/*- * Copyright (c) 1999, 2000, 2001, 2002 Robert N. M. Watson * Copyright (c) 2001 Ilmar S. Habibulin * Copyright (c) 2001, 2002 Networks Associates Technology, Inc. * All rights reserved. * * This software was developed by Robert Watson and Ilmar Habibulin for the * TrustedBSD Project. * * This software was developed for the FreeBSD Project in part by NAI Labs, * the Security Research Division of Network Associates, Inc. under * DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS"), as part of the DARPA * CHATS research program. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. The names of the authors may not be used to endorse or promote * products derived from this software without specific prior written * permission. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ /* * Developed by the TrustedBSD Project. * * Framework for extensible kernel access control. Kernel and userland * interface to the framework, policy registration and composition. */ #include "opt_mac.h" #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 #ifdef MAC /* * Declare that the kernel provides MAC support, version 1. This permits * modules to refuse to be loaded if the necessary support isn't present, * even if it's pre-boot. */ MODULE_VERSION(kernel_mac_support, 1); SYSCTL_DECL(_security); SYSCTL_NODE(_security, OID_AUTO, mac, CTLFLAG_RW, 0, "TrustedBSD MAC policy controls"); SYSCTL_NODE(_security_mac, OID_AUTO, debug, CTLFLAG_RW, 0, "TrustedBSD MAC debug info"); static int mac_debug_label_fallback = 0; SYSCTL_INT(_security_mac_debug, OID_AUTO, label_fallback, CTLFLAG_RW, &mac_debug_label_fallback, 0, "Filesystems should fall back to fs label" "when label is corrupted."); TUNABLE_INT("security.mac.debug_label_fallback", &mac_debug_label_fallback); #ifndef MAC_MAX_POLICIES #define MAC_MAX_POLICIES 8 #endif #if MAC_MAX_POLICIES > 32 #error "MAC_MAX_POLICIES too large" #endif static unsigned int mac_max_policies = MAC_MAX_POLICIES; static unsigned int mac_policy_offsets_free = (1 << MAC_MAX_POLICIES) - 1; SYSCTL_UINT(_security_mac, OID_AUTO, max_policies, CTLFLAG_RD, &mac_max_policies, 0, ""); static int mac_late = 0; static int mac_enforce_fs = 1; SYSCTL_INT(_security_mac, OID_AUTO, enforce_fs, CTLFLAG_RW, &mac_enforce_fs, 0, "Enforce MAC policy on file system objects"); TUNABLE_INT("security.mac.enforce_fs", &mac_enforce_fs); static int mac_enforce_network = 1; SYSCTL_INT(_security_mac, OID_AUTO, enforce_network, CTLFLAG_RW, &mac_enforce_network, 0, "Enforce MAC policy on network packets"); TUNABLE_INT("security.mac.enforce_network", &mac_enforce_network); static int mac_enforce_process = 1; SYSCTL_INT(_security_mac, OID_AUTO, enforce_process, CTLFLAG_RW, &mac_enforce_process, 0, "Enforce MAC policy on inter-process operations"); TUNABLE_INT("security.mac.enforce_process", &mac_enforce_process); static int mac_enforce_socket = 1; SYSCTL_INT(_security_mac, OID_AUTO, enforce_socket, CTLFLAG_RW, &mac_enforce_socket, 0, "Enforce MAC policy on socket operations"); TUNABLE_INT("security.mac.enforce_socket", &mac_enforce_socket); static int mac_enforce_pipe = 1; SYSCTL_INT(_security_mac, OID_AUTO, enforce_pipe, CTLFLAG_RW, &mac_enforce_pipe, 0, "Enforce MAC policy on pipe operations"); static int mac_label_size = sizeof(struct mac); SYSCTL_INT(_security_mac, OID_AUTO, label_size, CTLFLAG_RD, &mac_label_size, 0, "Pre-compiled MAC label size"); static int mac_cache_fslabel_in_vnode = 1; SYSCTL_INT(_security_mac, OID_AUTO, cache_fslabel_in_vnode, CTLFLAG_RW, &mac_cache_fslabel_in_vnode, 0, "Cache mount fslabel in vnode"); TUNABLE_INT("security.mac.cache_fslabel_in_vnode", &mac_cache_fslabel_in_vnode); static int mac_vnode_label_cache_hits = 0; SYSCTL_INT(_security_mac, OID_AUTO, vnode_label_cache_hits, CTLFLAG_RD, &mac_vnode_label_cache_hits, 0, "Cache hits on vnode labels"); static int mac_vnode_label_cache_misses = 0; SYSCTL_INT(_security_mac, OID_AUTO, vnode_label_cache_misses, CTLFLAG_RD, &mac_vnode_label_cache_misses, 0, "Cache misses on vnode labels"); static int mac_mmap_revocation_via_cow = 0; SYSCTL_INT(_security_mac, OID_AUTO, mmap_revocation_via_cow, CTLFLAG_RW, &mac_mmap_revocation_via_cow, 0, "Revoke mmap access to files via " "copy-on-write semantics, or by removing all write access"); #ifdef MAC_DEBUG static unsigned int nmacmbufs, nmaccreds, nmacifnets, nmacbpfdescs, nmacsockets, nmacmounts, nmactemp, nmacvnodes, nmacdevfsdirents, nmacipqs, nmacpipes; SYSCTL_UINT(_security_mac_debug, OID_AUTO, mbufs, CTLFLAG_RD, &nmacmbufs, 0, "number of mbufs in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, creds, CTLFLAG_RD, &nmaccreds, 0, "number of ucreds in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, ifnets, CTLFLAG_RD, &nmacifnets, 0, "number of ifnets in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, ipqs, CTLFLAG_RD, &nmacipqs, 0, "number of ipqs in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, bpfdescs, CTLFLAG_RD, &nmacbpfdescs, 0, "number of bpfdescs in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, sockets, CTLFLAG_RD, &nmacsockets, 0, "number of sockets in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, pipes, CTLFLAG_RD, &nmacpipes, 0, "number of pipes in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, mounts, CTLFLAG_RD, &nmacmounts, 0, "number of mounts in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, temp, CTLFLAG_RD, &nmactemp, 0, "number of temporary labels in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, vnodes, CTLFLAG_RD, &nmacvnodes, 0, "number of vnodes in use"); SYSCTL_UINT(_security_mac_debug, OID_AUTO, devfsdirents, CTLFLAG_RD, &nmacdevfsdirents, 0, "number of devfs dirents inuse"); #endif static int error_select(int error1, int error2); static int mac_externalize(struct label *label, struct mac *mac); static int mac_policy_register(struct mac_policy_conf *mpc); static int mac_policy_unregister(struct mac_policy_conf *mpc); static int mac_stdcreatevnode_ea(struct vnode *vp); static void mac_cred_mmapped_drop_perms(struct thread *td, struct ucred *cred); static void mac_cred_mmapped_drop_perms_recurse(struct thread *td, struct ucred *cred, struct vm_map *map); MALLOC_DEFINE(M_MACOPVEC, "macopvec", "MAC policy operation vector"); MALLOC_DEFINE(M_MACPIPELABEL, "macpipelabel", "MAC labels for pipes"); /* * mac_policy_list_lock protects the consistency of 'mac_policy_list', * the linked list of attached policy modules. Read-only consumers of * the list must acquire a shared lock for the duration of their use; * writers must acquire an exclusive lock. Note that for compound * operations, locks should be held for the entire compound operation, * and that this is not yet done for relabel requests. */ static struct mtx mac_policy_list_lock; static LIST_HEAD(, mac_policy_conf) mac_policy_list; static int mac_policy_list_busy; #define MAC_POLICY_LIST_LOCKINIT() mtx_init(&mac_policy_list_lock, \ "mac_policy_list_lock", NULL, MTX_DEF); #define MAC_POLICY_LIST_LOCK() mtx_lock(&mac_policy_list_lock); #define MAC_POLICY_LIST_UNLOCK() mtx_unlock(&mac_policy_list_lock); #define MAC_POLICY_LIST_BUSY() do { \ MAC_POLICY_LIST_LOCK(); \ mac_policy_list_busy++; \ MAC_POLICY_LIST_UNLOCK(); \ } while (0) #define MAC_POLICY_LIST_UNBUSY() do { \ MAC_POLICY_LIST_LOCK(); \ mac_policy_list_busy--; \ if (mac_policy_list_busy < 0) \ panic("Extra mac_policy_list_busy--"); \ MAC_POLICY_LIST_UNLOCK(); \ } while (0) /* * MAC_CHECK performs the designated check by walking the policy * module list and checking with each as to how it feels about the * request. Note that it returns its value via 'error' in the scope * of the caller. */ #define MAC_CHECK(check, args...) do { \ struct mac_policy_conf *mpc; \ \ error = 0; \ MAC_POLICY_LIST_BUSY(); \ LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \ if (mpc->mpc_ops->mpo_ ## check != NULL) \ error = error_select( \ mpc->mpc_ops->mpo_ ## check (args), \ error); \ } \ MAC_POLICY_LIST_UNBUSY(); \ } while (0) /* * MAC_BOOLEAN performs the designated boolean composition by walking * the module list, invoking each instance of the operation, and * combining the results using the passed C operator. Note that it * returns its value via 'result' in the scope of the caller, which * should be initialized by the caller in a meaningful way to get * a meaningful result. */ #define MAC_BOOLEAN(operation, composition, args...) do { \ struct mac_policy_conf *mpc; \ \ MAC_POLICY_LIST_BUSY(); \ LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \ if (mpc->mpc_ops->mpo_ ## operation != NULL) \ result = result composition \ mpc->mpc_ops->mpo_ ## operation (args); \ } \ MAC_POLICY_LIST_UNBUSY(); \ } while (0) /* * MAC_PERFORM performs the designated operation by walking the policy * module list and invoking that operation for each policy. */ #define MAC_PERFORM(operation, args...) do { \ struct mac_policy_conf *mpc; \ \ MAC_POLICY_LIST_BUSY(); \ LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \ if (mpc->mpc_ops->mpo_ ## operation != NULL) \ mpc->mpc_ops->mpo_ ## operation (args); \ } \ MAC_POLICY_LIST_UNBUSY(); \ } while (0) /* * Initialize the MAC subsystem, including appropriate SMP locks. */ static void mac_init(void) { LIST_INIT(&mac_policy_list); MAC_POLICY_LIST_LOCKINIT(); } /* * For the purposes of modules that want to know if they were loaded * "early", set the mac_late flag once we've processed modules either * linked into the kernel, or loaded before the kernel startup. */ static void mac_late_init(void) { mac_late = 1; } /* * Allow MAC policy modules to register during boot, etc. */ int mac_policy_modevent(module_t mod, int type, void *data) { struct mac_policy_conf *mpc; int error; error = 0; mpc = (struct mac_policy_conf *) data; switch (type) { case MOD_LOAD: if (mpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_NOTLATE && mac_late) { printf("mac_policy_modevent: can't load %s policy " "after booting\n", mpc->mpc_name); error = EBUSY; break; } error = mac_policy_register(mpc); break; case MOD_UNLOAD: /* Don't unregister the module if it was never registered. */ if ((mpc->mpc_runtime_flags & MPC_RUNTIME_FLAG_REGISTERED) != 0) error = mac_policy_unregister(mpc); else error = 0; break; default: break; } return (error); } static int mac_policy_register(struct mac_policy_conf *mpc) { struct mac_policy_conf *tmpc; struct mac_policy_ops *ops; struct mac_policy_op_entry *mpe; int slot; MALLOC(mpc->mpc_ops, struct mac_policy_ops *, sizeof(*ops), M_MACOPVEC, M_WAITOK | M_ZERO); for (mpe = mpc->mpc_entries; mpe->mpe_constant != MAC_OP_LAST; mpe++) { switch (mpe->mpe_constant) { case MAC_OP_LAST: /* * Doesn't actually happen, but this allows checking * that all enumerated values are handled. */ break; case MAC_DESTROY: mpc->mpc_ops->mpo_destroy = mpe->mpe_function; break; case MAC_INIT: mpc->mpc_ops->mpo_init = mpe->mpe_function; break; case MAC_INIT_BPFDESC: mpc->mpc_ops->mpo_init_bpfdesc = mpe->mpe_function; break; case MAC_INIT_CRED: mpc->mpc_ops->mpo_init_cred = mpe->mpe_function; break; case MAC_INIT_DEVFSDIRENT: mpc->mpc_ops->mpo_init_devfsdirent = mpe->mpe_function; break; case MAC_INIT_IFNET: mpc->mpc_ops->mpo_init_ifnet = mpe->mpe_function; break; case MAC_INIT_IPQ: mpc->mpc_ops->mpo_init_ipq = mpe->mpe_function; break; case MAC_INIT_MBUF: mpc->mpc_ops->mpo_init_mbuf = mpe->mpe_function; break; case MAC_INIT_MOUNT: mpc->mpc_ops->mpo_init_mount = mpe->mpe_function; break; case MAC_INIT_PIPE: mpc->mpc_ops->mpo_init_pipe = mpe->mpe_function; break; case MAC_INIT_SOCKET: mpc->mpc_ops->mpo_init_socket = mpe->mpe_function; break; case MAC_INIT_TEMP: mpc->mpc_ops->mpo_init_temp = mpe->mpe_function; break; case MAC_INIT_VNODE: mpc->mpc_ops->mpo_init_vnode = mpe->mpe_function; break; case MAC_DESTROY_BPFDESC: mpc->mpc_ops->mpo_destroy_bpfdesc = mpe->mpe_function; break; case MAC_DESTROY_CRED: mpc->mpc_ops->mpo_destroy_cred = mpe->mpe_function; break; case MAC_DESTROY_DEVFSDIRENT: mpc->mpc_ops->mpo_destroy_devfsdirent = mpe->mpe_function; break; case MAC_DESTROY_IFNET: mpc->mpc_ops->mpo_destroy_ifnet = mpe->mpe_function; break; case MAC_DESTROY_IPQ: mpc->mpc_ops->mpo_destroy_ipq = mpe->mpe_function; break; case MAC_DESTROY_MBUF: mpc->mpc_ops->mpo_destroy_mbuf = mpe->mpe_function; break; case MAC_DESTROY_MOUNT: mpc->mpc_ops->mpo_destroy_mount = mpe->mpe_function; break; case MAC_DESTROY_PIPE: mpc->mpc_ops->mpo_destroy_pipe = mpe->mpe_function; break; case MAC_DESTROY_SOCKET: mpc->mpc_ops->mpo_destroy_socket = mpe->mpe_function; break; case MAC_DESTROY_TEMP: mpc->mpc_ops->mpo_destroy_temp = mpe->mpe_function; break; case MAC_DESTROY_VNODE: mpc->mpc_ops->mpo_destroy_vnode = mpe->mpe_function; break; case MAC_EXTERNALIZE: mpc->mpc_ops->mpo_externalize = mpe->mpe_function; break; case MAC_INTERNALIZE: mpc->mpc_ops->mpo_internalize = mpe->mpe_function; break; case MAC_CREATE_DEVFS_DEVICE: mpc->mpc_ops->mpo_create_devfs_device = mpe->mpe_function; break; case MAC_CREATE_DEVFS_DIRECTORY: mpc->mpc_ops->mpo_create_devfs_directory = mpe->mpe_function; break; case MAC_CREATE_DEVFS_VNODE: mpc->mpc_ops->mpo_create_devfs_vnode = mpe->mpe_function; break; case MAC_STDCREATEVNODE_EA: mpc->mpc_ops->mpo_stdcreatevnode_ea = mpe->mpe_function; break; case MAC_CREATE_VNODE: mpc->mpc_ops->mpo_create_vnode = mpe->mpe_function; break; case MAC_CREATE_MOUNT: mpc->mpc_ops->mpo_create_mount = mpe->mpe_function; break; case MAC_CREATE_ROOT_MOUNT: mpc->mpc_ops->mpo_create_root_mount = mpe->mpe_function; break; case MAC_RELABEL_VNODE: mpc->mpc_ops->mpo_relabel_vnode = mpe->mpe_function; break; case MAC_UPDATE_DEVFSDIRENT: mpc->mpc_ops->mpo_update_devfsdirent = mpe->mpe_function; break; case MAC_UPDATE_PROCFSVNODE: mpc->mpc_ops->mpo_update_procfsvnode = mpe->mpe_function; break; case MAC_UPDATE_VNODE_FROM_EXTATTR: mpc->mpc_ops->mpo_update_vnode_from_extattr = mpe->mpe_function; break; case MAC_UPDATE_VNODE_FROM_EXTERNALIZED: mpc->mpc_ops->mpo_update_vnode_from_externalized = mpe->mpe_function; break; case MAC_UPDATE_VNODE_FROM_MOUNT: mpc->mpc_ops->mpo_update_vnode_from_mount = mpe->mpe_function; break; case MAC_CREATE_MBUF_FROM_SOCKET: mpc->mpc_ops->mpo_create_mbuf_from_socket = mpe->mpe_function; break; case MAC_CREATE_PIPE: mpc->mpc_ops->mpo_create_pipe = mpe->mpe_function; break; case MAC_CREATE_SOCKET: mpc->mpc_ops->mpo_create_socket = mpe->mpe_function; break; case MAC_CREATE_SOCKET_FROM_SOCKET: mpc->mpc_ops->mpo_create_socket_from_socket = mpe->mpe_function; break; case MAC_RELABEL_PIPE: mpc->mpc_ops->mpo_relabel_pipe = mpe->mpe_function; break; case MAC_RELABEL_SOCKET: mpc->mpc_ops->mpo_relabel_socket = mpe->mpe_function; break; case MAC_SET_SOCKET_PEER_FROM_MBUF: mpc->mpc_ops->mpo_set_socket_peer_from_mbuf = mpe->mpe_function; break; case MAC_SET_SOCKET_PEER_FROM_SOCKET: mpc->mpc_ops->mpo_set_socket_peer_from_socket = mpe->mpe_function; break; case MAC_CREATE_BPFDESC: mpc->mpc_ops->mpo_create_bpfdesc = mpe->mpe_function; break; case MAC_CREATE_DATAGRAM_FROM_IPQ: mpc->mpc_ops->mpo_create_datagram_from_ipq = mpe->mpe_function; break; case MAC_CREATE_FRAGMENT: mpc->mpc_ops->mpo_create_fragment = mpe->mpe_function; break; case MAC_CREATE_IFNET: mpc->mpc_ops->mpo_create_ifnet = mpe->mpe_function; break; case MAC_CREATE_IPQ: mpc->mpc_ops->mpo_create_ipq = mpe->mpe_function; break; case MAC_CREATE_MBUF_FROM_MBUF: mpc->mpc_ops->mpo_create_mbuf_from_mbuf = mpe->mpe_function; break; case MAC_CREATE_MBUF_LINKLAYER: mpc->mpc_ops->mpo_create_mbuf_linklayer = mpe->mpe_function; break; case MAC_CREATE_MBUF_FROM_BPFDESC: mpc->mpc_ops->mpo_create_mbuf_from_bpfdesc = mpe->mpe_function; break; case MAC_CREATE_MBUF_FROM_IFNET: mpc->mpc_ops->mpo_create_mbuf_from_ifnet = mpe->mpe_function; break; case MAC_CREATE_MBUF_MULTICAST_ENCAP: mpc->mpc_ops->mpo_create_mbuf_multicast_encap = mpe->mpe_function; break; case MAC_CREATE_MBUF_NETLAYER: mpc->mpc_ops->mpo_create_mbuf_netlayer = mpe->mpe_function; break; case MAC_FRAGMENT_MATCH: mpc->mpc_ops->mpo_fragment_match = mpe->mpe_function; break; case MAC_RELABEL_IFNET: mpc->mpc_ops->mpo_relabel_ifnet = mpe->mpe_function; break; case MAC_UPDATE_IPQ: mpc->mpc_ops->mpo_update_ipq = mpe->mpe_function; break; case MAC_CREATE_CRED: mpc->mpc_ops->mpo_create_cred = mpe->mpe_function; break; case MAC_EXECVE_TRANSITION: mpc->mpc_ops->mpo_execve_transition = mpe->mpe_function; break; case MAC_EXECVE_WILL_TRANSITION: mpc->mpc_ops->mpo_execve_will_transition = mpe->mpe_function; break; case MAC_CREATE_PROC0: mpc->mpc_ops->mpo_create_proc0 = mpe->mpe_function; break; case MAC_CREATE_PROC1: mpc->mpc_ops->mpo_create_proc1 = mpe->mpe_function; break; case MAC_RELABEL_CRED: mpc->mpc_ops->mpo_relabel_cred = mpe->mpe_function; break; case MAC_CHECK_BPFDESC_RECEIVE: mpc->mpc_ops->mpo_check_bpfdesc_receive = mpe->mpe_function; break; case MAC_CHECK_CRED_RELABEL: mpc->mpc_ops->mpo_check_cred_relabel = mpe->mpe_function; break; case MAC_CHECK_CRED_VISIBLE: mpc->mpc_ops->mpo_check_cred_visible = mpe->mpe_function; break; case MAC_CHECK_IFNET_RELABEL: mpc->mpc_ops->mpo_check_ifnet_relabel = mpe->mpe_function; break; case MAC_CHECK_IFNET_TRANSMIT: mpc->mpc_ops->mpo_check_ifnet_transmit = mpe->mpe_function; break; case MAC_CHECK_MOUNT_STAT: mpc->mpc_ops->mpo_check_mount_stat = mpe->mpe_function; break; case MAC_CHECK_PIPE_IOCTL: mpc->mpc_ops->mpo_check_pipe_ioctl = mpe->mpe_function; break; case MAC_CHECK_PIPE_OP: mpc->mpc_ops->mpo_check_pipe_op = mpe->mpe_function; break; case MAC_CHECK_PIPE_RELABEL: mpc->mpc_ops->mpo_check_pipe_relabel = mpe->mpe_function; break; case MAC_CHECK_PROC_DEBUG: mpc->mpc_ops->mpo_check_proc_debug = mpe->mpe_function; break; case MAC_CHECK_PROC_SCHED: mpc->mpc_ops->mpo_check_proc_sched = mpe->mpe_function; break; case MAC_CHECK_PROC_SIGNAL: mpc->mpc_ops->mpo_check_proc_signal = mpe->mpe_function; break; case MAC_CHECK_SOCKET_BIND: mpc->mpc_ops->mpo_check_socket_bind = mpe->mpe_function; break; case MAC_CHECK_SOCKET_CONNECT: mpc->mpc_ops->mpo_check_socket_connect = mpe->mpe_function; break; case MAC_CHECK_SOCKET_DELIVER: mpc->mpc_ops->mpo_check_socket_deliver = mpe->mpe_function; break; case MAC_CHECK_SOCKET_LISTEN: mpc->mpc_ops->mpo_check_socket_listen = mpe->mpe_function; break; case MAC_CHECK_SOCKET_RELABEL: mpc->mpc_ops->mpo_check_socket_relabel = mpe->mpe_function; break; case MAC_CHECK_SOCKET_VISIBLE: mpc->mpc_ops->mpo_check_socket_visible = mpe->mpe_function; break; case MAC_CHECK_VNODE_ACCESS: mpc->mpc_ops->mpo_check_vnode_access = mpe->mpe_function; break; case MAC_CHECK_VNODE_CHDIR: mpc->mpc_ops->mpo_check_vnode_chdir = mpe->mpe_function; break; case MAC_CHECK_VNODE_CHROOT: mpc->mpc_ops->mpo_check_vnode_chroot = mpe->mpe_function; break; case MAC_CHECK_VNODE_CREATE: mpc->mpc_ops->mpo_check_vnode_create = mpe->mpe_function; break; case MAC_CHECK_VNODE_DELETE: mpc->mpc_ops->mpo_check_vnode_delete = mpe->mpe_function; break; case MAC_CHECK_VNODE_DELETEACL: mpc->mpc_ops->mpo_check_vnode_deleteacl = mpe->mpe_function; break; case MAC_CHECK_VNODE_EXEC: mpc->mpc_ops->mpo_check_vnode_exec = mpe->mpe_function; break; case MAC_CHECK_VNODE_GETACL: mpc->mpc_ops->mpo_check_vnode_getacl = mpe->mpe_function; break; case MAC_CHECK_VNODE_GETEXTATTR: mpc->mpc_ops->mpo_check_vnode_getextattr = mpe->mpe_function; break; case MAC_CHECK_VNODE_LOOKUP: mpc->mpc_ops->mpo_check_vnode_lookup = mpe->mpe_function; break; case MAC_CHECK_VNODE_MMAP_PERMS: mpc->mpc_ops->mpo_check_vnode_mmap_perms = mpe->mpe_function; break; case MAC_CHECK_VNODE_OP: mpc->mpc_ops->mpo_check_vnode_op = mpe->mpe_function; break; case MAC_CHECK_VNODE_OPEN: mpc->mpc_ops->mpo_check_vnode_open = mpe->mpe_function; break; case MAC_CHECK_VNODE_READDIR: mpc->mpc_ops->mpo_check_vnode_readdir = mpe->mpe_function; break; case MAC_CHECK_VNODE_READLINK: mpc->mpc_ops->mpo_check_vnode_readlink = mpe->mpe_function; break; case MAC_CHECK_VNODE_RELABEL: mpc->mpc_ops->mpo_check_vnode_relabel = mpe->mpe_function; break; case MAC_CHECK_VNODE_RENAME_FROM: mpc->mpc_ops->mpo_check_vnode_rename_from = mpe->mpe_function; break; case MAC_CHECK_VNODE_RENAME_TO: mpc->mpc_ops->mpo_check_vnode_rename_to = mpe->mpe_function; break; case MAC_CHECK_VNODE_REVOKE: mpc->mpc_ops->mpo_check_vnode_revoke = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETACL: mpc->mpc_ops->mpo_check_vnode_setacl = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETEXTATTR: mpc->mpc_ops->mpo_check_vnode_setextattr = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETFLAGS: mpc->mpc_ops->mpo_check_vnode_setflags = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETMODE: mpc->mpc_ops->mpo_check_vnode_setmode = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETOWNER: mpc->mpc_ops->mpo_check_vnode_setowner = mpe->mpe_function; break; case MAC_CHECK_VNODE_SETUTIMES: mpc->mpc_ops->mpo_check_vnode_setutimes = mpe->mpe_function; break; case MAC_CHECK_VNODE_STAT: mpc->mpc_ops->mpo_check_vnode_stat = mpe->mpe_function; break; /* default: printf("MAC policy `%s': unknown operation %d\n", mpc->mpc_name, mpe->mpe_constant); return (EINVAL); */ } } MAC_POLICY_LIST_LOCK(); if (mac_policy_list_busy > 0) { MAC_POLICY_LIST_UNLOCK(); FREE(mpc->mpc_ops, M_MACOPVEC); mpc->mpc_ops = NULL; return (EBUSY); } LIST_FOREACH(tmpc, &mac_policy_list, mpc_list) { if (strcmp(tmpc->mpc_name, mpc->mpc_name) == 0) { MAC_POLICY_LIST_UNLOCK(); FREE(mpc->mpc_ops, M_MACOPVEC); mpc->mpc_ops = NULL; return (EEXIST); } } if (mpc->mpc_field_off != NULL) { slot = ffs(mac_policy_offsets_free); if (slot == 0) { MAC_POLICY_LIST_UNLOCK(); FREE(mpc->mpc_ops, M_MACOPVEC); mpc->mpc_ops = NULL; return (ENOMEM); } slot--; mac_policy_offsets_free &= ~(1 << slot); *mpc->mpc_field_off = slot; } mpc->mpc_runtime_flags |= MPC_RUNTIME_FLAG_REGISTERED; LIST_INSERT_HEAD(&mac_policy_list, mpc, mpc_list); /* Per-policy initialization. */ if (mpc->mpc_ops->mpo_init != NULL) (*(mpc->mpc_ops->mpo_init))(mpc); MAC_POLICY_LIST_UNLOCK(); printf("Security policy loaded: %s (%s)\n", mpc->mpc_fullname, mpc->mpc_name); return (0); } static int mac_policy_unregister(struct mac_policy_conf *mpc) { #if 0 /* * Don't allow unloading modules with private data. */ if (mpc->mpc_field_off != NULL) return (EBUSY); #endif if ((mpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_UNLOADOK) == 0) return (EBUSY); MAC_POLICY_LIST_LOCK(); if (mac_policy_list_busy > 0) { MAC_POLICY_LIST_UNLOCK(); return (EBUSY); } if (mpc->mpc_ops->mpo_destroy != NULL) (*(mpc->mpc_ops->mpo_destroy))(mpc); LIST_REMOVE(mpc, mpc_list); MAC_POLICY_LIST_UNLOCK(); FREE(mpc->mpc_ops, M_MACOPVEC); mpc->mpc_ops = NULL; printf("Security policy unload: %s (%s)\n", mpc->mpc_fullname, mpc->mpc_name); return (0); } /* * Define an error value precedence, and given two arguments, selects the * value with the higher precedence. */ static int error_select(int error1, int error2) { /* Certain decision-making errors take top priority. */ if (error1 == EDEADLK || error2 == EDEADLK) return (EDEADLK); /* Invalid arguments should be reported where possible. */ if (error1 == EINVAL || error2 == EINVAL) return (EINVAL); /* Precedence goes to "visibility", with both process and file. */ if (error1 == ESRCH || error2 == ESRCH) return (ESRCH); if (error1 == ENOENT || error2 == ENOENT) return (ENOENT); /* Precedence goes to DAC/MAC protections. */ if (error1 == EACCES || error2 == EACCES) return (EACCES); /* Precedence goes to privilege. */ if (error1 == EPERM || error2 == EPERM) return (EPERM); /* Precedence goes to error over success; otherwise, arbitrary. */ if (error1 != 0) return (error1); return (error2); } void mac_update_devfsdirent(struct devfs_dirent *de, struct vnode *vp) { MAC_PERFORM(update_devfsdirent, de, &de->de_label, vp, &vp->v_label); } void mac_update_procfsvnode(struct vnode *vp, struct ucred *cred) { MAC_PERFORM(update_procfsvnode, vp, &vp->v_label, cred); } /* * Support callout for policies that manage their own externalization * using extended attributes. */ static int mac_update_vnode_from_extattr(struct vnode *vp, struct mount *mp) { int error; MAC_CHECK(update_vnode_from_extattr, vp, &vp->v_label, mp, &mp->mnt_fslabel); return (error); } /* * Given an externalized mac label, internalize it and stamp it on a * vnode. */ static int mac_update_vnode_from_externalized(struct vnode *vp, struct mac *extmac) { int error; MAC_CHECK(update_vnode_from_externalized, vp, &vp->v_label, extmac); return (error); } /* * Call out to individual policies to update the label in a vnode from * the mountpoint. */ void mac_update_vnode_from_mount(struct vnode *vp, struct mount *mp) { MAC_PERFORM(update_vnode_from_mount, vp, &vp->v_label, mp, &mp->mnt_fslabel); ASSERT_VOP_LOCKED(vp, "mac_update_vnode_from_mount"); if (mac_cache_fslabel_in_vnode) vp->v_vflag |= VV_CACHEDLABEL; } /* * Implementation of VOP_REFRESHLABEL() that relies on extended attributes * to store label data. Can be referenced by filesystems supporting * extended attributes. */ int vop_stdrefreshlabel_ea(struct vop_refreshlabel_args *ap) { struct vnode *vp = ap->a_vp; struct mac extmac; int buflen, error; ASSERT_VOP_LOCKED(vp, "vop_stdrefreshlabel_ea"); /* * Call out to external policies first. Order doesn't really * matter, as long as failure of one assures failure of all. */ error = mac_update_vnode_from_extattr(vp, vp->v_mount); if (error) return (error); buflen = sizeof(extmac); error = vn_extattr_get(vp, IO_NODELOCKED, FREEBSD_MAC_EXTATTR_NAMESPACE, FREEBSD_MAC_EXTATTR_NAME, &buflen, (char *)&extmac, curthread); switch (error) { case 0: /* Got it */ break; case ENOATTR: /* * Use the label from the mount point. */ mac_update_vnode_from_mount(vp, vp->v_mount); return (0); case EOPNOTSUPP: default: /* Fail horribly. */ return (error); } if (buflen != sizeof(extmac)) error = EPERM; /* Fail very closed. */ if (error == 0) error = mac_update_vnode_from_externalized(vp, &extmac); if (error == 0) vp->v_vflag |= VV_CACHEDLABEL; else { struct vattr va; printf("Corrupted label on %s", vp->v_mount->mnt_stat.f_mntonname); if (VOP_GETATTR(vp, &va, curthread->td_ucred, curthread) == 0) printf(" inum %ld", va.va_fileid); if (mac_debug_label_fallback) { printf(", falling back.\n"); mac_update_vnode_from_mount(vp, vp->v_mount); error = 0; } else { printf(".\n"); error = EPERM; } } return (error); } /* * Make sure the vnode label is up-to-date. If EOPNOTSUPP, then we handle * the labeling activity outselves. Filesystems should be careful not * to change their minds regarding whether they support vop_refreshlabel() * for a vnode or not. Don't cache the vnode here, allow the file * system code to determine if it's safe to cache. If we update from * the mount, don't cache since a change to the mount label should affect * all vnodes. */ static int vn_refreshlabel(struct vnode *vp, struct ucred *cred) { int error; ASSERT_VOP_LOCKED(vp, "vn_refreshlabel"); if (vp->v_mount == NULL) { /* Eventually, we probably want to special-case refreshing of deadfs vnodes, and if there's a lock-free race somewhere, that case might be handled here. mac_update_vnode_deadfs(vp); return (0); */ /* printf("vn_refreshlabel: null v_mount\n"); */ if (vp->v_tag != VT_NON) printf( "vn_refreshlabel: null v_mount with non-VT_NON\n"); return (EBADF); } if (vp->v_vflag & VV_CACHEDLABEL) { mac_vnode_label_cache_hits++; return (0); } else mac_vnode_label_cache_misses++; if ((vp->v_mount->mnt_flag & MNT_MULTILABEL) == 0) { mac_update_vnode_from_mount(vp, vp->v_mount); return (0); } error = VOP_REFRESHLABEL(vp, cred, curthread); switch (error) { case EOPNOTSUPP: /* * If labels are not supported on this vnode, fall back to * the label in the mount and propagate it to the vnode. * There should probably be some sort of policy/flag/decision * about doing this. */ mac_update_vnode_from_mount(vp, vp->v_mount); error = 0; default: return (error); } } /* * Helper function for file systems using the vop_std*_ea() calls. This * function must be called after EA service is available for the vnode, * but before it's hooked up to the namespace so that the node persists * if there's a crash, or before it can be accessed. On successful * commit of the label to disk (etc), do cache the label. */ int vop_stdcreatevnode_ea(struct vnode *dvp, struct vnode *tvp, struct ucred *cred) { struct mac extmac; int error; ASSERT_VOP_LOCKED(tvp, "vop_stdcreatevnode_ea"); if ((dvp->v_mount->mnt_flag & MNT_MULTILABEL) == 0) { mac_update_vnode_from_mount(tvp, tvp->v_mount); } else { error = vn_refreshlabel(dvp, cred); if (error) return (error); /* * Stick the label in the vnode. Then try to write to * disk. If we fail, return a failure to abort the * create operation. Really, this failure shouldn't * happen except in fairly unusual circumstances (out * of disk, etc). */ mac_create_vnode(cred, dvp, tvp); error = mac_stdcreatevnode_ea(tvp); if (error) return (error); /* * XXX: Eventually this will go away and all policies will * directly manage their extended attributes. */ error = mac_externalize(&tvp->v_label, &extmac); if (error) return (error); error = vn_extattr_set(tvp, IO_NODELOCKED, FREEBSD_MAC_EXTATTR_NAMESPACE, FREEBSD_MAC_EXTATTR_NAME, sizeof(extmac), (char *)&extmac, curthread); if (error == 0) tvp->v_vflag |= VV_CACHEDLABEL; else { #if 0 /* * In theory, we could have fall-back behavior here. * It would probably be incorrect. */ #endif return (error); } } return (0); } void mac_execve_transition(struct ucred *old, struct ucred *new, struct vnode *vp) { int error; ASSERT_VOP_LOCKED(vp, "mac_execve_transition"); error = vn_refreshlabel(vp, old); if (error) { printf("mac_execve_transition: vn_refreshlabel returned %d\n", error); printf("mac_execve_transition: using old vnode label\n"); } MAC_PERFORM(execve_transition, old, new, vp, &vp->v_label); } int mac_execve_will_transition(struct ucred *old, struct vnode *vp) { int error, result; error = vn_refreshlabel(vp, old); if (error) return (error); result = 0; MAC_BOOLEAN(execve_will_transition, ||, old, vp, &vp->v_label); return (result); } static void mac_init_label(struct label *label) { bzero(label, sizeof(*label)); label->l_flags = MAC_FLAG_INITIALIZED; } static void mac_init_structmac(struct mac *mac) { bzero(mac, sizeof(*mac)); mac->m_macflags = MAC_FLAG_INITIALIZED; } static void mac_destroy_label(struct label *label) { KASSERT(label->l_flags & MAC_FLAG_INITIALIZED, ("destroying uninitialized label")); bzero(label, sizeof(*label)); /* implicit: label->l_flags &= ~MAC_FLAG_INITIALIZED; */ } int mac_init_mbuf(struct mbuf *m, int how) { KASSERT(m->m_flags & M_PKTHDR, ("mac_init_mbuf on non-header mbuf")); /* "how" is one of M_(TRY|DONT)WAIT */ mac_init_label(&m->m_pkthdr.label); MAC_PERFORM(init_mbuf, m, how, &m->m_pkthdr.label); #ifdef MAC_DEBUG atomic_add_int(&nmacmbufs, 1); #endif return (0); } void mac_destroy_mbuf(struct mbuf *m) { MAC_PERFORM(destroy_mbuf, m, &m->m_pkthdr.label); mac_destroy_label(&m->m_pkthdr.label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacmbufs, 1); #endif } void mac_init_cred(struct ucred *cr) { mac_init_label(&cr->cr_label); MAC_PERFORM(init_cred, cr, &cr->cr_label); #ifdef MAC_DEBUG atomic_add_int(&nmaccreds, 1); #endif } void mac_destroy_cred(struct ucred *cr) { MAC_PERFORM(destroy_cred, cr, &cr->cr_label); mac_destroy_label(&cr->cr_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmaccreds, 1); #endif } void mac_init_ifnet(struct ifnet *ifp) { mac_init_label(&ifp->if_label); MAC_PERFORM(init_ifnet, ifp, &ifp->if_label); #ifdef MAC_DEBUG atomic_add_int(&nmacifnets, 1); #endif } void mac_destroy_ifnet(struct ifnet *ifp) { MAC_PERFORM(destroy_ifnet, ifp, &ifp->if_label); mac_destroy_label(&ifp->if_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacifnets, 1); #endif } void mac_init_ipq(struct ipq *ipq) { mac_init_label(&ipq->ipq_label); MAC_PERFORM(init_ipq, ipq, &ipq->ipq_label); #ifdef MAC_DEBUG atomic_add_int(&nmacipqs, 1); #endif } void mac_destroy_ipq(struct ipq *ipq) { MAC_PERFORM(destroy_ipq, ipq, &ipq->ipq_label); mac_destroy_label(&ipq->ipq_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacipqs, 1); #endif } void mac_init_socket(struct socket *socket) { mac_init_label(&socket->so_label); mac_init_label(&socket->so_peerlabel); MAC_PERFORM(init_socket, socket, &socket->so_label, &socket->so_peerlabel); #ifdef MAC_DEBUG atomic_add_int(&nmacsockets, 1); #endif } void mac_destroy_socket(struct socket *socket) { MAC_PERFORM(destroy_socket, socket, &socket->so_label, &socket->so_peerlabel); mac_destroy_label(&socket->so_label); mac_destroy_label(&socket->so_peerlabel); #ifdef MAC_DEBUG atomic_subtract_int(&nmacsockets, 1); #endif } void mac_init_pipe(struct pipe *pipe) { struct label *label; label = malloc(sizeof(struct label), M_MACPIPELABEL, M_ZERO|M_WAITOK); mac_init_label(label); pipe->pipe_label = label; pipe->pipe_peer->pipe_label = label; MAC_PERFORM(init_pipe, pipe, pipe->pipe_label); #ifdef MAC_DEBUG atomic_add_int(&nmacpipes, 1); #endif } void mac_destroy_pipe(struct pipe *pipe) { MAC_PERFORM(destroy_pipe, pipe, pipe->pipe_label); mac_destroy_label(pipe->pipe_label); free(pipe->pipe_label, M_MACPIPELABEL); #ifdef MAC_DEBUG atomic_subtract_int(&nmacpipes, 1); #endif } void mac_init_bpfdesc(struct bpf_d *bpf_d) { mac_init_label(&bpf_d->bd_label); MAC_PERFORM(init_bpfdesc, bpf_d, &bpf_d->bd_label); #ifdef MAC_DEBUG atomic_add_int(&nmacbpfdescs, 1); #endif } void mac_destroy_bpfdesc(struct bpf_d *bpf_d) { MAC_PERFORM(destroy_bpfdesc, bpf_d, &bpf_d->bd_label); mac_destroy_label(&bpf_d->bd_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacbpfdescs, 1); #endif } void mac_init_mount(struct mount *mp) { mac_init_label(&mp->mnt_mntlabel); mac_init_label(&mp->mnt_fslabel); MAC_PERFORM(init_mount, mp, &mp->mnt_mntlabel, &mp->mnt_fslabel); #ifdef MAC_DEBUG atomic_add_int(&nmacmounts, 1); #endif } void mac_destroy_mount(struct mount *mp) { MAC_PERFORM(destroy_mount, mp, &mp->mnt_mntlabel, &mp->mnt_fslabel); mac_destroy_label(&mp->mnt_fslabel); mac_destroy_label(&mp->mnt_mntlabel); #ifdef MAC_DEBUG atomic_subtract_int(&nmacmounts, 1); #endif } static void mac_init_temp(struct label *label) { mac_init_label(label); MAC_PERFORM(init_temp, label); #ifdef MAC_DEBUG atomic_add_int(&nmactemp, 1); #endif } static void mac_destroy_temp(struct label *label) { MAC_PERFORM(destroy_temp, label); mac_destroy_label(label); #ifdef MAC_DEBUG atomic_subtract_int(&nmactemp, 1); #endif } void mac_init_vnode(struct vnode *vp) { mac_init_label(&vp->v_label); MAC_PERFORM(init_vnode, vp, &vp->v_label); #ifdef MAC_DEBUG atomic_add_int(&nmacvnodes, 1); #endif } void mac_destroy_vnode(struct vnode *vp) { MAC_PERFORM(destroy_vnode, vp, &vp->v_label); mac_destroy_label(&vp->v_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacvnodes, 1); #endif } void mac_init_devfsdirent(struct devfs_dirent *de) { mac_init_label(&de->de_label); MAC_PERFORM(init_devfsdirent, de, &de->de_label); #ifdef MAC_DEBUG atomic_add_int(&nmacdevfsdirents, 1); #endif } void mac_destroy_devfsdirent(struct devfs_dirent *de) { MAC_PERFORM(destroy_devfsdirent, de, &de->de_label); mac_destroy_label(&de->de_label); #ifdef MAC_DEBUG atomic_subtract_int(&nmacdevfsdirents, 1); #endif } static int mac_externalize(struct label *label, struct mac *mac) { int error; mac_init_structmac(mac); MAC_CHECK(externalize, label, mac); return (error); } static int mac_internalize(struct label *label, struct mac *mac) { int error; mac_init_temp(label); MAC_CHECK(internalize, label, mac); if (error) mac_destroy_temp(label); return (error); } /* * Initialize MAC label for the first kernel process, from which other * kernel processes and threads are spawned. */ void mac_create_proc0(struct ucred *cred) { MAC_PERFORM(create_proc0, cred); } /* * Initialize MAC label for the first userland process, from which other * userland processes and threads are spawned. */ void mac_create_proc1(struct ucred *cred) { MAC_PERFORM(create_proc1, cred); } /* * When a new process is created, its label must be initialized. Generally, * this involves inheritence from the parent process, modulo possible * deltas. This function allows that processing to take place. */ void mac_create_cred(struct ucred *parent_cred, struct ucred *child_cred) { MAC_PERFORM(create_cred, parent_cred, child_cred); } int mac_check_vnode_access(struct ucred *cred, struct vnode *vp, int flags) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_access"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_access, cred, vp, &vp->v_label, flags); return (error); } int mac_check_vnode_chdir(struct ucred *cred, struct vnode *dvp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_chdir"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); MAC_CHECK(check_vnode_chdir, cred, dvp, &dvp->v_label); return (error); } int mac_check_vnode_chroot(struct ucred *cred, struct vnode *dvp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_chroot"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); MAC_CHECK(check_vnode_chroot, cred, dvp, &dvp->v_label); return (error); } int mac_check_vnode_create(struct ucred *cred, struct vnode *dvp, struct componentname *cnp, struct vattr *vap) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_create"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); MAC_CHECK(check_vnode_create, cred, dvp, &dvp->v_label, cnp, vap); return (error); } int mac_check_vnode_delete(struct ucred *cred, struct vnode *dvp, struct vnode *vp, struct componentname *cnp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_delete"); ASSERT_VOP_LOCKED(vp, "mac_check_vnode_delete"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_delete, cred, dvp, &dvp->v_label, vp, &vp->v_label, cnp); return (error); } int mac_check_vnode_deleteacl(struct ucred *cred, struct vnode *vp, acl_type_t type) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_deleteacl"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_deleteacl, cred, vp, &vp->v_label, type); return (error); } int mac_check_vnode_exec(struct ucred *cred, struct vnode *vp) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_exec"); if (!mac_enforce_process && !mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_exec, cred, vp, &vp->v_label); return (error); } int mac_check_vnode_getacl(struct ucred *cred, struct vnode *vp, acl_type_t type) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_getacl"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_getacl, cred, vp, &vp->v_label, type); return (error); } int mac_check_vnode_getextattr(struct ucred *cred, struct vnode *vp, int attrnamespace, const char *name, struct uio *uio) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_getextattr"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_getextattr, cred, vp, &vp->v_label, attrnamespace, name, uio); return (error); } int mac_check_vnode_lookup(struct ucred *cred, struct vnode *dvp, struct componentname *cnp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_lookup"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); MAC_CHECK(check_vnode_lookup, cred, dvp, &dvp->v_label, cnp); return (error); } vm_prot_t mac_check_vnode_mmap_prot(struct ucred *cred, struct vnode *vp, int newmapping) { vm_prot_t result = VM_PROT_ALL; /* * This should be some sort of MAC_BITWISE, maybe :) */ ASSERT_VOP_LOCKED(vp, "mac_check_vnode_mmap_perms"); MAC_BOOLEAN(check_vnode_mmap_perms, &, cred, vp, &vp->v_label, newmapping); return (result); } int mac_check_vnode_op(struct ucred *cred, struct vnode *vp, int op) { int error; if (!mac_enforce_fs) return (0); ASSERT_VOP_LOCKED(vp, "mac_check_vnode_op"); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_op, cred, vp, &vp->v_label, op); return (error); } int mac_check_vnode_open(struct ucred *cred, struct vnode *vp, mode_t acc_mode) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_open"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_open, cred, vp, &vp->v_label, acc_mode); return (error); } int mac_check_vnode_readdir(struct ucred *cred, struct vnode *dvp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_readdir"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); MAC_CHECK(check_vnode_readdir, cred, dvp, &dvp->v_label); return (error); } int mac_check_vnode_readlink(struct ucred *cred, struct vnode *vp) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_readlink"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_readlink, cred, vp, &vp->v_label); return (error); } static int mac_check_vnode_relabel(struct ucred *cred, struct vnode *vp, struct label *newlabel) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_relabel"); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_relabel, cred, vp, &vp->v_label, newlabel); return (error); } int mac_check_vnode_rename_from(struct ucred *cred, struct vnode *dvp, struct vnode *vp, struct componentname *cnp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_rename_from"); ASSERT_VOP_LOCKED(vp, "mac_check_vnode_rename_from"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_rename_from, cred, dvp, &dvp->v_label, vp, &vp->v_label, cnp); return (error); } int mac_check_vnode_rename_to(struct ucred *cred, struct vnode *dvp, struct vnode *vp, int samedir, struct componentname *cnp) { int error; ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_rename_to"); ASSERT_VOP_LOCKED(vp, "mac_check_vnode_rename_to"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(dvp, cred); if (error) return (error); if (vp != NULL) { error = vn_refreshlabel(vp, cred); if (error) return (error); } MAC_CHECK(check_vnode_rename_to, cred, dvp, &dvp->v_label, vp, vp != NULL ? &vp->v_label : NULL, samedir, cnp); return (error); } int mac_check_vnode_revoke(struct ucred *cred, struct vnode *vp) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_revoke"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_revoke, cred, vp, &vp->v_label); return (error); } int mac_check_vnode_setacl(struct ucred *cred, struct vnode *vp, acl_type_t type, struct acl *acl) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setacl"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setacl, cred, vp, &vp->v_label, type, acl); return (error); } int mac_check_vnode_setextattr(struct ucred *cred, struct vnode *vp, int attrnamespace, const char *name, struct uio *uio) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setextattr"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setextattr, cred, vp, &vp->v_label, attrnamespace, name, uio); return (error); } int mac_check_vnode_setflags(struct ucred *cred, struct vnode *vp, u_long flags) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setflags"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setflags, cred, vp, &vp->v_label, flags); return (error); } int mac_check_vnode_setmode(struct ucred *cred, struct vnode *vp, mode_t mode) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setmode"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setmode, cred, vp, &vp->v_label, mode); return (error); } int mac_check_vnode_setowner(struct ucred *cred, struct vnode *vp, uid_t uid, gid_t gid) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setowner"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setowner, cred, vp, &vp->v_label, uid, gid); return (error); } int mac_check_vnode_setutimes(struct ucred *cred, struct vnode *vp, struct timespec atime, struct timespec mtime) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setutimes"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_setutimes, cred, vp, &vp->v_label, atime, mtime); return (error); } int mac_check_vnode_stat(struct ucred *cred, struct vnode *vp) { int error; ASSERT_VOP_LOCKED(vp, "mac_check_vnode_stat"); if (!mac_enforce_fs) return (0); error = vn_refreshlabel(vp, cred); if (error) return (error); MAC_CHECK(check_vnode_stat, cred, vp, &vp->v_label); return (error); } /* * When relabeling a process, call out to the policies for the maximum * permission allowed for each object type we know about in its * memory space, and revoke access (in the least surprising ways we * know) when necessary. The process lock is not held here. */ static void mac_cred_mmapped_drop_perms(struct thread *td, struct ucred *cred) { /* XXX freeze all other threads */ mtx_lock(&Giant); mac_cred_mmapped_drop_perms_recurse(td, cred, &td->td_proc->p_vmspace->vm_map); mtx_unlock(&Giant); /* XXX allow other threads to continue */ } static __inline const char * prot2str(vm_prot_t prot) { switch (prot & VM_PROT_ALL) { case VM_PROT_READ: return ("r--"); case VM_PROT_READ | VM_PROT_WRITE: return ("rw-"); case VM_PROT_READ | VM_PROT_EXECUTE: return ("r-x"); case VM_PROT_READ | VM_PROT_WRITE | VM_PROT_EXECUTE: return ("rwx"); case VM_PROT_WRITE: return ("-w-"); case VM_PROT_EXECUTE: return ("--x"); case VM_PROT_WRITE | VM_PROT_EXECUTE: return ("-wx"); default: return ("---"); } } static void mac_cred_mmapped_drop_perms_recurse(struct thread *td, struct ucred *cred, struct vm_map *map) { struct vm_map_entry *vme; vm_prot_t result, revokeperms; vm_object_t object; vm_ooffset_t offset; struct vnode *vp; vm_map_lock_read(map); for (vme = map->header.next; vme != &map->header; vme = vme->next) { if (vme->eflags & MAP_ENTRY_IS_SUB_MAP) { mac_cred_mmapped_drop_perms_recurse(td, cred, vme->object.sub_map); continue; } /* * Skip over entries that obviously are not shared. */ if (vme->eflags & (MAP_ENTRY_COW | MAP_ENTRY_NOSYNC) || !vme->max_protection) continue; /* * Drill down to the deepest backing object. */ offset = vme->offset; object = vme->object.vm_object; if (object == NULL) continue; while (object->backing_object != NULL) { object = object->backing_object; offset += object->backing_object_offset; } /* * At the moment, vm_maps and objects aren't considered * by the MAC system, so only things with backing by a * normal object (read: vnodes) are checked. */ if (object->type != OBJT_VNODE) continue; vp = (struct vnode *)object->handle; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); result = mac_check_vnode_mmap_prot(cred, vp, 0); VOP_UNLOCK(vp, 0, td); /* * Find out what maximum protection we may be allowing * now but a policy needs to get removed. */ revokeperms = vme->max_protection & ~result; if (!revokeperms) continue; printf("pid %d: revoking %s perms from %#lx:%d " "(max %s/cur %s)\n", td->td_proc->p_pid, prot2str(revokeperms), vme->start, vme->end - vme->start, prot2str(vme->max_protection), prot2str(vme->protection)); vm_map_lock_upgrade(map); /* * This is the really simple case: if a map has more * max_protection than is allowed, but it's not being * actually used (that is, the current protection is * still allowed), we can just wipe it out and do * nothing more. */ if ((vme->protection & revokeperms) == 0) { vme->max_protection -= revokeperms; } else { if (revokeperms & VM_PROT_WRITE) { /* * In the more complicated case, flush out all * pending changes to the object then turn it * copy-on-write. */ vm_object_reference(object); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); vm_object_page_clean(object, OFF_TO_IDX(offset), OFF_TO_IDX(offset + vme->end - vme->start + PAGE_MASK), OBJPC_SYNC); VOP_UNLOCK(vp, 0, td); vm_object_deallocate(object); /* * Why bother if there's no read permissions * anymore? For the rest, we need to leave * the write permissions on for COW, or * remove them entirely if configured to. */ if (!mac_mmap_revocation_via_cow) { vme->max_protection &= ~VM_PROT_WRITE; vme->protection &= ~VM_PROT_WRITE; } if ((revokeperms & VM_PROT_READ) == 0) vme->eflags |= MAP_ENTRY_COW | MAP_ENTRY_NEEDS_COPY; } if (revokeperms & VM_PROT_EXECUTE) { vme->max_protection &= ~VM_PROT_EXECUTE; vme->protection &= ~VM_PROT_EXECUTE; } if (revokeperms & VM_PROT_READ) { vme->max_protection = 0; vme->protection = 0; } pmap_protect(map->pmap, vme->start, vme->end, vme->protection & ~revokeperms); vm_map_simplify_entry(map, vme); } vm_map_lock_downgrade(map); } vm_map_unlock_read(map); } /* * When the subject's label changes, it may require revocation of privilege * to mapped objects. This can't be done on-the-fly later with a unified * buffer cache. */ static void mac_relabel_cred(struct ucred *cred, struct label *newlabel) { MAC_PERFORM(relabel_cred, cred, newlabel); mac_cred_mmapped_drop_perms(curthread, cred); } void mac_relabel_vnode(struct ucred *cred, struct vnode *vp, struct label *newlabel) { MAC_PERFORM(relabel_vnode, cred, vp, &vp->v_label, newlabel); } void mac_create_ifnet(struct ifnet *ifnet) { MAC_PERFORM(create_ifnet, ifnet, &ifnet->if_label); } void mac_create_bpfdesc(struct ucred *cred, struct bpf_d *bpf_d) { MAC_PERFORM(create_bpfdesc, cred, bpf_d, &bpf_d->bd_label); } void mac_create_socket(struct ucred *cred, struct socket *socket) { MAC_PERFORM(create_socket, cred, socket, &socket->so_label); } void mac_create_pipe(struct ucred *cred, struct pipe *pipe) { MAC_PERFORM(create_pipe, cred, pipe, pipe->pipe_label); } void mac_create_socket_from_socket(struct socket *oldsocket, struct socket *newsocket) { MAC_PERFORM(create_socket_from_socket, oldsocket, &oldsocket->so_label, newsocket, &newsocket->so_label); } static void mac_relabel_socket(struct ucred *cred, struct socket *socket, struct label *newlabel) { MAC_PERFORM(relabel_socket, cred, socket, &socket->so_label, newlabel); } static void mac_relabel_pipe(struct ucred *cred, struct pipe *pipe, struct label *newlabel) { MAC_PERFORM(relabel_pipe, cred, pipe, pipe->pipe_label, newlabel); } void mac_set_socket_peer_from_mbuf(struct mbuf *mbuf, struct socket *socket) { MAC_PERFORM(set_socket_peer_from_mbuf, mbuf, &mbuf->m_pkthdr.label, socket, &socket->so_peerlabel); } void mac_set_socket_peer_from_socket(struct socket *oldsocket, struct socket *newsocket) { MAC_PERFORM(set_socket_peer_from_socket, oldsocket, &oldsocket->so_label, newsocket, &newsocket->so_peerlabel); } void mac_create_datagram_from_ipq(struct ipq *ipq, struct mbuf *datagram) { MAC_PERFORM(create_datagram_from_ipq, ipq, &ipq->ipq_label, datagram, &datagram->m_pkthdr.label); } void mac_create_fragment(struct mbuf *datagram, struct mbuf *fragment) { MAC_PERFORM(create_fragment, datagram, &datagram->m_pkthdr.label, fragment, &fragment->m_pkthdr.label); } void mac_create_ipq(struct mbuf *fragment, struct ipq *ipq) { MAC_PERFORM(create_ipq, fragment, &fragment->m_pkthdr.label, ipq, &ipq->ipq_label); } void mac_create_mbuf_from_mbuf(struct mbuf *oldmbuf, struct mbuf *newmbuf) { MAC_PERFORM(create_mbuf_from_mbuf, oldmbuf, &oldmbuf->m_pkthdr.label, newmbuf, &newmbuf->m_pkthdr.label); } void mac_create_mbuf_from_bpfdesc(struct bpf_d *bpf_d, struct mbuf *mbuf) { MAC_PERFORM(create_mbuf_from_bpfdesc, bpf_d, &bpf_d->bd_label, mbuf, &mbuf->m_pkthdr.label); } void mac_create_mbuf_linklayer(struct ifnet *ifnet, struct mbuf *mbuf) { MAC_PERFORM(create_mbuf_linklayer, ifnet, &ifnet->if_label, mbuf, &mbuf->m_pkthdr.label); } void mac_create_mbuf_from_ifnet(struct ifnet *ifnet, struct mbuf *mbuf) { MAC_PERFORM(create_mbuf_from_ifnet, ifnet, &ifnet->if_label, mbuf, &mbuf->m_pkthdr.label); } void mac_create_mbuf_multicast_encap(struct mbuf *oldmbuf, struct ifnet *ifnet, struct mbuf *newmbuf) { MAC_PERFORM(create_mbuf_multicast_encap, oldmbuf, &oldmbuf->m_pkthdr.label, ifnet, &ifnet->if_label, newmbuf, &newmbuf->m_pkthdr.label); } void mac_create_mbuf_netlayer(struct mbuf *oldmbuf, struct mbuf *newmbuf) { MAC_PERFORM(create_mbuf_netlayer, oldmbuf, &oldmbuf->m_pkthdr.label, newmbuf, &newmbuf->m_pkthdr.label); } int mac_fragment_match(struct mbuf *fragment, struct ipq *ipq) { int result; result = 1; MAC_BOOLEAN(fragment_match, &&, fragment, &fragment->m_pkthdr.label, ipq, &ipq->ipq_label); return (result); } void mac_update_ipq(struct mbuf *fragment, struct ipq *ipq) { MAC_PERFORM(update_ipq, fragment, &fragment->m_pkthdr.label, ipq, &ipq->ipq_label); } void mac_create_mbuf_from_socket(struct socket *socket, struct mbuf *mbuf) { MAC_PERFORM(create_mbuf_from_socket, socket, &socket->so_label, mbuf, &mbuf->m_pkthdr.label); } void mac_create_mount(struct ucred *cred, struct mount *mp) { MAC_PERFORM(create_mount, cred, mp, &mp->mnt_mntlabel, &mp->mnt_fslabel); } void mac_create_root_mount(struct ucred *cred, struct mount *mp) { MAC_PERFORM(create_root_mount, cred, mp, &mp->mnt_mntlabel, &mp->mnt_fslabel); } int mac_check_bpfdesc_receive(struct bpf_d *bpf_d, struct ifnet *ifnet) { int error; if (!mac_enforce_network) return (0); MAC_CHECK(check_bpfdesc_receive, bpf_d, &bpf_d->bd_label, ifnet, &ifnet->if_label); return (error); } static int mac_check_cred_relabel(struct ucred *cred, struct label *newlabel) { int error; MAC_CHECK(check_cred_relabel, cred, newlabel); return (error); } int mac_check_cred_visible(struct ucred *u1, struct ucred *u2) { int error; if (!mac_enforce_process) return (0); MAC_CHECK(check_cred_visible, u1, u2); return (error); } int mac_check_ifnet_transmit(struct ifnet *ifnet, struct mbuf *mbuf) { int error; if (!mac_enforce_network) return (0); KASSERT(mbuf->m_flags & M_PKTHDR, ("packet has no pkthdr")); if (!(mbuf->m_pkthdr.label.l_flags & MAC_FLAG_INITIALIZED)) printf("%s%d: not initialized\n", ifnet->if_name, ifnet->if_unit); MAC_CHECK(check_ifnet_transmit, ifnet, &ifnet->if_label, mbuf, &mbuf->m_pkthdr.label); return (error); } int mac_check_mount_stat(struct ucred *cred, struct mount *mount) { int error; if (!mac_enforce_fs) return (0); MAC_CHECK(check_mount_stat, cred, mount, &mount->mnt_mntlabel); return (error); } int mac_check_pipe_ioctl(struct ucred *cred, struct pipe *pipe, unsigned long cmd, void *data) { int error; MAC_CHECK(check_pipe_ioctl, cred, pipe, pipe->pipe_label, cmd, data); return (error); } int mac_check_pipe_op(struct ucred *cred, struct pipe *pipe, int op) { int error; MAC_CHECK(check_pipe_op, cred, pipe, pipe->pipe_label, op); return (error); } static int mac_check_pipe_relabel(struct ucred *cred, struct pipe *pipe, struct label *newlabel) { int error; MAC_CHECK(check_pipe_relabel, cred, pipe, pipe->pipe_label, newlabel); return (error); } int mac_check_proc_debug(struct ucred *cred, struct proc *proc) { int error; PROC_LOCK_ASSERT(proc, MA_OWNED); if (!mac_enforce_process) return (0); MAC_CHECK(check_proc_debug, cred, proc); return (error); } int mac_check_proc_sched(struct ucred *cred, struct proc *proc) { int error; PROC_LOCK_ASSERT(proc, MA_OWNED); if (!mac_enforce_process) return (0); MAC_CHECK(check_proc_sched, cred, proc); return (error); } int mac_check_proc_signal(struct ucred *cred, struct proc *proc, int signum) { int error; PROC_LOCK_ASSERT(proc, MA_OWNED); if (!mac_enforce_process) return (0); MAC_CHECK(check_proc_signal, cred, proc, signum); return (error); } int mac_check_socket_bind(struct ucred *ucred, struct socket *socket, struct sockaddr *sockaddr) { int error; if (!mac_enforce_socket) return (0); MAC_CHECK(check_socket_bind, ucred, socket, &socket->so_label, sockaddr); return (error); } int mac_check_socket_connect(struct ucred *cred, struct socket *socket, struct sockaddr *sockaddr) { int error; if (!mac_enforce_socket) return (0); MAC_CHECK(check_socket_connect, cred, socket, &socket->so_label, sockaddr); return (error); } int mac_check_socket_deliver(struct socket *socket, struct mbuf *mbuf) { int error; if (!mac_enforce_socket) return (0); MAC_CHECK(check_socket_deliver, socket, &socket->so_label, mbuf, &mbuf->m_pkthdr.label); return (error); } int mac_check_socket_listen(struct ucred *cred, struct socket *socket) { int error; if (!mac_enforce_socket) return (0); MAC_CHECK(check_socket_listen, cred, socket, &socket->so_label); return (error); } static int mac_check_socket_relabel(struct ucred *cred, struct socket *socket, struct label *newlabel) { int error; MAC_CHECK(check_socket_relabel, cred, socket, &socket->so_label, newlabel); return (error); } int mac_check_socket_visible(struct ucred *cred, struct socket *socket) { int error; if (!mac_enforce_socket) return (0); MAC_CHECK(check_socket_visible, cred, socket, &socket->so_label); return (error); } int mac_ioctl_ifnet_get(struct ucred *cred, struct ifreq *ifr, struct ifnet *ifnet) { struct mac label; int error; error = mac_externalize(&ifnet->if_label, &label); if (error) return (error); return (copyout(&label, ifr->ifr_ifru.ifru_data, sizeof(label))); } int mac_ioctl_ifnet_set(struct ucred *cred, struct ifreq *ifr, struct ifnet *ifnet) { struct mac newlabel; struct label intlabel; int error; error = copyin(ifr->ifr_ifru.ifru_data, &newlabel, sizeof(newlabel)); if (error) return (error); error = mac_internalize(&intlabel, &newlabel); if (error) return (error); /* * XXX: Note that this is a redundant privilege check, since * policies impose this check themselves if required by the * policy. Eventually, this should go away. */ error = suser_cred(cred, 0); if (error) goto out; MAC_CHECK(check_ifnet_relabel, cred, ifnet, &ifnet->if_label, &intlabel); if (error) goto out; MAC_PERFORM(relabel_ifnet, cred, ifnet, &ifnet->if_label, &intlabel); out: mac_destroy_temp(&intlabel); return (error); } void mac_create_devfs_vnode(struct devfs_dirent *de, struct vnode *vp) { MAC_PERFORM(create_devfs_vnode, de, &de->de_label, vp, &vp->v_label); } void mac_create_devfs_device(dev_t dev, struct devfs_dirent *de) { MAC_PERFORM(create_devfs_device, dev, de, &de->de_label); } static int mac_stdcreatevnode_ea(struct vnode *vp) { int error; MAC_CHECK(stdcreatevnode_ea, vp, &vp->v_label); return (error); } void mac_create_devfs_directory(char *dirname, int dirnamelen, struct devfs_dirent *de) { MAC_PERFORM(create_devfs_directory, dirname, dirnamelen, de, &de->de_label); } /* * When a new vnode is created, this call will initialize its label. */ void mac_create_vnode(struct ucred *cred, struct vnode *parent, struct vnode *child) { int error; ASSERT_VOP_LOCKED(parent, "mac_create_vnode"); ASSERT_VOP_LOCKED(child, "mac_create_vnode"); error = vn_refreshlabel(parent, cred); if (error) { printf("mac_create_vnode: vn_refreshlabel returned %d\n", error); printf("mac_create_vnode: using old vnode label\n"); } MAC_PERFORM(create_vnode, cred, parent, &parent->v_label, child, &child->v_label); } int mac_setsockopt_label_set(struct ucred *cred, struct socket *so, struct mac *extmac) { struct label intlabel; int error; error = mac_internalize(&intlabel, extmac); if (error) return (error); mac_check_socket_relabel(cred, so, &intlabel); if (error) { mac_destroy_temp(&intlabel); return (error); } mac_relabel_socket(cred, so, &intlabel); mac_destroy_temp(&intlabel); return (0); } int mac_pipe_label_set(struct ucred *cred, struct pipe *pipe, struct label *label) { int error; error = mac_check_pipe_relabel(cred, pipe, label); if (error) return (error); mac_relabel_pipe(cred, pipe, label); return (0); } int mac_getsockopt_label_get(struct ucred *cred, struct socket *so, struct mac *extmac) { return (mac_externalize(&so->so_label, extmac)); } int mac_getsockopt_peerlabel_get(struct ucred *cred, struct socket *so, struct mac *extmac) { return (mac_externalize(&so->so_peerlabel, extmac)); } /* * Implementation of VOP_SETLABEL() that relies on extended attributes * to store label data. Can be referenced by filesystems supporting * extended attributes. */ int vop_stdsetlabel_ea(struct vop_setlabel_args *ap) { struct vnode *vp = ap->a_vp; struct label *intlabel = ap->a_label; struct mac extmac; int error; ASSERT_VOP_LOCKED(vp, "vop_stdsetlabel_ea"); /* * XXX: Eventually call out to EA check/set calls here. * Be particularly careful to avoid race conditions, * consistency problems, and stability problems when * dealing with multiple EAs. In particular, we require * the ability to write multiple EAs on the same file in * a single transaction, which the current EA interface * does not provide. */ error = mac_externalize(intlabel, &extmac); if (error) return (error); error = vn_extattr_set(vp, IO_NODELOCKED, FREEBSD_MAC_EXTATTR_NAMESPACE, FREEBSD_MAC_EXTATTR_NAME, sizeof(extmac), (char *)&extmac, curthread); if (error) return (error); mac_relabel_vnode(ap->a_cred, vp, intlabel); vp->v_vflag |= VV_CACHEDLABEL; return (0); } static int vn_setlabel(struct vnode *vp, struct label *intlabel, struct ucred *cred) { int error; if (vp->v_mount == NULL) { /* printf("vn_setlabel: null v_mount\n"); */ if (vp->v_tag != VT_NON) printf("vn_setlabel: null v_mount with non-VT_NON\n"); return (EBADF); } if ((vp->v_mount->mnt_flag & MNT_MULTILABEL) == 0) return (EOPNOTSUPP); /* * Multi-phase commit. First check the policies to confirm the * change is OK. Then commit via the filesystem. Finally, * update the actual vnode label. Question: maybe the filesystem * should update the vnode at the end as part of VOP_SETLABEL()? */ error = mac_check_vnode_relabel(cred, vp, intlabel); if (error) return (error); /* * VADMIN provides the opportunity for the filesystem to make * decisions about who is and is not able to modify labels * and protections on files. This might not be right. We can't * assume VOP_SETLABEL() will do it, because we might implement * that as part of vop_stdsetlabel_ea(). */ error = VOP_ACCESS(vp, VADMIN, cred, curthread); if (error) return (error); error = VOP_SETLABEL(vp, intlabel, cred, curthread); if (error) return (error); return (0); } /* * MPSAFE */ int __mac_get_proc(struct thread *td, struct __mac_get_proc_args *uap) { struct mac extmac; int error; error = mac_externalize(&td->td_ucred->cr_label, &extmac); if (error == 0) error = copyout(&extmac, SCARG(uap, mac_p), sizeof(extmac)); return (error); } /* * MPSAFE * * XXX: Needs to be re-written for proc locking. */ int __mac_set_proc(struct thread *td, struct __mac_set_proc_args *uap) { struct ucred *newcred, *oldcred; struct proc *p; struct mac extmac; struct label intlabel; int error; error = copyin(SCARG(uap, mac_p), &extmac, sizeof(extmac)); if (error) return (error); error = mac_internalize(&intlabel, &extmac); if (error) return (error); newcred = crget(); p = td->td_proc; PROC_LOCK(p); oldcred = p->p_ucred; error = mac_check_cred_relabel(oldcred, &intlabel); if (error) { PROC_UNLOCK(p); mac_destroy_temp(&intlabel); crfree(newcred); return (error); } setsugid(p); crcopy(newcred, oldcred); PROC_UNLOCK(p); mac_relabel_cred(newcred, &intlabel); PROC_LOCK(p); p->p_ucred = newcred; PROC_UNLOCK(p); crfree(oldcred); mac_destroy_temp(&intlabel); return (0); } /* * MPSAFE */ int __mac_get_fd(struct thread *td, struct __mac_get_fd_args *uap) { struct file *fp; struct mac extmac; struct vnode *vp; struct pipe *pipe; int error; mtx_lock(&Giant); error = fget(td, SCARG(uap, fd), &fp); if (error) goto out; switch (fp->f_type) { case DTYPE_FIFO: case DTYPE_VNODE: vp = (struct vnode *)fp->f_data; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); error = vn_refreshlabel(vp, td->td_ucred); if (error == 0) error = mac_externalize(&vp->v_label, &extmac); VOP_UNLOCK(vp, 0, td); break; case DTYPE_PIPE: pipe = (struct pipe *)fp->f_data; error = mac_externalize(pipe->pipe_label, &extmac); break; default: error = EINVAL; } if (error == 0) error = copyout(&extmac, SCARG(uap, mac_p), sizeof(extmac)); fdrop(fp, td); out: mtx_unlock(&Giant); return (error); } /* * MPSAFE */ int __mac_get_file(struct thread *td, struct __mac_get_file_args *uap) { struct nameidata nd; struct mac extmac; int error; mtx_lock(&Giant); NDINIT(&nd, LOOKUP, LOCKLEAF | FOLLOW, UIO_USERSPACE, SCARG(uap, path_p), td); error = namei(&nd); if (error) goto out; error = vn_refreshlabel(nd.ni_vp, td->td_ucred); if (error == 0) error = mac_externalize(&nd.ni_vp->v_label, &extmac); NDFREE(&nd, 0); if (error) goto out; error = copyout(&extmac, SCARG(uap, mac_p), sizeof(extmac)); out: mtx_unlock(&Giant); return (error); } /* * MPSAFE */ int __mac_set_fd(struct thread *td, struct __mac_set_fd_args *uap) { struct file *fp; struct mac extmac; struct label intlabel; struct mount *mp; struct vnode *vp; struct pipe *pipe; int error; mtx_lock(&Giant); error = fget(td, SCARG(uap, fd), &fp); if (error) goto out1; error = copyin(SCARG(uap, mac_p), &extmac, sizeof(extmac)); if (error) goto out2; error = mac_internalize(&intlabel, &extmac); if (error) goto out2; switch (fp->f_type) { case DTYPE_FIFO: case DTYPE_VNODE: vp = (struct vnode *)fp->f_data; error = vn_start_write(vp, &mp, V_WAIT | PCATCH); if (error != 0) break; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); error = vn_setlabel(vp, &intlabel, td->td_ucred); VOP_UNLOCK(vp, 0, td); vn_finished_write(mp); mac_destroy_temp(&intlabel); break; case DTYPE_PIPE: pipe = (struct pipe *)fp->f_data; error = mac_pipe_label_set(td->td_ucred, pipe, &intlabel); break; default: error = EINVAL; } out2: fdrop(fp, td); out1: mtx_unlock(&Giant); return (error); } /* * MPSAFE */ int __mac_set_file(struct thread *td, struct __mac_set_file_args *uap) { struct nameidata nd; struct mac extmac; struct label intlabel; struct mount *mp; int error; mtx_lock(&Giant); error = copyin(SCARG(uap, mac_p), &extmac, sizeof(extmac)); if (error) goto out; error = mac_internalize(&intlabel, &extmac); if (error) goto out; NDINIT(&nd, LOOKUP, LOCKLEAF | FOLLOW, UIO_USERSPACE, SCARG(uap, path_p), td); error = namei(&nd); if (error) goto out2; error = vn_start_write(nd.ni_vp, &mp, V_WAIT | PCATCH); if (error) goto out2; error = vn_setlabel(nd.ni_vp, &intlabel, td->td_ucred); vn_finished_write(mp); out2: mac_destroy_temp(&intlabel); NDFREE(&nd, 0); out: mtx_unlock(&Giant); return (error); } SYSINIT(mac, SI_SUB_MAC, SI_ORDER_FIRST, mac_init, NULL); SYSINIT(mac_late, SI_SUB_MAC_LATE, SI_ORDER_FIRST, mac_late_init, NULL); #else /* !MAC */ int __mac_get_proc(struct thread *td, struct __mac_get_proc_args *uap) { return (ENOSYS); } int __mac_set_proc(struct thread *td, struct __mac_set_proc_args *uap) { return (ENOSYS); } int __mac_get_fd(struct thread *td, struct __mac_get_fd_args *uap) { return (ENOSYS); } int __mac_get_file(struct thread *td, struct __mac_get_file_args *uap) { return (ENOSYS); } int __mac_set_fd(struct thread *td, struct __mac_set_fd_args *uap) { return (ENOSYS); } int __mac_set_file(struct thread *td, struct __mac_set_file_args *uap) { return (ENOSYS); } #endif /* !MAC */