Priveleged operations should be allowed on loop devices within a
devloop mount by root within the user namespace which owns the
mount. Stash away the namespace at mount time and allow
CAP_SYS_ADMIN within this namespace to perform priveleged
operations on loop devices.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Add limited capability for use of loop devices in non-root
containers via a loopfs psuedo fs. When mounted this filesystem
will contain only a loop-control device node. This can be used
to request free loop devices which will be "owned" by that mount.
Device nodes appear automatically for these devices, and the same
device will not be given to another loopfs mount. Privileged loop
ioctls (for encrypted loop) will be allowed within the namespace
which mounted the loopfs.

Privileged block ioctls are not permitted, so features such as
partitions are not supported for unprivileged users.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

loop_clr_fd defers teardown if the loop device itself is
referenced but proceeds when a partition of the loop device is
open. When this happens partition scanning will fail, leaving
behind the block devices for the loop device's partitions.
Prevent this by returning -EBUSY when attempting to tear down
the loop device while any partition block device is open.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Unprivileged users are normally restricted from mounting with the
allow_other option by system policy, but this could be bypassed
for a mount done with user namespace root permissions. In such
cases allow_other should not allow users outside the userns
to access the mount as doing so would give the unprivileged user
the ability to manipulate processes it would otherwise be unable
to manipulate. Restrict allow_other to apply to users in the same
userns used at mount or a descendant of that namespace.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Update fuse to translate uids and gids to/from the user namspace
of the process servicing requests on /dev/fuse. Any ids which do
not map into the namespace will result in errors. inodes will
also be marked bad when unmappable ids are received from the
userspace fuse process.

Currently no use cases are known for letting the userspace fuse
daemon switch namespaces after opening /dev/fuse. Given this
fact, and in order to keep the implementation as simple as
possible and ease security auditing, the user namespace from
which /dev/fuse is opened is used for all id translations. This
is required to be the same namespace as s_user_ns to maintain
behavior consistent with other filesystems which can be mounted
in user namespaces.

For cuse the namespace used for the connection is also simply
current_user_ns() at the time /dev/cuse is opened.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

If the userspace process servicing fuse requests is running in
a pid namespace then pids passed via the fuse fd need to be
translated relative to that namespace. Capture the pid namespace
in use when the filesystem is mounted and use this for pid
translation.

Since no use case currently exists for changing namespaces all
translations are done relative to the pid namespace in use when
/dev/fuse is opened. Mounting or /dev/fuse IO from another
namespace will return errors.

Requests from processes whose pid cannot be translated into the
target namespace are not permitted, except for requests
allocated via fuse_get_req_nofail_nopages. For no-fail requests
in.h.pid will be 0 if the pid translation fails.

File locking changes based on previous work done by Eric
Biederman.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>

Support unprivileged mounting of ext4 volumes from user
namespaces. This requires the following changes:

 - Perform all uid and gid conversions to/from disk relative to
   s_user_ns. In many cases this will already be handled by the
   vfs helper functions. This also requires updates to handle
   cases where ids may not map into s_user_ns.

 - Update most capability checks to check for capabilities in
   s_user_ns rather than init_user_ns. These mostly reflect
   changes to the filesystem that a user in s_user_ns could
   already make externally by virtue of having write access to
   the backing device.

 - Restrict unsafe options in either the mount options or the
   ext4 superblock. Currently the only concerning option is
   errors=panic, and this is made to require CAP_SYS_ADMIN in
   init_user_ns.

 - Verify that unprivileged users have the required access to the
   journal device at the path passed via the journal_path mount
   option.

   Note that for the journal_path and the journal_dev mount
   options, and for external journal devices specified in the
   ext4 superblock, devcgroup restrictions will be enforced by
   __blkdev_get(), (via blkdev_get_by_dev()), ensuring that the
   user has been granted appropriate access to the block device.

 - Set the FS_USERNS_MOUNT flag on the filesystem types supported
   by ext4.

sysfs attributes for ext4 mounts remain writable only by real
root.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

A privileged user in a super block's s_user_ns is privileged
towards that file system and thus should be allowed to set file
capabilities. The file capabilities will not be trusted outside
of s_user_ns, so an unprivileged user cannot use this to gain
privileges in a user namespace where they are not already
privileged.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

A user privileged towards s_user_ns is privileged towards the
superblock and should be allowed to perform privileged quota
operations on that superblock. Convert check_quotactl_permission
to check for CAP_SYS_ADMIN in s_user_ns for these operations.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Superblock level remounts are currently restricted to global
CAP_SYS_ADMIN, as is the path for changing the root mount to
read only on umount. Loosen both of these permission checks to
also allow CAP_SYS_ADMIN in any namespace which is privileged
towards the userns which originally mounted the filesystem.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>

The user in control of a super block should be allowed to freeze
and thaw it. Relax the restrictions on the FIFREEZE and FITHAW
ioctls to require CAP_SYS_ADMIN in s_user_ns.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Privileged users in the namespace which controls a super block
should not be prevented from creating hard links. Expand the
check in may_linkat() to allow CAP_FOWNER in s_user_ns to set
any hardlink.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Expand the check in should_remove_suid() to keep privileges for
CAP_FSETID in s_user_ns rather than init_user_ns.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

The mounter of a filesystem should be privileged towards the
inodes of that filesystem. Extend the checks in
inode_owner_or_capable() and capable_wrt_inode_uidgid() to
permit access by users priviliged in the user namespace of the
inode's superblock.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

The EVM HMAC should be calculated using the on disk user and
group ids, so the k[ug]ids in the inode must be translated
relative to the s_user_ns of the inode's super block.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

ids in on-disk ACLs should be converted to s_user_ns instead of
init_user_ns as is done now. This introduces the possibility for
id mappings to fail, and when this happens syscalls will return
EOVERFLOW.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

When dealing with mounts from user namespaces quota user ids must
be translated relative to s_user_ns, especially when they will be
stored on disk. For purposes of the in-kernel hash table using
init_user_ns is still okay and may help reduce hash collisions,
so continue using init_user_ns there.

These changes introduce the possibility that the conversion
operations in struct qtree_fmt_operations could fail if an id
cannot be legitimately converted. Change these operations to
return int rather than void, and update the callers to handle
failures.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Add checks to inode_change_ok to verify that uid and gid changes
will map into the superblock's user namespace. If they do not
fail with -EOVERFLOW. This cannot be overriden with ATTR_FORCE.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

For filesystems mounted from a user namespace on-disk ids should
be translated relative to s_users_ns rather than init_user_ns.

When an id in the filesystem doesn't exist in s_user_ns the
associated id in the inode will be set to INVALID_[UG]ID, which
turns these into de facto "nobody" ids. This actually maps pretty
well into the way most code already works, and those places where
it didn't were fixed in previous patches. Moving forward vfs code
needs to be careful to handle instances where ids in inodes may
be invalid.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Using INVALID_[UG]ID for the LSM file creation context doesn't
make sense, so return an error if the inode passed to
set_create_file_as() has an invalid id.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Filesystem uids which don't map into a user namespace may result
in inode->i_uid being INVALID_UID. A symlink and its parent
could have different owners in the filesystem can both get
mapped to INVALID_UID, which may result in following a symlink
when this would not have otherwise been permitted when protected
symlinks are enabled.

Add a new helper function, uid_valid_eq(), and use this to
validate that the ids in may_follow_link() are both equal and
valid. Also add an equivalent helper for gids, which is
currently unused.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

The SMACK64, SMACK64EXEC, and SMACK64MMAP labels are all handled
differently in untrusted mounts. This is confusing and
potentically problematic. Change this to handle them all the same
way that SMACK64 is currently handled; that is, read the label
from disk and check it at use time. For SMACK64 and SMACK64MMAP
access is denied if the label does not match smk_root. To be
consistent with suid, a SMACK64EXEC label which does not match
smk_root will still allow execution of the file but will not run
with the label supplied in the xattr.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

All current callers of in_userns pass current_user_ns as the
first argument. Simplify by replacing in_userns with
current_in_userns which checks whether current_user_ns is in the
namespace supplied as an argument.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Security labels from unprivileged mounts in user namespaces must
be ignored. Force superblocks from user namespaces whose labeling
behavior is to use xattrs to use mountpoint labeling instead.
For the mountpoint label, default to converting the current task
context into a form suitable for file objects, but also allow the
policy writer to specify a different label through policy
transition rules.

Pieced together from code snippets provided by Stephen Smalley.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>

If a process gets access to a mount from a different user
namespace, that process should not be able to take advantage of
setuid files or selinux entrypoints from that filesystem.  Prevent
this by treating mounts from other mount namespaces and those not
owned by current_user_ns() or an ancestor as nosuid.

This will make it safer to allow more complex filesystems to be
mounted in non-root user namespaces.

This does not remove the need for MNT_LOCK_NOSUID.  The setuid,
setgid, and file capability bits can no longer be abused if code in
a user namespace were to clear nosuid on an untrusted filesystem,
but this patch, by itself, is insufficient to protect the system
from abuse of files that, when execed, would increase MAC privilege.

As a more concrete explanation, any task that can manipulate a
vfsmount associated with a given user namespace already has
capabilities in that namespace and all of its descendents.  If they
can cause a malicious setuid, setgid, or file-caps executable to
appear in that mount, then that executable will only allow them to
elevate privileges in exactly the set of namespaces in which they
are already privileges.

On the other hand, if they can cause a malicious executable to
appear with a dangerous MAC label, running it could change the
caller's security context in a way that should not have been
possible, even inside the namespace in which the task is confined.

As a hardening measure, this would have made CVE-2014-5207 much
more difficult to exploit.
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Unprivileged users should not be able to mount mtd block devices
when they lack sufficient privileges towards the block device
inode.  Update mount_mtd() to validate that the user has the
required access to the inode at the specified path. The check
will be skipped for CAP_SYS_ADMIN, so privileged mounts will
continue working as before.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Unprivileged users should not be able to mount block devices when
they lack sufficient privileges towards the block device inode.
Update blkdev_get_by_path() to validate that the user has the
required access to the inode at the specified path. The check
will be skipped for CAP_SYS_ADMIN, so privileged mounts will
continue working as before.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

When looking up a block device by path no permission check is
done to verify that the user has access to the block device inode
at the specified path. In some cases it may be necessary to
check permissions towards the inode, such as allowing
unprivileged users to mount block devices in user namespaces.

Add an argument to lookup_bdev() to optionally perform this
permission check. A value of 0 skips the permission check and
behaves the same as before. A non-zero value specifies the mask
of access rights required towards the inode at the specified
path. The check is always skipped if the user has CAP_SYS_ADMIN.

All callers of lookup_bdev() currently pass a mask of 0, so this
patch results in no functional change. Subsequent patches will
add permission checks where appropriate.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

Security labels from unprivileged mounts cannot be trusted.
Ideally for these mounts we would assign the objects in the
filesystem the same label as the inode for the backing device
passed to mount. Unfortunately it's currently impossible to
determine which inode this is from the LSM mount hooks, so we
settle for the label of the process doing the mount.

This label is assigned to s_root, and also to smk_default to
ensure that new inodes receive this label. The transmute property
is also set on s_root to make this behavior more explicit, even
though it is technically not necessary.

If a filesystem has existing security labels, access to inodes is
permitted if the label is the same as smk_root, otherwise access
is denied. The SMACK64EXEC xattr is completely ignored.

Explicit setting of security labels continues to require
CAP_MAC_ADMIN in init_user_ns.

Altogether, this ensures that filesystem objects are not
accessible to subjects which cannot already access the backing
store, that MAC is not violated for any objects in the fileystem
which are already labeled, and that a user cannot use an
unprivileged mount to gain elevated MAC privileges.

sysfs, tmpfs, and ramfs are already mountable from user
namespaces and support security labels. We can't rule out the
possibility that these filesystems may already be used in mounts
from user namespaces with security lables set from the init
namespace, so failing to trust lables in these filesystems may
introduce regressions. It is safe to trust labels from these
filesystems, since the unprivileged user does not control the
backing store and thus cannot supply security labels, so an
explicit exception is made to trust labels from these
filesystems.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>

Capability sets attached to files must be ignored except in the
user namespaces where the mounter is privileged, i.e. s_user_ns
and its descendants. Otherwise a vector exists for gaining
privileges in namespaces where a user is not already privileged.

Add a new helper function, in_user_ns(), to test whether a user
namespace is the same as or a descendant of another namespace.
Use this helper to determine whether a file's capability set
should be applied to the caps constructed during exec.
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>

- Consolidate the testing if a device node may be opened in a new
  function may_open_dev.

- Move the check for allowing access to device nodes on filesystems
  not mounted in the initial user namespace from mount time to open
  time and include it in may_open_dev.

This set of changes removes the implicit adding of MNT_NODEV which
simplifies the logic in fs/namespace.c and removes a potentially
problematic difference in how normal and unprivileged mount
namespaces work.  This is a user visible change in behavior for
remount in unpriviliged mount namespaces but is unlikely to cause
problems for existing software.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>

Initially this will be used to eliminate the implicit MNT_NODEV
flag for mounts from user namespaces. In the future it will also
be used for translating ids and checking capabilities for
filesystems mounted from user namespaces.

s_user_ns is initialized in alloc_super() and is generally set to
current_user_ns(). To avoid security and corruption issues, two
additional mount checks are also added:

 - do_new_mount() gains a check that the user has CAP_SYS_ADMIN
   in current_user_ns().

 - sget() will fail with EBUSY when the filesystem it's looking
   for is already mounted from another user namespace.

proc requires some special handling. The user namespace of
current isn't appropriate when forking as a result of clone (2)
with CLONE_NEWPID|CLONE_NEWUSER, as it will set s_user_ns to the
namespace of the parent and make proc unmountable in the new user
namespace. Instead, the user namespace which owns the new pid
namespace is used. sget_userns() is allowed to allow passing in
a namespace other than that of current, and sget becomes a
wrapper around sget_userns() which passes current_user_ns().

Changes to original version of this patch
  * Documented @user_ns in sget_userns, alloc_super and fs.h
  * Kept an blank line in fs.h
  * Removed unncessary include of user_namespace.h from fs.h
  * Tweaked the location of get_user_ns and put_user_ns so
    the security modules can (if they wish) depend on it.
  -- EWB
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>

Pull CRIS updates from Jesper Nilsson:
 "Mostly removal of old cruft of which we can use a generic version, or
  fixes for code not commonly run in the cris port, but also additions
  to enable some good debug"

* tag 'cris-for-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/jesper/cris: (25 commits)
  CRISv10: delete unused lib/dmacopy.c
  CRISv10: delete unused lib/old_checksum.c
  CRIS: fix switch_mm() lockdep splat
  CRISv32: enable LOCKDEP_SUPPORT
  CRIS: add STACKTRACE_SUPPORT
  CRISv32: annotate irq enable in idle loop
  CRISv32: add support for irqflags tracing
  CRIS: UAPI: use generic types.h
  CRIS: UAPI: use generic shmbuf.h
  CRIS: UAPI: use generic msgbuf.h
  CRIS: UAPI: use generic socket.h
  CRIS: UAPI: use generic sembuf.h
  CRIS: UAPI: use generic sockios.h
  CRIS: UAPI: use generic auxvec.h
  CRIS: UAPI: use generic headers via Kbuild
  CRIS: UAPI: fix elf.h export
  CRIS: don't make asm/elf.h depend on asm/user.h
  CRIS: UAPI: fix ptrace.h
  CRISv32: Squash compile warnings for axisflashmap
  CRISv32: Add GPIO driver to the default configs
  ...

rq_data_dir() returns either READ or WRITE (0 == READ, 1 == WRITE), not
a boolean value.

Now, admittedly the "!= 0" doesn't really change the value (0 stays as
zero, 1 stays as one), but it's not only redundant, it confuses gcc, and
causes gcc to warn about the construct

    switch (rq_data_dir(req)) {
        case READ:
            ...
        case WRITE:
            ...

that we have in a few drivers.

Now, the gcc warning is silly and stupid (it seems to warn not about the
switch value having a different type from the case statements, but about
_any_ boolean switch value), but in this case the code itself is silly
and stupid too, so let's just change it, and get rid of warnings like
this:

  drivers/block/hd.c: In function ‘hd_request’:
  drivers/block/hd.c:630:11: warning: switch condition has boolean value [-Wswitch-bool]
     switch (rq_data_dir(req)) {

The odd '!= 0' came in when "cmd_flags" got turned into a "u64" in
commit 5953316d ("block: make rq->cmd_flags be 64-bit") and is
presumably because the old code (that just did a logical 'and' with 1)
would then end up making the type of rq_data_dir() be u64 too.

But if we want to retain the old regular integer type, let's just cast
the result to 'int' rather than use that rather odd '!= 0'.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix up the writeback plugging introduced in commit d353d758
("writeback: plug writeback at a high level") that then caused problems
due to the unplug happening with a spinlock held.

* writeback-plugging:
  writeback: plug writeback in wb_writeback() and writeback_inodes_wb()
  Revert "writeback: plug writeback at a high level"

We had to revert the pluggin in writeback_sb_inodes() because the
wb->list_lock is held, but we could easily plug at a higher level before
taking that lock, and unplug after releasing it.  This does that.

Chris will run performance numbers, just to verify that this approach is
comparable to the alternative (we could just drop and re-take the lock
around the blk_finish_plug() rather than these two commits.

I'd have preferred waiting for actual performance numbers before picking
one approach over the other, but I don't want to release rc1 with the
known "sleeping function called from invalid context" issue, so I'll
pick this cleanup version for now.  But if the numbers show that we
really want to plug just at the writeback_sb_inodes() level, and we
should just play ugly games with the spinlock, we'll switch to that.

Cc: Chris Mason <clm@fb.com>
Cc: Josef Bacik <jbacik@fb.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Neil Brown <neilb@suse.de>
Cc: Jan Kara <jack@suse.cz>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>