof (READDIRPLUS3args), xdr_READDIRPLUS3res, NULL_xdrproc_t, sizeof (READDIRPLUS3res), rfs3_readdirplus_free, RPC_AVOIDWORK, rfs3_readdirplus_getfh}, /* RFS3_FSSTAT = 18 */ {rfs3_fsstat, xdr_nfs_fh3_server, NULL_xdrproc_t, sizeof (FSSTAT3args), xdr_FSSTAT3res, NULL_xdrproc_t, sizeof (FSSTAT3res), nullfree, RPC_IDEMPOTENT, rfs3_fsstat_getfh}, /* RFS3_FSINFO = 19 */ {rfs3_fsinfo, xdr_nfs_fh3_server, NULL_xdrproc_t, sizeof (FSINFO3args), xdr_FSINFO3res, NULL_xdrproc_t, sizeof (FSINFO3res), nullfree, RPC_IDEMPOTENT|RPC_ALLOWANON, rfs3_fsinfo_getfh}, /* RFS3_PATHCONF = 20 */ {rfs3_pathconf, xdr_nfs_fh3_server, NULL_xdrproc_t, sizeof (PATHCONF3args), xdr_PATHCONF3res, NULL_xdrproc_t, sizeof (PATHCONF3res), nullfree, RPC_IDEMPOTENT, rfs3_pathconf_getfh}, /* RFS3_COMMIT = 21 */ {rfs3_commit, xdr_COMMIT3args, NULL_xdrproc_t, sizeof (COMMIT3args), xdr_COMMIT3res, NULL_xdrproc_t, sizeof (COMMIT3res), nullfree, RPC_IDEMPOTENT, rfs3_commit_getfh}, }; static char *rfscallnames_v4[] = { "RFS4_NULL", "RFS4_COMPOUND", "RFS4_NULL", "RFS4_NULL", "RFS4_NULL", "RFS4_NULL", "RFS4_NULL", "RFS4_NULL", "RFS4_CREATE" }; static struct rpcdisp rfsdisptab_v4[] = { /* * NFS VERSION 4 */ /* RFS_NULL = 0 */ {rpc_null, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* RFS4_compound = 1 */ {rfs4_compound, xdr_COMPOUND4args_srv, NULL_xdrproc_t, sizeof (COMPOUND4args), xdr_COMPOUND4res_srv, NULL_xdrproc_t, sizeof (COMPOUND4res), rfs4_compound_free, 0, 0}, }; union rfs_args { /* * NFS VERSION 2 */ /* RFS_NULL = 0 */ /* RFS_GETATTR = 1 */ fhandle_t nfs2_getattr_args; /* RFS_SETATTR = 2 */ struct nfssaargs nfs2_setattr_args; /* RFS_ROOT = 3 *** NO LONGER SUPPORTED *** */ /* RFS_LOOKUP = 4 */ struct nfsdiropargs nfs2_lookup_args; /* RFS_READLINK = 5 */ fhandle_t nfs2_readlink_args; /* RFS_READ = 6 */ struct nfsreadargs nfs2_read_args; /* RFS_WRITECACHE = 7 *** NO LONGER SUPPORTED *** */ /* RFS_WRITE = 8 */ struct nfswriteargs nfs2_write_args; /* RFS_CREATE = 9 */ struct nfscreatargs nfs2_create_args; /* RFS_REMOVE = 10 */ struct nfsdiropargs nfs2_remove_args; /* RFS_RENAME = 11 */ struct nfsrnmargs nfs2_rename_args; /* RFS_LINK = 12 */ struct nfslinkargs nfs2_link_args; /* RFS_SYMLINK = 13 */ struct nfsslargs nfs2_symlink_args; /* RFS_MKDIR = 14 */ struct nfscreatargs nfs2_mkdir_args; /* RFS_RMDIR = 15 */ struct nfsdiropargs nfs2_rmdir_args; /* RFS_READDIR = 16 */ struct nfsrddirargs nfs2_readdir_args; /* RFS_STATFS = 17 */ fhandle_t nfs2_statfs_args; /* * NFS VERSION 3 */ /* RFS_NULL = 0 */ /* RFS3_GETATTR = 1 */ GETATTR3args nfs3_getattr_args; /* RFS3_SETATTR = 2 */ SETATTR3args nfs3_setattr_args; /* RFS3_LOOKUP = 3 */ LOOKUP3args nfs3_lookup_args; /* RFS3_ACCESS = 4 */ ACCESS3args nfs3_access_args; /* RFS3_READLINK = 5 */ READLINK3args nfs3_readlink_args; /* RFS3_READ = 6 */ READ3args nfs3_read_args; /* RFS3_WRITE = 7 */ WRITE3args nfs3_write_args; /* RFS3_CREATE = 8 */ CREATE3args nfs3_create_args; /* RFS3_MKDIR = 9 */ MKDIR3args nfs3_mkdir_args; /* RFS3_SYMLINK = 10 */ SYMLINK3args nfs3_symlink_args; /* RFS3_MKNOD = 11 */ MKNOD3args nfs3_mknod_args; /* RFS3_REMOVE = 12 */ REMOVE3args nfs3_remove_args; /* RFS3_RMDIR = 13 */ RMDIR3args nfs3_rmdir_args; /* RFS3_RENAME = 14 */ RENAME3args nfs3_rename_args; /* RFS3_LINK = 15 */ LINK3args nfs3_link_args; /* RFS3_READDIR = 16 */ READDIR3args nfs3_readdir_args; /* RFS3_READDIRPLUS = 17 */ READDIRPLUS3args nfs3_readdirplus_args; /* RFS3_FSSTAT = 18 */ FSSTAT3args nfs3_fsstat_args; /* RFS3_FSINFO = 19 */ FSINFO3args nfs3_fsinfo_args; /* RFS3_PATHCONF = 20 */ PATHCONF3args nfs3_pathconf_args; /* RFS3_COMMIT = 21 */ COMMIT3args nfs3_commit_args; /* * NFS VERSION 4 */ /* RFS_NULL = 0 */ /* COMPUND = 1 */ COMPOUND4args nfs4_compound_args; }; union rfs_res { /* * NFS VERSION 2 */ /* RFS_NULL = 0 */ /* RFS_GETATTR = 1 */ struct nfsattrstat nfs2_getattr_res; /* RFS_SETATTR = 2 */ struct nfsattrstat nfs2_setattr_res; /* RFS_ROOT = 3 *** NO LONGER SUPPORTED *** */ /* RFS_LOOKUP = 4 */ struct nfsdiropres nfs2_lookup_res; /* RFS_READLINK = 5 */ struct nfsrdlnres nfs2_readlink_res; /* RFS_READ = 6 */ struct nfsrdresult nfs2_read_res; /* RFS_WRITECACHE = 7 *** NO LONGER SUPPORTED *** */ /* RFS_WRITE = 8 */ struct nfsattrstat nfs2_write_res; /* RFS_CREATE = 9 */ struct nfsdiropres nfs2_create_res; /* RFS_REMOVE = 10 */ enum nfsstat nfs2_remove_res; /* RFS_RENAME = 11 */ enum nfsstat nfs2_rename_res; /* RFS_LINK = 12 */ enum nfsstat nfs2_link_res; /* RFS_SYMLINK = 13 */ enum nfsstat nfs2_symlink_res; /* RFS_MKDIR = 14 */ struct nfsdiropres nfs2_mkdir_res; /* RFS_RMDIR = 15 */ enum nfsstat nfs2_rmdir_res; /* RFS_READDIR = 16 */ struct nfsrddirres nfs2_readdir_res; /* RFS_STATFS = 17 */ struct nfsstatfs nfs2_statfs_res; /* * NFS VERSION 3 */ /* RFS_NULL = 0 */ /* RFS3_GETATTR = 1 */ GETATTR3res nfs3_getattr_res; /* RFS3_SETATTR = 2 */ SETATTR3res nfs3_setattr_res; /* RFS3_LOOKUP = 3 */ LOOKUP3res nfs3_lookup_res; /* RFS3_ACCESS = 4 */ ACCESS3res nfs3_access_res; /* RFS3_READLINK = 5 */ READLINK3res nfs3_readlink_res; /* RFS3_READ = 6 */ READ3res nfs3_read_res; /* RFS3_WRITE = 7 */ WRITE3res nfs3_write_res; /* RFS3_CREATE = 8 */ CREATE3res nfs3_create_res; /* RFS3_MKDIR = 9 */ MKDIR3res nfs3_mkdir_res; /* RFS3_SYMLINK = 10 */ SYMLINK3res nfs3_symlink_res; /* RFS3_MKNOD = 11 */ MKNOD3res nfs3_mknod_res; /* RFS3_REMOVE = 12 */ REMOVE3res nfs3_remove_res; /* RFS3_RMDIR = 13 */ RMDIR3res nfs3_rmdir_res; /* RFS3_RENAME = 14 */ RENAME3res nfs3_rename_res; /* RFS3_LINK = 15 */ LINK3res nfs3_link_res; /* RFS3_READDIR = 16 */ READDIR3res nfs3_readdir_res; /* RFS3_READDIRPLUS = 17 */ READDIRPLUS3res nfs3_readdirplus_res; /* RFS3_FSSTAT = 18 */ FSSTAT3res nfs3_fsstat_res; /* RFS3_FSINFO = 19 */ FSINFO3res nfs3_fsinfo_res; /* RFS3_PATHCONF = 20 */ PATHCONF3res nfs3_pathconf_res; /* RFS3_COMMIT = 21 */ COMMIT3res nfs3_commit_res; /* * NFS VERSION 4 */ /* RFS_NULL = 0 */ /* RFS4_COMPOUND = 1 */ COMPOUND4res nfs4_compound_res; }; static struct rpc_disptable rfs_disptable[] = { {sizeof (rfsdisptab_v2) / sizeof (rfsdisptab_v2[0]), rfscallnames_v2, &rfsproccnt_v2_ptr, rfsdisptab_v2}, {sizeof (rfsdisptab_v3) / sizeof (rfsdisptab_v3[0]), rfscallnames_v3, &rfsproccnt_v3_ptr, rfsdisptab_v3}, {sizeof (rfsdisptab_v4) / sizeof (rfsdisptab_v4[0]), rfscallnames_v4, &rfsproccnt_v4_ptr, rfsdisptab_v4}, }; /* * If nfs_portmon is set, then clients are required to use privileged * ports (ports < IPPORT_RESERVED) in order to get NFS services. * * N.B.: this attempt to carry forward the already ill-conceived notion * of privileged ports for TCP/UDP is really quite ineffectual. Not only * is it transport-dependent, it's laughably easy to spoof. If you're * really interested in security, you must start with secure RPC instead. */ static int nfs_portmon = 0; #ifdef DEBUG static int cred_hits = 0; static int cred_misses = 0; #endif #ifdef DEBUG /* * Debug code to allow disabling of rfs_dispatch() use of * fastxdrargs() and fastxdrres() calls for testing purposes. */ static int rfs_no_fast_xdrargs = 0; static int rfs_no_fast_xdrres = 0; #endif union acl_args { /* * ACL VERSION 2 */ /* ACL2_NULL = 0 */ /* ACL2_GETACL = 1 */ GETACL2args acl2_getacl_args; /* ACL2_SETACL = 2 */ SETACL2args acl2_setacl_args; /* ACL2_GETATTR = 3 */ GETATTR2args acl2_getattr_args; /* ACL2_ACCESS = 4 */ ACCESS2args acl2_access_args; /* ACL2_GETXATTRDIR = 5 */ GETXATTRDIR2args acl2_getxattrdir_args; /* * ACL VERSION 3 */ /* ACL3_NULL = 0 */ /* ACL3_GETACL = 1 */ GETACL3args acl3_getacl_args; /* ACL3_SETACL = 2 */ SETACL3args acl3_setacl; /* ACL3_GETXATTRDIR = 3 */ GETXATTRDIR3args acl3_getxattrdir_args; }; union acl_res { /* * ACL VERSION 2 */ /* ACL2_NULL = 0 */ /* ACL2_GETACL = 1 */ GETACL2res acl2_getacl_res; /* ACL2_SETACL = 2 */ SETACL2res acl2_setacl_res; /* ACL2_GETATTR = 3 */ GETATTR2res acl2_getattr_res; /* ACL2_ACCESS = 4 */ ACCESS2res acl2_access_res; /* ACL2_GETXATTRDIR = 5 */ GETXATTRDIR2args acl2_getxattrdir_res; /* * ACL VERSION 3 */ /* ACL3_NULL = 0 */ /* ACL3_GETACL = 1 */ GETACL3res acl3_getacl_res; /* ACL3_SETACL = 2 */ SETACL3res acl3_setacl_res; /* ACL3_GETXATTRDIR = 3 */ GETXATTRDIR3res acl3_getxattrdir_res; }; static bool_t auth_tooweak(struct svc_req *req, char *res) { if (req->rq_vers == NFS_VERSION && req->rq_proc == RFS_LOOKUP) { struct nfsdiropres *dr = (struct nfsdiropres *)res; if (dr->dr_status == WNFSERR_CLNT_FLAVOR) return (TRUE); } else if (req->rq_vers == NFS_V3 && req->rq_proc == NFSPROC3_LOOKUP) { LOOKUP3res *resp = (LOOKUP3res *)res; if (resp->status == WNFSERR_CLNT_FLAVOR) return (TRUE); } return (FALSE); } static void common_dispatch(struct svc_req *req, SVCXPRT *xprt, rpcvers_t min_vers, rpcvers_t max_vers, char *pgmname, struct rpc_disptable *disptable) { int which; rpcvers_t vers; char *args; union { union rfs_args ra; union acl_args aa; } args_buf; char *res; union { union rfs_res rr; union acl_res ar; } res_buf; struct rpcdisp *disp = NULL; int dis_flags = 0; cred_t *cr; int error = 0; int anon_ok; struct exportinfo *exi = NULL; unsigned int nfslog_rec_id; int dupstat; struct dupreq *dr; int authres; bool_t publicfh_ok = FALSE; enum_t auth_flavor; bool_t dupcached = FALSE; struct netbuf nb; bool_t logging_enabled = FALSE; struct exportinfo *nfslog_exi = NULL; char **procnames; char cbuf[INET6_ADDRSTRLEN]; /* to hold both IPv4 and IPv6 addr */ vers = req->rq_vers; if (vers < min_vers || vers > max_vers) { svcerr_progvers(req->rq_xprt, min_vers, max_vers); error++; cmn_err(CE_NOTE, "%s: bad version number %u", pgmname, vers); goto done; } vers -= min_vers; which = req->rq_proc; if (which < 0 || which >= disptable[(int)vers].dis_nprocs) { svcerr_noproc(req->rq_xprt); error++; goto done; } (*(disptable[(int)vers].dis_proccntp))[which].value.ui64++; disp = &disptable[(int)vers].dis_table[which]; procnames = disptable[(int)vers].dis_procnames; auth_flavor = req->rq_cred.oa_flavor; /* * Deserialize into the args struct. */ args = (char *)&args_buf; #ifdef DEBUG if (rfs_no_fast_xdrargs || (auth_flavor == RPCSEC_GSS) || disp->dis_fastxdrargs == NULL_xdrproc_t || !SVC_GETARGS(xprt, disp->dis_fastxdrargs, (char *)&args)) #else if ((auth_flavor == RPCSEC_GSS) || disp->dis_fastxdrargs == NULL_xdrproc_t || !SVC_GETARGS(xprt, disp->dis_fastxdrargs, (char *)&args)) #endif { bzero(args, disp->dis_argsz); if (!SVC_GETARGS(xprt, disp->dis_xdrargs, args)) { error++; /* * Check if we are outside our capabilities. */ if (rfs4_minorvers_mismatch(req, xprt, (void *)args)) goto done; svcerr_decode(xprt); cmn_err(CE_NOTE, "Failed to decode arguments for %s version %u " "procedure %s client %s%s", pgmname, vers + min_vers, procnames[which], client_name(req), client_addr(req, cbuf)); goto done; } } /* * If Version 4 use that specific dispatch function. */ if (req->rq_vers == 4) { error += rfs4_dispatch(disp, req, xprt, args); goto done; } dis_flags = disp->dis_flags; /* * Find export information and check authentication, * setting the credential if everything is ok. */ if (disp->dis_getfh != NULL) { void *fh; fsid_t *fsid; fid_t *fid, *xfid; fhandle_t *fh2; nfs_fh3 *fh3; fh = (*disp->dis_getfh)(args); switch (req->rq_vers) { case NFS_VERSION: fh2 = (fhandle_t *)fh; fsid = &fh2->fh_fsid; fid = (fid_t *)&fh2->fh_len; xfid = (fid_t *)&fh2->fh_xlen; break; case NFS_V3: fh3 = (nfs_fh3 *)fh; fsid = &fh3->fh3_fsid; fid = FH3TOFIDP(fh3); xfid = FH3TOXFIDP(fh3); break; } /* * Fix for bug 1038302 - corbin * There is a problem here if anonymous access is * disallowed. If the current request is part of the * client's mount process for the requested filesystem, * then it will carry root (uid 0) credentials on it, and * will be denied by checkauth if that client does not * have explicit root=0 permission. This will cause the * client's mount operation to fail. As a work-around, * we check here to see if the request is a getattr or * statfs operation on the exported vnode itself, and * pass a flag to checkauth with the result of this test. * * The filehandle refers to the mountpoint itself if * the fh_data and fh_xdata portions of the filehandle * are equal. * * Added anon_ok argument to checkauth(). */ if ((dis_flags & RPC_ALLOWANON) && EQFID(fid, xfid)) anon_ok = 1; else anon_ok = 0; cr = xprt->xp_cred; ASSERT(cr != NULL); #ifdef DEBUG if (crgetref(cr) != 1) { crfree(cr); cr = crget(); xprt->xp_cred = cr; cred_misses++; } else cred_hits++; #else if (crgetref(cr) != 1) { crfree(cr); cr = crget(); xprt->xp_cred = cr; } #endif exi = checkexport(fsid, xfid); if (exi != NULL) { publicfh_ok = PUBLICFH_CHECK(disp, exi, fsid, xfid); /* * Don't allow non-V4 clients access * to pseudo exports */ if (PSEUDO(exi)) { svcerr_weakauth(xprt); error++; goto done; } authres = checkauth(exi, req, cr, anon_ok, publicfh_ok); /* * authres > 0: authentication OK - proceed * authres == 0: authentication weak - return error * authres < 0: authentication timeout - drop */ if (authres <= 0) { if (authres == 0) { svcerr_weakauth(xprt); error++; } goto done; } } } else cr = NULL; if ((dis_flags & RPC_MAPRESP) && (auth_flavor != RPCSEC_GSS)) { res = (char *)SVC_GETRES(xprt, disp->dis_ressz); if (res == NULL) res = (char *)&res_buf; } else res = (char *)&res_buf; if (!(dis_flags & RPC_IDEMPOTENT)) { dupstat = SVC_DUP_EXT(xprt, req, res, disp->dis_ressz, &dr, &dupcached); switch (dupstat) { case DUP_ERROR: svcerr_systemerr(xprt); error++; goto done; /* NOTREACHED */ case DUP_INPROGRESS: if (res != (char *)&res_buf) SVC_FREERES(xprt); error++; goto done; /* NOTREACHED */ case DUP_NEW: case DUP_DROP: curthread->t_flag |= T_DONTPEND; (*disp->dis_proc)(args, res, exi, req, cr); curthread->t_flag &= ~T_DONTPEND; if (curthread->t_flag & T_WOULDBLOCK) { curthread->t_flag &= ~T_WOULDBLOCK; SVC_DUPDONE_EXT(xprt, dr, res, NULL, disp->dis_ressz, DUP_DROP); if (res != (char *)&res_buf) SVC_FREERES(xprt); error++; goto done; } if (dis_flags & RPC_AVOIDWORK) { SVC_DUPDONE_EXT(xprt, dr, res, NULL, disp->dis_ressz, DUP_DROP); } else { SVC_DUPDONE_EXT(xprt, dr, res, disp->dis_resfree == nullfree ? NULL : disp->dis_resfree, disp->dis_ressz, DUP_DONE); dupcached = TRUE; } break; case DUP_DONE: break; } } else { curthread->t_flag |= T_DONTPEND; (*disp->dis_proc)(args, res, exi, req, cr); curthread->t_flag &= ~T_DONTPEND; if (curthread->t_flag & T_WOULDBLOCK) { curthread->t_flag &= ~T_WOULDBLOCK; if (res != (char *)&res_buf) SVC_FREERES(xprt); error++; goto done; } } if (auth_tooweak(req, res)) { svcerr_weakauth(xprt); error++; goto done; } /* * Check to see if logging has been enabled on the server. * If so, then obtain the export info struct to be used for * the later writing of the log record. This is done for * the case that a lookup is done across a non-logged public * file system. */ if (nfslog_buffer_list != NULL) { nfslog_exi = nfslog_get_exi(exi, req, res, &nfslog_rec_id); /* * Is logging enabled? */ logging_enabled = (nfslog_exi != NULL); /* * Copy the netbuf for logging purposes, before it is * freed by svc_sendreply(). */ if (logging_enabled) { NFSLOG_COPY_NETBUF(nfslog_exi, xprt, &nb); /* * If RPC_MAPRESP flag set (i.e. in V2 ops) the * res gets copied directly into the mbuf and * may be freed soon after the sendreply. So we * must copy it here to a safe place... */ if (res != (char *)&res_buf) { bcopy(res, (char *)&res_buf, disp->dis_ressz); } } } /* * Serialize and send results struct */ #ifdef DEBUG if (rfs_no_fast_xdrres == 0 && res != (char *)&res_buf) #else if (res != (char *)&res_buf) #endif { if (!svc_sendreply(xprt, disp->dis_fastxdrres, res)) { cmn_err(CE_NOTE, "%s: bad sendreply", pgmname); error++; } } else { if (!svc_sendreply(xprt, disp->dis_xdrres, res)) { cmn_err(CE_NOTE, "%s: bad sendreply", pgmname); error++; } } /* * Log if needed */ if (logging_enabled) { nfslog_write_record(nfslog_exi, req, args, (char *)&res_buf, cr, &nb, nfslog_rec_id, NFSLOG_ONE_BUFFER); exi_rele(nfslog_exi); kmem_free((&nb)->buf, (&nb)->len); } /* * Free results struct. With the addition of NFS V4 we can * have non-idempotent procedures with functions. */ if (disp->dis_resfree != nullfree && dupcached == FALSE) { (*disp->dis_resfree)(res); } done: /* * Free arguments struct */ if (disp) { if (!SVC_FREEARGS(xprt, disp->dis_xdrargs, args)) { cmn_err(CE_NOTE, "%s: bad freeargs", pgmname); error++; } } else { if (!SVC_FREEARGS(xprt, (xdrproc_t)0, (caddr_t)0)) { cmn_err(CE_NOTE, "%s: bad freeargs", pgmname); error++; } } if (exi != NULL) exi_rele(exi); global_svstat_ptr[req->rq_vers][NFS_BADCALLS].value.ui64 += error; global_svstat_ptr[req->rq_vers][NFS_CALLS].value.ui64++; } static void rfs_dispatch(struct svc_req *req, SVCXPRT *xprt) { common_dispatch(req, xprt, NFS_VERSMIN, NFS_VERSMAX, "NFS", rfs_disptable); } static char *aclcallnames_v2[] = { "ACL2_NULL", "ACL2_GETACL", "ACL2_SETACL", "ACL2_GETATTR", "ACL2_ACCESS", "ACL2_GETXATTRDIR" }; static struct rpcdisp acldisptab_v2[] = { /* * ACL VERSION 2 */ /* ACL2_NULL = 0 */ {rpc_null, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* ACL2_GETACL = 1 */ {acl2_getacl, xdr_GETACL2args, xdr_fastGETACL2args, sizeof (GETACL2args), xdr_GETACL2res, NULL_xdrproc_t, sizeof (GETACL2res), acl2_getacl_free, RPC_IDEMPOTENT, acl2_getacl_getfh}, /* ACL2_SETACL = 2 */ {acl2_setacl, xdr_SETACL2args, NULL_xdrproc_t, sizeof (SETACL2args), #ifdef _LITTLE_ENDIAN xdr_SETACL2res, xdr_fastSETACL2res, sizeof (SETACL2res), #else xdr_SETACL2res, NULL_xdrproc_t, sizeof (SETACL2res), #endif nullfree, RPC_MAPRESP, acl2_setacl_getfh}, /* ACL2_GETATTR = 3 */ {acl2_getattr, xdr_GETATTR2args, xdr_fastGETATTR2args, sizeof (GETATTR2args), #ifdef _LITTLE_ENDIAN xdr_GETATTR2res, xdr_fastGETATTR2res, sizeof (GETATTR2res), #else xdr_GETATTR2res, NULL_xdrproc_t, sizeof (GETATTR2res), #endif nullfree, RPC_IDEMPOTENT|RPC_ALLOWANON|RPC_MAPRESP, acl2_getattr_getfh}, /* ACL2_ACCESS = 4 */ {acl2_access, xdr_ACCESS2args, xdr_fastACCESS2args, sizeof (ACCESS2args), #ifdef _LITTLE_ENDIAN xdr_ACCESS2res, xdr_fastACCESS2res, sizeof (ACCESS2res), #else xdr_ACCESS2res, NULL_xdrproc_t, sizeof (ACCESS2res), #endif nullfree, RPC_IDEMPOTENT|RPC_MAPRESP, acl2_access_getfh}, /* ACL2_GETXATTRDIR = 5 */ {acl2_getxattrdir, xdr_GETXATTRDIR2args, NULL_xdrproc_t, sizeof (GETXATTRDIR2args), xdr_GETXATTRDIR2res, NULL_xdrproc_t, sizeof (GETXATTRDIR2res), nullfree, RPC_IDEMPOTENT, acl2_getxattrdir_getfh}, }; static char *aclcallnames_v3[] = { "ACL3_NULL", "ACL3_GETACL", "ACL3_SETACL", "ACL3_GETXATTRDIR" }; static struct rpcdisp acldisptab_v3[] = { /* * ACL VERSION 3 */ /* ACL3_NULL = 0 */ {rpc_null, xdr_void, NULL_xdrproc_t, 0, xdr_void, NULL_xdrproc_t, 0, nullfree, RPC_IDEMPOTENT, 0}, /* ACL3_GETACL = 1 */ {acl3_getacl, xdr_GETACL3args, NULL_xdrproc_t, sizeof (GETACL3args), xdr_GETACL3res, NULL_xdrproc_t, sizeof (GETACL3res), acl3_getacl_free, RPC_IDEMPOTENT, acl3_getacl_getfh}, /* ACL3_SETACL = 2 */ {acl3_setacl, xdr_SETACL3args, NULL_xdrproc_t, sizeof (SETACL3args), xdr_SETACL3res, NULL_xdrproc_t, sizeof (SETACL3res), nullfree, 0, acl3_setacl_getfh}, /* ACL3_GETXATTRDIR = 3 */ {acl3_getxattrdir, xdr_GETXATTRDIR3args, NULL_xdrproc_t, sizeof (GETXATTRDIR3args), xdr_GETXATTRDIR3res, NULL_xdrproc_t, sizeof (GETXATTRDIR3res), nullfree, RPC_IDEMPOTENT, acl3_getxattrdir_getfh}, }; static struct rpc_disptable acl_disptable[] = { {sizeof (acldisptab_v2) / sizeof (acldisptab_v2[0]), aclcallnames_v2, &aclproccnt_v2_ptr, acldisptab_v2}, {sizeof (acldisptab_v3) / sizeof (acldisptab_v3[0]), aclcallnames_v3, &aclproccnt_v3_ptr, acldisptab_v3}, }; static void acl_dispatch(struct svc_req *req, SVCXPRT *xprt) { common_dispatch(req, xprt, NFS_ACL_VERSMIN, NFS_ACL_VERSMAX, "ACL", acl_disptable); } int checkwin(int flavor, int window, struct svc_req *req) { struct authdes_cred *adc; switch (flavor) { case AUTH_DES: adc = (struct authdes_cred *)req->rq_clntcred; if (adc->adc_fullname.window > window) return (0); break; default: break; } return (1); } /* * checkauth() will check the access permission against the export * information. Then map root uid/gid to appropriate uid/gid. * * This routine is used by NFS V3 and V2 code. */ static int checkauth(struct exportinfo *exi, struct svc_req *req, cred_t *cr, int anon_ok, bool_t publicfh_ok) { int i, nfsflavor, rpcflavor, stat, access; struct secinfo *secp; caddr_t principal; char buf[INET6_ADDRSTRLEN]; /* to hold both IPv4 and IPv6 addr */ int anon_res = 0; /* * Check for privileged port number * N.B.: this assumes that we know the format of a netbuf. */ if (nfs_portmon) { struct sockaddr *ca; ca = (struct sockaddr *)svc_getrpccaller(req->rq_xprt)->buf; if (ca == NULL) return (0); if ((ca->sa_family == AF_INET && ntohs(((struct sockaddr_in *)ca)->sin_port) >= IPPORT_RESERVED) || (ca->sa_family == AF_INET6 && ntohs(((struct sockaddr_in6 *)ca)->sin6_port) >= IPPORT_RESERVED)) { cmn_err(CE_NOTE, "nfs_server: client %s%ssent NFS request from " "unprivileged port", client_name(req), client_addr(req, buf)); return (0); } } /* * return 1 on success or 0 on failure */ stat = sec_svc_getcred(req, cr, &principal, &nfsflavor); /* * A failed AUTH_UNIX svc_get_cred() implies we couldn't set * the credentials; below we map that to anonymous. */ if (!stat && nfsflavor != AUTH_UNIX) { cmn_err(CE_NOTE, "nfs_server: couldn't get unix cred for %s", client_name(req)); return (0); } /* * Short circuit checkauth() on operations that support the * public filehandle, and if the request for that operation * is using the public filehandle. Note that we must call * sec_svc_getcred() first so that xp_cookie is set to the * right value. Normally xp_cookie is just the RPC flavor * of the the request, but in the case of RPCSEC_GSS it * could be a pseudo flavor. */ if (publicfh_ok) return (1); rpcflavor = req->rq_cred.oa_flavor; /* * Check if the auth flavor is valid for this export */ access = nfsauth_access(exi, req); if (access & NFSAUTH_DROP) return (-1); /* drop the request */ if (access & NFSAUTH_DENIED) { /* * If anon_ok == 1 and we got NFSAUTH_DENIED, it was * probably due to the flavor not matching during the * the mount attempt. So map the flavor to AUTH_NONE * so that the credentials get mapped to the anonymous * user. */ if (anon_ok == 1) rpcflavor = AUTH_NONE; else return (0); /* deny access */ } else if (access & NFSAUTH_MAPNONE) { /* * Access was granted even though the flavor mismatched * because AUTH_NONE was one of the exported flavors. */ rpcflavor = AUTH_NONE; } else if (access & NFSAUTH_WRONGSEC) { /* * NFSAUTH_WRONGSEC is used for NFSv4. Since V2/V3 already * negotiates the security flavor thru MOUNT protocol, the * only way it can get NFSAUTH_WRONGSEC here is from * NFS_ACL for V4. This could be for a limited view, so * map it to RO access. V4 lookup/readdir will take care * of the limited view portion. */ access |= NFSAUTH_RO; access &= ~NFSAUTH_WRONGSEC; } switch (rpcflavor) { case AUTH_NONE: anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); break; case AUTH_UNIX: if (!stat || crgetuid(cr) == 0 && !(access & NFSAUTH_ROOT)) { anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); } break; case AUTH_DES: case RPCSEC_GSS: /* * Find the secinfo structure. We should be able * to find it by the time we reach here. * nfsauth_access() has done the checking. */ secp = NULL; for (i = 0; i < exi->exi_export.ex_seccnt; i++) { if (exi->exi_export.ex_secinfo[i].s_secinfo.sc_nfsnum == nfsflavor) { secp = &exi->exi_export.ex_secinfo[i]; break; } } if (!secp) { cmn_err(CE_NOTE, "nfs_server: client %s%shad " "no secinfo data for flavor %d", client_name(req), client_addr(req, buf), nfsflavor); return (0); } if (!checkwin(rpcflavor, secp->s_window, req)) { cmn_err(CE_NOTE, "nfs_server: client %s%sused invalid " "auth window value", client_name(req), client_addr(req, buf)); return (0); } /* * Map root principals listed in the share's root= list to root, * and map any others principals that were mapped to root by RPC * to anon. */ if (principal && sec_svc_inrootlist(rpcflavor, principal, secp->s_rootcnt, secp->s_rootnames)) { if (crgetuid(cr) == 0) return (1); (void) crsetugid(cr, 0, 0); /* * NOTE: If and when kernel-land privilege tracing is * added this may have to be replaced with code that * retrieves root's supplementary groups (e.g., using * kgss_get_group_info(). In the meantime principals * mapped to uid 0 get all privileges, so setting cr's * supplementary groups for them does nothing. */ (void) crsetgroups(cr, 0, NULL); return (1); } /* * Not a root princ, or not in root list, map UID 0/nobody to * the anon ID for the share. (RPC sets cr's UIDs and GIDs to * UID_NOBODY and GID_NOBODY, respectively.) */ if (crgetuid(cr) != 0 && (crgetuid(cr) != UID_NOBODY || crgetgid(cr) != GID_NOBODY)) return (1); anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); break; default: return (0); } /* switch on rpcflavor */ /* * Even if anon access is disallowed via ex_anon == -1, we allow * this access if anon_ok is set. So set creds to the default * "nobody" id. */ if (anon_res != 0) { if (anon_ok == 0) { cmn_err(CE_NOTE, "nfs_server: client %s%ssent wrong " "authentication for %s", client_name(req), client_addr(req, buf), exi->exi_export.ex_path ? exi->exi_export.ex_path : "?"); return (0); } if (crsetugid(cr, UID_NOBODY, GID_NOBODY) != 0) return (0); } return (1); } /* * returns 0 on failure, -1 on a drop, -2 on wrong security flavor, * and 1 on success */ int checkauth4(struct compound_state *cs, struct svc_req *req) { int i, rpcflavor, access; struct secinfo *secp; char buf[MAXHOST + 1]; int anon_res = 0, nfsflavor; struct exportinfo *exi; cred_t *cr; caddr_t principal; exi = cs->exi; cr = cs->cr; principal = cs->principal; nfsflavor = cs->nfsflavor; ASSERT(cr != NULL); rpcflavor = req->rq_cred.oa_flavor; cs->access &= ~CS_ACCESS_LIMITED; /* * Check the access right per auth flavor on the vnode of * this export for the given request. */ access = nfsauth4_access(cs->exi, cs->vp, req); if (access & NFSAUTH_WRONGSEC) return (-2); /* no access for this security flavor */ if (access & NFSAUTH_DROP) return (-1); /* drop the request */ if (access & NFSAUTH_DENIED) { if (exi->exi_export.ex_seccnt > 0) return (0); /* deny access */ } else if (access & NFSAUTH_LIMITED) { cs->access |= CS_ACCESS_LIMITED; } else if (access & NFSAUTH_MAPNONE) { /* * Access was granted even though the flavor mismatched * because AUTH_NONE was one of the exported flavors. */ rpcflavor = AUTH_NONE; } /* * XXX probably need to redo some of it for nfsv4? * return 1 on success or 0 on failure */ switch (rpcflavor) { case AUTH_NONE: anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); break; case AUTH_UNIX: if (crgetuid(cr) == 0 && !(access & NFSAUTH_ROOT)) { anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); } break; default: /* * Find the secinfo structure. We should be able * to find it by the time we reach here. * nfsauth_access() has done the checking. */ secp = NULL; for (i = 0; i < exi->exi_export.ex_seccnt; i++) { if (exi->exi_export.ex_secinfo[i].s_secinfo.sc_nfsnum == nfsflavor) { secp = &exi->exi_export.ex_secinfo[i]; break; } } if (!secp) { cmn_err(CE_NOTE, "nfs_server: client %s%shad " "no secinfo data for flavor %d", client_name(req), client_addr(req, buf), nfsflavor); return (0); } if (!checkwin(rpcflavor, secp->s_window, req)) { cmn_err(CE_NOTE, "nfs_server: client %s%sused invalid " "auth window value", client_name(req), client_addr(req, buf)); return (0); } /* * Map root principals listed in the share's root= list to root, * and map any others principals that were mapped to root by RPC * to anon. */ if (principal && sec_svc_inrootlist(rpcflavor, principal, secp->s_rootcnt, secp->s_rootnames)) { if (crgetuid(cr) == 0) return (1); (void) crsetugid(cr, 0, 0); /* * NOTE: If and when kernel-land privilege tracing is * added this may have to be replaced with code that * retrieves root's supplementary groups (e.g., using * kgss_get_group_info(). In the meantime principals * mapped to uid 0 get all privileges, so setting cr's * supplementary groups for them does nothing. */ (void) crsetgroups(cr, 0, NULL); return (1); } /* * Not a root princ, or not in root list, map UID 0/nobody to * the anon ID for the share. (RPC sets cr's UIDs and GIDs to * UID_NOBODY and GID_NOBODY, respectively.) */ if (crgetuid(cr) != 0 && (crgetuid(cr) != UID_NOBODY || crgetgid(cr) != GID_NOBODY)) return (1); anon_res = crsetugid(cr, exi->exi_export.ex_anon, exi->exi_export.ex_anon); (void) crsetgroups(cr, 0, NULL); break; } /* switch on rpcflavor */ /* * Even if anon access is disallowed via ex_anon == -1, we allow * this access if anon_ok is set. So set creds to the default * "nobody" id. */ if (anon_res != 0) { cmn_err(CE_NOTE, "nfs_server: client %s%ssent wrong " "authentication for %s", client_name(req), client_addr(req, buf), exi->exi_export.ex_path ? exi->exi_export.ex_path : "?"); return (0); } return (1); } static char * client_name(struct svc_req *req) { char *hostname = NULL; /* * If it's a Unix cred then use the * hostname from the credential. */ if (req->rq_cred.oa_flavor == AUTH_UNIX) { hostname = ((struct authunix_parms *) req->rq_clntcred)->aup_machname; } if (hostname == NULL) hostname = ""; return (hostname); } static char * client_addr(struct svc_req *req, char *buf) { struct sockaddr *ca; uchar_t *b; char *frontspace = ""; /* * We assume we are called in tandem with client_name and the * format string looks like "...client %s%sblah blah..." * * If it's a Unix cred then client_name returned * a host name, so we need insert a space between host name * and IP address. */ if (req->rq_cred.oa_flavor == AUTH_UNIX) frontspace = " "; /* * Convert the caller's IP address to a dotted string */ ca = (struct sockaddr *)svc_getrpccaller(req->rq_xprt)->buf; if (ca->sa_family == AF_INET) { b = (uchar_t *)&((struct sockaddr_in *)ca)->sin_addr; (void) sprintf(buf, "%s(%d.%d.%d.%d) ", frontspace, b[0] & 0xFF, b[1] & 0xFF, b[2] & 0xFF, b[3] & 0xFF); } else if (ca->sa_family == AF_INET6) { struct sockaddr_in6 *sin6; sin6 = (struct sockaddr_in6 *)ca; (void) kinet_ntop6((uchar_t *)&sin6->sin6_addr, buf, INET6_ADDRSTRLEN); } else { /* * No IP address to print. If there was a host name * printed, then we print a space. */ (void) sprintf(buf, frontspace); } return (buf); } /* * NFS Server initialization routine. This routine should only be called * once. It performs the following tasks: * - Call sub-initialization routines (localize access to variables) * - Initialize all locks * - initialize the version 3 write verifier */ int nfs_srvinit(void) { int error; error = nfs_exportinit(); if (error != 0) return (error); error = rfs4_srvrinit(); if (error != 0) { nfs_exportfini(); return (error); } rfs_srvrinit(); rfs3_srvrinit(); nfsauth_init(); /* Init the stuff to control start/stop */ nfs_server_upordown = NFS_SERVER_STOPPED; mutex_init(&nfs_server_upordown_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&nfs_server_upordown_cv, NULL, CV_DEFAULT, NULL); mutex_init(&rdma_wait_mutex, NULL, MUTEX_DEFAULT, NULL); cv_init(&rdma_wait_cv, NULL, CV_DEFAULT, NULL); return (0); } /* * NFS Server finalization routine. This routine is called to cleanup the * initialization work previously performed if the NFS server module could * not be loaded correctly. */ void nfs_srvfini(void) { nfsauth_fini(); rfs3_srvrfini(); rfs_srvrfini(); nfs_exportfini(); mutex_destroy(&nfs_server_upordown_lock); cv_destroy(&nfs_server_upordown_cv); mutex_destroy(&rdma_wait_mutex); cv_destroy(&rdma_wait_cv); } /* * Set up an iovec array of up to cnt pointers. */ void mblk_to_iov(mblk_t *m, int cnt, struct iovec *iovp) { while (m != NULL && cnt-- > 0) { iovp->iov_base = (caddr_t)m->b_rptr; iovp->iov_len = (m->b_wptr - m->b_rptr); iovp++; m = m->b_cont; } } /* * Common code between NFS Version 2 and NFS Version 3 for the public * filehandle multicomponent lookups. */ /* * Public filehandle evaluation of a multi-component lookup, following * symbolic links, if necessary. This may result in a vnode in another * filesystem, which is OK as long as the other filesystem is exported. * * Note that the exi will be set either to NULL or a new reference to the * exportinfo struct that corresponds to the vnode of the multi-component path. * It is the callers responsibility to release this reference. */ int rfs_publicfh_mclookup(char *p, vnode_t *dvp, cred_t *cr, vnode_t **vpp, struct exportinfo **exi, struct sec_ol *sec) { int pathflag; vnode_t *mc_dvp = NULL; vnode_t *realvp; int error; *exi = NULL; /* * check if the given path is a url or native path. Since p is * modified by MCLpath(), it may be empty after returning from * there, and should be checked. */ if ((pathflag = MCLpath(&p)) == -1) return (EIO); /* * If pathflag is SECURITY_QUERY, turn the SEC_QUERY bit * on in sec->sec_flags. This bit will later serve as an * indication in makefh_ol() or makefh3_ol() to overload the * filehandle to contain the sec modes used by the server for * the path. */ if (pathflag == SECURITY_QUERY) { if ((sec->sec_index = (uint_t)(*p)) > 0) { sec->sec_flags |= SEC_QUERY; p++; if ((pathflag = MCLpath(&p)) == -1) return (EIO); } else { cmn_err(CE_NOTE, "nfs_server: invalid security index %d, " "violating WebNFS SNEGO protocol.", sec->sec_index); return (EIO); } } if (p[0] == '\0') { error = ENOENT; goto publicfh_done; } error = rfs_pathname(p, &mc_dvp, vpp, dvp, cr, pathflag); /* * If name resolves to "/" we get EINVAL since we asked for * the vnode of the directory that the file is in. Try again * with NULL directory vnode. */ if (error == EINVAL) { error = rfs_pathname(p, NULL, vpp, dvp, cr, pathflag); if (!error) { ASSERT(*vpp != NULL); if ((*vpp)->v_type == VDIR) { VN_HOLD(*vpp); mc_dvp = *vpp; } else { /* * This should not happen, the filesystem is * in an inconsistent state. Fail the lookup * at this point. */ VN_RELE(*vpp); error = EINVAL; } } } if (error) goto publicfh_done; if (*vpp == NULL) { error = ENOENT; goto publicfh_done; } ASSERT(mc_dvp != NULL); ASSERT(*vpp != NULL); if ((*vpp)->v_type == VDIR) { do { /* * *vpp may be an AutoFS node, so we perform * a VOP_ACCESS() to trigger the mount of the intended * filesystem, so we can perform the lookup in the * intended filesystem. */ (void) VOP_ACCESS(*vpp, 0, 0, cr, NULL); /* * If vnode is covered, get the * the topmost vnode. */ if (vn_mountedvfs(*vpp) != NULL) { error = traverse(vpp); if (error) { VN_RELE(*vpp); goto publicfh_done; } } if (VOP_REALVP(*vpp, &realvp, NULL) == 0 && realvp != *vpp) { /* * If realvp is different from *vpp * then release our reference on *vpp, so that * the export access check be performed on the * real filesystem instead. */ VN_HOLD(realvp); VN_RELE(*vpp); *vpp = realvp; } else break; /* LINTED */ } while (TRUE); /* * Let nfs_vptexi() figure what the real parent is. */ VN_RELE(mc_dvp); mc_dvp = NULL; } else { /* * If vnode is covered, get the * the topmost vnode. */ if (vn_mountedvfs(mc_dvp) != NULL) { error = traverse(&mc_dvp); if (error) { VN_RELE(*vpp); goto publicfh_done; } } if (VOP_REALVP(mc_dvp, &realvp, NULL) == 0 && realvp != mc_dvp) { /* * *vpp is a file, obtain realvp of the parent * directory vnode. */ VN_HOLD(realvp); VN_RELE(mc_dvp); mc_dvp = realvp; } } /* * The pathname may take us from the public filesystem to another. * If that's the case then just set the exportinfo to the new export * and build filehandle for it. Thanks to per-access checking there's * no security issues with doing this. If the client is not allowed * access to this new export then it will get an access error when it * tries to use the filehandle */ if (error = nfs_check_vpexi(mc_dvp, *vpp, kcred, exi)) { VN_RELE(*vpp); goto publicfh_done; } /* * Not allowed access to pseudo exports. */ if (PSEUDO(*exi)) { error = ENOENT; VN_RELE(*vpp); goto publicfh_done; } /* * Do a lookup for the index file. We know the index option doesn't * allow paths through handling in the share command, so mc_dvp will * be the parent for the index file vnode, if its present. Use * temporary pointers to preserve and reuse the vnode pointers of the * original directory in case there's no index file. Note that the * index file is a native path, and should not be interpreted by * the URL parser in rfs_pathname() */ if (((*exi)->exi_export.ex_flags & EX_INDEX) && ((*vpp)->v_type == VDIR) && (pathflag == URLPATH)) { vnode_t *tvp, *tmc_dvp; /* temporary vnode pointers */ tmc_dvp = mc_dvp; mc_dvp = tvp = *vpp; error = rfs_pathname((*exi)->exi_export.ex_index, NULL, vpp, mc_dvp, cr, NATIVEPATH); if (error == ENOENT) { *vpp = tvp; mc_dvp = tmc_dvp; error = 0; } else { /* ok or error other than ENOENT */ if (tmc_dvp) VN_RELE(tmc_dvp); if (error) goto publicfh_done; /* * Found a valid vp for index "filename". Sanity check * for odd case where a directory is provided as index * option argument and leads us to another filesystem */ /* Release the reference on the old exi value */ ASSERT(*exi != NULL); exi_rele(*exi); if (error = nfs_check_vpexi(mc_dvp, *vpp, kcred, exi)) { VN_RELE(*vpp); goto publicfh_done; } } } publicfh_done: if (mc_dvp) VN_RELE(mc_dvp); return (error); } /* * Evaluate a multi-component path */ int rfs_pathname( char *path, /* pathname to evaluate */ vnode_t **dirvpp, /* ret for ptr to parent dir vnode */ vnode_t **compvpp, /* ret for ptr to component vnode */ vnode_t *startdvp, /* starting vnode */ cred_t *cr, /* user's credential */ int pathflag) /* flag to identify path, e.g. URL */ { char namebuf[TYPICALMAXPATHLEN]; struct pathname pn; int error; /* * If pathname starts with '/', then set startdvp to root. */ if (*path == '/') { while (*path == '/') path++; startdvp = rootdir; } error = pn_get_buf(path, UIO_SYSSPACE, &pn, namebuf, sizeof (namebuf)); if (error == 0) { /* * Call the URL parser for URL paths to modify the original * string to handle any '%' encoded characters that exist. * Done here to avoid an extra bcopy in the lookup. * We need to be careful about pathlen's. We know that * rfs_pathname() is called with a non-empty path. However, * it could be emptied due to the path simply being all /'s, * which is valid to proceed with the lookup, or due to the * URL parser finding an encoded null character at the * beginning of path which should not proceed with the lookup. */ if (pn.pn_pathlen != 0 && pathflag == URLPATH) { URLparse(pn.pn_path); if ((pn.pn_pathlen = strlen(pn.pn_path)) == 0) return (ENOENT); } VN_HOLD(startdvp); error = lookuppnvp(&pn, NULL, NO_FOLLOW, dirvpp, compvpp, rootdir, startdvp, cr); } if (error == ENAMETOOLONG) { /* * This thread used a pathname > TYPICALMAXPATHLEN bytes long. */ if (error = pn_get(path, UIO_SYSSPACE, &pn)) return (error); if (pn.pn_pathlen != 0 && pathflag == URLPATH) { URLparse(pn.pn_path); if ((pn.pn_pathlen = strlen(pn.pn_path)) == 0) { pn_free(&pn); return (ENOENT); } } VN_HOLD(startdvp); error = lookuppnvp(&pn, NULL, NO_FOLLOW, dirvpp, compvpp, rootdir, startdvp, cr); pn_free(&pn); } return (error); } /* * Adapt the multicomponent lookup path depending on the pathtype */ static int MCLpath(char **path) { unsigned char c = (unsigned char)**path; /* * If the MCL path is between 0x20 and 0x7E (graphic printable * character of the US-ASCII coded character set), its a URL path, * per RFC 1738. */ if (c >= 0x20 && c <= 0x7E) return (URLPATH); /* * If the first octet of the MCL path is not an ASCII character * then it must be interpreted as a tag value that describes the * format of the remaining octets of the MCL path. * * If the first octet of the MCL path is 0x81 it is a query * for the security info. */ switch (c) { case 0x80: /* native path, i.e. MCL via mount protocol */ (*path)++; return (NATIVEPATH); case 0x81: /* security query */ (*path)++; return (SECURITY_QUERY); default: return (-1); } } #define fromhex(c) ((c >= '0' && c <= '9') ? (c - '0') : \ ((c >= 'A' && c <= 'F') ? (c - 'A' + 10) :\ ((c >= 'a' && c <= 'f') ? (c - 'a' + 10) : 0))) /* * The implementation of URLparse guarantees that the final string will * fit in the original one. Replaces '%' occurrences followed by 2 characters * with its corresponding hexadecimal character. */ static void URLparse(char *str) { char *p, *q; p = q = str; while (*p) { *q = *p; if (*p++ == '%') { if (*p) { *q = fromhex(*p) * 16; p++; if (*p) { *q += fromhex(*p); p++; } } } q++; } *q = '\0'; } /* * Get the export information for the lookup vnode, and verify its * useable. */ int nfs_check_vpexi(vnode_t *mc_dvp, vnode_t *vp, cred_t *cr, struct exportinfo **exi) { int walk; int error = 0; *exi = nfs_vptoexi(mc_dvp, vp, cr, &walk, NULL, FALSE); if (*exi == NULL) error = EACCES; else { /* * If nosub is set for this export then * a lookup relative to the public fh * must not terminate below the * exported directory. */ if ((*exi)->exi_export.ex_flags & EX_NOSUB && walk > 0) error = EACCES; } return (error); } /* * Do the main work of handling HA-NFSv4 Resource Group failover on * Sun Cluster. * We need to detect whether any RG admin paths have been added or removed, * and adjust resources accordingly. * Currently we're using a very inefficient algorithm, ~ 2 * O(n**2). In * order to scale, the list and array of paths need to be held in more * suitable data structures. */ static void hanfsv4_failover(void) { int i, start_grace, numadded_paths = 0; char **added_paths = NULL; rfs4_dss_path_t *dss_path; /* * Note: currently, rfs4_dss_pathlist cannot be NULL, since * it will always include an entry for NFS4_DSS_VAR_DIR. If we * make the latter dynamically specified too, the following will * need to be adjusted. */ /* * First, look for removed paths: RGs that have been failed-over * away from this node. * Walk the "currently-serving" rfs4_dss_pathlist and, for each * path, check if it is on the "passed-in" rfs4_dss_newpaths array * from nfsd. If not, that RG path has been removed. * * Note that nfsd has sorted rfs4_dss_newpaths for us, and removed * any duplicates. */ dss_path = rfs4_dss_pathlist; do { int found = 0; char *path = dss_path->path; /* used only for non-HA so may not be removed */ if (strcmp(path, NFS4_DSS_VAR_DIR) == 0) { dss_path = dss_path->next; continue; } for (i = 0; i < rfs4_dss_numnewpaths; i++) { int cmpret; char *newpath = rfs4_dss_newpaths[i]; /* * Since nfsd has sorted rfs4_dss_newpaths for us, * once the return from strcmp is negative we know * we've passed the point where "path" should be, * and can stop searching: "path" has been removed. */ cmpret = strcmp(path, newpath); if (cmpret < 0) break; if (cmpret == 0) { found = 1; break; } } if (found == 0) { unsigned index = dss_path->index; rfs4_servinst_t *sip = dss_path->sip; rfs4_dss_path_t *path_next = dss_path->next; /* * This path has been removed. * We must clear out the servinst reference to * it, since it's now owned by another * node: we should not attempt to touch it. */ ASSERT(dss_path == sip->dss_paths[index]); sip->dss_paths[index] = NULL; /* remove from "currently-serving" list, and destroy */ remque(dss_path); /* allow for NUL */ kmem_free(dss_path->path, strlen(dss_path->path) + 1); kmem_free(dss_path, sizeof (rfs4_dss_path_t)); dss_path = path_next; } else { /* path was found; not removed */ dss_path = dss_path->next; } } while (dss_path != rfs4_dss_pathlist); /* * Now, look for added paths: RGs that have been failed-over * to this node. * Walk the "passed-in" rfs4_dss_newpaths array from nfsd and, * for each path, check if it is on the "currently-serving" * rfs4_dss_pathlist. If not, that RG path has been added. * * Note: we don't do duplicate detection here; nfsd does that for us. * * Note: numadded_paths <= rfs4_dss_numnewpaths, which gives us * an upper bound for the size needed for added_paths[numadded_paths]. */ /* probably more space than we need, but guaranteed to be enough */ if (rfs4_dss_numnewpaths > 0) { size_t sz = rfs4_dss_numnewpaths * sizeof (char *); added_paths = kmem_zalloc(sz, KM_SLEEP); } /* walk the "passed-in" rfs4_dss_newpaths array from nfsd */ for (i = 0; i < rfs4_dss_numnewpaths; i++) { int found = 0; char *newpath = rfs4_dss_newpaths[i]; dss_path = rfs4_dss_pathlist; do { char *path = dss_path->path; /* used only for non-HA */ if (strcmp(path, NFS4_DSS_VAR_DIR) == 0) { dss_path = dss_path->next; continue; } if (strncmp(path, newpath, strlen(path)) == 0) { found = 1; break; } dss_path = dss_path->next; } while (dss_path != rfs4_dss_pathlist); if (found == 0) { added_paths[numadded_paths] = newpath; numadded_paths++; } } /* did we find any added paths? */ if (numadded_paths > 0) { /* create a new server instance, and start its grace period */ start_grace = 1; rfs4_servinst_create(start_grace, numadded_paths, added_paths); /* read in the stable storage state from these paths */ rfs4_dss_readstate(numadded_paths, added_paths); /* * Multiple failovers during a grace period will cause * clients of the same resource group to be partitioned * into different server instances, with different * grace periods. Since clients of the same resource * group must be subject to the same grace period, * we need to reset all currently active grace periods. */ rfs4_grace_reset_all(); } if (rfs4_dss_numnewpaths > 0) kmem_free(added_paths, rfs4_dss_numnewpaths * sizeof (char *)); }