From 92f28d973cce45ef5823209aab3138eb45d8b349 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 15 Mar 2013 01:03:33 -0700 Subject: scm: Require CAP_SYS_ADMIN over the current pidns to spoof pids. Don't allow spoofing pids over unix domain sockets in the corner cases where a user has created a user namespace but has not yet created a pid namespace. Cc: stable@vger.kernel.org Reported-by: Andy Lutomirski Signed-off-by: "Eric W. Biederman" --- net/core/scm.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/net/core/scm.c b/net/core/scm.c index 905dcc6ad1e3..2dc6cdaaae8a 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -52,7 +53,8 @@ static __inline__ int scm_check_creds(struct ucred *creds) if (!uid_valid(uid) || !gid_valid(gid)) return -EINVAL; - if ((creds->pid == task_tgid_vnr(current) || nsown_capable(CAP_SYS_ADMIN)) && + if ((creds->pid == task_tgid_vnr(current) || + ns_capable(current->nsproxy->pid_ns->user_ns, CAP_SYS_ADMIN)) && ((uid_eq(uid, cred->uid) || uid_eq(uid, cred->euid) || uid_eq(uid, cred->suid)) || nsown_capable(CAP_SETUID)) && ((gid_eq(gid, cred->gid) || gid_eq(gid, cred->egid) || -- cgit v1.2.3-58-ga151 From 751c644b95bb48aaa8825f0c66abbcc184d92051 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Tue, 26 Mar 2013 02:27:11 -0700 Subject: pid: Handle the exit of a multi-threaded init. When a multi-threaded init exits and the initial thread is not the last thread to exit the initial thread hangs around as a zombie until the last thread exits. In that case zap_pid_ns_processes needs to wait until there are only 2 hashed pids in the pid namespace not one. v2. Replace thread_pid_vnr(me) == 1 with the test thread_group_leader(me) as suggested by Oleg. Cc: stable@vger.kernel.org Cc: Oleg Nesterov Reported-by: Caj Larsson Signed-off-by: "Eric W. Biederman" --- kernel/pid_namespace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index c1c3dc1c6023..bea15bdf82b0 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -181,6 +181,7 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) int nr; int rc; struct task_struct *task, *me = current; + int init_pids = thread_group_leader(me) ? 1 : 2; /* Don't allow any more processes into the pid namespace */ disable_pid_allocation(pid_ns); @@ -230,7 +231,7 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) */ for (;;) { set_current_state(TASK_UNINTERRUPTIBLE); - if (pid_ns->nr_hashed == 1) + if (pid_ns->nr_hashed == init_pids) break; schedule(); } -- cgit v1.2.3-58-ga151 From eddc0a3abff273842a94784d2d022bbc36dc9015 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 21 Mar 2013 02:30:41 -0700 Subject: yama: Better permission check for ptraceme Change the permission check for yama_ptrace_ptracee to the standard ptrace permission check, testing if the traceer has CAP_SYS_PTRACE in the tracees user namespace. Reviewed-by: Kees Cook Signed-off-by: "Eric W. Biederman" --- security/yama/yama_lsm.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c index 23414b93771f..13c88fbcf037 100644 --- a/security/yama/yama_lsm.c +++ b/security/yama/yama_lsm.c @@ -347,10 +347,8 @@ int yama_ptrace_traceme(struct task_struct *parent) /* Only disallow PTRACE_TRACEME on more aggressive settings. */ switch (ptrace_scope) { case YAMA_SCOPE_CAPABILITY: - rcu_read_lock(); - if (!ns_capable(__task_cred(parent)->user_ns, CAP_SYS_PTRACE)) + if (!has_ns_capability(parent, current_user_ns(), CAP_SYS_PTRACE)) rc = -EPERM; - rcu_read_unlock(); break; case YAMA_SCOPE_NO_ATTACH: rc = -EPERM; -- cgit v1.2.3-58-ga151 From 3151527ee007b73a0ebd296010f1c0454a919c7d Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 15 Mar 2013 01:45:51 -0700 Subject: userns: Don't allow creation if the user is chrooted Guarantee that the policy of which files may be access that is established by setting the root directory will not be violated by user namespaces by verifying that the root directory points to the root of the mount namespace at the time of user namespace creation. Changing the root is a privileged operation, and as a matter of policy it serves to limit unprivileged processes to files below the current root directory. For reasons of simplicity and comprehensibility the privilege to change the root directory is gated solely on the CAP_SYS_CHROOT capability in the user namespace. Therefore when creating a user namespace we must ensure that the policy of which files may be access can not be violated by changing the root directory. Anyone who runs a processes in a chroot and would like to use user namespace can setup the same view of filesystems with a mount namespace instead. With this result that this is not a practical limitation for using user namespaces. Cc: stable@vger.kernel.org Acked-by: Serge Hallyn Reported-by: Andy Lutomirski Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 24 ++++++++++++++++++++++++ include/linux/fs_struct.h | 2 ++ kernel/user_namespace.c | 9 +++++++++ 3 files changed, 35 insertions(+) diff --git a/fs/namespace.c b/fs/namespace.c index 50ca17d3cb45..a3035223d421 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -2732,6 +2732,30 @@ bool our_mnt(struct vfsmount *mnt) return check_mnt(real_mount(mnt)); } +bool current_chrooted(void) +{ + /* Does the current process have a non-standard root */ + struct path ns_root; + struct path fs_root; + bool chrooted; + + /* Find the namespace root */ + ns_root.mnt = ¤t->nsproxy->mnt_ns->root->mnt; + ns_root.dentry = ns_root.mnt->mnt_root; + path_get(&ns_root); + while (d_mountpoint(ns_root.dentry) && follow_down_one(&ns_root)) + ; + + get_fs_root(current->fs, &fs_root); + + chrooted = !path_equal(&fs_root, &ns_root); + + path_put(&fs_root); + path_put(&ns_root); + + return chrooted; +} + static void *mntns_get(struct task_struct *task) { struct mnt_namespace *ns = NULL; diff --git a/include/linux/fs_struct.h b/include/linux/fs_struct.h index 729eded4b24f..2b93a9a5a1e6 100644 --- a/include/linux/fs_struct.h +++ b/include/linux/fs_struct.h @@ -50,4 +50,6 @@ static inline void get_fs_root_and_pwd(struct fs_struct *fs, struct path *root, spin_unlock(&fs->lock); } +extern bool current_chrooted(void); + #endif /* _LINUX_FS_STRUCT_H */ diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index b14f4d342043..0f1e42884577 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -61,6 +61,15 @@ int create_user_ns(struct cred *new) kgid_t group = new->egid; int ret; + /* + * Verify that we can not violate the policy of which files + * may be accessed that is specified by the root directory, + * by verifing that the root directory is at the root of the + * mount namespace which allows all files to be accessed. + */ + if (current_chrooted()) + return -EPERM; + /* The creator needs a mapping in the parent user namespace * or else we won't be able to reasonably tell userspace who * created a user_namespace. -- cgit v1.2.3-58-ga151 From 90563b198e4c6674c63672fae1923da467215f45 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 22 Mar 2013 03:10:15 -0700 Subject: vfs: Add a mount flag to lock read only bind mounts When a read-only bind mount is copied from mount namespace in a higher privileged user namespace to a mount namespace in a lesser privileged user namespace, it should not be possible to remove the the read-only restriction. Add a MNT_LOCK_READONLY mount flag to indicate that a mount must remain read-only. CC: stable@vger.kernel.org Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 3 +++ include/linux/mount.h | 2 ++ 2 files changed, 5 insertions(+) diff --git a/fs/namespace.c b/fs/namespace.c index a3035223d421..8505b5ece5de 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1713,6 +1713,9 @@ static int change_mount_flags(struct vfsmount *mnt, int ms_flags) if (readonly_request == __mnt_is_readonly(mnt)) return 0; + if (mnt->mnt_flags & MNT_LOCK_READONLY) + return -EPERM; + if (readonly_request) error = mnt_make_readonly(real_mount(mnt)); else diff --git a/include/linux/mount.h b/include/linux/mount.h index d7029f4a191a..73005f9957ea 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -47,6 +47,8 @@ struct mnt_namespace; #define MNT_INTERNAL 0x4000 +#define MNT_LOCK_READONLY 0x400000 + struct vfsmount { struct dentry *mnt_root; /* root of the mounted tree */ struct super_block *mnt_sb; /* pointer to superblock */ -- cgit v1.2.3-58-ga151 From 132c94e31b8bca8ea921f9f96a57d684fa4ae0a9 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 22 Mar 2013 04:08:05 -0700 Subject: vfs: Carefully propogate mounts across user namespaces As a matter of policy MNT_READONLY should not be changable if the original mounter had more privileges than creator of the mount namespace. Add the flag CL_UNPRIVILEGED to note when we are copying a mount from a mount namespace that requires more privileges to a mount namespace that requires fewer privileges. When the CL_UNPRIVILEGED flag is set cause clone_mnt to set MNT_NO_REMOUNT if any of the mnt flags that should never be changed are set. This protects both mount propagation and the initial creation of a less privileged mount namespace. Cc: stable@vger.kernel.org Acked-by: Serge Hallyn Reported-by: Andy Lutomirski Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 6 +++++- fs/pnode.c | 6 ++++++ fs/pnode.h | 1 + 3 files changed, 12 insertions(+), 1 deletion(-) diff --git a/fs/namespace.c b/fs/namespace.c index 8505b5ece5de..968d4c5eae03 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -798,6 +798,10 @@ static struct mount *clone_mnt(struct mount *old, struct dentry *root, } mnt->mnt.mnt_flags = old->mnt.mnt_flags & ~MNT_WRITE_HOLD; + /* Don't allow unprivileged users to change mount flags */ + if ((flag & CL_UNPRIVILEGED) && (mnt->mnt.mnt_flags & MNT_READONLY)) + mnt->mnt.mnt_flags |= MNT_LOCK_READONLY; + atomic_inc(&sb->s_active); mnt->mnt.mnt_sb = sb; mnt->mnt.mnt_root = dget(root); @@ -2342,7 +2346,7 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, /* First pass: copy the tree topology */ copy_flags = CL_COPY_ALL | CL_EXPIRE; if (user_ns != mnt_ns->user_ns) - copy_flags |= CL_SHARED_TO_SLAVE; + copy_flags |= CL_SHARED_TO_SLAVE | CL_UNPRIVILEGED; new = copy_tree(old, old->mnt.mnt_root, copy_flags); if (IS_ERR(new)) { up_write(&namespace_sem); diff --git a/fs/pnode.c b/fs/pnode.c index 3e000a51ac0d..8b29d2164da6 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -9,6 +9,7 @@ #include #include #include +#include #include "internal.h" #include "pnode.h" @@ -220,6 +221,7 @@ static struct mount *get_source(struct mount *dest, int propagate_mnt(struct mount *dest_mnt, struct dentry *dest_dentry, struct mount *source_mnt, struct list_head *tree_list) { + struct user_namespace *user_ns = current->nsproxy->mnt_ns->user_ns; struct mount *m, *child; int ret = 0; struct mount *prev_dest_mnt = dest_mnt; @@ -237,6 +239,10 @@ int propagate_mnt(struct mount *dest_mnt, struct dentry *dest_dentry, source = get_source(m, prev_dest_mnt, prev_src_mnt, &type); + /* Notice when we are propagating across user namespaces */ + if (m->mnt_ns->user_ns != user_ns) + type |= CL_UNPRIVILEGED; + child = copy_tree(source, source->mnt.mnt_root, type); if (IS_ERR(child)) { ret = PTR_ERR(child); diff --git a/fs/pnode.h b/fs/pnode.h index 19b853a3445c..a0493d5ebfbf 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -23,6 +23,7 @@ #define CL_MAKE_SHARED 0x08 #define CL_PRIVATE 0x10 #define CL_SHARED_TO_SLAVE 0x20 +#define CL_UNPRIVILEGED 0x40 static inline void set_mnt_shared(struct mount *mnt) { -- cgit v1.2.3-58-ga151 From a636b702ed1805e988ad3d8ff8b52c060f8b341c Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 21 Mar 2013 18:13:15 -0700 Subject: ipc: Restrict mounting the mqueue filesystem Only allow mounting the mqueue filesystem if the caller has CAP_SYS_ADMIN rights over the ipc namespace. The principle here is if you create or have capabilities over it you can mount it, otherwise you get to live with what other people have mounted. This information is not particularly sensitive and mqueue essentially only reports which posix messages queues exist. Still when creating a restricted environment for an application to live any extra information may be of use to someone with sufficient creativity. The historical if imperfect way this information has been restricted has been not to allow mounts and restricting this to ipc namespace creators maintains the spirit of the historical restriction. Cc: stable@vger.kernel.org Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- ipc/mqueue.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/ipc/mqueue.c b/ipc/mqueue.c index e5c4f609f22c..c4ae32ec6c6b 100644 --- a/ipc/mqueue.c +++ b/ipc/mqueue.c @@ -330,8 +330,16 @@ static struct dentry *mqueue_mount(struct file_system_type *fs_type, int flags, const char *dev_name, void *data) { - if (!(flags & MS_KERNMOUNT)) - data = current->nsproxy->ipc_ns; + if (!(flags & MS_KERNMOUNT)) { + struct ipc_namespace *ns = current->nsproxy->ipc_ns; + /* Don't allow mounting unless the caller has CAP_SYS_ADMIN + * over the ipc namespace. + */ + if (!ns_capable(ns->user_ns, CAP_SYS_ADMIN)) + return ERR_PTR(-EPERM); + + data = ns; + } return mount_ns(fs_type, flags, data, mqueue_fill_super); } -- cgit v1.2.3-58-ga151 From 87a8ebd637dafc255070f503909a053cf0d98d3f Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sun, 24 Mar 2013 14:28:27 -0700 Subject: userns: Restrict when proc and sysfs can be mounted Only allow unprivileged mounts of proc and sysfs if they are already mounted when the user namespace is created. proc and sysfs are interesting because they have content that is per namespace, and so fresh mounts are needed when new namespaces are created while at the same time proc and sysfs have content that is shared between every instance. Respect the policy of who may see the shared content of proc and sysfs by only allowing new mounts if there was an existing mount at the time the user namespace was created. In practice there are only two interesting cases: proc and sysfs are mounted at their usual places, proc and sysfs are not mounted at all (some form of mount namespace jail). Cc: stable@vger.kernel.org Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 21 +++++++++++++++++++++ fs/proc/root.c | 4 ++++ fs/sysfs/mount.c | 4 ++++ include/linux/user_namespace.h | 4 ++++ kernel/user.c | 2 ++ kernel/user_namespace.c | 2 ++ 6 files changed, 37 insertions(+) diff --git a/fs/namespace.c b/fs/namespace.c index 968d4c5eae03..d581e45c0a9f 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -2763,6 +2763,27 @@ bool current_chrooted(void) return chrooted; } +void update_mnt_policy(struct user_namespace *userns) +{ + struct mnt_namespace *ns = current->nsproxy->mnt_ns; + struct mount *mnt; + + down_read(&namespace_sem); + list_for_each_entry(mnt, &ns->list, mnt_list) { + switch (mnt->mnt.mnt_sb->s_magic) { + case SYSFS_MAGIC: + userns->may_mount_sysfs = true; + break; + case PROC_SUPER_MAGIC: + userns->may_mount_proc = true; + break; + } + if (userns->may_mount_sysfs && userns->may_mount_proc) + break; + } + up_read(&namespace_sem); +} + static void *mntns_get(struct task_struct *task) { struct mnt_namespace *ns = NULL; diff --git a/fs/proc/root.c b/fs/proc/root.c index c6e9fac26bac..9c7fab1d23f0 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include #include @@ -108,6 +109,9 @@ static struct dentry *proc_mount(struct file_system_type *fs_type, } else { ns = task_active_pid_ns(current); options = data; + + if (!current_user_ns()->may_mount_proc) + return ERR_PTR(-EPERM); } sb = sget(fs_type, proc_test_super, proc_set_super, flags, ns); diff --git a/fs/sysfs/mount.c b/fs/sysfs/mount.c index 8d924b5ec733..afd83273e6ce 100644 --- a/fs/sysfs/mount.c +++ b/fs/sysfs/mount.c @@ -19,6 +19,7 @@ #include #include #include +#include #include "sysfs.h" @@ -111,6 +112,9 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type, struct super_block *sb; int error; + if (!(flags & MS_KERNMOUNT) && !current_user_ns()->may_mount_sysfs) + return ERR_PTR(-EPERM); + info = kzalloc(sizeof(*info), GFP_KERNEL); if (!info) return ERR_PTR(-ENOMEM); diff --git a/include/linux/user_namespace.h b/include/linux/user_namespace.h index 4ce009324933..b6b215f13b45 100644 --- a/include/linux/user_namespace.h +++ b/include/linux/user_namespace.h @@ -26,6 +26,8 @@ struct user_namespace { kuid_t owner; kgid_t group; unsigned int proc_inum; + bool may_mount_sysfs; + bool may_mount_proc; }; extern struct user_namespace init_user_ns; @@ -82,4 +84,6 @@ static inline void put_user_ns(struct user_namespace *ns) #endif +void update_mnt_policy(struct user_namespace *userns); + #endif /* _LINUX_USER_H */ diff --git a/kernel/user.c b/kernel/user.c index e81978e8c03b..8e635a18ab52 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -51,6 +51,8 @@ struct user_namespace init_user_ns = { .owner = GLOBAL_ROOT_UID, .group = GLOBAL_ROOT_GID, .proc_inum = PROC_USER_INIT_INO, + .may_mount_sysfs = true, + .may_mount_proc = true, }; EXPORT_SYMBOL_GPL(init_user_ns); diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 0f1e42884577..a54f26f82eb2 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -96,6 +96,8 @@ int create_user_ns(struct cred *new) set_cred_user_ns(new, ns); + update_mnt_policy(ns); + return 0; } -- cgit v1.2.3-58-ga151