@@ -1043,6 +1043,8 @@ static int __init fcntl_init(void)
HWEIGHT32(
(VALID_OPEN_FLAGS & ~(O_NONBLOCK | O_NDELAY)) |
__FMODE_EXEC | __FMODE_NONOTIFY));
+ BUILD_BUG_ON(HWEIGHT32(VALID_OPENAT2_FLAGS) !=
+ HWEIGHT32(VALID_OPEN_FLAGS) + 1);
fasync_cache = kmem_cache_create("fasync_cache",
sizeof(struct fasync_struct), 0,
@@ -3776,6 +3776,43 @@ static int do_o_path(struct nameidata *nd, unsigned flags, struct file *file)
return error;
}
+static const struct cred *openat2_init_creds(int dfd)
+{
+ struct cred *cred;
+ struct fd f;
+
+ if (dfd == AT_FDCWD)
+ return ERR_PTR(-EINVAL);
+
+ f = fdget_raw(dfd);
+ if (!f.file)
+ return ERR_PTR(-EBADF);
+
+ cred = ERR_PTR(-EPERM);
+ if (!(f.file->f_flags & O_CRED_ALLOW))
+ goto done;
+
+ cred = prepare_creds();
+ if (!cred) {
+ cred = ERR_PTR(-ENOMEM);
+ goto done;
+ }
+
+ cred->fsuid = f.file->f_cred->fsuid;
+ cred->fsgid = f.file->f_cred->fsgid;
+ cred->group_info = get_group_info(f.file->f_cred->group_info);
+
+done:
+ fdput(f);
+ return cred;
+}
+
+static void openat2_done_creds(const struct cred *cred)
+{
+ put_group_info(cred->group_info);
+ put_cred(cred);
+}
+
static struct file *path_openat(struct nameidata *nd,
const struct open_flags *op, unsigned flags)
{
@@ -3793,18 +3830,33 @@ static struct file *path_openat(struct nameidata *nd,
error = do_o_path(nd, flags, file);
} else {
const char *s;
+ const struct cred *old_cred = NULL, *cred = NULL;
- file = alloc_empty_file(open_flags, current_cred());
- if (IS_ERR(file))
+ if (open_flags & OA2_CRED_INHERIT) {
+ cred = openat2_init_creds(nd->dfd);
+ if (IS_ERR(cred))
+ return ERR_CAST(cred);
+ }
+ file = alloc_empty_file(open_flags, cred ?: current_cred());
+ if (IS_ERR(file)) {
+ if (cred)
+ openat2_done_creds(cred);
return file;
+ }
s = path_init(nd, flags);
+ if (cred)
+ old_cred = override_creds(cred);
while (!(error = link_path_walk(s, nd)) &&
(s = open_last_lookups(nd, file, op)) != NULL)
;
if (!error)
error = do_open(nd, file, op);
+ if (old_cred)
+ revert_creds(old_cred);
terminate_walk(nd);
+ if (cred)
+ openat2_done_creds(cred);
}
if (likely(!error)) {
if (likely(file->f_mode & FMODE_OPENED))
@@ -1225,7 +1225,7 @@ inline int build_open_flags(const struct open_how *how, struct open_flags *op)
* values before calling build_open_flags(), but openat2(2) checks all
* of its arguments.
*/
- if (flags & ~VALID_OPEN_FLAGS)
+ if (flags & ~VALID_OPENAT2_FLAGS)
return -EINVAL;
if (how->resolve & ~VALID_RESOLVE_FLAGS)
return -EINVAL;
@@ -1326,6 +1326,14 @@ inline int build_open_flags(const struct open_how *how, struct open_flags *op)
lookup_flags |= LOOKUP_CACHED;
}
+ if (flags & OA2_CRED_INHERIT) {
+ /* Inherit creds only with scoped look-up modes. */
+ if (!(lookup_flags & LOOKUP_IS_SCOPED))
+ return -EPERM;
+ /* Reject /proc "magic" links if inheriting creds. */
+ lookup_flags |= LOOKUP_NO_MAGICLINKS;
+ }
+
op->lookup_flags = lookup_flags;
return 0;
}
@@ -12,6 +12,8 @@
FASYNC | O_DIRECT | O_LARGEFILE | O_DIRECTORY | O_NOFOLLOW | \
O_NOATIME | O_CLOEXEC | O_PATH | __O_TMPFILE | O_CRED_ALLOW)
+#define VALID_OPENAT2_FLAGS (VALID_OPEN_FLAGS | OA2_CRED_INHERIT)
+
/* List of all valid flags for the how->resolve argument: */
#define VALID_RESOLVE_FLAGS \
(RESOLVE_NO_XDEV | RESOLVE_NO_MAGICLINKS | RESOLVE_NO_SYMLINKS | \
@@ -40,4 +40,6 @@ struct open_how {
return -EAGAIN if that's not
possible. */
+#define OA2_CRED_INHERIT (1UL << 28)
+
#endif /* _UAPI_LINUX_OPENAT2_H */
This flag performs the open operation with the fs credentials (fsuid, fsgid, group_info) that were in effect when dir_fd was opened. dir_fd must be opened with O_CRED_ALLOW flag for this to work. This allows the process to pre-open some directories and then change eUID (and all other UIDs/GIDs) to a less-privileged user, retaining the ability to open/create files within these directories. Design goal: The idea is to provide a very light-weight sandboxing, where the process, without the use of any heavy-weight techniques like chroot within namespaces, can restrict the access to the set of pre-opened directories. This patch is just a first step to such sandboxing. If things go well, in the future the same extension can be added to more syscalls. These should include at least unlinkat(), renameat2() and the not-yet-upstreamed setxattrat(). Security considerations: - Only the bare minimal set of credentials is overridden: fsuid, fsgid and group_info. The rest, for example capabilities, are not overridden to avoid unneeded security risks. - To avoid sandboxing escape, this patch makes sure the restricted lookup modes are used. Namely, RESOLVE_BENEATH or RESOLVE_IN_ROOT. - Magic /proc symlinks are discarded, as suggested by Andy Lutomirski <luto@kernel.org> - O_CRED_ALLOW fds cannot be passed via unix socket and are always closed on exec() to prevent "unsuspecting userspace" from not being able to fully drop privs. Use cases: Virtual machines that deal with untrusted code, can use that instead of a more heavy-weighted approaches. Currently the approach is being tested on a dosemu2 VM. Signed-off-by: Stas Sergeev <stsp2@yandex.ru> CC: Stefan Metzmacher <metze@samba.org> CC: Eric Biederman <ebiederm@xmission.com> CC: Alexander Viro <viro@zeniv.linux.org.uk> CC: Andy Lutomirski <luto@kernel.org> CC: Christian Brauner <brauner@kernel.org> CC: Jan Kara <jack@suse.cz> CC: Jeff Layton <jlayton@kernel.org> CC: Chuck Lever <chuck.lever@oracle.com> CC: Alexander Aring <alex.aring@gmail.com> CC: linux-fsdevel@vger.kernel.org CC: linux-kernel@vger.kernel.org CC: Paolo Bonzini <pbonzini@redhat.com> CC: Christian Göttsche <cgzones@googlemail.com> --- fs/fcntl.c | 2 ++ fs/namei.c | 56 ++++++++++++++++++++++++++++++++++-- fs/open.c | 10 ++++++- include/linux/fcntl.h | 2 ++ include/uapi/linux/openat2.h | 2 ++ 5 files changed, 69 insertions(+), 3 deletions(-)