Commit graph

925 commits

Author SHA1 Message Date
James Morris
d254117099 Merge branch 'master' into next 2009-05-08 17:56:47 +10:00
Mimi Zohar
07ff7a0b18 integrity: remove __setup auditing msgs
Remove integrity audit messages from __setup()

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-05-06 10:35:59 +10:00
Mimi Zohar
e5e520a715 integrity: use audit_log_string
Based on a request from Eric Paris to simplify parsing, replace
audit_log_format statements containing "%s" with audit_log_string().

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-05-06 10:35:56 +10:00
Mimi Zohar
53fc0e2259 integrity: lsm audit rule matching fix
An audit subsystem change replaced AUDIT_EQUAL with Audit_equal.
Update calls to security_filter_rule_init()/match() to reflect
the change.

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-05-06 10:35:53 +10:00
Stephen Smalley
65c90bca0d selinux: Fix send_sigiotask hook
The CRED patch incorrectly converted the SELinux send_sigiotask hook to
use the current task SID rather than the target task SID in its
permission check, yielding the wrong permission check.  This fixes the
hook function.  Detected by the ltp selinux testsuite and confirmed to
correct the test failure.

Signed-off-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-05-05 08:31:03 +10:00
Oleg Nesterov
ecd6de3c88 selinux: selinux_bprm_committed_creds() should wake up ->real_parent, not ->parent.
We shouldn't worry about the tracer if current is ptraced, exec() must not
succeed if the tracer has no rights to trace this task after cred changing.
But we should notify ->real_parent which is, well, real parent.

Also, we don't need _irq to take tasklist, and we don't need parent's
->siglock to wake_up_interruptible(real_parent->signal->wait_chldexit).
Since we hold tasklist, real_parent->signal must be stable. Otherwise
spin_lock(siglock) is not safe too and can't help anyway.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-30 09:08:48 +10:00
David Howells
3bcac0263f SELinux: Don't flush inherited SIGKILL during execve()
Don't flush inherited SIGKILL during execve() in SELinux's post cred commit
hook.  This isn't really a security problem: if the SIGKILL came before the
credentials were changed, then we were right to receive it at the time, and
should honour it; if it came after the creds were changed, then we definitely
should honour it; and in any case, all that will happen is that the process
will be scrapped before it ever returns to userspace.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-30 09:07:13 +10:00
Eric Paris
88c48db978 SELinux: drop secondary_ops->sysctl
We are still calling secondary_ops->sysctl even though the capabilities
module does not define a sysctl operation.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-30 08:45:56 +10:00
Etienne Basset
defc433ba3 Smack: check for SMACK xattr validity in smack_inode_setxattr
the following patch moves checks for SMACK xattr validity
from smack_inode_post_setxattr (which cannot return an error to the user)
to smack_inode_setxattr (which can return an error).

Signed-off-by: Etienne Basset <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-18 12:58:25 +10:00
Jiri Pirko
05725f7eb4 rculist: use list_entry_rcu in places where it's appropriate
Use previously introduced list_entry_rcu instead of an open-coded
list_entry + rcu_dereference combination.

Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: dipankar@in.ibm.com
LKML-Reference: <20090414181715.GA3634@psychotron.englab.brq.redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-15 12:05:25 +02:00
Tetsuo Handa
39826a1e17 tomoyo: version bump to 2.2.0.
Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-14 09:15:02 +10:00
Etienne Basset
ecfcc53fef smack: implement logging V3
the following patch, add logging of Smack security decisions.
This is of course very useful to understand what your current smack policy does.
As suggested by Casey, it also now forbids labels with ', " or \

It introduces a '/smack/logging' switch :
0: no logging
1: log denied (default)
2: log accepted
3: log denied&accepted

Signed-off-by: Etienne Basset <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-14 09:00:23 +10:00
Etienne Basset
6e837fb152 smack: implement logging V3
This patch creates auditing functions usable by LSM to audit security
events. It provides standard dumping of FS, NET, task etc ... events
(code borrowed from SELinux)
and provides 2 callbacks to define LSM specific auditing, which should be
flexible enough to convert SELinux too.

Signed-off-by: Etienne Basset <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
cked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-14 09:00:19 +10:00
David Howells
34574dd10b keys: Handle there being no fallback destination keyring for request_key()
When request_key() is called, without there being any standard process
keyrings on which to fall back if a destination keyring is not specified, an
oops is liable to occur when construct_alloc_key() calls down_write() on
dest_keyring's semaphore.

Due to function inlining this may be seen as an oops in down_write() as called
from request_key_and_link().

This situation crops up during boot, where request_key() is called from within
the kernel (such as in CIFS mounts) where nobody is actually logged in, and so
PAM has not had a chance to create a session keyring and user keyrings to act
as the fallback.

To fix this, make construct_alloc_key() not attempt to cache a key if there is
no fallback key if no destination keyring is given specifically.

Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-09 10:41:19 -07:00
Serge E. Hallyn
5bf37ec3e0 cap_prctl: don't set error to 0 at 'no_change'
One-liner: capsh --print is broken without this patch.

In certain cases, cap_prctl returns error > 0 for success.  However,
the 'no_change' label was always setting error to 0.  As a result,
for example, 'prctl(CAP_BSET_READ, N)' would always return 0.
It should return 1 if a process has N in its bounding set (as
by default it does).

I'm keeping the no_change label even though it's now functionally
the same as 'error'.

Signed-off-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-09 09:12:03 +10:00
Tetsuo Handa
a0558fc349 tomoyo: remove "undelete domain" command.
Since TOMOYO's policy management tools does not use the "undelete domain"
command, we decided to remove that command.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-07 16:08:56 +10:00
Tetsuo Handa
7ba5779533 tomoyo: remove "undelete domain" command.
Since TOMOYO's policy management tools does not use the "undelete domain"
command, we decided to remove that command.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-07 08:17:43 +10:00
David Howells
800a964787 CacheFiles: Export things for CacheFiles
Export a number of functions for CacheFiles's use.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Steve Dickson <steved@redhat.com>
Acked-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Tested-by: Daire Byrne <Daire.Byrne@framestore.com>
2009-04-03 16:42:40 +01:00
Linus Torvalds
8fe74cf053 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  Remove two unneeded exports and make two symbols static in fs/mpage.c
  Cleanup after commit 585d3bc06f
  Trim includes of fdtable.h
  Don't crap into descriptor table in binfmt_som
  Trim includes in binfmt_elf
  Don't mess with descriptor table in load_elf_binary()
  Get rid of indirect include of fs_struct.h
  New helper - current_umask()
  check_unsafe_exec() doesn't care about signal handlers sharing
  New locking/refcounting for fs_struct
  Take fs_struct handling to new file (fs/fs_struct.c)
  Get rid of bumping fs_struct refcount in pivot_root(2)
  Kill unsharing fs_struct in __set_personality()
2009-04-02 21:09:10 -07:00
Li Zefan
b4046f00ee devcgroup: avoid using cgroup_lock
There is nothing special that has to be protected by cgroup_lock,
so introduce devcgroup_mtuex for it's own use.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-02 19:04:55 -07:00
Serge E. Hallyn
b5f22a59c0 don't raise all privs on setuid-root file with fE set (v2)
Distributions face a backward compatibility problem with starting to use
file capabilities.  For instance, removing setuid root from ping and
doing setcap cap_net_raw=pe means that booting with an older kernel
or one compiled without file capabilities means ping won't work for
non-root users.

In order to replace the setuid root bit on a capability-unaware
program, one has to set the effective, or legacy, file capability,
which makes the capability effective immediately.  This patch
uses the legacy bit as a queue to not automatically add full
privilege to a setuid-root program.

So, with this patch, an ordinary setuid-root program will run with
privilege.  But if /bin/ping has both setuid-root and cap_net_raw in
fP and fE, then ping (when run by non-root user) will not run
with only cap_net_raw.

Changelog:
	Apr 2 2009: Print a message once when such a binary is loaded,
		as per James Morris' suggestion.
	Apr 2 2009: Fix the condition to only catch uid!=0 && euid==0.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-04-03 11:49:31 +11:00
KaiGai Kohei
8a6f83afd0 Permissive domain in userspace object manager
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>
2009-04-02 09:23:45 +11:00
Al Viro
5ad4e53bd5 Get rid of indirect include of fs_struct.h
Don't pull it in sched.h; very few files actually need it and those
can include directly.  sched.h itself only needs forward declaration
of struct fs_struct;

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2009-03-31 23:00:27 -04:00
Etienne Basset
4303154e86 smack: Add a new '-CIPSO' option to the network address label configuration
This patch adds a new special option '-CIPSO' to the Smack subsystem. When used
in the netlabel list, it means "use CIPSO networking". A use case is when your
local network speaks CIPSO and you want also to connect to the unlabeled
Internet. This patch also add some documentation describing that. The patch
also corrects an oops when setting a '' SMACK64 xattr to a file.

Signed-off-by: Etienne Basset <etienne.basset@numericable.fr>
Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-28 15:01:37 +11:00
Paul Moore
07feee8f81 netlabel: Cleanup the Smack/NetLabel code to fix incoming TCP connections
This patch cleans up a lot of the Smack network access control code.  The
largest changes are to fix the labeling of incoming TCP connections in a
manner similar to the recent SELinux changes which use the
security_inet_conn_request() hook to label the request_sock and let the label
move to the child socket via the normal network stack mechanisms.  In addition
to the incoming TCP connection fixes this patch also removes the smk_labled
field from the socket_smack struct as the minor optimization advantage was
outweighed by the difficulty in maintaining it's proper state.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-28 15:01:37 +11:00
Paul Moore
8651d5c0b1 lsm: Remove the socket_post_accept() hook
The socket_post_accept() hook is not currently used by any in-tree modules
and its existence continues to cause problems by confusing people about
what can be safely accomplished using this hook.  If a legitimate need for
this hook arises in the future it can always be reintroduced.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-28 15:01:37 +11:00
Paul Moore
58bfbb51ff selinux: Remove the "compat_net" compatibility code
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>
2009-03-28 15:01:37 +11:00
Paul Moore
389fb800ac netlabel: Label incoming TCP connections correctly in SELinux
The current NetLabel/SELinux behavior for incoming TCP connections works but
only through a series of happy coincidences that rely on the limited nature of
standard CIPSO (only able to convey MLS attributes) and the write equality
imposed by the SELinux MLS constraints.  The problem is that network sockets
created as the result of an incoming TCP connection were not on-the-wire
labeled based on the security attributes of the parent socket but rather based
on the wire label of the remote peer.  The issue had to do with how IP options
were managed as part of the network stack and where the LSM hooks were in
relation to the code which set the IP options on these newly created child
sockets.  While NetLabel/SELinux did correctly set the socket's on-the-wire
label it was promptly cleared by the network stack and reset based on the IP
options of the remote peer.

This patch, in conjunction with a prior patch that adjusted the LSM hook
locations, works to set the correct on-the-wire label format for new incoming
connections through the security_inet_conn_request() hook.  Besides the
correct behavior there are many advantages to this change, the most significant
is that all of the NetLabel socket labeling code in SELinux now lives in hooks
which can return error codes to the core stack which allows us to finally get
ride of the selinux_netlbl_inode_permission() logic which greatly simplfies
the NetLabel/SELinux glue code.  In the process of developing this patch I
also ran into a small handful of AF_INET6 cleanliness issues that have been
fixed which should make the code safer and easier to extend in the future.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-28 15:01:36 +11:00
Tetsuo Handa
a106cbfd1f TOMOYO: Fix a typo.
Fix a typo.

Reported-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-27 19:03:44 +11:00
Etienne Basset
7198e2eeb4 smack: convert smack to standard linux lists
the following patch (on top of 2.6.29) converts Smack lists to standard linux lists
Please review and consider for inclusion in 2.6.30-rc

regards,
Etienne

Signed-off-by: Etienne Basset <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
2009-03-26 09:17:04 +11:00
James Morris
703a3cd728 Merge branch 'master' into next 2009-03-24 10:52:46 +11:00
Eric Paris
df7f54c012 SELinux: inode_doinit_with_dentry drop no dentry printk
Drop the printk message when an inode is found without an associated
dentry.  This should only happen when userspace can't be accessing those
inodes and those labels will get set correctly on the next d_instantiate.
Thus there is no reason to send this message.

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-10 08:40:02 +11:00
Eric Paris
dd34b5d75a SELinux: new permission between tty audit and audit socket
New selinux permission to separate the ability to turn on tty auditing from
the ability to set audit rules.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-06 08:50:21 +11:00
Eric Paris
6a25b27d60 SELinux: open perm for sock files
When I did open permissions I didn't think any sockets would have an open.
Turns out AF_UNIX sockets can have an open when they are bound to the
filesystem namespace.  This patch adds a new SOCK_FILE__OPEN permission.
It's safe to add this as the open perms are already predicated on
capabilities and capabilities means we have unknown perm handling so
systems should be as backwards compatible as the policy wants them to
be.

https://bugzilla.redhat.com/show_bug.cgi?id=475224

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-06 08:50:18 +11:00
etienne
211a40c087 smack: fixes for unlabeled host support
The following patch (against 2.6.29rc5) fixes a few issues in the
smack/netlabel "unlabeled host support" functionnality that was added in
2.6.29rc.  It should go in before -final.

1) smack_host_label disregard a "0.0.0.0/0 @" rule (or other label),
preventing 'tagged' tasks to access Internet (many systems drop packets with
IP options)

2) netmasks were not handled correctly, they were stored in a way _not
equivalent_ to conversion to be32 (it was equivalent for /0, /8, /16, /24,
/32 masks but not other masks)

3) smack_netlbladdr prefixes (IP/mask) were not consistent (mask&IP was not
done), so there could have been different list entries for the same IP
prefix; if those entries had different labels, well ...

4) they were not sorted

1) 2) 3) are bugs, 4) is a more cosmetic issue.
The patch :

-creates a new helper smk_netlbladdr_insert to insert a smk_netlbladdr,
-sorted by netmask length

-use the new sorted nature of  smack_netlbladdrs list to simplify
 smack_host_label : the first match _will_ be the more specific

-corrects endianness issues in smk_write_netlbladdr &  netlbladdr_seq_show

Signed-off-by: <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-05 08:36:34 +11:00
etienne
113a0e4590 smack: fixes for unlabeled host support
The following patch (against 2.6.29rc5) fixes a few issues in the
smack/netlabel "unlabeled host support" functionnality that was added in
2.6.29rc.  It should go in before -final.

1) smack_host_label disregard a "0.0.0.0/0 @" rule (or other label),
preventing 'tagged' tasks to access Internet (many systems drop packets with
IP options)

2) netmasks were not handled correctly, they were stored in a way _not
equivalent_ to conversion to be32 (it was equivalent for /0, /8, /16, /24,
/32 masks but not other masks)

3) smack_netlbladdr prefixes (IP/mask) were not consistent (mask&IP was not
done), so there could have been different list entries for the same IP
prefix; if those entries had different labels, well ...

4) they were not sorted

1) 2) 3) are bugs, 4) is a more cosmetic issue.
The patch :

-creates a new helper smk_netlbladdr_insert to insert a smk_netlbladdr,
-sorted by netmask length

-use the new sorted nature of  smack_netlbladdrs list to simplify
 smack_host_label : the first match _will_ be the more specific

-corrects endianness issues in smk_write_netlbladdr &  netlbladdr_seq_show

Signed-off-by: <etienne.basset@numericable.fr>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-05 08:30:01 +11:00
Paul Moore
d7f59dc464 selinux: Fix a panic in selinux_netlbl_inode_permission()
Rick McNeal from LSI identified a panic in selinux_netlbl_inode_permission()
caused by a certain sequence of SUNRPC operations.  The problem appears to be
due to the lack of NULL pointer checking in the function; this patch adds the
pointer checks so the function will exit safely in the cases where the socket
is not completely initialized.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-03-02 09:30:04 +11:00
Serge E. Hallyn
454804ab03 keys: make procfiles per-user-namespace
Restrict the /proc/keys and /proc/key-users output to keys
belonging to the same user namespace as the reading task.

We may want to make this more complicated - so that any
keys in a user-namespace which is belongs to the reading
task are also shown.  But let's see if anyone wants that
first.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-27 12:35:15 +11:00
Serge E. Hallyn
2ea190d0a0 keys: skip keys from another user namespace
When listing keys, do not return keys belonging to the
same uid in another user namespace.  Otherwise uid 500
in another user namespace will return keyrings called
uid.500 for another user namespace.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-27 12:35:12 +11:00
Serge E. Hallyn
8ff3bc3138 keys: consider user namespace in key_permission
If a key is owned by another user namespace, then treat the
key as though it is owned by both another uid and gid.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-27 12:35:09 +11:00
Serge E. Hallyn
1d1e97562e keys: distinguish per-uid keys in different namespaces
per-uid keys were looked by uid only.  Use the user namespace
to distinguish the same uid in different namespaces.

This does not address key_permission.  So a task can for instance
try to join a keyring owned by the same uid in another namespace.
That will be handled by a separate patch.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-27 12:35:06 +11:00
Paul Moore
09c50b4a52 selinux: Fix the NetLabel glue code for setsockopt()
At some point we (okay, I) managed to break the ability for users to use the
setsockopt() syscall to set IPv4 options when NetLabel was not active on the
socket in question.  The problem was noticed by someone trying to use the
"-R" (record route) option of ping:

 # ping -R 10.0.0.1
 ping: record route: No message of desired type

The solution is relatively simple, we catch the unlabeled socket case and
clear the error code, allowing the operation to succeed.  Please note that we
still deny users the ability to override IPv4 options on socket's which have
NetLabel labeling active; this is done to ensure the labeling remains intact.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-23 10:05:55 +11:00
Mimi Zohar
be38e0fd5f integrity: ima iint radix_tree_lookup locking fix
Based on Andrew Morton's comments:
- add missing locks around radix_tree_lookup in ima_iint_insert()

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-23 09:54:53 +11:00
Tetsuo Handa
1581e7ddbd TOMOYO: Do not call tomoyo_realpath_init unless registered.
tomoyo_realpath_init() is unconditionally called by security_initcall().
But nobody will use realpath related functions if TOMOYO is not registered.

So, let tomoyo_init() call tomoyo_realpath_init().

This patch saves 4KB of memory allocation if TOMOYO is not registered.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-23 09:45:05 +11:00
Mimi Zohar
0da0a420bb integrity: ima scatterlist bug fix
Based on Alexander Beregalov's post http://lkml.org/lkml/2009/2/19/198

- replaced sg_set_buf() with sg_init_one()

 kernel BUG at include/linux/scatterlist.h:65!
 invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
 last sysfs file:
 CPU 2
 Modules linked in:
 Pid: 1, comm: swapper Not tainted 2.6.29-rc5-next-20090219 #5 PowerEdge 1950
 RIP: 0010:[<ffffffff8045ec70>]  [<ffffffff8045ec70>] ima_calc_hash+0xc0/0x160
 RSP: 0018:ffff88007f46bc40  EFLAGS: 00010286
 RAX: ffffe200032c45e8 RBX: 00000000fffffff4 RCX: 0000000087654321
 RDX: 0000000000000002 RSI: 0000000000000001 RDI: ffff88007cf71048
 RBP: ffff88007f46bcd0 R08: 0000000000000000 R09: 0000000000000163
 R10: ffff88007f4707a8 R11: 0000000000000000 R12: ffff88007cf71048
 R13: 0000000000001000 R14: 0000000000000000 R15: 0000000000009d98
 FS:  0000000000000000(0000) GS:ffff8800051ac000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
 CR2: 0000000000000000 CR3: 0000000000201000 CR4: 00000000000006e0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Tested-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-21 00:29:59 +11:00
Randy Dunlap
251a2a958b smack: fix lots of kernel-doc notation
Fix/add kernel-doc notation and fix typos in security/smack/.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-19 15:51:10 +11:00
Tetsuo Handa
e5a3b95f58 TOMOYO: Don't create securityfs entries unless registered.
TOMOYO should not create /sys/kernel/security/tomoyo/ interface unless
TOMOYO is registered.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-16 09:01:48 +11:00
Tetsuo Handa
33043cbb9f TOMOYO: Fix exception policy read failure.
Due to wrong initialization, "cat /sys/kernel/security/tomoyo/exception_policy"
returned nothing.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 12:33:30 +11:00
Eric Paris
26036651c5 SELinux: convert the avc cache hash list to an hlist
We do not need O(1) access to the tail of the avc cache lists and so we are
wasting lots of space using struct list_head instead of struct hlist_head.
This patch converts the avc cache to use hlists in which there is a single
pointer from the head which saves us about 4k of global memory.

Resulted in about a 1.5% decrease in time spent in avc_has_perm_noaudit based
on oprofile sampling of tbench.  Although likely within the noise....

Signed-off-by: Eric Paris <eparis@redhat.com>
Reviewed-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:23:48 +11:00
Eric Paris
edf3d1aecd SELinux: code readability with avc_cache
The code making use of struct avc_cache was not easy to read thanks to liberal
use of &avc_cache.{slots_lock,slots}[hvalue] throughout.  This patch simply
creates local pointers and uses those instead of the long global names.

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:23:45 +11:00
Eric Paris
f1c6381a6e SELinux: remove unused av.decided field
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>
2009-02-14 09:23:08 +11:00
Eric Paris
21193dcd1f SELinux: more careful use of avd in avc_has_perm_noaudit
we are often needlessly jumping through hoops when it comes to avd
entries in avc_has_perm_noaudit and we have extra initialization and memcpy
which are just wasting performance.  Try to clean the function up a bit.

This patch resulted in a 13% drop in time spent in avc_has_perm_noaudit in my
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>
2009-02-14 09:23:04 +11:00
Eric Paris
906d27d9d2 SELinux: remove the unused ae.used
Currently SELinux code has an atomic which was intended to track how many
times an avc entry was used and to evict entries when they haven't been
used recently.  Instead we never let this atomic get above 1 and evict when
it is first checked for eviction since it hits zero.  This is a total waste
of time so I'm completely dropping ae.used.

This change resulted in about a 3% faster avc_has_perm_noaudit when running
oprofile against 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>
2009-02-14 09:22:37 +11:00
Eric Paris
a5dda68332 SELinux: check seqno when updating an avc_node
The avc update node callbacks do not check the seqno of the caller with the
seqno of the node found.  It is possible that a policy change could happen
(although almost impossibly unlikely) in which a permissive or
permissive_domain decision is not valid for the entry found.  Simply pass
and check that the seqno of the caller and the seqno of the node found
match.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:22:34 +11:00
Eric Paris
4cb912f1d1 SELinux: NULL terminate al contexts from disk
When a context is pulled in from disk we don't know that it is null
terminated.  This patch forecebly null terminates contexts when we pull
them from disk.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:22:30 +11:00
Eric Paris
4ba0a8ad63 SELinux: better printk when file with invalid label found
Currently when an inode is read into the kernel with an invalid label
string (can often happen with removable media) we output a string like:

SELinux: inode_doinit_with_dentry:  context_to_sid([SOME INVALID LABEL])
returned -22 dor dev=[blah] ino=[blah]

Which is all but incomprehensible to all but a couple of us.  Instead, on
EINVAL only, I plan to output a much more user friendly string and I plan to
ratelimit the printk since many of these could be generated very rapidly.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:22:27 +11:00
Eric Paris
200ac532a4 SELinux: call capabilities code directory
For cleanliness and efficiency remove all calls to secondary-> and instead
call capabilities code directly.  capabilities are the only module that
selinux stacks with and so the code should not indicate that other stacking
might be possible.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-14 09:22:24 +11:00
Randy Dunlap
b53fab9d48 ima: fix build error
IMA_LSM_RULES requires AUDIT.  This is automatic if SECURITY_SELINUX=y
but not when SECURITY_SMACK=y (and SECURITY_SELINUX=n), so make the
dependency explicit.  This fixes the following build error:

security/integrity/ima/ima_policy.c:111:error: implicit declaration of function 'security_audit_rule_match'
security/integrity/ima/ima_policy.c:230:error: implicit declaration of function 'security_audit_rule_init'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-13 09:27:56 +11:00
Tetsuo Handa
35d50e60e8 tomoyo: fix sparse warning
Fix sparse warning.

$ make C=2 SUBDIRS=security/tomoyo CF="-D__cold__="
 CHECK   security/tomoyo/common.c
 CHECK   security/tomoyo/realpath.c
 CHECK   security/tomoyo/tomoyo.c
security/tomoyo/tomoyo.c:110:8: warning: symbol 'buf' shadows an earlier one
security/tomoyo/tomoyo.c💯7: originally declared here

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 20:21:10 +11:00
James Morris
42d5aaa2d8 security: change link order of LSMs so security=tomoyo works
LSMs need to be linked before root_plug to ensure the security=
boot parameter works with them.  Do this for Tomoyo.

(root_plug probably needs to be taken out and shot at some point,
too).

Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 16:29:04 +11:00
Kentaro Takeda
00d7d6f840 Kconfig and Makefile
TOMOYO uses LSM hooks for pathname based access control and securityfs support.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:19:00 +11:00
Kentaro Takeda
f743324377 LSM adapter functions.
DAC's permissions and TOMOYO's permissions are not one-to-one mapping.

Regarding DAC, there are "read", "write", "execute" permissions.
Regarding TOMOYO, there are "allow_read", "allow_write", "allow_read/write",
"allow_execute", "allow_create", "allow_unlink", "allow_mkdir", "allow_rmdir",
"allow_mkfifo", "allow_mksock", "allow_mkblock", "allow_mkchar",
"allow_truncate", "allow_symlink", "allow_rewrite", "allow_link",
"allow_rename" permissions.

+----------------------------------+----------------------------------+
| requested operation              | required TOMOYO's permission     |
+----------------------------------+----------------------------------+
| sys_open(O_RDONLY)               | allow_read                       |
+----------------------------------+----------------------------------+
| sys_open(O_WRONLY)               | allow_write                      |
+----------------------------------+----------------------------------+
| sys_open(O_RDWR)                 | allow_read/write                 |
+----------------------------------+----------------------------------+
| open_exec() from do_execve()     | allow_execute                    |
+----------------------------------+----------------------------------+
| open_exec() from !do_execve()    | allow_read                       |
+----------------------------------+----------------------------------+
| sys_read()                       | (none)                           |
+----------------------------------+----------------------------------+
| sys_write()                      | (none)                           |
+----------------------------------+----------------------------------+
| sys_mmap()                       | (none)                           |
+----------------------------------+----------------------------------+
| sys_uselib()                     | allow_read                       |
+----------------------------------+----------------------------------+
| sys_open(O_CREAT)                | allow_create                     |
+----------------------------------+----------------------------------+
| sys_open(O_TRUNC)                | allow_truncate                   |
+----------------------------------+----------------------------------+
| sys_truncate()                   | allow_truncate                   |
+----------------------------------+----------------------------------+
| sys_ftruncate()                  | allow_truncate                   |
+----------------------------------+----------------------------------+
| sys_open() without O_APPEND      | allow_rewrite                    |
+----------------------------------+----------------------------------+
| setfl() without O_APPEND         | allow_rewrite                    |
+----------------------------------+----------------------------------+
| sys_sysctl() for writing         | allow_write                      |
+----------------------------------+----------------------------------+
| sys_sysctl() for reading         | allow_read                       |
+----------------------------------+----------------------------------+
| sys_unlink()                     | allow_unlink                     |
+----------------------------------+----------------------------------+
| sys_mknod(S_IFREG)               | allow_create                     |
+----------------------------------+----------------------------------+
| sys_mknod(0)                     | allow_create                     |
+----------------------------------+----------------------------------+
| sys_mknod(S_IFIFO)               | allow_mkfifo                     |
+----------------------------------+----------------------------------+
| sys_mknod(S_IFSOCK)              | allow_mksock                     |
+----------------------------------+----------------------------------+
| sys_bind(AF_UNIX)                | allow_mksock                     |
+----------------------------------+----------------------------------+
| sys_mknod(S_IFBLK)               | allow_mkblock                    |
+----------------------------------+----------------------------------+
| sys_mknod(S_IFCHR)               | allow_mkchar                     |
+----------------------------------+----------------------------------+
| sys_symlink()                    | allow_symlink                    |
+----------------------------------+----------------------------------+
| sys_mkdir()                      | allow_mkdir                      |
+----------------------------------+----------------------------------+
| sys_rmdir()                      | allow_rmdir                      |
+----------------------------------+----------------------------------+
| sys_link()                       | allow_link                       |
+----------------------------------+----------------------------------+
| sys_rename()                     | allow_rename                     |
+----------------------------------+----------------------------------+

TOMOYO requires "allow_execute" permission of a pathname passed to do_execve()
but does not require "allow_read" permission of that pathname.
Let's consider 3 patterns (statically linked, dynamically linked,
shell script). This description is to some degree simplified.

  $ cat hello.c
  #include <stdio.h>
  int main() {
          printf("Hello\n");
          return 0;
  }
  $ cat hello.sh
  #! /bin/sh
  echo "Hello"
  $ gcc -static -o hello-static hello.c
  $ gcc -o hello-dynamic hello.c
  $ chmod 755 hello.sh

Case 1 -- Executing hello-static from bash.

  (1) The bash process calls fork() and the child process requests
      do_execve("hello-static").

  (2) The kernel checks "allow_execute hello-static" from "bash" domain.

  (3) The kernel calculates "bash hello-static" as the domain to transit to.

  (4) The kernel overwrites the child process by "hello-static".

  (5) The child process transits to "bash hello-static" domain.

  (6) The "hello-static" starts and finishes.

Case 2 -- Executing hello-dynamic from bash.

  (1) The bash process calls fork() and the child process requests
      do_execve("hello-dynamic").

  (2) The kernel checks "allow_execute hello-dynamic" from "bash" domain.

  (3) The kernel calculates "bash hello-dynamic" as the domain to transit to.

  (4) The kernel checks "allow_read ld-linux.so" from "bash hello-dynamic"
      domain. I think permission to access ld-linux.so should be charged
      hello-dynamic program, for "hello-dynamic needs ld-linux.so" is not
      a fault of bash program.

  (5) The kernel overwrites the child process by "hello-dynamic".

  (6) The child process transits to "bash hello-dynamic" domain.

  (7) The "hello-dynamic" starts and finishes.

Case 3 -- Executing hello.sh from bash.

  (1) The bash process calls fork() and the child process requests
      do_execve("hello.sh").

  (2) The kernel checks "allow_execute hello.sh" from "bash" domain.

  (3) The kernel calculates "bash hello.sh" as the domain to transit to.

  (4) The kernel checks "allow_read /bin/sh" from "bash hello.sh" domain.
      I think permission to access /bin/sh should be charged hello.sh program,
      for "hello.sh needs /bin/sh" is not a fault of bash program.

  (5) The kernel overwrites the child process by "/bin/sh".

  (6) The child process transits to "bash hello.sh" domain.

  (7) The "/bin/sh" requests open("hello.sh").

  (8) The kernel checks "allow_read hello.sh" from  "bash hello.sh" domain.

  (9) The "/bin/sh" starts and finishes.

Whether a file is interpreted as a program or not depends on an application.
The kernel cannot know whether the file is interpreted as a program or not.
Thus, TOMOYO treats "hello-static" "hello-dynamic" "ld-linux.so" "hello.sh"
"/bin/sh" equally as merely files; no distinction between executable and
non-executable. Therefore, TOMOYO doesn't check DAC's execute permission.
TOMOYO checks "allow_read" permission instead.

Calling do_execve() is a bold gesture that an old program's instance (i.e.
current process) is ready to be overwritten by a new program and is ready to
transfer control to the new program. To split purview of programs, TOMOYO
requires "allow_execute" permission of the new program against the old
program's instance and performs domain transition. If do_execve() succeeds,
the old program is no longer responsible against the consequence of the new
program's behavior. Only the new program is responsible for all consequences.

But TOMOYO doesn't require "allow_read" permission of the new program.
If TOMOYO requires "allow_read" permission of the new program, TOMOYO will
allow an attacker (who hijacked the old program's instance) to open the new
program and steal data from the new program. Requiring "allow_read" permission
will widen purview of the old program.

Not requiring "allow_read" permission of the new program against the old
program's instance is my design for reducing purview of the old program.
To be able to know whether the current process is in do_execve() or not,
I want to add in_execve flag to "task_struct".

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:15:05 +11:00
Kentaro Takeda
26a2a1c9eb Domain transition handler.
This file controls domain creation/deletion/transition.

Every process belongs to a domain in TOMOYO Linux.
Domain transition occurs when execve(2) is called
and the domain is expressed as 'process invocation history',
such as '<kernel> /sbin/init /etc/init.d/rc'.
Domain information is stored in current->cred->security field.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:15:05 +11:00
Kentaro Takeda
b69a54ee58 File operation restriction part.
This file controls file related operations of TOMOYO Linux.

tomoyo/tomoyo.c calls the following six functions in this file.
Each function handles the following access types.

 * tomoyo_check_file_perm
sysctl()'s "read" and "write".

 * tomoyo_check_exec_perm
"execute".

 * tomoyo_check_open_permission
open(2) for "read" and "write".

 * tomoyo_check_1path_perm
"create", "unlink", "mkdir", "rmdir", "mkfifo",
"mksock", "mkblock", "mkchar", "truncate" and "symlink".

 * tomoyo_check_2path_perm
"rename" and "unlink".

 * tomoyo_check_rewrite_permission
"rewrite".
("rewrite" are operations which may lose already recorded data of a file,
i.e. open(!O_APPEND) || open(O_TRUNC) || truncate() || ftruncate())

The functions which actually checks ACLs are the following three functions.
Each function handles the following access types.
ACL directive is expressed by "allow_<access type>".

 * tomoyo_check_file_acl
Open() operation and execve() operation.
("read", "write", "read/write" and "execute")

 * tomoyo_check_single_write_acl
Directory modification operations with 1 pathname.
("create", "unlink", "mkdir", "rmdir", "mkfifo", "mksock",
 "mkblock", "mkchar", "truncate", "symlink" and "rewrite")

 * tomoyo_check_double_write_acl
Directory modification operations with 2 pathname.
("link" and "rename")

Also, this file contains handlers of some utility directives
for file related operations.

 * "allow_read":   specifies globally (for all domains) readable files.
 * "path_group":   specifies pathname macro.
 * "deny_rewrite": restricts rewrite operation.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:15:05 +11:00
Kentaro Takeda
9590837b89 Common functions for TOMOYO Linux.
This file contains common functions (e.g. policy I/O, pattern matching).

-------------------- About pattern matching --------------------

Since TOMOYO Linux is a name based access control, TOMOYO Linux seriously
considers "safe" string representation.

TOMOYO Linux's string manipulation functions make reviewers feel crazy,
but there are reasons why TOMOYO Linux needs its own string manipulation
functions.

----- Part 1 : preconditions -----

People definitely want to use wild card.

  To support pattern matching, we have to support wild card characters.

  In a typical Linux system, filenames are likely consists of only alphabets,
  numbers, and some characters (e.g. + - ~ . / ).
  But theoretically, the Linux kernel accepts all characters but NUL character
  (which is used as a terminator of a string).

    Some Linux systems can have filenames which contain * ? ** etc.

Therefore, we have to somehow modify string so that we can distinguish
wild card characters and normal characters.

  It might be possible for some application's configuration files to restrict
  acceptable characters.
  It is impossible for kernel to restrict acceptable characters.

    We can't accept approaches which will cause troubles for applications.

----- Part 2 : commonly used approaches -----

Text formatted strings separated by space character (0x20) and new line
character (0x0A) is more preferable for users over array of NUL-terminated
string.

  Thus, people use text formatted configuration files separated by space
  character and new line.

We sometimes need to handle non-printable characters.

  Thus, people use \ character (0x5C) as escape character and represent
  non-printable characters using octal or hexadecimal format.

At this point, we remind (at least) 3 approaches.

  (1) Shell glob style expression
  (2) POSIX regular expression (UNIX style regular expression)
  (3) Maverick wild card expression

On the surface, (1) and (2) sound good choices. But they have a big pitfall.
All meta-characters in (1) and (2) are legal characters for representing
a pathname, and users easily write incorrect expression. What is worse, users
unlikely notice incorrect expressions because characters used for regular
pathnames unlikely contain meta-characters. This incorrect use of
meta-characters in pathname representation reveals vulnerability
(e.g. unexpected results) only when irregular pathname is specified.

The authors of TOMOYO Linux think that approaches which adds some character
for interpreting meta-characters as normal characters (i.e. (1) and (2)) are
not suitable for security use.

Therefore, the authors of TOMOYO Linux propose (3).

----- Part 3: consideration points -----

We need to solve encoding problem.

  A single character can be represented in several ways using encodings.

    For Japanese language, there are "ShiftJIS", "ISO-2022-JP", "EUC-JP",
    "UTF-8" and more.

  Some languages (e.g. Japanese language) supports multi-byte characters
  (where a single character is represented using several bytes).

    Some multi-byte characters may match the escape character.

    For Japanese language, some characters in "ShiftJIS" encoding match
    \ character, and bothering Web's CGI developers.

  It is important that the kernel string is not bothered by encoding problem.

    Linus said, "I really would expect that kernel strings don't have
    an encoding. They're just C strings: a NUL-terminated stream of bytes."
    http://lkml.org/lkml/2007/11/6/142

    Yes. The kernel strings are just C strings.
    We are talking about how to store and carry "kernel strings" safely.

  If we store "kernel string" into policy file as-is, the "kernel string" will
  be interpreted differently depending on application's encoding settings.
  One application may interpret "kernel string" as "UTF-8",
  another application may interpret "kernel string" as "ShiftJIS".

    Therefore, we propose to represent strings using ASCII encoding.
    In this way, we are no longer bothered by encoding problems.

We need to avoid information loss caused by display.

  It is difficult to input and display non-printable characters, but we have to
  be able to handle such characters because the kernel string is a C string.

  If we use only ASCII printable characters (from 0x21 to 0x7E) and space
  character (0x20) and new line character (0x0A), it is easy to input from
  keyboard and display on all terminals which is running Linux.

  Therefore, we propose to represent strings using only characters which value
  is one of "from 0x21 to 0x7E", "0x20", "0x0A".

We need to consider ease of splitting strings from a line.

  If we use an approach which uses "\ " for representing a space character
  within a string, we have to count the string from the beginning to check
  whether this space character is accompanied with \ character or not.
  As a result, we cannot monotonically split a line using space character.

  If we use an approach which uses "\040" for representing a space character
  within a string, we can monotonically split a line using space character.

  If we use an approach which uses NUL character as a delimiter, we cannot
  use string manipulation functions for splitting strings from a line.

  Therefore, we propose that we represent space character as "\040".

We need to avoid wrong designations (incorrect use of special characters).

  Not all users can understand and utilize POSIX's regular expressions
  correctly and perfectly.

  If a character acts as a wild card by default, the user will get unexpected
  result if that user didn't know the meaning of that character.

    Therefore, we propose that all characters but \ character act as
    a normal character and let the user add \ character to make a character
    act as a wild card.

    In this way, users needn't to know all wild card characters beforehand.
    They can learn when they encountered an unseen wild card character
    for their first time.

----- Part 4: supported wild card expressions -----

At this point, we have wild card expressions listed below.

  +-----------+--------------------------------------------------------------+
  | Wild card | Meaning and example                                          |
  +-----------+--------------------------------------------------------------+
  |   \*      | More than or equals to 0 character other than '/'.           |
  |           |           /var/log/samba/\*                                  |
  +-----------+--------------------------------------------------------------+
  |   \@      | More than or equals to 0 character other than '/' or '.'.    |
  |           |           /var/www/html/\@.html                              |
  +-----------+--------------------------------------------------------------+
  |   \?      | 1 byte character other than '/'.                             |
  |           |           /tmp/mail.\?\?\?\?\?\?                             |
  +-----------+--------------------------------------------------------------+
  |   \$      | More than or equals to 1 decimal digit.                      |
  |           |           /proc/\$/cmdline                                   |
  +-----------+--------------------------------------------------------------+
  |   \+      | 1 decimal digit.                                             |
  |           |           /var/tmp/my_work.\+                                |
  +-----------+--------------------------------------------------------------+
  |   \X      | More than or equals to 1 hexadecimal digit.                  |
  |           |           /var/tmp/my-work.\X                                |
  +-----------+--------------------------------------------------------------+
  |   \x      | 1 hexadecimal digit.                                         |
  |           |           /tmp/my-work.\x                                    |
  +-----------+--------------------------------------------------------------+
  |   \A      | More than or equals to 1 alphabet character.                 |
  |           |           /var/log/my-work/\$-\A-\$.log                      |
  +-----------+--------------------------------------------------------------+
  |   \a      | 1 alphabet character.                                        |
  |           |           /home/users/\a/\*/public_html/\*.html              |
  +-----------+--------------------------------------------------------------+
  |   \-      | Pathname subtraction operator.                               |
  |           | +---------------------+------------------------------------+ |
  |           | | Example             | Meaning                            | |
  |           | +---------------------+------------------------------------+ |
  |           | | /etc/\*             | All files in /etc/ directory.      | |
  |           | +---------------------+------------------------------------+ |
  |           | | /etc/\*\-\*shadow\* | /etc/\* other than /etc/\*shadow\* | |
  |           | +---------------------+------------------------------------+ |
  |           | | /\*\-proc\-sys/     | /\*/ other than /proc/ /sys/       | |
  |           | +---------------------+------------------------------------+ |
  +-----------+--------------------------------------------------------------+

  +----------------+---------------------------------------------------------+
  | Representation | Meaning and example                                     |
  +----------------+---------------------------------------------------------+
  |   \\           | backslash character itself.                             |
  +----------------+---------------------------------------------------------+
  |   \ooo         | 1 byte character.                                       |
  |                | ooo is 001 <= ooo <= 040 || 177 <= ooo <= 377.          |
  |                |                                                         |
  |                |           \040 for space character.                     |
  |                |           \177 for del character.                       |
  |                |                                                         |
  +----------------+---------------------------------------------------------+

----- Part 5: Advantages -----

We can obtain extensibility.

  Since our proposed approach adds \ to a character to interpret as a wild
  card, we can introduce new wild card in future while maintaining backward
  compatibility.

We can process monotonically.

  Since our proposed approach separates strings using a space character,
  we can split strings using existing string manipulation functions.

We can reliably analyze access logs.

  It is guaranteed that a string doesn't contain space character (0x20) and
  new line character (0x0A).

  It is guaranteed that a string won't be converted by FTP and won't be damaged
  by a terminal's settings.

  It is guaranteed that a string won't be affected by encoding converters
  (except encodings which insert NUL character (e.g. UTF-16)).

----- Part 6: conclusion -----

TOMOYO Linux is using its own encoding with reasons described above.
There is a disadvantage that we need to introduce a series of new string
manipulation functions. But TOMOYO Linux's encoding is useful for all users
(including audit and AppArmor) who want to perform pattern matching and
safely exchange string information between the kernel and the userspace.

-------------------- About policy interface --------------------

TOMOYO Linux creates the following files on securityfs (normally
mounted on /sys/kernel/security) as interfaces between kernel and
userspace. These files are for TOMOYO Linux management tools *only*,
not for general programs.

  * profile
  * exception_policy
  * domain_policy
  * manager
  * meminfo
  * self_domain
  * version
  * .domain_status
  * .process_status

** /sys/kernel/security/tomoyo/profile **

This file is used to read or write profiles.

"profile" means a running mode of process. A profile lists up
functions and their modes in "$number-$variable=$value" format. The
$number is profile number between 0 and 255. Each domain is assigned
one profile. To assign profile to domains, use "ccs-setprofile" or
"ccs-editpolicy" or "ccs-loadpolicy" commands.

(Example)
[root@tomoyo]# cat /sys/kernel/security/tomoyo/profile
0-COMMENT=-----Disabled Mode-----
0-MAC_FOR_FILE=disabled
0-MAX_ACCEPT_ENTRY=2048
0-TOMOYO_VERBOSE=disabled
1-COMMENT=-----Learning Mode-----
1-MAC_FOR_FILE=learning
1-MAX_ACCEPT_ENTRY=2048
1-TOMOYO_VERBOSE=disabled
2-COMMENT=-----Permissive Mode-----
2-MAC_FOR_FILE=permissive
2-MAX_ACCEPT_ENTRY=2048
2-TOMOYO_VERBOSE=enabled
3-COMMENT=-----Enforcing Mode-----
3-MAC_FOR_FILE=enforcing
3-MAX_ACCEPT_ENTRY=2048
3-TOMOYO_VERBOSE=enabled

- MAC_FOR_FILE:
Specifies access control level regarding file access requests.
- MAX_ACCEPT_ENTRY:
Limits the max number of ACL entries that are automatically appended
during learning mode. Default is 2048.
- TOMOYO_VERBOSE:
Specifies whether to print domain policy violation messages or not.

** /sys/kernel/security/tomoyo/manager **

This file is used to read or append the list of programs or domains
that can write to /sys/kernel/security/tomoyo interface. By default,
only processes with both UID = 0 and EUID = 0 can modify policy via
/sys/kernel/security/tomoyo interface. You can use keyword
"manage_by_non_root" to allow policy modification by non root user.

(Example)
[root@tomoyo]# cat /sys/kernel/security/tomoyo/manager
/usr/lib/ccs/loadpolicy
/usr/lib/ccs/editpolicy
/usr/lib/ccs/setlevel
/usr/lib/ccs/setprofile
/usr/lib/ccs/ld-watch
/usr/lib/ccs/ccs-queryd

** /sys/kernel/security/tomoyo/exception_policy **

This file is used to read and write system global settings. Each line
has a directive and operand pair. Directives are listed below.

- initialize_domain:
To initialize domain transition when specific program is executed,
use initialize_domain directive.
  * initialize_domain "program" from "domain"
  * initialize_domain "program" from "the last program part of domain"
  * initialize_domain "program"
If the part "from" and after is not given, the entry is applied to
all domain. If the "domain" doesn't start with "<kernel>", the entry
is applied to all domain whose domainname ends with "the last program
part of domain".
This directive is intended to aggregate domain transitions for daemon
program and program that are invoked by the kernel on demand, by
transiting to different domain.

- keep_domain
To prevent domain transition when program is executed from specific
domain, use keep_domain directive.
  * keep_domain "program" from "domain"
  * keep_domain "program" from "the last program part of domain"
  * keep_domain "domain"
  * keep_domain "the last program part of domain"
If the part "from" and before is not given, this entry is applied to
all program. If the "domain" doesn't start with "<kernel>", the entry
is applied to all domain whose domainname ends with "the last program
part of domain".
This directive is intended to reduce total number of domains and
memory usage by suppressing unneeded domain transitions.
To declare domain keepers, use keep_domain directive followed by
domain definition.
Any process that belongs to any domain declared with this directive,
the process stays at the same domain unless any program registered
with initialize_domain directive is executed.

In order to control domain transition in detail, you can use
no_keep_domain/no_initialize_domain keywrods.

- alias:
To allow executing programs using the name of symbolic links, use
alias keyword followed by dereferenced pathname and reference
pathname. For example, /sbin/pidof is a symbolic link to
/sbin/killall5 . In normal case, if /sbin/pidof is executed, the
domain is defined as if /sbin/killall5 is executed. By specifying
"alias /sbin/killall5 /sbin/pidof", you can run /sbin/pidof in the
domain for /sbin/pidof .
(Example)
alias /sbin/killall5 /sbin/pidof

- allow_read:
To grant unconditionally readable permissions, use allow_read keyword
followed by canonicalized file. This keyword is intended to reduce
size of domain policy by granting read access to library files such
as GLIBC and locale files. Exception is, if ignore_global_allow_read
keyword is given to a domain, entries specified by this keyword are
ignored.
(Example)
allow_read /lib/libc-2.5.so

- file_pattern:
To declare pathname pattern, use file_pattern keyword followed by
pathname pattern. The pathname pattern must be a canonicalized
Pathname. This keyword is not applicable to neither granting execute
permissions nor domain definitions.
For example, canonicalized pathname that contains a process ID
(i.e. /proc/PID/ files) needs to be grouped in order to make access
control work well.
(Example)
file_pattern /proc/\$/cmdline

- path_group
To declare pathname group, use path_group keyword followed by name of
the group and pathname pattern. For example, if you want to group all
files under home directory, you can define
   path_group HOME-DIR-FILE /home/\*/\*
   path_group HOME-DIR-FILE /home/\*/\*/\*
   path_group HOME-DIR-FILE /home/\*/\*/\*/\*
in the exception policy and use like
   allow_read @HOME-DIR-FILE
to grant file access permission.

- deny_rewrite:
To deny overwriting already written contents of file (such as log
files) by default, use deny_rewrite keyword followed by pathname
pattern. Files whose pathname match the patterns are not permitted to
open for writing without append mode or truncate unless the pathnames
are explicitly granted using allow_rewrite keyword in domain policy.
(Example)
deny_rewrite /var/log/\*

- aggregator
To deal multiple programs as a single program, use aggregator keyword
followed by name of original program and aggregated program. This
keyword is intended to aggregate similar programs.
For example, /usr/bin/tac and /bin/cat are similar. By specifying
"aggregator /usr/bin/tac /bin/cat", you can run /usr/bin/tac in the
domain for /bin/cat .
For example, /usr/sbin/logrotate for Fedora Core 3 generates programs
like /tmp/logrotate.\?\?\?\?\?\? and run them, but TOMOYO Linux
doesn't allow using patterns for granting execute permission and
defining domains. By specifying
"aggregator /tmp/logrotate.\?\?\?\?\?\? /tmp/logrotate.tmp", you can
run /tmp/logrotate.\?\?\?\?\?\? as if /tmp/logrotate.tmp is running.

** /sys/kernel/security/tomoyo/domain_policy **

This file contains definition of all domains and permissions that are
granted to each domain.

Lines from the next line to a domain definition ( any lines starting
with "<kernel>") to the previous line to the next domain definitions
are interpreted as access permissions for that domain.

** /sys/kernel/security/tomoyo/meminfo **

This file is to show the total RAM used to keep policy in the kernel
by TOMOYO Linux in bytes.
(Example)
[root@tomoyo]# cat /sys/kernel/security/tomoyo/meminfo
Shared:       61440
Private:      69632
Dynamic:        768
Total:       131840

You can set memory quota by writing to this file.
(Example)
[root@tomoyo]# echo Shared: 2097152 > /sys/kernel/security/tomoyo/meminfo
[root@tomoyo]# echo Private: 2097152 > /sys/kernel/security/tomoyo/meminfo

** /sys/kernel/security/tomoyo/self_domain **

This file is to show the name of domain the caller process belongs to.
(Example)
[root@etch]# cat /sys/kernel/security/tomoyo/self_domain
<kernel> /usr/sbin/sshd /bin/zsh /bin/cat

** /sys/kernel/security/tomoyo/version **

This file is used for getting TOMOYO Linux's version.
(Example)
[root@etch]# cat /sys/kernel/security/tomoyo/version
2.2.0-pre

** /sys/kernel/security/tomoyo/.domain_status **

This is a view (of a DBMS) that contains only profile number and
domainnames of domain so that "ccs-setprofile" command can do
line-oriented processing easily.

** /sys/kernel/security/tomoyo/.process_status **

This file is used by "ccs-ccstree" command to show "list of processes
currently running" and "domains which each process belongs to" and
"profile number which the domain is currently assigned" like "pstree"
command. This file is writable by programs that aren't registered as
policy manager.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:15:04 +11:00
Kentaro Takeda
c73bd6d473 Memory and pathname management functions.
TOMOYO Linux performs pathname based access control.
To remove factors that make pathname based access control difficult
(e.g. symbolic links, "..", "//" etc.), TOMOYO Linux derives realpath
of requested pathname from "struct dentry" and "struct vfsmount".

The maximum length of string data is limited to 4000 including trailing '\0'.
Since TOMOYO Linux uses '\ooo' style representation for non ASCII printable
characters, maybe TOMOYO Linux should be able to support 16336 (which means
(NAME_MAX * (PATH_MAX / (NAME_MAX + 1)) * 4 + (PATH_MAX / (NAME_MAX + 1)))
including trailing '\0'), but I think 4000 is enough for practical use.

TOMOYO uses only 0x21 - 0x7E (as printable characters) and 0x20 (as word
delimiter) and 0x0A (as line delimiter).
0x01 - 0x20 and 0x80 - 0xFF is handled in \ooo style representation.
The reason to use \ooo is to guarantee that "%s" won't damage logs.
Userland program can request

 open("/tmp/file granted.\nAccess /tmp/file ", O_WRONLY | O_CREAT, 0600)

and logging such crazy pathname using "Access %s denied.\n" format will cause
"fabrication of logs" like

 Access /tmp/file granted.
 Access /tmp/file denied.

TOMOYO converts such characters to \ooo so that the logs will become

 Access /tmp/file\040granted.\012Access\040/tmp/file denied.

and the administrator can read the logs safely using /bin/cat .
Likewise, a crazy request like

 open("/tmp/\x01\x02\x03\x04\x05\x06\x07\x08\x09", O_WRONLY | O_CREAT, 0600)

will be processed safely by converting to

 Access /tmp/\001\002\003\004\005\006\007\010\011 denied.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 15:15:04 +11:00
Mimi Zohar
523979adfa integrity: audit update
Based on discussions on linux-audit, as per Steve Grubb's request
http://lkml.org/lkml/2009/2/6/269, the following changes were made:
- forced audit result to be either 0 or 1.
- made template names const
- Added new stand-alone message type: AUDIT_INTEGRITY_RULE

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Steve Grubb <sgrubb@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-12 09:40:14 +11:00
James Morris
cb5629b10d Merge branch 'master' into next
Conflicts:
	fs/namei.c

Manually merged per:

diff --cc fs/namei.c
index 734f2b5,bbc15c2..0000000
--- a/fs/namei.c
+++ b/fs/namei.c
@@@ -860,9 -848,8 +849,10 @@@ static int __link_path_walk(const char
  		nd->flags |= LOOKUP_CONTINUE;
  		err = exec_permission_lite(inode);
  		if (err == -EAGAIN)
- 			err = vfs_permission(nd, MAY_EXEC);
+ 			err = inode_permission(nd->path.dentry->d_inode,
+ 					       MAY_EXEC);
 +		if (!err)
 +			err = ima_path_check(&nd->path, MAY_EXEC);
   		if (err)
  			break;

@@@ -1525,14 -1506,9 +1509,14 @@@ int may_open(struct path *path, int acc
  		flag &= ~O_TRUNC;
  	}

- 	error = vfs_permission(nd, acc_mode);
+ 	error = inode_permission(inode, acc_mode);
  	if (error)
  		return error;
 +
- 	error = ima_path_check(&nd->path,
++	error = ima_path_check(path,
 +			       acc_mode & (MAY_READ | MAY_WRITE | MAY_EXEC));
 +	if (error)
 +		return error;
  	/*
  	 * An append-only file must be opened in append mode for writing.
  	 */

Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 11:01:45 +11:00
James Morris
64c61d80a6 IMA: fix ima_delete_rules() definition
Fix ima_delete_rules() definition so sparse doesn't complain.

Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:34 +11:00
Mimi Zohar
1df9f0a731 Integrity: IMA file free imbalance
The number of calls to ima_path_check()/ima_file_free()
should be balanced.  An extra call to fput(), indicates
the file could have been accessed without first being
measured.

Although f_count is incremented/decremented in places other
than fget/fput, like fget_light/fput_light and get_file, the
current task must already hold a file refcnt.  The call to
__fput() is delayed until the refcnt becomes 0, resulting
in ima_file_free() flagging any changes.

- add hook to increment opencount for IPC shared memory(SYSV),
  shmat files, and /dev/zero
- moved NULL iint test in opencount_get()

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:33 +11:00
Mimi Zohar
f4bd857bc8 integrity: IMA policy open
Sequentialize access to the policy file
- permit multiple attempts to replace default policy with a valid policy

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:32 +11:00
Mimi Zohar
4af4662fa4 integrity: IMA policy
Support for a user loadable policy through securityfs
with support for LSM specific policy data.
- free invalid rule in ima_parse_add_rule()

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:31 +11:00
Mimi Zohar
bab7393787 integrity: IMA display
Make the measurement lists available through securityfs.
- removed test for NULL return code from securityfs_create_file/dir

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:31 +11:00
Mimi Zohar
3323eec921 integrity: IMA as an integrity service provider
IMA provides hardware (TPM) based measurement and attestation for
file measurements. As the Trusted Computing (TPM) model requires,
IMA measures all files before they are accessed in any way (on the
integrity_bprm_check, integrity_path_check and integrity_file_mmap
hooks), and commits the measurements to the TPM. Once added to the
TPM, measurements can not be removed.

In addition, IMA maintains a list of these file measurements, which
can be used to validate the aggregate value stored in the TPM.  The
TPM can sign these measurements, and thus the system can prove, to
itself and to a third party, the system's integrity in a way that
cannot be circumvented by malicious or compromised software.

- alloc ima_template_entry before calling ima_store_template()
- log ima_add_boot_aggregate() failure
- removed unused IMA_TEMPLATE_NAME_LEN
- replaced hard coded string length with #define name

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-06 09:05:30 +11:00
Serge E. Hallyn
faa3aad75a securityfs: fix long-broken securityfs_create_file comment
If there is an error creating a file through securityfs_create_file,
NULL is not returned, rather the error is propagated.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-02-03 11:02:51 +11:00
James Morris
5626d3e861 selinux: remove hooks which simply defer to capabilities
Remove SELinux hooks which do nothing except defer to the capabilites
hooks (or in one case, replicates the function).

Signed-off-by: James Morris <jmorris@namei.org>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
2009-02-02 09:20:34 +11:00
James Morris
95c14904b6 selinux: remove secondary ops call to shm_shmat
Remove secondary ops call to shm_shmat, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:16 +11:00
James Morris
5c4054ccfa selinux: remove secondary ops call to unix_stream_connect
Remove secondary ops call to unix_stream_connect, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:15 +11:00
James Morris
2cbbd19812 selinux: remove secondary ops call to task_kill
Remove secondary ops call to task_kill, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:14 +11:00
James Morris
ef76e748fa selinux: remove secondary ops call to task_setrlimit
Remove secondary ops call to task_setrlimit, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:13 +11:00
James Morris
ca5143d3ff selinux: remove unused cred_commit hook
Remove unused cred_commit hook from SELinux.   This
currently calls into the capabilities hook, which is a noop.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:12 +11:00
James Morris
af294e41d0 selinux: remove secondary ops call to task_create
Remove secondary ops call to task_create, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:11 +11:00
James Morris
d541bbee69 selinux: remove secondary ops call to file_mprotect
Remove secondary ops call to file_mprotect, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:11 +11:00
James Morris
438add6b32 selinux: remove secondary ops call to inode_setattr
Remove secondary ops call to inode_setattr, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:10 +11:00
James Morris
188fbcca9d selinux: remove secondary ops call to inode_permission
Remove secondary ops call to inode_permission, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:09 +11:00
James Morris
f51115b9ab selinux: remove secondary ops call to inode_follow_link
Remove secondary ops call to inode_follow_link, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:08 +11:00
James Morris
dd4907a6d4 selinux: remove secondary ops call to inode_mknod
Remove secondary ops call to inode_mknod, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:07 +11:00
James Morris
e4737250b7 selinux: remove secondary ops call to inode_unlink
Remove secondary ops call to inode_unlink, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:06 +11:00
James Morris
efdfac4376 selinux: remove secondary ops call to inode_link
Remove secondary ops call to inode_link, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:06 +11:00
James Morris
97422ab9ef selinux: remove secondary ops call to sb_umount
Remove secondary ops call to sb_umount, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:05 +11:00
James Morris
ef935b9136 selinux: remove secondary ops call to sb_mount
Remove secondary ops call to sb_mount, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:04 +11:00
James Morris
5565b0b865 selinux: remove secondary ops call to bprm_committed_creds
Remove secondary ops call to bprm_committed_creds, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:03 +11:00
James Morris
2ec5dbe23d selinux: remove secondary ops call to bprm_committing_creds
Remove secondary ops call to bprm_committing_creds, which is
a noop in capabilities.

Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:02 +11:00
James Morris
bc05595845 selinux: remove unused bprm_check_security hook
Remove unused bprm_check_security hook from SELinux.   This
currently calls into the capabilities hook, which is a noop.

Acked-by: Eric Paris <eparis@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-30 08:55:01 +11:00
Casey Schaufler
152a649b64 smackfs load append mode fix
Given just how hard it is to find the code that uses MAY_APPEND
it's probably not a big surprise that this went unnoticed for so
long. The Smack rules loading code is incorrectly setting the
MAY_READ bit when MAY_APPEND is requested.

Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-27 20:13:32 -08:00
David P. Quigley
cd89596f0c SELinux: Unify context mount and genfs behavior
Context mounts and genfs labeled file systems behave differently with respect to
setting file system labels. This patch brings genfs labeled file systems in line
with context mounts in that setxattr calls to them should return EOPNOTSUPP and
fscreate calls will be ignored.

Signed-off-by: David P. Quigley <dpquigl@tycho.nsa.gov>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@macbook.localdomain>
2009-01-19 09:47:14 +11:00
David P. Quigley
11689d47f0 SELinux: Add new security mount option to indicate security label support.
There is no easy way to tell if a file system supports SELinux security labeling.
Because of this a new flag is being added to the super block security structure
to indicate that the particular super block supports labeling. This flag is set
for file systems using the xattr, task, and transition labeling methods unless
that behavior is overridden by context mounts.

Signed-off-by: David P. Quigley <dpquigl@tycho.nsa.gov>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@macbook.localdomain>
2009-01-19 09:47:06 +11:00
David P. Quigley
0d90a7ec48 SELinux: Condense super block security structure flags and cleanup necessary code.
The super block security structure currently has three fields for what are
essentially flags.  The flags field is used for mount options while two other
char fields are used for initialization and proc flags. These latter two fields are
essentially bit fields since the only used values are 0 and 1.  These fields
have been collapsed into the flags field and new bit masks have been added for
them. The code is also fixed to work with these new flags.

Signed-off-by: David P. Quigley <dpquigl@tycho.nsa.gov>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@macbook.localdomain>
2009-01-19 09:46:40 +11:00
Vegard Nossum
0d54ee1c78 security: introduce missing kfree
Plug this leak.

Acked-by: David Howells <dhowells@redhat.com>
Cc: James Morris <jmorris@namei.org>
Cc: <stable@kernel.org>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-17 14:24:46 -08:00
Heiko Carstens
938bb9f5e8 [CVE-2009-0029] System call wrappers part 28
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2009-01-14 14:15:30 +01:00
Heiko Carstens
1e7bfb2134 [CVE-2009-0029] System call wrappers part 27
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2009-01-14 14:15:29 +01:00
Fernando Carrijo
c19a28e119 remove lots of double-semicolons
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Theodore Ts'o <tytso@mit.edu>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Acked-by: David S. Miller <davem@davemloft.net>
Cc: James Morris <jmorris@namei.org>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Acked-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 08:31:14 -08:00
Serge E. Hallyn
0b82ac37b8 devices cgroup: allow mkfifo
The devcgroup_inode_permission() hook in the devices whitelist cgroup has
always bypassed access checks on fifos.  But the mknod hook did not.  The
devices whitelist is only about block and char devices, and fifos can't
even be added to the whitelist, so fifos can't be created at all except by
tasks which have 'a' in their whitelist (meaning they have access to all
devices).

Fix the behavior by bypassing access checks to mkfifo.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Paul Menage <menage@google.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: James Morris <jmorris@namei.org>
Reported-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Cc: <stable@kernel.org>		[2.6.27.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 08:31:03 -08:00
Lai Jiangshan
116e057512 devcgroup: use list_for_each_entry_rcu()
We should use list_for_each_entry_rcu in RCU read site.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Balbir Singh <balbir@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 08:31:03 -08:00
James Morris
ac8cc0fa53 Merge branch 'next' into for-linus 2009-01-07 09:58:22 +11:00
David Howells
3699c53c48 CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #3]
Fix a regression in cap_capable() due to:

	commit 3b11a1dece
	Author: David Howells <dhowells@redhat.com>
	Date:   Fri Nov 14 10:39:26 2008 +1100

	    CRED: Differentiate objective and effective subjective credentials on a task

The problem is that the above patch allows a process to have two sets of
credentials, and for the most part uses the subjective credentials when
accessing current's creds.

There is, however, one exception: cap_capable(), and thus capable(), uses the
real/objective credentials of the target task, whether or not it is the current
task.

Ordinarily this doesn't matter, since usually the two cred pointers in current
point to the same set of creds.  However, sys_faccessat() makes use of this
facility to override the credentials of the calling process to make its test,
without affecting the creds as seen from other processes.

One of the things sys_faccessat() does is to make an adjustment to the
effective capabilities mask, which cap_capable(), as it stands, then ignores.

The affected capability check is in generic_permission():

	if (!(mask & MAY_EXEC) || execute_ok(inode))
		if (capable(CAP_DAC_OVERRIDE))
			return 0;

This change passes the set of credentials to be tested down into the commoncap
and SELinux code.  The security functions called by capable() and
has_capability() select the appropriate set of credentials from the process
being checked.

This can be tested by compiling the following program from the XFS testsuite:

/*
 *  t_access_root.c - trivial test program to show permission bug.
 *
 *  Written by Michael Kerrisk - copyright ownership not pursued.
 *  Sourced from: http://linux.derkeiler.com/Mailing-Lists/Kernel/2003-10/6030.html
 */
#include <limits.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <fcntl.h>
#include <sys/stat.h>

#define UID 500
#define GID 100
#define PERM 0
#define TESTPATH "/tmp/t_access"

static void
errExit(char *msg)
{
    perror(msg);
    exit(EXIT_FAILURE);
} /* errExit */

static void
accessTest(char *file, int mask, char *mstr)
{
    printf("access(%s, %s) returns %d\n", file, mstr, access(file, mask));
} /* accessTest */

int
main(int argc, char *argv[])
{
    int fd, perm, uid, gid;
    char *testpath;
    char cmd[PATH_MAX + 20];

    testpath = (argc > 1) ? argv[1] : TESTPATH;
    perm = (argc > 2) ? strtoul(argv[2], NULL, 8) : PERM;
    uid = (argc > 3) ? atoi(argv[3]) : UID;
    gid = (argc > 4) ? atoi(argv[4]) : GID;

    unlink(testpath);

    fd = open(testpath, O_RDWR | O_CREAT, 0);
    if (fd == -1) errExit("open");

    if (fchown(fd, uid, gid) == -1) errExit("fchown");
    if (fchmod(fd, perm) == -1) errExit("fchmod");
    close(fd);

    snprintf(cmd, sizeof(cmd), "ls -l %s", testpath);
    system(cmd);

    if (seteuid(uid) == -1) errExit("seteuid");

    accessTest(testpath, 0, "0");
    accessTest(testpath, R_OK, "R_OK");
    accessTest(testpath, W_OK, "W_OK");
    accessTest(testpath, X_OK, "X_OK");
    accessTest(testpath, R_OK | W_OK, "R_OK | W_OK");
    accessTest(testpath, R_OK | X_OK, "R_OK | X_OK");
    accessTest(testpath, W_OK | X_OK, "W_OK | X_OK");
    accessTest(testpath, R_OK | W_OK | X_OK, "R_OK | W_OK | X_OK");

    exit(EXIT_SUCCESS);
} /* main */

This can be run against an Ext3 filesystem as well as against an XFS
filesystem.  If successful, it will show:

	[root@andromeda src]# ./t_access_root /tmp/xxx 0 4043 4043
	---------- 1 dhowells dhowells 0 2008-12-31 03:00 /tmp/xxx
	access(/tmp/xxx, 0) returns 0
	access(/tmp/xxx, R_OK) returns 0
	access(/tmp/xxx, W_OK) returns 0
	access(/tmp/xxx, X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK) returns 0
	access(/tmp/xxx, R_OK | X_OK) returns -1
	access(/tmp/xxx, W_OK | X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK | X_OK) returns -1

If unsuccessful, it will show:

	[root@andromeda src]# ./t_access_root /tmp/xxx 0 4043 4043
	---------- 1 dhowells dhowells 0 2008-12-31 02:56 /tmp/xxx
	access(/tmp/xxx, 0) returns 0
	access(/tmp/xxx, R_OK) returns -1
	access(/tmp/xxx, W_OK) returns -1
	access(/tmp/xxx, X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK) returns -1
	access(/tmp/xxx, R_OK | X_OK) returns -1
	access(/tmp/xxx, W_OK | X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK | X_OK) returns -1

I've also tested the fix with the SELinux and syscalls LTP testsuites.

Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: J. Bruce Fields <bfields@citi.umich.edu>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-07 09:38:48 +11:00
James Morris
29881c4502 Revert "CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #2]"
This reverts commit 14eaddc967.

David has a better version to come.
2009-01-07 09:21:54 +11:00
Linus Torvalds
520c853466 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  inotify: fix type errors in interfaces
  fix breakage in reiserfs_new_inode()
  fix the treatment of jfs special inodes
  vfs: remove duplicate code in get_fs_type()
  add a vfs_fsync helper
  sys_execve and sys_uselib do not call into fsnotify
  zero i_uid/i_gid on inode allocation
  inode->i_op is never NULL
  ntfs: don't NULL i_op
  isofs check for NULL ->i_op in root directory is dead code
  affs: do not zero ->i_op
  kill suid bit only for regular files
  vfs: lseek(fd, 0, SEEK_CUR) race condition
2009-01-05 18:32:06 -08:00
Al Viro
56ff5efad9 zero i_uid/i_gid on inode allocation
... 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>
2009-01-05 11:54:28 -05:00
Al Viro
acfa4380ef inode->i_op is never NULL
We used to have rather schizophrenic set of checks for NULL ->i_op even
though it had been eliminated years ago.  You'd need to go out of your
way to set it to NULL explicitly _and_ a bunch of code would die on
such inodes anyway.  After killing two remaining places that still
did that bogosity, all that crap can go away.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2009-01-05 11:54:28 -05:00
Eric Paris
76f7ba35d4 SELinux: shrink sizeof av_inhert selinux_class_perm and context
I started playing with pahole today and decided to put it against the
selinux structures.  Found we could save a little bit of space on x86_64
(and no harm on i686) just reorganizing some structs.

Object size changes:
av_inherit: 24 -> 16
selinux_class_perm: 48 -> 40
context: 80 -> 72

Admittedly there aren't many of av_inherit or selinux_class_perm's in
the kernel (33 and 1 respectively) But the change to the size of struct
context reverberate out a bit.  I can get some hard number if they are
needed, but I don't see why they would be.  We do change which cacheline
context->len and context->str would be on, but I don't see that as a
problem since we are clearly going to have to load both if the context
is to be of any value.  I've run with the patch and don't seem to be
having any problems.

An example of what's going on using struct av_inherit would be:

form: to:
struct av_inherit {			struct av_inherit {
	u16 tclass;				const char **common_pts;
	const char **common_pts;		u32 common_base;
	u32 common_base;			u16 tclass;
};

(notice all I did was move u16 tclass to the end of the struct instead
of the beginning)

Memory layout before the change:
struct av_inherit {
	u16 tclass; /* 2 */
	/* 6 bytes hole */
	const char** common_pts; /* 8 */
	u32 common_base; /* 4 */
	/* 4 byes padding */

	/* size: 24, cachelines: 1 */
	/* sum members: 14, holes: 1, sum holes: 6 */
	/* padding: 4 */
};

Memory layout after the change:
struct av_inherit {
	const char ** common_pts; /* 8 */
	u32 common_base; /* 4 */
	u16 tclass; /* 2 */
	/* 2 bytes padding */

	/* size: 16, cachelines: 1 */
	/* sum members: 14, holes: 0, sum holes: 0 */
	/* padding: 2 */
};

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-05 19:19:55 +11:00
David Howells
14eaddc967 CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #2]
Fix a regression in cap_capable() due to:

	commit 5ff7711e635b32f0a1e558227d030c7e45b4a465
	Author: David Howells <dhowells@redhat.com>
	Date:   Wed Dec 31 02:52:28 2008 +0000

	    CRED: Differentiate objective and effective subjective credentials on a task

The problem is that the above patch allows a process to have two sets of
credentials, and for the most part uses the subjective credentials when
accessing current's creds.

There is, however, one exception: cap_capable(), and thus capable(), uses the
real/objective credentials of the target task, whether or not it is the current
task.

Ordinarily this doesn't matter, since usually the two cred pointers in current
point to the same set of creds.  However, sys_faccessat() makes use of this
facility to override the credentials of the calling process to make its test,
without affecting the creds as seen from other processes.

One of the things sys_faccessat() does is to make an adjustment to the
effective capabilities mask, which cap_capable(), as it stands, then ignores.

The affected capability check is in generic_permission():

	if (!(mask & MAY_EXEC) || execute_ok(inode))
		if (capable(CAP_DAC_OVERRIDE))
			return 0;

This change splits capable() from has_capability() down into the commoncap and
SELinux code.  The capable() security op now only deals with the current
process, and uses the current process's subjective creds.  A new security op -
task_capable() - is introduced that can check any task's objective creds.

strictly the capable() security op is superfluous with the presence of the
task_capable() op, however it should be faster to call the capable() op since
two fewer arguments need be passed down through the various layers.

This can be tested by compiling the following program from the XFS testsuite:

/*
 *  t_access_root.c - trivial test program to show permission bug.
 *
 *  Written by Michael Kerrisk - copyright ownership not pursued.
 *  Sourced from: http://linux.derkeiler.com/Mailing-Lists/Kernel/2003-10/6030.html
 */
#include <limits.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <fcntl.h>
#include <sys/stat.h>

#define UID 500
#define GID 100
#define PERM 0
#define TESTPATH "/tmp/t_access"

static void
errExit(char *msg)
{
    perror(msg);
    exit(EXIT_FAILURE);
} /* errExit */

static void
accessTest(char *file, int mask, char *mstr)
{
    printf("access(%s, %s) returns %d\n", file, mstr, access(file, mask));
} /* accessTest */

int
main(int argc, char *argv[])
{
    int fd, perm, uid, gid;
    char *testpath;
    char cmd[PATH_MAX + 20];

    testpath = (argc > 1) ? argv[1] : TESTPATH;
    perm = (argc > 2) ? strtoul(argv[2], NULL, 8) : PERM;
    uid = (argc > 3) ? atoi(argv[3]) : UID;
    gid = (argc > 4) ? atoi(argv[4]) : GID;

    unlink(testpath);

    fd = open(testpath, O_RDWR | O_CREAT, 0);
    if (fd == -1) errExit("open");

    if (fchown(fd, uid, gid) == -1) errExit("fchown");
    if (fchmod(fd, perm) == -1) errExit("fchmod");
    close(fd);

    snprintf(cmd, sizeof(cmd), "ls -l %s", testpath);
    system(cmd);

    if (seteuid(uid) == -1) errExit("seteuid");

    accessTest(testpath, 0, "0");
    accessTest(testpath, R_OK, "R_OK");
    accessTest(testpath, W_OK, "W_OK");
    accessTest(testpath, X_OK, "X_OK");
    accessTest(testpath, R_OK | W_OK, "R_OK | W_OK");
    accessTest(testpath, R_OK | X_OK, "R_OK | X_OK");
    accessTest(testpath, W_OK | X_OK, "W_OK | X_OK");
    accessTest(testpath, R_OK | W_OK | X_OK, "R_OK | W_OK | X_OK");

    exit(EXIT_SUCCESS);
} /* main */

This can be run against an Ext3 filesystem as well as against an XFS
filesystem.  If successful, it will show:

	[root@andromeda src]# ./t_access_root /tmp/xxx 0 4043 4043
	---------- 1 dhowells dhowells 0 2008-12-31 03:00 /tmp/xxx
	access(/tmp/xxx, 0) returns 0
	access(/tmp/xxx, R_OK) returns 0
	access(/tmp/xxx, W_OK) returns 0
	access(/tmp/xxx, X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK) returns 0
	access(/tmp/xxx, R_OK | X_OK) returns -1
	access(/tmp/xxx, W_OK | X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK | X_OK) returns -1

If unsuccessful, it will show:

	[root@andromeda src]# ./t_access_root /tmp/xxx 0 4043 4043
	---------- 1 dhowells dhowells 0 2008-12-31 02:56 /tmp/xxx
	access(/tmp/xxx, 0) returns 0
	access(/tmp/xxx, R_OK) returns -1
	access(/tmp/xxx, W_OK) returns -1
	access(/tmp/xxx, X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK) returns -1
	access(/tmp/xxx, R_OK | X_OK) returns -1
	access(/tmp/xxx, W_OK | X_OK) returns -1
	access(/tmp/xxx, R_OK | W_OK | X_OK) returns -1

I've also tested the fix with the SELinux and syscalls LTP testsuites.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2009-01-05 11:17:04 +11:00
James Morris
5c8c40be4b Merge branch 'master' of git://git.infradead.org/users/pcmoore/lblnet-2.6_next into next 2009-01-05 08:56:01 +11:00
Al Viro
5af75d8d58 audit: validate comparison operations, store them in sane form
Don't store the field->op in the messy (and very inconvenient for e.g.
audit_comparator()) form; translate to dense set of values and do full
validation of userland-submitted value while we are at it.

->audit_init_rule() and ->audit_match_rule() get new values now; in-tree
instances updated.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2009-01-04 15:14:42 -05:00
Linus Torvalds
7d3b56ba37 Merge branch 'cpus4096-for-linus-3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'cpus4096-for-linus-3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (77 commits)
  x86: setup_per_cpu_areas() cleanup
  cpumask: fix compile error when CONFIG_NR_CPUS is not defined
  cpumask: use alloc_cpumask_var_node where appropriate
  cpumask: convert shared_cpu_map in acpi_processor* structs to cpumask_var_t
  x86: use cpumask_var_t in acpi/boot.c
  x86: cleanup some remaining usages of NR_CPUS where s/b nr_cpu_ids
  sched: put back some stack hog changes that were undone in kernel/sched.c
  x86: enable cpus display of kernel_max and offlined cpus
  ia64: cpumask fix for is_affinity_mask_valid()
  cpumask: convert RCU implementations, fix
  xtensa: define __fls
  mn10300: define __fls
  m32r: define __fls
  h8300: define __fls
  frv: define __fls
  cris: define __fls
  cpumask: CONFIG_DISABLE_OBSOLETE_CPUMASK_FUNCTIONS
  cpumask: zero extra bits in alloc_cpumask_var_node
  cpumask: replace for_each_cpu_mask_nr with for_each_cpu in kernel/time/
  cpumask: convert mm/
  ...
2009-01-03 12:04:39 -08:00
Rusty Russell
4f4b6c1a94 cpumask: prepare for iterators to only go to nr_cpu_ids/nr_cpumask_bits.: core
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>
2009-01-01 10:12:15 +10:30
James Morris
90bd49ab66 keys: fix sparse warning by adding __user annotation to cast
Fix the following sparse warning:

      CC      security/keys/key.o
    security/keys/keyctl.c:1297:10: warning: incorrect type in argument 2 (different address spaces)
    security/keys/keyctl.c:1297:10:    expected char [noderef] <asn:1>*buffer
    security/keys/keyctl.c:1297:10:    got char *<noident>

which appears to be caused by lack of __user annotation to the cast of
a syscall argument.

Signed-off-by: James Morris <jmorris@namei.org>
Acked-by: David Howells <dhowells@redhat.com>
2009-01-01 10:32:44 +11:00
Kentaro Takeda
be6d3e56a6 introduce new LSM hooks where vfsmount is available.
Add new LSM hooks for path-based checks.  Call them on directory-modifying
operations at the points where we still know the vfsmount involved.

Signed-off-by: Kentaro Takeda <takedakn@nttdata.co.jp>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Toshiharu Harada <haradats@nttdata.co.jp>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2008-12-31 18:07:37 -05:00
Casey Schaufler
6d3dc07cbb smack: Add support for unlabeled network hosts and networks
Add support for unlabeled network hosts and networks.
Relies heavily on Paul Moore's netlabel support.

Creates a new entry in /smack called netlabel. Writes to /smack/netlabel
take the form:

    A.B.C.D LABEL
or
    A.B.C.D/N LABEL

where A.B.C.D is a network address, N is an integer between 0-32,
and LABEL is the Smack label to be used. If /N is omitted /32 is
assumed. N designates the netmask for the address. Entries are
matched by the most specific address/mask pair. 0.0.0.0/0 will
match everything, while 192.168.1.117/32 will match exactly one
host.

A new system label "@", pronounced "web", is defined. Processes
can not be assigned the web label. An address assigned the web
label can be written to by any process, and packets coming from
a web address can be written to any socket. Use of the web label
is a violation of any strict MAC policy, but the web label has
been requested many times.

The nltype entry has been removed from /smack. It did not work right
and the netlabel interface can be used to specify that all hosts
be treated as unlabeled.

CIPSO labels on incoming packets will be honored, even from designated
single label hosts. Single label hosts can only be written to by
processes with labels that can write to the label of the host.
Packets sent to single label hosts will always be unlabeled.

Once added a single label designation cannot be removed, however
the label may be changed.

The behavior of the ambient label remains unchanged.


Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <paul.moore@hp.com>
2008-12-31 12:54:12 -05:00
Paul Moore
277d342fc4 selinux: Deprecate and schedule the removal of the the compat_net functionality
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>
2008-12-31 12:54:11 -05:00
Paul Moore
6c2e8ac095 netlabel: Update kernel configuration API
Update the NetLabel kernel API to expose the new features added in kernel
releases 2.6.25 and 2.6.28: the static/fallback label functionality and network
address based selectors.

Signed-off-by: Paul Moore <paul.moore@hp.com>
2008-12-31 12:54:11 -05:00
David Howells
eca1bf5b4f KEYS: Fix variable uninitialisation warnings
Fix variable uninitialisation warnings introduced in:

	commit 8bbf4976b5
	Author: David Howells <dhowells@redhat.com>
	Date:   Fri Nov 14 10:39:14 2008 +1100

	KEYS: Alter use of key instantiation link-to-keyring argument

As:

  security/keys/keyctl.c: In function 'keyctl_negate_key':
  security/keys/keyctl.c:976: warning: 'dest_keyring' may be used uninitialized in this function
  security/keys/keyctl.c: In function 'keyctl_instantiate_key':
  security/keys/keyctl.c:898: warning: 'dest_keyring' may be used uninitialized in this function

Some versions of gcc notice that get_instantiation_key() doesn't always set
*_dest_keyring, but fail to observe that if this happens then *_dest_keyring
will not be read by the caller.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-12-29 14:24:43 +11:00
James Morris
54d2f649a6 Merge branch 'next' into for-linus 2008-12-29 09:57:38 +11:00
Linus Torvalds
0191b625ca Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next-2.6: (1429 commits)
  net: Allow dependancies of FDDI & Tokenring to be modular.
  igb: Fix build warning when DCA is disabled.
  net: Fix warning fallout from recent NAPI interface changes.
  gro: Fix potential use after free
  sfc: If AN is enabled, always read speed/duplex from the AN advertising bits
  sfc: When disabling the NIC, close the device rather than unregistering it
  sfc: SFT9001: Add cable diagnostics
  sfc: Add support for multiple PHY self-tests
  sfc: Merge top-level functions for self-tests
  sfc: Clean up PHY mode management in loopback self-test
  sfc: Fix unreliable link detection in some loopback modes
  sfc: Generate unique names for per-NIC workqueues
  802.3ad: use standard ethhdr instead of ad_header
  802.3ad: generalize out mac address initializer
  802.3ad: initialize ports LACPDU from const initializer
  802.3ad: remove typedef around ad_system
  802.3ad: turn ports is_individual into a bool
  802.3ad: turn ports is_enabled into a bool
  802.3ad: make ntt bool
  ixgbe: Fix set_ringparam in ixgbe to use the same memory pools.
  ...

Fixed trivial IPv4/6 address printing conflicts in fs/cifs/connect.c due
to the conversion to %pI (in this networking merge) and the addition of
doing IPv6 addresses (from the earlier merge of CIFS).
2008-12-28 12:49:40 -08:00
Sergio Luis
81ea714bf1 smackfs: check for allocation failures in smk_set_access()
smackfs: check for allocation failures in smk_set_access()

 While adding a new subject/object pair to smack_list, smk_set_access()
 didn't check the return of kzalloc().

 This patch changes smk_set_access() to return 0 or -ENOMEM, based on
 kzalloc()'s return. It also updates its caller, smk_write_load(), to
 check for smk_set_access()'s return, given it is no longer a void
 return function.

 Signed-off-by: Sergio Luis <sergio@larces.uece.br>
 To: Casey Schaufler <casey@schaufler-ca.com>
 Cc: Ahmed S. Darwish <darwish.07@gmail.com>
 Cc: LSM <linux-security-module@vger.kernel.org>
 Cc: LKLM <linux-kernel@vger.kernel.org>

Acked-by: Casey Schaufler <casey@schaufler-ca.com>
2008-12-25 12:14:55 +11:00
James Morris
7419224691 SELinux: don't check permissions for kernel mounts
Don't bother checking permissions when the kernel performs an
internal mount, as this should always be allowed.

Signed-off-by: James Morris <jmorris@namei.org>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
2008-12-20 09:03:39 +11:00
James Morris
12204e24b1 security: pass mount flags to security_sb_kern_mount()
Pass mount flags to security_sb_kern_mount(), so security modules
can determine if a mount operation is being performed by the kernel.

Signed-off-by: James Morris <jmorris@namei.org>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
2008-12-20 09:02:39 +11:00
Stephen Smalley
459c19f524 SELinux: correctly detect proc filesystems of the form "proc/foo"
Map all of these proc/ filesystem types to "proc" for the policy lookup at
filesystem mount time.

Signed-off-by: James Morris <jmorris@namei.org>
2008-12-20 09:01:03 +11:00
Hannes Eder
200036ca9b CRED: fix sparse warnings
Impact: fix sparse warnings

Fix the following sparse warnings:

  security/security.c:228:2: warning: returning void-valued expression
  security/security.c:233:2: warning: returning void-valued expression
  security/security.c:616:2: warning: returning void-valued expression

Signed-off-by: Hannes Eder <hannes@hanneseder.net>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-25 06:33:17 +05:30
Eric Paris
e50a906e02 capabilities: define get_vfs_caps_from_disk when file caps are not enabled
When CONFIG_SECURITY_FILE_CAPABILITIES is not set the audit system may
try to call into the capabilities function vfs_cap_from_file.  This
patch defines that function so kernels can build and work.

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-15 08:50:52 +11:00
David Howells
3a3b7ce933 CRED: Allow kernel services to override LSM settings for task actions
Allow kernel services to override LSM settings appropriate to the actions
performed by a task by duplicating a set of credentials, modifying it and then
using task_struct::cred to point to it when performing operations on behalf of
a task.

This is used, for example, by CacheFiles which has to transparently access the
cache on behalf of a process that thinks it is doing, say, NFS accesses with a
potentially inappropriate (with respect to accessing the cache) set of
credentials.

This patch provides two LSM hooks for modifying a task security record:

 (*) security_kernel_act_as() which allows modification of the security datum
     with which a task acts on other objects (most notably files).

 (*) security_kernel_create_files_as() which allows modification of the
     security datum that is used to initialise the security data on a file that
     a task creates.

The patch also provides four new credentials handling functions, which wrap the
LSM functions:

 (1) prepare_kernel_cred()

     Prepare a set of credentials for a kernel service to use, based either on
     a daemon's credentials or on init_cred.  All the keyrings are cleared.

 (2) set_security_override()

     Set the LSM security ID in a set of credentials to a specific security
     context, assuming permission from the LSM policy.

 (3) set_security_override_from_ctx()

     As (2), but takes the security context as a string.

 (4) set_create_files_as()

     Set the file creation LSM security ID in a set of credentials to be the
     same as that on a particular inode.

Signed-off-by: Casey Schaufler <casey@schaufler-ca.com> [Smack changes]
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:28 +11:00
David Howells
1bfdc75ae0 CRED: Add a kernel_service object class to SELinux
Add a 'kernel_service' object class to SELinux and give this object class two
access vectors: 'use_as_override' and 'create_files_as'.

The first vector is used to grant a process the right to nominate an alternate
process security ID for the kernel to use as an override for the SELinux
subjective security when accessing stuff on behalf of another process.

For example, CacheFiles when accessing the cache on behalf on a process
accessing an NFS file needs to use a subjective security ID appropriate to the
cache rather then the one the calling process is using.  The cachefilesd
daemon will nominate the security ID to be used.

The second vector is used to grant a process the right to nominate a file
creation label for a kernel service to use.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:27 +11:00
David Howells
3b11a1dece CRED: Differentiate objective and effective subjective credentials on a task
Differentiate the objective and real subjective credentials from the effective
subjective credentials on a task by introducing a second credentials pointer
into the task_struct.

task_struct::real_cred then refers to the objective and apparent real
subjective credentials of a task, as perceived by the other tasks in the
system.

task_struct::cred then refers to the effective subjective credentials of a
task, as used by that task when it's actually running.  These are not visible
to the other tasks in the system.

__task_cred(task) then refers to the objective/real credentials of the task in
question.

current_cred() refers to the effective subjective credentials of the current
task.

prepare_creds() uses the objective creds as a base and commit_creds() changes
both pointers in the task_struct (indeed commit_creds() requires them to be the
same).

override_creds() and revert_creds() change the subjective creds pointer only,
and the former returns the old subjective creds.  These are used by NFSD,
faccessat() and do_coredump(), and will by used by CacheFiles.

In SELinux, current_has_perm() is provided as an alternative to
task_has_perm().  This uses the effective subjective context of current,
whereas task_has_perm() uses the objective/real context of the subject.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:26 +11:00
David Howells
1d045980e1 CRED: Prettify commoncap.c
Prettify commoncap.c.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:24 +11:00
David Howells
a6f76f23d2 CRED: Make execve() take advantage of copy-on-write credentials
Make execve() take advantage of copy-on-write credentials, allowing it to set
up the credentials in advance, and then commit the whole lot after the point
of no return.

This patch and the preceding patches have been tested with the LTP SELinux
testsuite.

This patch makes several logical sets of alteration:

 (1) execve().

     The credential bits from struct linux_binprm are, for the most part,
     replaced with a single credentials pointer (bprm->cred).  This means that
     all the creds can be calculated in advance and then applied at the point
     of no return with no possibility of failure.

     I would like to replace bprm->cap_effective with:

	cap_isclear(bprm->cap_effective)

     but this seems impossible due to special behaviour for processes of pid 1
     (they always retain their parent's capability masks where normally they'd
     be changed - see cap_bprm_set_creds()).

     The following sequence of events now happens:

     (a) At the start of do_execve, the current task's cred_exec_mutex is
     	 locked to prevent PTRACE_ATTACH from obsoleting the calculation of
     	 creds that we make.

     (a) prepare_exec_creds() is then called to make a copy of the current
     	 task's credentials and prepare it.  This copy is then assigned to
     	 bprm->cred.

  	 This renders security_bprm_alloc() and security_bprm_free()
     	 unnecessary, and so they've been removed.

     (b) The determination of unsafe execution is now performed immediately
     	 after (a) rather than later on in the code.  The result is stored in
     	 bprm->unsafe for future reference.

     (c) prepare_binprm() is called, possibly multiple times.

     	 (i) This applies the result of set[ug]id binaries to the new creds
     	     attached to bprm->cred.  Personality bit clearance is recorded,
     	     but now deferred on the basis that the exec procedure may yet
     	     fail.

         (ii) This then calls the new security_bprm_set_creds().  This should
	     calculate the new LSM and capability credentials into *bprm->cred.

	     This folds together security_bprm_set() and parts of
	     security_bprm_apply_creds() (these two have been removed).
	     Anything that might fail must be done at this point.

         (iii) bprm->cred_prepared is set to 1.

	     bprm->cred_prepared is 0 on the first pass of the security
	     calculations, and 1 on all subsequent passes.  This allows SELinux
	     in (ii) to base its calculations only on the initial script and
	     not on the interpreter.

     (d) flush_old_exec() is called to commit the task to execution.  This
     	 performs the following steps with regard to credentials:

	 (i) Clear pdeath_signal and set dumpable on certain circumstances that
	     may not be covered by commit_creds().

         (ii) Clear any bits in current->personality that were deferred from
             (c.i).

     (e) install_exec_creds() [compute_creds() as was] is called to install the
     	 new credentials.  This performs the following steps with regard to
     	 credentials:

         (i) Calls security_bprm_committing_creds() to apply any security
             requirements, such as flushing unauthorised files in SELinux, that
             must be done before the credentials are changed.

	     This is made up of bits of security_bprm_apply_creds() and
	     security_bprm_post_apply_creds(), both of which have been removed.
	     This function is not allowed to fail; anything that might fail
	     must have been done in (c.ii).

         (ii) Calls commit_creds() to apply the new credentials in a single
             assignment (more or less).  Possibly pdeath_signal and dumpable
             should be part of struct creds.

	 (iii) Unlocks the task's cred_replace_mutex, thus allowing
	     PTRACE_ATTACH to take place.

         (iv) Clears The bprm->cred pointer as the credentials it was holding
             are now immutable.

         (v) Calls security_bprm_committed_creds() to apply any security
             alterations that must be done after the creds have been changed.
             SELinux uses this to flush signals and signal handlers.

     (f) If an error occurs before (d.i), bprm_free() will call abort_creds()
     	 to destroy the proposed new credentials and will then unlock
     	 cred_replace_mutex.  No changes to the credentials will have been
     	 made.

 (2) LSM interface.

     A number of functions have been changed, added or removed:

     (*) security_bprm_alloc(), ->bprm_alloc_security()
     (*) security_bprm_free(), ->bprm_free_security()

     	 Removed in favour of preparing new credentials and modifying those.

     (*) security_bprm_apply_creds(), ->bprm_apply_creds()
     (*) security_bprm_post_apply_creds(), ->bprm_post_apply_creds()

     	 Removed; split between security_bprm_set_creds(),
     	 security_bprm_committing_creds() and security_bprm_committed_creds().

     (*) security_bprm_set(), ->bprm_set_security()

     	 Removed; folded into security_bprm_set_creds().

     (*) security_bprm_set_creds(), ->bprm_set_creds()

     	 New.  The new credentials in bprm->creds should be checked and set up
     	 as appropriate.  bprm->cred_prepared is 0 on the first call, 1 on the
     	 second and subsequent calls.

     (*) security_bprm_committing_creds(), ->bprm_committing_creds()
     (*) security_bprm_committed_creds(), ->bprm_committed_creds()

     	 New.  Apply the security effects of the new credentials.  This
     	 includes closing unauthorised files in SELinux.  This function may not
     	 fail.  When the former is called, the creds haven't yet been applied
     	 to the process; when the latter is called, they have.

 	 The former may access bprm->cred, the latter may not.

 (3) SELinux.

     SELinux has a number of changes, in addition to those to support the LSM
     interface changes mentioned above:

     (a) The bprm_security_struct struct has been removed in favour of using
     	 the credentials-under-construction approach.

     (c) flush_unauthorized_files() now takes a cred pointer and passes it on
     	 to inode_has_perm(), file_has_perm() and dentry_open().

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>
2008-11-14 10:39:24 +11:00
David Howells
d84f4f992c CRED: Inaugurate COW credentials
Inaugurate copy-on-write credentials management.  This uses RCU to manage the
credentials pointer in the task_struct with respect to accesses by other tasks.
A process may only modify its own credentials, and so does not need locking to
access or modify its own credentials.

A mutex (cred_replace_mutex) is added to the task_struct to control the effect
of PTRACE_ATTACHED on credential calculations, particularly with respect to
execve().

With this patch, the contents of an active credentials struct may not be
changed directly; rather a new set of credentials must be prepared, modified
and committed using something like the following sequence of events:

	struct cred *new = prepare_creds();
	int ret = blah(new);
	if (ret < 0) {
		abort_creds(new);
		return ret;
	}
	return commit_creds(new);

There are some exceptions to this rule: the keyrings pointed to by the active
credentials may be instantiated - keyrings violate the COW rule as managing
COW keyrings is tricky, given that it is possible for a task to directly alter
the keys in a keyring in use by another task.

To help enforce this, various pointers to sets of credentials, such as those in
the task_struct, are declared const.  The purpose of this is compile-time
discouragement of altering credentials through those pointers.  Once a set of
credentials has been made public through one of these pointers, it may not be
modified, except under special circumstances:

  (1) Its reference count may incremented and decremented.

  (2) The keyrings to which it points may be modified, but not replaced.

The only safe way to modify anything else is to create a replacement and commit
using the functions described in Documentation/credentials.txt (which will be
added by a later patch).

This patch and the preceding patches have been tested with the LTP SELinux
testsuite.

This patch makes several logical sets of alteration:

 (1) execve().

     This now prepares and commits credentials in various places in the
     security code rather than altering the current creds directly.

 (2) Temporary credential overrides.

     do_coredump() and sys_faccessat() now prepare their own credentials and
     temporarily override the ones currently on the acting thread, whilst
     preventing interference from other threads by holding cred_replace_mutex
     on the thread being dumped.

     This will be replaced in a future patch by something that hands down the
     credentials directly to the functions being called, rather than altering
     the task's objective credentials.

 (3) LSM interface.

     A number of functions have been changed, added or removed:

     (*) security_capset_check(), ->capset_check()
     (*) security_capset_set(), ->capset_set()

     	 Removed in favour of security_capset().

     (*) security_capset(), ->capset()

     	 New.  This is passed a pointer to the new creds, a pointer to the old
     	 creds and the proposed capability sets.  It should fill in the new
     	 creds or return an error.  All pointers, barring the pointer to the
     	 new creds, are now const.

     (*) security_bprm_apply_creds(), ->bprm_apply_creds()

     	 Changed; now returns a value, which will cause the process to be
     	 killed if it's an error.

     (*) security_task_alloc(), ->task_alloc_security()

     	 Removed in favour of security_prepare_creds().

     (*) security_cred_free(), ->cred_free()

     	 New.  Free security data attached to cred->security.

     (*) security_prepare_creds(), ->cred_prepare()

     	 New. Duplicate any security data attached to cred->security.

     (*) security_commit_creds(), ->cred_commit()

     	 New. Apply any security effects for the upcoming installation of new
     	 security by commit_creds().

     (*) security_task_post_setuid(), ->task_post_setuid()

     	 Removed in favour of security_task_fix_setuid().

     (*) security_task_fix_setuid(), ->task_fix_setuid()

     	 Fix up the proposed new credentials for setuid().  This is used by
     	 cap_set_fix_setuid() to implicitly adjust capabilities in line with
     	 setuid() changes.  Changes are made to the new credentials, rather
     	 than the task itself as in security_task_post_setuid().

     (*) security_task_reparent_to_init(), ->task_reparent_to_init()

     	 Removed.  Instead the task being reparented to init is referred
     	 directly to init's credentials.

	 NOTE!  This results in the loss of some state: SELinux's osid no
	 longer records the sid of the thread that forked it.

     (*) security_key_alloc(), ->key_alloc()
     (*) security_key_permission(), ->key_permission()

     	 Changed.  These now take cred pointers rather than task pointers to
     	 refer to the security context.

 (4) sys_capset().

     This has been simplified and uses less locking.  The LSM functions it
     calls have been merged.

 (5) reparent_to_kthreadd().

     This gives the current thread the same credentials as init by simply using
     commit_thread() to point that way.

 (6) __sigqueue_alloc() and switch_uid()

     __sigqueue_alloc() can't stop the target task from changing its creds
     beneath it, so this function gets a reference to the currently applicable
     user_struct which it then passes into the sigqueue struct it returns if
     successful.

     switch_uid() is now called from commit_creds(), and possibly should be
     folded into that.  commit_creds() should take care of protecting
     __sigqueue_alloc().

 (7) [sg]et[ug]id() and co and [sg]et_current_groups.

     The set functions now all use prepare_creds(), commit_creds() and
     abort_creds() to build and check a new set of credentials before applying
     it.

     security_task_set[ug]id() is called inside the prepared section.  This
     guarantees that nothing else will affect the creds until we've finished.

     The calling of set_dumpable() has been moved into commit_creds().

     Much of the functionality of set_user() has been moved into
     commit_creds().

     The get functions all simply access the data directly.

 (8) security_task_prctl() and cap_task_prctl().

     security_task_prctl() has been modified to return -ENOSYS if it doesn't
     want to handle a function, or otherwise return the return value directly
     rather than through an argument.

     Additionally, cap_task_prctl() now prepares a new set of credentials, even
     if it doesn't end up using it.

 (9) Keyrings.

     A number of changes have been made to the keyrings code:

     (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have
     	 all been dropped and built in to the credentials functions directly.
     	 They may want separating out again later.

     (b) key_alloc() and search_process_keyrings() now take a cred pointer
     	 rather than a task pointer to specify the security context.

     (c) copy_creds() gives a new thread within the same thread group a new
     	 thread keyring if its parent had one, otherwise it discards the thread
     	 keyring.

     (d) The authorisation key now points directly to the credentials to extend
     	 the search into rather pointing to the task that carries them.

     (e) Installing thread, process or session keyrings causes a new set of
     	 credentials to be created, even though it's not strictly necessary for
     	 process or session keyrings (they're shared).

(10) Usermode helper.

     The usermode helper code now carries a cred struct pointer in its
     subprocess_info struct instead of a new session keyring pointer.  This set
     of credentials is derived from init_cred and installed on the new process
     after it has been cloned.

     call_usermodehelper_setup() allocates the new credentials and
     call_usermodehelper_freeinfo() discards them if they haven't been used.  A
     special cred function (prepare_usermodeinfo_creds()) is provided
     specifically for call_usermodehelper_setup() to call.

     call_usermodehelper_setkeys() adjusts the credentials to sport the
     supplied keyring as the new session keyring.

(11) SELinux.

     SELinux has a number of changes, in addition to those to support the LSM
     interface changes mentioned above:

     (a) selinux_setprocattr() no longer does its check for whether the
     	 current ptracer can access processes with the new SID inside the lock
     	 that covers getting the ptracer's SID.  Whilst this lock ensures that
     	 the check is done with the ptracer pinned, the result is only valid
     	 until the lock is released, so there's no point doing it inside the
     	 lock.

(12) is_single_threaded().

     This function has been extracted from selinux_setprocattr() and put into
     a file of its own in the lib/ directory as join_session_keyring() now
     wants to use it too.

     The code in SELinux just checked to see whether a task shared mm_structs
     with other tasks (CLONE_VM), but that isn't good enough.  We really want
     to know if they're part of the same thread group (CLONE_THREAD).

(13) nfsd.

     The NFS server daemon now has to use the COW credentials to set the
     credentials it is going to use.  It really needs to pass the credentials
     down to the functions it calls, but it can't do that until other patches
     in this series have been applied.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:23 +11:00
David Howells
745ca2475a CRED: Pass credentials through dentry_open()
Pass credentials through dentry_open() so that the COW creds patch can have
SELinux's flush_unauthorized_files() pass the appropriate creds back to itself
when it opens its null chardev.

The security_dentry_open() call also now takes a creds pointer, as does the
dentry_open hook in struct security_operations.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:22 +11:00
David Howells
88e67f3b88 CRED: Make inode_has_perm() and file_has_perm() take a cred pointer
Make inode_has_perm() and file_has_perm() take a cred pointer rather than a
task pointer.

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>
2008-11-14 10:39:21 +11:00
David Howells
bb952bb98a CRED: Separate per-task-group keyrings from signal_struct
Separate per-task-group keyrings from signal_struct and dangle their anchor
from the cred struct rather than the signal_struct.

Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:20 +11:00
David Howells
275bb41e9d CRED: Wrap access to SELinux's task SID
Wrap access to SELinux's task SID, using task_sid() and current_sid() as
appropriate.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:19 +11:00
David Howells
c69e8d9c01 CRED: Use RCU to access another task's creds and to release a task's own creds
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>
2008-11-14 10:39:19 +11:00
David Howells
86a264abe5 CRED: Wrap current->cred and a few other accessors
Wrap current->cred and a few other accessors to hide their actual
implementation.

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>
2008-11-14 10:39:18 +11:00
David Howells
f1752eec61 CRED: Detach the credentials from task_struct
Detach the credentials from task_struct, duplicating them in copy_process()
and releasing them in __put_task_struct().

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>
2008-11-14 10:39:17 +11:00
David Howells
b6dff3ec5e CRED: Separate task security context from task_struct
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>
2008-11-14 10:39:16 +11:00
David Howells
15a2460ed0 CRED: Constify the kernel_cap_t arguments to the capset LSM hooks
Constify the kernel_cap_t arguments to the capset LSM hooks.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:15 +11:00
David Howells
1cdcbec1a3 CRED: Neuter sys_capset()
Take away the ability for sys_capset() to affect processes other than current.

This means that current will not need to lock its own credentials when reading
them against interference by other processes.

This has effectively been the case for a while anyway, since:

 (1) Without LSM enabled, sys_capset() is disallowed.

 (2) With file-based capabilities, sys_capset() is neutered.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:14 +11:00
David Howells
8bbf4976b5 KEYS: Alter use of key instantiation link-to-keyring argument
Alter the use of the key instantiation and negation functions' link-to-keyring
arguments.  Currently this specifies a keyring in the target process to link
the key into, creating the keyring if it doesn't exist.  This, however, can be
a problem for copy-on-write credentials as it means that the instantiating
process can alter the credentials of the requesting process.

This patch alters the behaviour such that:

 (1) If keyctl_instantiate_key() or keyctl_negate_key() are given a specific
     keyring by ID (ringid >= 0), then that keyring will be used.

 (2) If keyctl_instantiate_key() or keyctl_negate_key() are given one of the
     special constants that refer to the requesting process's keyrings
     (KEY_SPEC_*_KEYRING, all <= 0), then:

     (a) If sys_request_key() was given a keyring to use (destringid) then the
     	 key will be attached to that keyring.

     (b) If sys_request_key() was given a NULL keyring, then the key being
     	 instantiated will be attached to the default keyring as set by
     	 keyctl_set_reqkey_keyring().

 (3) No extra link will be made.

Decision point (1) follows current behaviour, and allows those instantiators
who've searched for a specifically named keyring in the requestor's keyring so
as to partition the keys by type to still have their named keyrings.

Decision point (2) allows the requestor to make sure that the key or keys that
get produced by request_key() go where they want, whilst allowing the
instantiator to request that the key is retained.  This is mainly useful for
situations where the instantiator makes a secondary request, the key for which
should be retained by the initial requestor:

	+-----------+        +--------------+        +--------------+
	|           |        |              |        |              |
	| Requestor |------->| Instantiator |------->| Instantiator |
	|           |        |              |        |              |
	+-----------+        +--------------+        +--------------+
	           request_key()           request_key()

This might be useful, for example, in Kerberos, where the requestor requests a
ticket, and then the ticket instantiator requests the TGT, which someone else
then has to go and fetch.  The TGT, however, should be retained in the
keyrings of the requestor, not the first instantiator.  To make this explict
an extra special keyring constant is also added.

Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:14 +11:00
David Howells
e9e349b051 KEYS: Disperse linux/key_ui.h
Disperse the bits of linux/key_ui.h as the reason they were put here (keyfs)
didn't get in.

Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:13 +11:00
David Howells
b103c59883 CRED: Wrap task credential accesses in the capabilities code
Wrap access to task credentials so that they can be separated more easily from
the task_struct during the introduction of COW creds.

Change most current->(|e|s|fs)[ug]id to current_(|e|s|fs)[ug]id().

Change some task->e?[ug]id to task_e?[ug]id().  In some places it makes more
sense to use RCU directly rather than a convenient wrapper; these will be
addressed by later patches.

Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Cc: Andrew G. Morgan <morgan@kernel.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:11 +11:00
David Howells
47d804bfa1 CRED: Wrap task credential accesses in the key management code
Wrap access to task credentials so that they can be separated more easily from
the task_struct during the introduction of COW creds.

Change most current->(|e|s|fs)[ug]id to current_(|e|s|fs)[ug]id().

Change some task->e?[ug]id to task_e?[ug]id().  In some places it makes more
sense to use RCU directly rather than a convenient wrapper; these will be
addressed by later patches.

Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-14 10:39:11 +11:00
David S. Miller
7e452baf6b Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Conflicts:

	drivers/message/fusion/mptlan.c
	drivers/net/sfc/ethtool.c
	net/mac80211/debugfs_sta.c
2008-11-11 15:43:02 -08:00
Eric Paris
066746796b Currently SELinux jumps through some ugly hoops to not audit a capbility
check when determining if a process has additional powers to override
memory limits or when trying to read/write illegal file labels.  Use
the new noaudit call instead.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 22:02:57 +11:00
Eric Paris
06112163f5 Add a new capable interface that will be used by systems that use audit to
make an A or B type decision instead of a security decision.  Currently
this is the case at least for filesystems when deciding if a process can use
the reserved 'root' blocks and for the case of things like the oom
algorithm determining if processes are root processes and should be less
likely to be killed.  These types of security system requests should not be
audited or logged since they are not really security decisions.  It would be
possible to solve this problem like the vm_enough_memory security check did
by creating a new LSM interface and moving all of the policy into that
interface but proves the needlessly bloat the LSM and provide complex
indirection.

This merely allows those decisions to be made where they belong and to not
flood logs or printk with denials for thing that are not security decisions.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 22:02:50 +11:00
Eric Paris
3fc689e96c Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a
non-zero pE we will crate a new audit record which contains the entire set
of known information about the executable in question, fP, fI, fE, fversion
and includes the process's pE, pI, pP.  Before and after the bprm capability
are applied.  This record type will only be emitted from execve syscalls.

an example of making ping use fcaps instead of setuid:

setcap "cat_net_raw+pe" /bin/ping

type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null)
type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000
type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1"
type=CWD msg=audit(1225742021.015:236):  cwd="/home/test"
type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2
type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 21:48:18 +11:00
Eric Paris
c0b004413a This patch add a generic cpu endian caps structure and externally available
functions which retrieve fcaps information from disk.  This information is
necessary so fcaps information can be collected and recorded by the audit
system.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 21:48:10 +11:00
David Howells
1f8f5cf6e4 KEYS: Make request key instantiate the per-user keyrings
Make request_key() instantiate the per-user keyrings so that it doesn't oops
if it needs to get hold of the user session keyring because there isn't a
session keyring in place.

Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: Steve French <smfrench@gmail.com>
Tested-by: Rutger Nijlunsing <rutger.nijlunsing@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-11-10 13:20:57 -08:00
Eric Paris
39c9aede2b SELinux: Use unknown perm handling to handle unknown netlink msg types
Currently when SELinux has not been updated to handle a netlink message
type the operation is denied with EINVAL.  This patch will leave the
audit/warning message so things get fixed but if policy chose to allow
unknowns this will allow the netlink operation.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-09 07:33:18 +08:00
David S. Miller
9eeda9abd1 Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Conflicts:

	drivers/net/wireless/ath5k/base.c
	net/8021q/vlan_core.c
2008-11-06 22:43:03 -08:00
Serge E. Hallyn
1f29fae297 file capabilities: add no_file_caps switch (v4)
Add a no_file_caps boot option when file capabilities are
compiled into the kernel (CONFIG_SECURITY_FILE_CAPABILITIES=y).

This allows distributions to ship a kernel with file capabilities
compiled in, without forcing users to use (and understand and
trust) them.

When no_file_caps is specified at boot, then when a process executes
a file, any file capabilities stored with that file will not be
used in the calculation of the process' new capability sets.

This means that booting with the no_file_caps boot option will
not be the same as booting a kernel with file capabilities
compiled out - in particular a task with  CAP_SETPCAP will not
have any chance of passing capabilities to another task (which
isn't "really" possible anyway, and which may soon by killed
altogether by David Howells in any case), and it will instead
be able to put new capabilities in its pI.  However since fI
will always be empty and pI is masked with fI, it gains the
task nothing.

We also support the extra prctl options, setting securebits and
dropping capabilities from the per-process bounding set.

The other remaining difference is that killpriv, task_setscheduler,
setioprio, and setnice will continue to be hooked.  That will
be noticable in the case where a root task changed its uid
while keeping some caps, and another task owned by the new uid
tries to change settings for the more privileged task.

Changelog:
	Nov 05 2008: (v4) trivial port on top of always-start-\
		with-clear-caps patch
	Sep 23 2008: nixed file_caps_enabled when file caps are
		not compiled in as it isn't used.
		Document no_file_caps in kernel-parameters.txt.

Signed-off-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-06 07:14:51 +08:00
James Morris
e21e696edb Merge branch 'master' into next 2008-11-06 07:12:34 +08:00
Michal Schmidt
2f99db28af selinux: recognize netlink messages for 'ip addrlabel'
In enforcing mode '/sbin/ip addrlabel' results in a SELinux error:
type=SELINUX_ERR msg=audit(1225698822.073:42): SELinux:  unrecognized
netlink message type=74 for sclass=43

The problem is missing RTM_*ADDRLABEL entries in SELinux's netlink
message types table.

Reported in https://bugzilla.redhat.com/show_bug.cgi?id=469423

Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-06 07:08:36 +08:00
Eric Paris
41d9f9c524 SELinux: hold tasklist_lock and siglock while waking wait_chldexit
SELinux has long been calling wake_up_interruptible() on
current->parent->signal->wait_chldexit without holding any locks.  It
appears that this operation should hold the tasklist_lock to dereference
current->parent and we should hold the siglock when waking up the
signal->wait_chldexit.

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-05 08:44:11 +11:00
Linus Torvalds
0a6d2fac61 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  SELinux: properly handle empty tty_files list
2008-11-01 09:50:38 -07:00
Serge Hallyn
3318a386e4 file caps: always start with clear bprm->caps_*
While Linux doesn't honor setuid on scripts.  However, it mistakenly
behaves differently for file capabilities.

This patch fixes that behavior by making sure that get_file_caps()
begins with empty bprm->caps_*.  That way when a script is loaded,
its bprm->caps_* may be filled when binfmt_misc calls prepare_binprm(),
but they will be cleared again when binfmt_elf calls prepare_binprm()
next to read the interpreter's file capabilities.

Signed-off-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-11-01 09:49:45 -07:00
Eric Paris
37dd0bd04a SELinux: properly handle empty tty_files list
SELinux has wrongly (since 2004) had an incorrect test for an empty
tty->tty_files list.  With an empty list selinux would be pointing to part
of the tty struct itself and would then proceed to dereference that value
and again dereference that result.  An F10 change to plymouth on a ppc64
system is actually currently triggering this bug.  This patch uses
list_empty() to handle empty lists rather than looking at a meaningless
location.

[note, this fixes the oops reported in
https://bugzilla.redhat.com/show_bug.cgi?id=469079]

Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-11-01 09:38:48 +11:00
Harvey Harrison
3685f25de1 misc: replace NIPQUAD()
Using NIPQUAD() with NIPQUAD_FMT, %d.%d.%d.%d or %u.%u.%u.%u
can be replaced with %pI4

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-31 00:56:49 -07:00
David S. Miller
a1744d3bee Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Conflicts:

	drivers/net/wireless/p54/p54common.c
2008-10-31 00:17:34 -07:00
Alan Cox
731572d39f nfsd: fix vm overcommit crash
Junjiro R.  Okajima reported a problem where knfsd crashes if you are
using it to export shmemfs objects and run strict overcommit.  In this
situation the current->mm based modifier to the overcommit goes through a
NULL pointer.

We could simply check for NULL and skip the modifier but we've caught
other real bugs in the past from mm being NULL here - cases where we did
need a valid mm set up (eg the exec bug about a year ago).

To preserve the checks and get the logic we want shuffle the checking
around and add a new helper to the vm_ security wrappers

Also fix a current->mm reference in nommu that should use the passed mm

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: fix build]
Reported-by: Junjiro R. Okajima <hooanon05@yahoo.co.jp>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-30 11:38:47 -07:00
Eric Paris
8b6a5a37f8 SELinux: check open perms in dentry_open not inode_permission
Some operations, like searching a directory path or connecting a unix domain
socket, make explicit calls into inode_permission.  Our choices are to
either try to come up with a signature for all of the explicit calls to
inode_permission and do not check open on those, or to move the open checks to
dentry_open where we know this is always an open operation.  This patch moves
the checks to dentry_open.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by:  Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-10-31 02:00:52 +11:00
Harvey Harrison
5b095d9892 net: replace %p6 with %pI6
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-29 12:52:50 -07:00
Harvey Harrison
1afa67f5e7 misc: replace NIP6_FMT with %p6 format specifier
The iscsi_ibft.c changes are almost certainly a bugfix as the
pointer 'ip' is a u8 *, so they never print the last 8 bytes
of the IPv6 address, and the eight bytes they do print have
a zero byte with them in each 16-bit word.

Other than that, this should cause no difference in functionality.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-28 16:06:44 -07:00
Alexey Dobriyan
def8b4faff net: reduce structures when XFRM=n
ifdef out
* struct sk_buff::sp		(pointer)
* struct dst_entry::xfrm	(pointer)
* struct sock::sk_policy	(2 pointers)

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-28 13:24:06 -07:00
Linus Torvalds
99ebcf8285 Merge branch 'v28-timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'v28-timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (36 commits)
  fix documentation of sysrq-q really
  Fix documentation of sysrq-q
  timer_list: add base address to clock base
  timer_list: print cpu number of clockevents device
  timer_list: print real timer address
  NOHZ: restart tick device from irq_enter()
  NOHZ: split tick_nohz_restart_sched_tick()
  NOHZ: unify the nohz function calls in irq_enter()
  timers: fix itimer/many thread hang, fix
  timers: fix itimer/many thread hang, v3
  ntp: improve adjtimex frequency rounding
  timekeeping: fix rounding problem during clock update
  ntp: let update_persistent_clock() sleep
  hrtimer: reorder struct hrtimer to save 8 bytes on 64bit builds
  posix-timers: lock_timer: make it readable
  posix-timers: lock_timer: kill the bogus ->it_id check
  posix-timers: kill ->it_sigev_signo and ->it_sigev_value
  posix-timers: sys_timer_create: cleanup the error handling
  posix-timers: move the initialization of timer->sigq from send to create path
  posix-timers: sys_timer_create: simplify and s/tasklist/rcu/
  ...

Fix trivial conflicts due to sysrq-q description clahes in
Documentation/sysrq.txt and drivers/char/sysrq.c
2008-10-20 13:19:56 -07:00
Lai Jiangshan
47c59803be devcgroup: remove spin_lock()
Since we introduced rcu for read side, spin_lock is used only for update.
But we always hold cgroup_lock() when update, so spin_lock() is not need.

Additional cleanup:
1) include linux/rcupdate.h explicitly
2) remove unused variable cur_devcgroup in devcgroup_update_access()

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: "Serge E. Hallyn" <serue@us.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:38 -07:00
Li Zefan
c012a54ae0 devcgroup: remove unused variable
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:38 -07:00
Li Zefan
2cdc7241a2 devcgroup: use kmemdup()
This saves 40 bytes on my x86_32 box.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-20 08:52:38 -07:00
Thomas Gleixner
c465a76af6 Merge branches 'timers/clocksource', 'timers/hrtimers', 'timers/nohz', 'timers/ntp', 'timers/posixtimers' and 'timers/debug' into v28-timers-for-linus 2008-10-20 13:14:06 +02:00
Steven Whitehouse
a447c09324 vfs: Use const for kernel parser table
This is a much better version of a previous patch to make the parser
tables constant. Rather than changing the typedef, we put the "const" in
all the various places where its required, allowing the __initconst
exception for nfsroot which was the cause of the previous trouble.

This was posted for review some time ago and I believe its been in -mm
since then.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Cc: Alexander Viro <aviro@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-13 10:10:37 -07:00
Linus Torvalds
8d71ff0bef Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6: (24 commits)
  integrity: special fs magic
  As pointed out by Jonathan Corbet, the timer must be deleted before
  ERROR: code indent should use tabs where possible
  The tpm_dev_release function is only called for platform devices, not pnp
  Protect tpm_chip_list when transversing it.
  Renames num_open to is_open, as only one process can open the file at a time.
  Remove the BKL calls from the TPM driver, which were added in the overall
  netlabel: Add configuration support for local labeling
  cipso: Add support for native local labeling and fixup mapping names
  netlabel: Changes to the NetLabel security attributes to allow LSMs to pass full contexts
  selinux: Cache NetLabel secattrs in the socket's security struct
  selinux: Set socket NetLabel based on connection endpoint
  netlabel: Add functionality to set the security attributes of a packet
  netlabel: Add network address selectors to the NetLabel/LSM domain mapping
  netlabel: Add a generic way to create ordered linked lists of network addrs
  netlabel: Replace protocol/NetLabel linking with refrerence counts
  smack: Fix missing calls to netlbl_skbuff_err()
  selinux: Fix missing calls to netlbl_skbuff_err()
  selinux: Fix a problem in security_netlbl_sid_to_secattr()
  selinux: Better local/forward check in selinux_ip_postroute()
  ...
2008-10-13 10:00:44 -07:00
Alan Cox
934e6ebf96 tty: Redo current tty locking
Currently it is sometimes locked by the tty mutex and sometimes by the
sighand lock. The latter is in fact correct and now we can hand back referenced
objects we can fix this up without problems around sleeping functions.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-13 09:51:41 -07:00
Alan Cox
452a00d2ee tty: Make get_current_tty use a kref
We now return a kref covered tty reference. That ensures the tty structure
doesn't go away when you have a return from get_current_tty. This is not
enough to protect you from most of the resources being freed behind your
back - yet.

[Updated to include fixes for SELinux problems found by Andrew Morton and
 an s390 leak found while debugging the former]

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-10-13 09:51:41 -07:00
Mimi Zohar
9256292782 integrity: special fs magic
Discussion on the mailing list questioned the use of these
magic values in userspace, concluding these values are already
exported to userspace via statfs and their correct/incorrect
usage is left up to the userspace application.

  - Move special fs magic number definitions to magic.h
  - Add magic.h include

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-10-13 09:47:43 +11:00
James Morris
0da939b005 Merge branch 'master' of git://git.infradead.org/users/pcmoore/lblnet-2.6_next into next 2008-10-11 09:26:14 +11:00
Paul Moore
8d75899d03 netlabel: Changes to the NetLabel security attributes to allow LSMs to pass full contexts
This patch provides support for including the LSM's secid in addition to
the LSM's MLS information in the NetLabel security attributes structure.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:33 -04:00
Paul Moore
6c5b3fc014 selinux: Cache NetLabel secattrs in the socket's security struct
Previous work enabled the use of address based NetLabel selectors, which
while highly useful, brought the potential for additional per-packet overhead
when used.  This patch attempts to mitigate some of that overhead by caching
the NetLabel security attribute struct within the SELinux socket security
structure.  This should help eliminate the need to recreate the NetLabel
secattr structure for each packet resulting in less overhead.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:33 -04:00
Paul Moore
014ab19a69 selinux: Set socket NetLabel based on connection endpoint
Previous work enabled the use of address based NetLabel selectors, which while
highly useful, brought the potential for additional per-packet overhead when
used.  This patch attempts to solve that by applying NetLabel socket labels
when sockets are connect()'d.  This should alleviate the per-packet NetLabel
labeling for all connected sockets (yes, it even works for connected DGRAM
sockets).

Signed-off-by: Paul Moore <paul.moore@hp.com>
Reviewed-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:33 -04:00
Paul Moore
948bf85c1b netlabel: Add functionality to set the security attributes of a packet
This patch builds upon the new NetLabel address selector functionality by
providing the NetLabel KAPI and CIPSO engine support needed to enable the
new packet-based labeling.  The only new addition to the NetLabel KAPI at
this point is shown below:

 * int netlbl_skbuff_setattr(skb, family, secattr)

... and is designed to be called from a Netfilter hook after the packet's
IP header has been populated such as in the FORWARD or LOCAL_OUT hooks.

This patch also provides the necessary SELinux hooks to support this new
functionality.  Smack support is not currently included due to uncertainty
regarding the permissions needed to expand the Smack network access controls.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Reviewed-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:32 -04:00
Paul Moore
b1edeb1023 netlabel: Replace protocol/NetLabel linking with refrerence counts
NetLabel has always had a list of backpointers in the CIPSO DOI definition
structure which pointed to the NetLabel LSM domain mapping structures which
referenced the CIPSO DOI struct.  The rationale for this was that when an
administrator removed a CIPSO DOI from the system all of the associated
NetLabel LSM domain mappings should be removed as well; a list of
backpointers made this a simple operation.

Unfortunately, while the backpointers did make the removal easier they were
a bit of a mess from an implementation point of view which was making
further development difficult.  Since the removal of a CIPSO DOI is a
realtively rare event it seems to make sense to remove this backpointer
list as the optimization was hurting us more then it was helping.  However,
we still need to be able to track when a CIPSO DOI definition is being used
so replace the backpointer list with a reference count.  In order to
preserve the current functionality of removing the associated LSM domain
mappings when a CIPSO DOI is removed we walk the LSM domain mapping table,
removing the relevant entries.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Reviewed-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:31 -04:00
Paul Moore
a8134296ba smack: Fix missing calls to netlbl_skbuff_err()
Smack needs to call netlbl_skbuff_err() to let NetLabel do the necessary
protocol specific error handling.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
2008-10-10 10:16:31 -04:00
Paul Moore
dfaebe9825 selinux: Fix missing calls to netlbl_skbuff_err()
At some point I think I messed up and dropped the calls to netlbl_skbuff_err()
which are necessary for CIPSO to send error notifications to remote systems.
This patch re-introduces the error handling calls into the SELinux code.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:31 -04:00
Paul Moore
99d854d231 selinux: Fix a problem in security_netlbl_sid_to_secattr()
Currently when SELinux fails to allocate memory in
security_netlbl_sid_to_secattr() the NetLabel LSM domain field is set to
NULL which triggers the default NetLabel LSM domain mapping which may not
always be the desired mapping.  This patch fixes this by returning an error
when the kernel is unable to allocate memory.  This could result in more
failures on a system with heavy memory pressure but it is the "correct"
thing to do.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:30 -04:00
Paul Moore
d8395c876b selinux: Better local/forward check in selinux_ip_postroute()
It turns out that checking to see if skb->sk is NULL is not a very good
indicator of a forwarded packet as some locally generated packets also have
skb->sk set to NULL.  Fix this by not only checking the skb->sk field but also
the IP[6]CB(skb)->flags field for the IP[6]SKB_FORWARDED flag.  While we are
at it, we are calling selinux_parse_skb() much earlier than we really should
resulting in potentially wasted cycles parsing packets for information we
might no use; so shuffle the code around a bit to fix this.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:30 -04:00
Paul Moore
aa86290089 selinux: Correctly handle IPv4 packets on IPv6 sockets in all cases
We did the right thing in a few cases but there were several areas where we
determined a packet's address family based on the socket's address family which
is not the right thing to do since we can get IPv4 packets on IPv6 sockets.
This patch fixes these problems by either taking the address family directly
from the packet.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:29 -04:00
Paul Moore
accc609322 selinux: Cleanup the NetLabel glue code
We were doing a lot of extra work in selinux_netlbl_sock_graft() what wasn't
necessary so this patch removes that code.  It also removes the redundant
second argument to selinux_netlbl_sock_setsid() which allows us to simplify a
few other functions.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: James Morris <jmorris@namei.org>
2008-10-10 10:16:29 -04:00
Paul Moore
3040a6d5a2 selinux: Fix an uninitialized variable BUG/panic in selinux_secattr_to_sid()
At some point during the 2.6.27 development cycle two new fields were added
to the SELinux context structure, a string pointer and a length field.  The
code in selinux_secattr_to_sid() was not modified and as a result these two
fields were left uninitialized which could result in erratic behavior,
including kernel panics, when NetLabel is used.  This patch fixes the
problem by fully initializing the context in selinux_secattr_to_sid() before
use and reducing the level of direct context manipulation done to help
prevent future problems.

Please apply this to the 2.6.27-rcX release stream.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-10-04 08:25:18 +10:00
Paul Moore
81990fbdd1 selinux: Fix an uninitialized variable BUG/panic in selinux_secattr_to_sid()
At some point during the 2.6.27 development cycle two new fields were added
to the SELinux context structure, a string pointer and a length field.  The
code in selinux_secattr_to_sid() was not modified and as a result these two
fields were left uninitialized which could result in erratic behavior,
including kernel panics, when NetLabel is used.  This patch fixes the
problem by fully initializing the context in selinux_secattr_to_sid() before
use and reducing the level of direct context manipulation done to help
prevent future problems.

Please apply this to the 2.6.27-rcX release stream.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: James Morris <jmorris@namei.org>
2008-10-04 08:18:18 +10:00
Stephen Smalley
ea6b184f7d selinux: use default proc sid on symlinks
As we are not concerned with fine-grained control over reading of
symlinks in proc, always use the default proc SID for all proc symlinks.
This should help avoid permission issues upon changes to the proc tree
as in the /proc/net -> /proc/self/net example.
This does not alter labeling of symlinks within /proc/pid directories.
ls -Zd /proc/net output before and after the patch should show the difference.

Signed-off-by:  Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
2008-09-30 00:26:53 +10:00
Serge E. Hallyn
de45e806a8 file capabilities: uninline cap_safe_nice
This reduces the kernel size by 289 bytes.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Signed-off-by: James Morris <jmorris@namei.org>
2008-09-27 15:07:56 +10:00
James Morris
ab2b49518e Merge branch 'master' into next
Conflicts:

	MAINTAINERS

Thanks for breaking my tree :-)

Signed-off-by: James Morris <jmorris@namei.org>
2008-09-21 17:41:56 -07:00
Frank Mayhar
f06febc96b timers: fix itimer/many thread hang
Overview

This patch reworks the handling of POSIX CPU timers, including the
ITIMER_PROF, ITIMER_VIRT timers and rlimit handling.  It was put together
with the help of Roland McGrath, the owner and original writer of this code.

The problem we ran into, and the reason for this rework, has to do with using
a profiling timer in a process with a large number of threads.  It appears
that the performance of the old implementation of run_posix_cpu_timers() was
at least O(n*3) (where "n" is the number of threads in a process) or worse.
Everything is fine with an increasing number of threads until the time taken
for that routine to run becomes the same as or greater than the tick time, at
which point things degrade rather quickly.

This patch fixes bug 9906, "Weird hang with NPTL and SIGPROF."

Code Changes

This rework corrects the implementation of run_posix_cpu_timers() to make it
run in constant time for a particular machine.  (Performance may vary between
one machine and another depending upon whether the kernel is built as single-
or multiprocessor and, in the latter case, depending upon the number of
running processors.)  To do this, at each tick we now update fields in
signal_struct as well as task_struct.  The run_posix_cpu_timers() function
uses those fields to make its decisions.

We define a new structure, "task_cputime," to contain user, system and
scheduler times and use these in appropriate places:

struct task_cputime {
	cputime_t utime;
	cputime_t stime;
	unsigned long long sum_exec_runtime;
};

This is included in the structure "thread_group_cputime," which is a new
substructure of signal_struct and which varies for uniprocessor versus
multiprocessor kernels.  For uniprocessor kernels, it uses "task_cputime" as
a simple substructure, while for multiprocessor kernels it is a pointer:

struct thread_group_cputime {
	struct task_cputime totals;
};

struct thread_group_cputime {
	struct task_cputime *totals;
};

We also add a new task_cputime substructure directly to signal_struct, to
cache the earliest expiration of process-wide timers, and task_cputime also
replaces the it_*_expires fields of task_struct (used for earliest expiration
of thread timers).  The "thread_group_cputime" structure contains process-wide
timers that are updated via account_user_time() and friends.  In the non-SMP
case the structure is a simple aggregator; unfortunately in the SMP case that
simplicity was not achievable due to cache-line contention between CPUs (in
one measured case performance was actually _worse_ on a 16-cpu system than
the same test on a 4-cpu system, due to this contention).  For SMP, the
thread_group_cputime counters are maintained as a per-cpu structure allocated
using alloc_percpu().  The timer functions update only the timer field in
the structure corresponding to the running CPU, obtained using per_cpu_ptr().

We define a set of inline functions in sched.h that we use to maintain the
thread_group_cputime structure and hide the differences between UP and SMP
implementations from the rest of the kernel.  The thread_group_cputime_init()
function initializes the thread_group_cputime structure for the given task.
The thread_group_cputime_alloc() is a no-op for UP; for SMP it calls the
out-of-line function thread_group_cputime_alloc_smp() to allocate and fill
in the per-cpu structures and fields.  The thread_group_cputime_free()
function, also a no-op for UP, in SMP frees the per-cpu structures.  The
thread_group_cputime_clone_thread() function (also a UP no-op) for SMP calls
thread_group_cputime_alloc() if the per-cpu structures haven't yet been
allocated.  The thread_group_cputime() function fills the task_cputime
structure it is passed with the contents of the thread_group_cputime fields;
in UP it's that simple but in SMP it must also safely check that tsk->signal
is non-NULL (if it is it just uses the appropriate fields of task_struct) and,
if so, sums the per-cpu values for each online CPU.  Finally, the three
functions account_group_user_time(), account_group_system_time() and
account_group_exec_runtime() are used by timer functions to update the
respective fields of the thread_group_cputime structure.

Non-SMP operation is trivial and will not be mentioned further.

The per-cpu structure is always allocated when a task creates its first new
thread, via a call to thread_group_cputime_clone_thread() from copy_signal().
It is freed at process exit via a call to thread_group_cputime_free() from
cleanup_signal().

All functions that formerly summed utime/stime/sum_sched_runtime values from
from all threads in the thread group now use thread_group_cputime() to
snapshot the values in the thread_group_cputime structure or the values in
the task structure itself if the per-cpu structure hasn't been allocated.

Finally, the code in kernel/posix-cpu-timers.c has changed quite a bit.
The run_posix_cpu_timers() function has been split into a fast path and a
slow path; the former safely checks whether there are any expired thread
timers and, if not, just returns, while the slow path does the heavy lifting.
With the dedicated thread group fields, timers are no longer "rebalanced" and
the process_timer_rebalance() function and related code has gone away.  All
summing loops are gone and all code that used them now uses the
thread_group_cputime() inline.  When process-wide timers are set, the new
task_cputime structure in signal_struct is used to cache the earliest
expiration; this is checked in the fast path.

Performance

The fix appears not to add significant overhead to existing operations.  It
generally performs the same as the current code except in two cases, one in
which it performs slightly worse (Case 5 below) and one in which it performs
very significantly better (Case 2 below).  Overall it's a wash except in those
two cases.

I've since done somewhat more involved testing on a dual-core Opteron system.

Case 1: With no itimer running, for a test with 100,000 threads, the fixed
	kernel took 1428.5 seconds, 513 seconds more than the unfixed system,
	all of which was spent in the system.  There were twice as many
	voluntary context switches with the fix as without it.

Case 2: With an itimer running at .01 second ticks and 4000 threads (the most
	an unmodified kernel can handle), the fixed kernel ran the test in
	eight percent of the time (5.8 seconds as opposed to 70 seconds) and
	had better tick accuracy (.012 seconds per tick as opposed to .023
	seconds per tick).

Case 3: A 4000-thread test with an initial timer tick of .01 second and an
	interval of 10,000 seconds (i.e. a timer that ticks only once) had
	very nearly the same performance in both cases:  6.3 seconds elapsed
	for the fixed kernel versus 5.5 seconds for the unfixed kernel.

With fewer threads (eight in these tests), the Case 1 test ran in essentially
the same time on both the modified and unmodified kernels (5.2 seconds versus
5.8 seconds).  The Case 2 test ran in about the same time as well, 5.9 seconds
versus 5.4 seconds but again with much better tick accuracy, .013 seconds per
tick versus .025 seconds per tick for the unmodified kernel.

Since the fix affected the rlimit code, I also tested soft and hard CPU limits.

Case 4: With a hard CPU limit of 20 seconds and eight threads (and an itimer
	running), the modified kernel was very slightly favored in that while
	it killed the process in 19.997 seconds of CPU time (5.002 seconds of
	wall time), only .003 seconds of that was system time, the rest was
	user time.  The unmodified kernel killed the process in 20.001 seconds
	of CPU (5.014 seconds of wall time) of which .016 seconds was system
	time.  Really, though, the results were too close to call.  The results
	were essentially the same with no itimer running.

Case 5: With a soft limit of 20 seconds and a hard limit of 2000 seconds
	(where the hard limit would never be reached) and an itimer running,
	the modified kernel exhibited worse tick accuracy than the unmodified
	kernel: .050 seconds/tick versus .028 seconds/tick.  Otherwise,
	performance was almost indistinguishable.  With no itimer running this
	test exhibited virtually identical behavior and times in both cases.

In times past I did some limited performance testing.  those results are below.

On a four-cpu Opteron system without this fix, a sixteen-thread test executed
in 3569.991 seconds, of which user was 3568.435s and system was 1.556s.  On
the same system with the fix, user and elapsed time were about the same, but
system time dropped to 0.007 seconds.  Performance with eight, four and one
thread were comparable.  Interestingly, the timer ticks with the fix seemed
more accurate:  The sixteen-thread test with the fix received 149543 ticks
for 0.024 seconds per tick, while the same test without the fix received 58720
for 0.061 seconds per tick.  Both cases were configured for an interval of
0.01 seconds.  Again, the other tests were comparable.  Each thread in this
test computed the primes up to 25,000,000.

I also did a test with a large number of threads, 100,000 threads, which is
impossible without the fix.  In this case each thread computed the primes only
up to 10,000 (to make the runtime manageable).  System time dominated, at
1546.968 seconds out of a total 2176.906 seconds (giving a user time of
629.938s).  It received 147651 ticks for 0.015 seconds per tick, still quite
accurate.  There is obviously no comparable test without the fix.

Signed-off-by: Frank Mayhar <fmayhar@google.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-09-14 16:25:35 +02:00