A long time ago, in v2.4, VM_RESERVED kept swapout process off VMA,
currently it lost original meaning but still has some effects:
| effect | alternative flags
-+------------------------+---------------------------------------------
1| account as reserved_vm | VM_IO
2| skip in core dump | VM_IO, VM_DONTDUMP
3| do not merge or expand | VM_IO, VM_DONTEXPAND, VM_HUGETLB, VM_PFNMAP
4| do not mlock | VM_IO, VM_DONTEXPAND, VM_HUGETLB, VM_PFNMAP
This patch removes reserved_vm counter from mm_struct. Seems like nobody
cares about it, it does not exported into userspace directly, it only
reduces total_vm showed in proc.
Thus VM_RESERVED can be replaced with VM_IO or pair VM_DONTEXPAND | VM_DONTDUMP.
remap_pfn_range() and io_remap_pfn_range() set VM_IO|VM_DONTEXPAND|VM_DONTDUMP.
remap_vmalloc_range() set VM_DONTEXPAND | VM_DONTDUMP.
[akpm@linux-foundation.org: drivers/vfio/pci/vfio_pci.c fixup]
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Carsten Otte <cotte@de.ibm.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Eric Paris <eparis@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Kentaro Takeda <takedakn@nttdata.co.jp>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Venkatesh Pallipadi <venki@google.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Eric Paris <eparis@parisplace.org>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
a) %d does _not_ produce a page worth of output
b) snprintf() doesn't return negatives - it used to in old glibc, but
that's the kernel...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
We reset the bool names and values array to NULL, but do not reset the
number of entries in these arrays to 0. If we error out and then get back
into this function we will walk these NULL pointers based on the belief
that they are non-zero length.
Signed-off-by: Eric Paris <eparis@redhat.com>
cc: stable@kernel.org
I'm not really sure what the idea behind the sel_div function is, but it's
useless. Since a and b are both unsigned, it's impossible for a % b < 0.
That means that part of the function never does anything. Thus it's just a
normal /. Just do that instead. I don't even understand what that operation
was supposed to mean in the signed case however....
If it was signed:
sel_div(-2, 4) == ((-2 / 4) - ((-2 % 4) < 0))
((0) - ((-2) < 0))
((0) - (1))
(-1)
What actually happens:
sel_div(-2, 4) == ((18446744073709551614 / 4) - ((18446744073709551614 % 4) < 0))
((4611686018427387903) - ((2 < 0))
(4611686018427387903 - 0)
((unsigned int)4611686018427387903)
(4294967295)
Neither makes a whole ton of sense to me. So I'm getting rid of the
function entirely.
Signed-off-by: Eric Paris <eparis@redhat.com>
There is no reason the DAC perms on reading the policy file need to be root
only. There are selinux checks which should control this access.
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Eric Paris <eparis@parisplace.org>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: James Morris <jmorris@namei.org>
My @hp.com will no longer be valid starting August 5, 2011 so an update is
necessary. My new email address is employer independent so we don't have
to worry about doing this again any time soon.
Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For a number of file systems that don't have a mount point (e.g. sockfs
and pipefs), they are not marked as long term. Therefore in
mntput_no_expire, all locks in vfs_mount lock are taken instead of just
local cpu's lock to aggregate reference counts when we release
reference to file objects. In fact, only local lock need to have been
taken to update ref counts as these file systems are in no danger of
going away until we are ready to unregister them.
The attached patch marks file systems using kern_mount without
mount point as long term. The contentions of vfs_mount lock
is now eliminated. Before un-registering such file system,
kern_unmount should be called to remove the long term flag and
make the mount point ready to be freed.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
I submit the patch again, according to patch submission convension.
This patch enables to accept percent-encoded object names as forth
argument of /selinux/create interface to avoid possible bugs when we
give an object name including whitespace or multibutes.
E.g) if and when a userspace object manager tries to create a new object
named as "resolve.conf but fake", it shall give this name as the forth
argument of the /selinux/create. But sscanf() logic in kernel space
fetches only the part earlier than the first whitespace.
In this case, selinux may unexpectedly answer a default security context
configured to "resolve.conf", but it is bug.
Although I could not test this patch on named TYPE_TRANSITION rules
actually, But debug printk() message seems to me the logic works
correctly.
I assume the libselinux provides an interface to apply this logic
transparently, so nothing shall not be changed from the viewpoint of
application.
Signed-off-by: KaiGai Kohei <kohei.kaigai@emea.nec.com>
Signed-off-by: Eric Paris <eparis@redhat.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (39 commits)
b43: fix comment typo reqest -> request
Haavard Skinnemoen has left Atmel
cris: typo in mach-fs Makefile
Kconfig: fix copy/paste-ism for dell-wmi-aio driver
doc: timers-howto: fix a typo ("unsgined")
perf: Only include annotate.h once in tools/perf/util/ui/browsers/annotate.c
md, raid5: Fix spelling error in comment ('Ofcourse' --> 'Of course').
treewide: fix a few typos in comments
regulator: change debug statement be consistent with the style of the rest
Revert "arm: mach-u300/gpio: Fix mem_region resource size miscalculations"
audit: acquire creds selectively to reduce atomic op overhead
rtlwifi: don't touch with treewide double semicolon removal
treewide: cleanup continuations and remove logging message whitespace
ath9k_hw: don't touch with treewide double semicolon removal
include/linux/leds-regulator.h: fix syntax in example code
tty: fix typo in descripton of tty_termios_encode_baud_rate
xtensa: remove obsolete BKL kernel option from defconfig
m68k: fix comment typo 'occcured'
arch:Kconfig.locks Remove unused config option.
treewide: remove extra semicolons
...
There is no point in counting hits - we can calculate it from the number
of lookups and misses.
This makes the avc statistics a bit smaller, and makes the code
generation better too.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the interest of keeping userspace from having to create new root
filesystems all the time, let's follow the lead of the other in-kernel
filesystems and provide a proper mount point for it in sysfs.
For selinuxfs, this mount point should be in /sys/fs/selinux/
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: James Morris <jmorris@namei.org>
Cc: Eric Paris <eparis@parisplace.org>
Cc: Lennart Poettering <mzerqung@0pointer.de>
Cc: Daniel J Walsh <dwalsh@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[include kobject.h - Eric Paris]
[use selinuxfs_obj throughout - Eric Paris]
Signed-off-by: Eric Paris <eparis@redhat.com>
The attached patch allows /selinux/create takes optional 4th argument
to support TYPE_TRANSITION with name extension for userspace object
managers.
If 4th argument is not supplied, it shall perform as existing kernel.
In fact, the regression test of SE-PostgreSQL works well on the patched
kernel.
Thanks,
Signed-off-by: KaiGai Kohei <kohei.kaigai@eu.nec.com>
[manually verify fuzz was not an issue, and it wasn't: eparis]
Signed-off-by: Eric Paris <eparis@redhat.com>
dget_locked was a shortcut to avoid the lazy lru manipulation when we already
held dcache_lock (lru manipulation was relatively cheap at that point).
However, how that the lru lock is an innermost one, we never hold it at any
caller, so the lock cost can now be avoided. We already have well working lazy
dcache LRU, so it should be fine to defer LRU manipulations to scan time.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Protect d_subdirs and d_child with d_lock, except in filesystems that aren't
using dcache_lock for these anyway (eg. using i_mutex).
Note: if we change the locking rule in future so that ->d_child protection is
provided only with ->d_parent->d_lock, it may allow us to reduce some locking.
But it would be an exception to an otherwise regular locking scheme, so we'd
have to see some good results. Probably not worthwhile.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
selinuxfs carefully uses i_ino to figure out what the inode refers to. The
VFS used to generically set this value and we would reset it to something
useable. After 85fe4025c6 each filesystem sets this value to a default
if needed. Since selinuxfs doesn't use the default value and it can only
lead to problems (I'd rather have 2 inodes with i_ino == 0 than one
pointing to the wrong data) lets just stop setting a default.
Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
selinuxfs.c has lots of different standards on how to handle return paths on
error. For the most part transition to
rc=errno
if (failure)
goto out;
[...]
out:
cleanup()
return rc;
Instead of doing cleanup mid function, or having multiple returns or other
options. This doesn't do that for every function, but most of the complex
functions which have cleanup routines on error.
Signed-off-by: Eric Paris <eparis@redhat.com>
Instead of always assigning an increasing inode number in new_inode
move the call to assign it into those callers that actually need it.
For now callers that need it is estimated conservatively, that is
the call is added to all filesystems that do not assign an i_ino
by themselves. For a few more filesystems we can avoid assigning
any inode number given that they aren't user visible, and for others
it could be done lazily when an inode number is actually needed,
but that's left for later patches.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
/selinux/policy allows a user to copy the policy back out of the kernel.
This patch allows userspace to actually mmap that file and use it directly.
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
There is interest in being able to see what the actual policy is that was
loaded into the kernel. The patch creates a new selinuxfs file
/selinux/policy which can be read by userspace. The actual policy that is
loaded into the kernel will be written back out to userspace.
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
This patch provides a new /selinux/status entry which allows applications
read-only mmap(2).
This region reflects selinux_kernel_status structure in kernel space.
struct selinux_kernel_status
{
u32 length; /* length of this structure */
u32 sequence; /* sequence number of seqlock logic */
u32 enforcing; /* current setting of enforcing mode */
u32 policyload; /* times of policy reloaded */
u32 deny_unknown; /* current setting of deny_unknown */
};
When userspace object manager caches access control decisions provided
by SELinux, it needs to invalidate the cache on policy reload and setenforce
to keep consistency.
However, the applications need to check the kernel state for each accesses
on userspace avc, or launch a background worker process.
In heuristic, frequency of invalidation is much less than frequency of
making access control decision, so it is annoying to invoke a system call
to check we don't need to invalidate the userspace cache.
If we can use a background worker thread, it allows to receive invalidation
messages from the kernel. But it requires us an invasive coding toward the
base application in some cases; E.g, when we provide a feature performing
with SELinux as a plugin module, it is unwelcome manner to launch its own
worker thread from the module.
If we could map /selinux/status to process memory space, application can
know updates of selinux status; policy reload or setenforce.
A typical application checks selinux_kernel_status::sequence when it tries
to reference userspace avc. If it was changed from the last time when it
checked userspace avc, it means something was updated in the kernel space.
Then, the application can reset userspace avc or update current enforcing
mode, without any system call invocations.
This sequence number is updated according to the seqlock logic, so we need
to wait for a while if it is odd number.
Signed-off-by: KaiGai Kohei <kaigai@ak.jp.nec.com>
Acked-by: Eric Paris <eparis@redhat.com>
--
security/selinux/include/security.h | 21 ++++++
security/selinux/selinuxfs.c | 56 +++++++++++++++
security/selinux/ss/Makefile | 2 +-
security/selinux/ss/services.c | 3 +
security/selinux/ss/status.c | 129 +++++++++++++++++++++++++++++++++++
5 files changed, 210 insertions(+), 1 deletions(-)
Signed-off-by: James Morris <jmorris@namei.org>
The default for llseek will change to no_llseek,
so selinuxfs needs to add explicit .llseek
assignments. Since we're dealing with regular
files from a VFS perspective, use generic_file_llseek.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: Eric Paris <eparis@parisplace.org>
Signed-off-by: James Morris <jmorris@namei.org>
In sel_make_bools, kernel allocates memory for bool_pending_names[i]
with security_get_bools. So if we just free bool_pending_names, those
memories for bool_pending_names[i] will be leaked.
This patch resolves dozens of following kmemleak report after resuming
from suspend:
unreferenced object 0xffff88022e4c7380 (size 32):
comm "init", pid 1, jiffies 4294677173
backtrace:
[<ffffffff810f76b5>] create_object+0x1a2/0x2a9
[<ffffffff810f78bb>] kmemleak_alloc+0x26/0x4b
[<ffffffff810ef3eb>] __kmalloc+0x18f/0x1b8
[<ffffffff811cd511>] security_get_bools+0xd7/0x16f
[<ffffffff811c48c0>] sel_write_load+0x12e/0x62b
[<ffffffff810f9a39>] vfs_write+0xae/0x10b
[<ffffffff810f9b56>] sys_write+0x4a/0x6e
[<ffffffff81011b82>] system_call_fastpath+0x16/0x1b
[<ffffffffffffffff>] 0xffffffffffffffff
Signed-off-by: Xiaotian Feng <dfeng@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
Allow runtime switching between different policy types (e.g. from a MLS/MCS
policy to a non-MLS/non-MCS policy or viceversa).
Signed-off-by: Guido Trentalancia <guido@trentalancia.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
If allow_unknown==deny, SELinux treats an undefined kernel security
class as an error condition rather than as a typical permission denial
and thus does not allow permissions on undefined classes even when in
permissive mode. Change the SELinux logic so that this case is handled
as a typical permission denial, subject to the usual permissive mode and
permissive domain handling.
Also drop the 'requested' argument from security_compute_av() and
helpers as it is a legacy of the original security server interface and
is unused.
Changes:
- Handle permissive domains consistently by moving up the test for a
permissive domain.
- Make security_compute_av_user() consistent with security_compute_av();
the only difference now is that security_compute_av() performs mapping
between the kernel-private class and permission indices and the policy
values. In the userspace case, this mapping is handled by libselinux.
- Moved avd_init inside the policy lock.
Based in part on a patch by Paul Moore <paul.moore@hp.com>.
Reported-by: Andrew Worsley <amworsley@gmail.com>
Signed-off-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Reviewed-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
Modify SELinux to dynamically discover class and permission values
upon policy load, based on the dynamic object class/perm discovery
logic from libselinux. A mapping is created between kernel-private
class and permission indices used outside the security server and the
policy values used within the security server.
The mappings are only applied upon kernel-internal computations;
similar mappings for the private indices of userspace object managers
is handled on a per-object manager basis by the userspace AVC. The
interfaces for compute_av and transition_sid are split for kernel
vs. userspace; the userspace functions are distinguished by a _user
suffix.
The kernel-private class indices are no longer tied to the policy
values and thus do not need to skip indices for userspace classes;
thus the kernel class index values are compressed. The flask.h
definitions were regenerated by deleting the userspace classes from
refpolicy's definitions and then regenerating the headers. Going
forward, we can just maintain the flask.h, av_permissions.h, and
classmap.h definitions separately from policy as they are no longer
tied to the policy values. The next patch introduces a utility to
automate generation of flask.h and av_permissions.h from the
classmap.h definitions.
The older kernel class and permission string tables are removed and
replaced by a single security class mapping table that is walked at
policy load to generate the mapping. The old kernel class validation
logic is completely replaced by the mapping logic.
The handle unknown logic is reworked. reject_unknown=1 is handled
when the mappings are computed at policy load time, similar to the old
handling by the class validation logic. allow_unknown=1 is handled
when computing and mapping decisions - if the permission was not able
to be mapped (i.e. undefined, mapped to zero), then it is
automatically added to the allowed vector. If the class was not able
to be mapped (i.e. undefined, mapped to zero), then all permissions
are allowed for it if allow_unknown=1.
avc_audit leverages the new security class mapping table to lookup the
class and permission names from the kernel-private indices.
The mdp program is updated to use the new table when generating the
class definitions and allow rules for a minimal boot policy for the
kernel. It should be noted that this policy will not include any
userspace classes, nor will its policy index values for the kernel
classes correspond with the ones in refpolicy (they will instead match
the kernel-private indices).
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
On Tue, 2009-05-19 at 00:05 -0400, Eamon Walsh wrote:
> Recent versions of coreutils have bumped the read buffer size from 4K to
> 32K in several of the utilities.
>
> This means that "cat /selinux/booleans/xserver_object_manager" no longer
> works, it returns "Invalid argument" on F11. getsebool works fine.
>
> sel_read_bool has a check for "count > PAGE_SIZE" that doesn't seem to
> be present in the other read functions. Maybe it could be removed?
Yes, that check is obsoleted by the conversion of those functions to
using simple_read_from_buffer(), which will reduce count if necessary to
what is available in the buffer.
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
This patch enables applications to handle permissive domain correctly.
Since the v2.6.26 kernel, SELinux has supported an idea of permissive
domain which allows certain processes to work as if permissive mode,
even if the global setting is enforcing mode.
However, we don't have an application program interface to inform
what domains are permissive one, and what domains are not.
It means applications focuses on SELinux (XACE/SELinux, SE-PostgreSQL
and so on) cannot handle permissive domain correctly.
This patch add the sixth field (flags) on the reply of the /selinux/access
interface which is used to make an access control decision from userspace.
If the first bit of the flags field is positive, it means the required
access control decision is on permissive domain, so application should
allow any required actions, as the kernel doing.
This patch also has a side benefit. The av_decision.flags is set at
context_struct_compute_av(). It enables to check required permissions
without read_lock(&policy_rwlock).
Signed-off-by: KaiGai Kohei <kaigai@ak.jp.nec.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
Acked-by: Eric Paris <eparis@redhat.com>
--
security/selinux/avc.c | 2 +-
security/selinux/include/security.h | 4 +++-
security/selinux/selinuxfs.c | 4 ++--
security/selinux/ss/services.c | 30 +++++-------------------------
4 files changed, 11 insertions(+), 29 deletions(-)
Signed-off-by: James Morris <jmorris@namei.org>
The SELinux "compat_net" is marked as deprecated, the time has come to
finally remove it from the kernel. Further code simplifications are
likely in the future, but this patch was intended to be a simple,
straight-up removal of the compat_net code.
Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
It appears there was an intention to have the security server only decide
certain permissions and leave other for later as some sort of a portential
performance win. We are currently always deciding all 32 bits of
permissions and this is a useless couple of branches and wasted space.
This patch completely drops the av.decided concept.
This in a 17% reduction in the time spent in avc_has_perm_noaudit
based on oprofile sampling of a tbench benchmark.
Signed-off-by: Eric Paris <eparis@redhat.com>
Reviewed-by: Paul Moore <paul.moore@hp.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
... and don't bother in callers. Don't bother with zeroing i_blocks,
while we are at it - it's already been zeroed.
i_mode is not worth the effort; it has no common default value.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Impact: cleanup
In future, all cpumask ops will only be valid (in general) for bit
numbers < nr_cpu_ids. So use that instead of NR_CPUS in iterators
and other comparisons.
This is always safe: no cpu number can be >= nr_cpu_ids, and
nr_cpu_ids is initialized to NR_CPUS at boot.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: James Morris <jmorris@namei.org>
Cc: Eric Biederman <ebiederm@xmission.com>
This patch is the first step towards removing the old "compat_net" code from
the kernel. Secmark, the "compat_net" replacement was first introduced in
2.6.18 (September 2006) and the major Linux distributions with SELinux support
have transitioned to Secmark so it is time to start deprecating the "compat_net"
mechanism. Testing a patched version of 2.6.28-rc6 with the initial release of
Fedora Core 5 did not show any problems when running in enforcing mode.
This patch adds an entry to the feature-removal-schedule.txt file and removes
the SECURITY_SELINUX_ENABLE_SECMARK_DEFAULT configuration option, forcing
Secmark on by default although it can still be disabled at runtime. The patch
also makes the Secmark permission checks "dynamic" in the sense that they are
only executed when Secmark is configured; this should help prevent problems
with older distributions that have not yet migrated to Secmark.
Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
Use RCU to access another task's creds and to release a task's own creds.
This means that it will be possible for the credentials of a task to be
replaced without another task (a) requiring a full lock to read them, and (b)
seeing deallocated memory.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
Separate the task security context from task_struct. At this point, the
security data is temporarily embedded in the task_struct with two pointers
pointing to it.
Note that the Alpha arch is altered as it refers to (E)UID and (E)GID in
entry.S via asm-offsets.
With comment fixes Signed-off-by: Marc Dionne <marc.c.dionne@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
