user/sven/linux.git/security, branch v4.9.181

selinux: never allow relabeling on context mounts

2019-05-08T05:19:12Z

commit a83d6ddaebe541570291205cb538e35ad4ff94f9 upstream. In the SECURITY_FS_USE_MNTPOINT case we never want to allow relabeling files/directories, so we should never set the SBLABEL_MNT flag. The 'special handling' in selinux_is_sblabel_mnt() is only intended for when the behavior is set to SECURITY_FS_USE_GENFS. While there, make the logic in selinux_is_sblabel_mnt() more explicit and add a BUILD_BUG_ON() to make sure that introducing a new SECURITY_FS_USE_* forces a review of the logic. Fixes: d5f3a5f6e7e7 ("selinux: add security in-core xattr support for pstore and debugfs") Signed-off-by: Ondrej Mosnacek Reviewed-by: Stephen Smalley Signed-off-by: Paul Moore Signed-off-by: Greg Kroah-Hartman

device_cgroup: fix RCU imbalance in error case

2019-04-27T07:34:46Z

commit 0fcc4c8c044e117ac126ab6df4138ea9a67fa2a9 upstream. When dev_exception_add() returns an error (due to a failed memory allocation), make sure that we move the RCU preemption count back to where it was before we were called. We dropped the RCU read lock inside the loop body, so we can't just "break". sparse complains about this, too: $ make -s C=2 security/device_cgroup.o ./include/linux/rcupdate.h:647:9: warning: context imbalance in 'propagate_exception' - unexpected unlock Fixes: d591fb56618f ("device_cgroup: simplify cgroup tree walk in propagate_exception()") Cc: stable@vger.kernel.org Signed-off-by: Jann Horn Acked-by: Michal Hocko Signed-off-by: Tejun Heo Signed-off-by: Greg Kroah-Hartman

selinux: do not override context on context mounts

2019-04-05T20:29:15Z

[ Upstream commit 53e0c2aa9a59a48e3798ef193d573ade85aa80f5 ] Ignore all selinux_inode_notifysecctx() calls on mounts with SBLABEL_MNT flag unset. This is achived by returning -EOPNOTSUPP for this case in selinux_inode_setsecurtity() (because that function should not be called in such case anyway) and translating this error to 0 in selinux_inode_notifysecctx(). This fixes behavior of kernfs-based filesystems when mounted with the 'context=' option. Before this patch, if a node's context had been explicitly set to a non-default value and later the filesystem has been remounted with the 'context=' option, then this node would show up as having the manually-set context and not the mount-specified one. Steps to reproduce: # mount -t cgroup2 cgroup2 /sys/fs/cgroup/unified # chcon unconfined_u:object_r:user_home_t:s0 /sys/fs/cgroup/unified/cgroup.stat # ls -lZ /sys/fs/cgroup/unified total 0 -r--r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.controllers -rw-r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.max.depth -rw-r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.max.descendants -rw-r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.procs -r--r--r--. 1 root root unconfined_u:object_r:user_home_t:s0 0 Dec 13 10:41 cgroup.stat -rw-r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.subtree_control -rw-r--r--. 1 root root system_u:object_r:cgroup_t:s0 0 Dec 13 10:41 cgroup.threads # umount /sys/fs/cgroup/unified # mount -o context=system_u:object_r:tmpfs_t:s0 -t cgroup2 cgroup2 /sys/fs/cgroup/unified Result before: # ls -lZ /sys/fs/cgroup/unified total 0 -r--r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.controllers -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.max.depth -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.max.descendants -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.procs -r--r--r--. 1 root root unconfined_u:object_r:user_home_t:s0 0 Dec 13 10:41 cgroup.stat -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.subtree_control -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.threads Result after: # ls -lZ /sys/fs/cgroup/unified total 0 -r--r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.controllers -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.max.depth -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.max.descendants -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.procs -r--r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.stat -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.subtree_control -rw-r--r--. 1 root root system_u:object_r:tmpfs_t:s0 0 Dec 13 10:41 cgroup.threads Signed-off-by: Ondrej Mosnacek Reviewed-by: Stephen Smalley Signed-off-by: Paul Moore Signed-off-by: Sasha Levin

missing barriers in some of unix_sock ->addr and ->path accesses

2019-03-19T12:14:10Z

[ Upstream commit ae3b564179bfd06f32d051b9e5d72ce4b2a07c37 ] Several u->addr and u->path users are not holding any locks in common with unix_bind(). unix_state_lock() is useless for those purposes. u->addr is assign-once and *(u->addr) is fully set up by the time we set u->addr (all under unix_table_lock). u->path is also set in the same critical area, also before setting u->addr, and any unix_sock with ->path filled will have non-NULL ->addr. So setting ->addr with smp_store_release() is all we need for those "lockless" users - just have them fetch ->addr with smp_load_acquire() and don't even bother looking at ->path if they see NULL ->addr. Users of ->addr and ->path fall into several classes now: 1) ones that do smp_load_acquire(u->addr) and access *(u->addr) and u->path only if smp_load_acquire() has returned non-NULL. 2) places holding unix_table_lock. These are guaranteed that *(u->addr) is seen fully initialized. If unix_sock is in one of the "bound" chains, so's ->path. 3) unix_sock_destructor() using ->addr is safe. All places that set u->addr are guaranteed to have seen all stores *(u->addr) while holding a reference to u and unix_sock_destructor() is called when (atomic) refcount hits zero. 4) unix_release_sock() using ->path is safe. unix_bind() is serialized wrt unix_release() (normally - by struct file refcount), and for the instances that had ->path set by unix_bind() unix_release_sock() comes from unix_release(), so they are fine. Instances that had it set in unix_stream_connect() either end up attached to a socket (in unix_accept()), in which case the call chain to unix_release_sock() and serialization are the same as in the previous case, or they never get accept'ed and unix_release_sock() is called when the listener is shut down and its queue gets purged. In that case the listener's queue lock provides the barriers needed - unix_stream_connect() shoves our unix_sock into listener's queue under that lock right after having set ->path and eventual unix_release_sock() caller picks them from that queue under the same lock right before calling unix_release_sock(). 5) unix_find_other() use of ->path is pointless, but safe - it happens with successful lookup by (abstract) name, so ->path.dentry is guaranteed to be NULL there. earlier-variant-reviewed-by: "Paul E. McKenney" Signed-off-by: Al Viro Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

KEYS: restrict /proc/keys by credentials at open time

2019-03-19T12:14:08Z

commit 4aa68e07d845562561f5e73c04aa521376e95252 upstream. When checking for permission to view keys whilst reading from /proc/keys, we should use the credentials with which the /proc/keys file was opened. This is because, in a classic type of exploit, it can be possible to bypass checks for the *current* credentials by passing the file descriptor to a suid program. Following commit 34dbbcdbf633 ("Make file credentials available to the seqfile interfaces") we can finally fix it. So let's do it. Signed-off-by: Eric Biggers Signed-off-by: David Howells Signed-off-by: Zubin Mithra Signed-off-by: Greg Kroah-Hartman

KEYS: always initialize keyring_index_key::desc_len

2019-02-27T09:07:00Z

commit ede0fa98a900e657d1fcd80b50920efc896c1a4c upstream. syzbot hit the 'BUG_ON(index_key->desc_len == 0);' in __key_link_begin() called from construct_alloc_key() during sys_request_key(), because the length of the key description was never calculated. The problem is that we rely on ->desc_len being initialized by search_process_keyrings(), specifically by search_nested_keyrings(). But, if the process isn't subscribed to any keyrings that never happens. Fix it by always initializing keyring_index_key::desc_len as soon as the description is set, like we already do in some places. The following program reproduces the BUG_ON() when it's run as root and no session keyring has been installed. If it doesn't work, try removing pam_keyinit.so from /etc/pam.d/login and rebooting. #include #include #include int main(void) { int id = add_key("keyring", "syz", NULL, 0, KEY_SPEC_USER_KEYRING); keyctl_setperm(id, KEY_OTH_WRITE); setreuid(5000, 5000); request_key("user", "desc", "", id); } Reported-by: syzbot+ec24e95ea483de0a24da@syzkaller.appspotmail.com Fixes: b2a4df200d57 ("KEYS: Expand the capacity of a keyring") Signed-off-by: Eric Biggers Signed-off-by: David Howells Cc: stable@vger.kernel.org Signed-off-by: James Morris Signed-off-by: Greg Kroah-Hartman

KEYS: allow reaching the keys quotas exactly

2019-02-27T09:06:58Z

commit a08bf91ce28ed3ae7b6fef35d843fef8dc8c2cd9 upstream. If the sysctl 'kernel.keys.maxkeys' is set to some number n, then actually users can only add up to 'n - 1' keys. Likewise for 'kernel.keys.maxbytes' and the root_* versions of these sysctls. But these sysctls are apparently supposed to be *maximums*, as per their names and all documentation I could find -- the keyrings(7) man page, Documentation/security/keys/core.rst, and all the mentions of EDQUOT meaning that the key quota was *exceeded* (as opposed to reached). Thus, fix the code to allow reaching the quotas exactly. Fixes: 0b77f5bfb45c ("keys: make the keyring quotas controllable through /proc/sys") Cc: stable@vger.kernel.org Signed-off-by: Eric Biggers Signed-off-by: David Howells Signed-off-by: James Morris Signed-off-by: Greg Kroah-Hartman

smack: fix access permissions for keyring

2019-02-12T18:44:54Z

[ Upstream commit 5b841bfab695e3b8ae793172a9ff7990f99cc3e2 ] Function smack_key_permission() only issues smack requests for the following operations: - KEY_NEED_READ (issues MAY_READ) - KEY_NEED_WRITE (issues MAY_WRITE) - KEY_NEED_LINK (issues MAY_WRITE) - KEY_NEED_SETATTR (issues MAY_WRITE) A blank smack request is issued in all other cases, resulting in smack access being granted if there is any rule defined between subject and object, or denied with -EACCES otherwise. Request MAY_READ access for KEY_NEED_SEARCH and KEY_NEED_VIEW. Fix the logic in the unlikely case when both MAY_READ and MAY_WRITE are needed. Validate access permission field for valid contents. Signed-off-by: Zoran Markovic Signed-off-by: Casey Schaufler Cc: Casey Schaufler Cc: James Morris Cc: "Serge E. Hallyn" Signed-off-by: Sasha Levin

selinux: always allow mounting submounts

2019-01-26T08:38:33Z

[ Upstream commit 2cbdcb882f97a45f7475c67ac6257bbc16277dfe ] If a superblock has the MS_SUBMOUNT flag set, we should always allow mounting it. These mounts are done automatically by the kernel either as part of mounting some parent mount (e.g. debugfs always mounts tracefs under "tracing" for compatibility) or they are mounted automatically as needed on subdirectory accesses (e.g. NFS crossmnt mounts). Since such automounts are either an implicit consequence of the parent mount (which is already checked) or they can happen during regular accesses (where it doesn't make sense to check against the current task's context), the mount permission check should be skipped for them. Without this patch, attempts to access contents of an automounted directory can cause unexpected SELinux denials. In the current kernel tree, the MS_SUBMOUNT flag is set only via vfs_submount(), which is called only from the following places: - AFS, when automounting special "symlinks" referencing other cells - CIFS, when automounting "referrals" - NFS, when automounting subtrees - debugfs, when automounting tracefs In all cases the submounts are meant to be transparent to the user and it makes sense that if mounting the master is allowed, then so should be the automounts. Note that CAP_SYS_ADMIN capability checking is already skipped for (SB_KERNMOUNT|SB_SUBMOUNT) in: - sget_userns() in fs/super.c: if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) && !(type->fs_flags & FS_USERNS_MOUNT) && !capable(CAP_SYS_ADMIN)) return ERR_PTR(-EPERM); - sget() in fs/super.c: /* Ensure the requestor has permissions over the target filesystem */ if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) && !ns_capable(user_ns, CAP_SYS_ADMIN)) return ERR_PTR(-EPERM); Verified internally on patched RHEL 7.6 with a reproducer using NFS+httpd and selinux-tesuite. Fixes: 93faccbbfa95 ("fs: Better permission checking for submounts") Signed-off-by: Ondrej Mosnacek Signed-off-by: Paul Moore Signed-off-by: Sasha Levin

selinux: fix GPF on invalid policy

2019-01-23T07:10:56Z

commit 5b0e7310a2a33c06edc7eb81ffc521af9b2c5610 upstream. levdatum->level can be NULL if we encounter an error while loading the policy during sens_read prior to initializing it. Make sure sens_destroy handles that case correctly. Reported-by: syzbot+6664500f0f18f07a5c0e@syzkaller.appspotmail.com Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore Signed-off-by: Greg Kroah-Hartman