This is the central patch of a cleanup series. In most cases there is no good
reason why someone would want to use a dentry for itself. This series reflects
that fact and embeds a struct path into nameidata.

Together with the other patches of this series
- it enforced the correct order of getting/releasing the reference count on
  <dentry,vfsmount> pairs
- it prepares the VFS for stacking support since it is essential to have a
  struct path in every place where the stack can be traversed
- it reduces the overall code size:

without patch series:
   text    data     bss     dec     hex filename
5321639  858418  715768 6895825  6938d1 vmlinux

with patch series:
   text    data     bss     dec     hex filename
5320026  858418  715768 6894212  693284 vmlinux

This patch:

Switch from nd->{dentry,mnt} to nd->path.{dentry,mnt} everywhere.

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: fix cifs]
[akpm@linux-foundation.org: fix smack]
Signed-off-by: NJan Blunck <jblunck@suse.de>
Signed-off-by: NAndreas Gruenbacher <agruen@suse.de>
Acked-by: NChristoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Some audit messages (namely configuration changes) are still emitted even if
the audit subsystem has been explicitly disabled.  This patch turns those
messages off as well.
Signed-off-by: NEric Paris <eparis@redhat.com>

New kind of audit rule predicates: "object is visible in given subtree".
The part that can be sanely implemented, that is.  Limitations:
	* if you have hardlink from outside of tree, you'd better watch
it too (or just watch the object itself, obviously)
	* if you mount something under a watched tree, tell audit
that new chunk should be added to watched subtrees
	* if you umount something in a watched tree and it's still mounted
elsewhere, you will get matches on events happening there.  New command
tells audit to recalculate the trees, trimming such sources of false
positives.

Note that it's _not_ about path - if something mounted in several places
(multiple mount, bindings, different namespaces, etc.), the match does
_not_ depend on which one we are using for access.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NDaniel Walker <dwalker@mvista.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Right now the audit filter can match on = != > < >= blah blah blah.
This allow the filter to also look at bitwise AND operations, &
Signed-off-by: NEric Paris <eparis@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The sanity check in audit_match_class() is wrong.  We are able to audit
2048 syscalls but in audit_match_class() we were accidentally using
sizeof(_u32) instead of number of bits in _u32 when deciding how many
syscalls were valid.  On ia64 in particular we were hitting syscall
numbers over the (wrong) limit of 256.  Fixing the audit_match_class
check takes care of the problem.
Signed-off-by: NKlaus Weidner <klaus@atsec.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Kill this warning...

kernel/auditfilter.c: In function ‘audit_receive_filter’:
kernel/auditfilter.c:1213: warning: ‘ndw’ may be used uninitialized in this function
kernel/auditfilter.c:1213: warning: ‘ndp’ may be used uninitialized in this function

...with a simplification of the code.  audit_put_nd() can accept NULL
arguments, just like kfree().  It is cleaner to init two existing vars
to NULL, remove the redundant test variable 'putnd_needed' branches, and call
audit_put_nd() directly.

As a desired side effect, the warning goes away.
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Removing a watched file will oops if audit is disabled (auditctl -e 0).

To reproduce:
- auditctl -e 1
- touch /tmp/foo
- auditctl -w /tmp/foo
- auditctl -e 0
- rm /tmp/foo (or mv)
Signed-off-by: NTony Jones <tonyj@suse.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When auditing syscalls that send signals, log the pid and security
context for each target process. Optimize the data collection by
adding a counter for signal-related rules, and avoiding allocating an
aux struct unless we have more than one target process. For process
groups, collect pid/context data in blocks of 16. Move the
audit_signal_info() hook up in check_kill_permission() so we audit
attempts where permission is denied.
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

I was looking at parsing some of these messages and found that I wanted what
it was doing next to an op= for the parser to key on. Also missing was the list
number and results.
Signed-off-by: NSteve Grubb <sgrubb@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

gcc emits this warning:

 kernel/auditfilter.c: In function 'audit_filter_user':
 kernel/auditfilter.c:1611: warning: 'state' is used uninitialized in this function

I tend to agree with gcc - there are a couple of plausible exit paths from
audit_filter_user_rules() where it does not set 'state', keeping the
variable uninitialized.  For example if a filter rule has an AUDIT_POSSIBLE
action.  Initialize to 'wont audit'.  Fix whitespace damage too.
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

kstrdup() returns NULL on error.

Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Currently the kernel audit system represents arch's as numbers and will
gladly accept comparisons between archs using >, <, >=, <= when the only
thing that makes sense is = or !=.  I'm told that the next revision of
auditctl will do this checking but this will provide enforcement in the
kernel even for old userspace.  A simple command to show the issue would
be to run

auditctl -d entry,always -F arch>i686 -S chmod

with this patch the kernel will reject this with -EINVAL

Please comment/ack/nak as soon as possible.

-Eric

 kernel/auditfilter.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Rename selinux_ctxid_to_string to selinux_sid_to_string to be
consistent with other interfaces.
Signed-off-by: NStephen Smalley <sds@tycho.nsa.gov>
Acked-by: NJames Morris <jmorris@namei.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

add support for AUDIT_PERM predicate
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Make the audit message for implicit rule removal more informative.
Make the rule update message consistent with other messages.
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Hello,

During some troubleshooting, I found that ppid was accidentally omitted from
the legacy rule section. This resulted in EINVAL for any rule with ppid sent
with AUDIT_ADD.
Signed-off-by: NSteve Grubb <sgrubb@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Michael C Thompson wrote:  [Tue Aug 01 2006, 02:36:36PM EDT]
> The trigger for this oops is:
> # auditctl -a exit,always -S pread64 -F 'inode<1'

Setting the err value will fix it.
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Allow to tie upper bits of syscall bitmap in audit rules to kernel-defined
sets of syscalls.  Infrastructure, a couple of classes (with 32bit counterparts
for biarch targets) and actual tie-in on i386, amd64 and ia64.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

This patch introduces object audit filters based on the elements
of the SELinux context.
Signed-off-by: NDarrel Goeddel <dgoeddel@trustedcs.com>
Acked-by: NStephen Smalley <sds@tycho.nsa.gov>

 kernel/auditfilter.c           |   25 +++++++++++++++++++++++++
 kernel/auditsc.c               |   40 ++++++++++++++++++++++++++++++++++++++++
 security/selinux/ss/services.c |   18 +++++++++++++++++-
 3 files changed, 82 insertions(+), 1 deletion(-)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

This patch renames some audit constant definitions and adds
additional definitions used by the following patch.  The renaming
avoids ambiguity with respect to the new definitions.
Signed-off-by: NDarrel Goeddel <dgoeddel@trustedcs.com>

 include/linux/audit.h          |   15 ++++++++----
 kernel/auditfilter.c           |   50 ++++++++++++++++++++---------------------
 kernel/auditsc.c               |   10 ++++----
 security/selinux/ss/services.c |   32 +++++++++++++-------------
 4 files changed, 56 insertions(+), 51 deletions(-)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Add support for a rule key, which can be used to tie audit records to audit
rules.  This is useful when a watched file is accessed through a link or
symlink, as well as for general audit log analysis.

Because this patch uses a string key instead of an integer key, there is a bit
of extra overhead to do the kstrdup() when a rule fires.  However, we're also
allocating memory for the audit record buffer, so it's probably not that
significant.  I went ahead with a string key because it seems more
user-friendly.

Note that the user must ensure that filterkeys are unique.  The kernel only
checks for duplicate rules.
Signed-off-by: NAmy Griffis <amy.griffis@hpd.com>

When an audit event involves changes to a directory entry, include
a PATH record for the directory itself.  A few other notable changes:

    - fixed audit_inode_child() hooks in fsnotify_move()
    - removed unused flags arg from audit_inode()
    - added audit log routines for logging a portion of a string

Here's some sample output.

before patch:
type=SYSCALL msg=audit(1149821605.320:26): arch=40000003 syscall=39 success=yes exit=0 a0=bf8d3c7c a1=1ff a2=804e1b8 a3=bf8d3c7c items=1 ppid=739 pid=800 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=ttyS0 comm="mkdir" exe="/bin/mkdir" subj=root:system_r:unconfined_t:s0-s0:c0.c255
type=CWD msg=audit(1149821605.320:26):  cwd="/root"
type=PATH msg=audit(1149821605.320:26): item=0 name="foo" parent=164068 inode=164010 dev=03:00 mode=040755 ouid=0 ogid=0 rdev=00:00 obj=root:object_r:user_home_t:s0

after patch:
type=SYSCALL msg=audit(1149822032.332:24): arch=40000003 syscall=39 success=yes exit=0 a0=bfdd9c7c a1=1ff a2=804e1b8 a3=bfdd9c7c items=2 ppid=714 pid=777 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=ttyS0 comm="mkdir" exe="/bin/mkdir" subj=root:system_r:unconfined_t:s0-s0:c0.c255
type=CWD msg=audit(1149822032.332:24):  cwd="/root"
type=PATH msg=audit(1149822032.332:24): item=0 name="/root" inode=164068 dev=03:00 mode=040750 ouid=0 ogid=0 rdev=00:00 obj=root:object_r:user_home_dir_t:s0
type=PATH msg=audit(1149822032.332:24): item=1 name="foo" inode=164010 dev=03:00 mode=040755 ouid=0 ogid=0 rdev=00:00 obj=root:object_r:user_home_t:s0
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Clear AUDIT_FILTER_PREPEND flag after adding rule to list.  This
fixes three problems when a rule is added with the -A syntax:

    - auditctl displays filter list as "(null)"
    - the rule cannot be removed using -d
    - a duplicate rule can be added with -a
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

In this implementation, audit registers inotify watches on the parent
directories of paths specified in audit rules.  When audit's inotify
event handler is called, it updates any affected rules based on the
filesystem event.  If the parent directory is renamed, removed, or its
filesystem is unmounted, audit removes all rules referencing that
inotify watch.

To keep things simple, this implementation limits location-based
auditing to the directory entries in an existing directory.  Given
a path-based rule for /foo/bar/passwd, the following table applies:

    passwd modified -- audit event logged
    passwd replaced -- audit event logged, rules list updated
    bar renamed     -- rule removed
    foo renamed     -- untracked, meaning that the rule now applies to
		       the new location

Audit users typically want to have many rules referencing filesystem
objects, which can significantly impact filtering performance.  This
patch also adds an inode-number-based rule hash to mitigate this
situation.

The patch is relative to the audit git tree:
http://kernel.org/git/?p=linux/kernel/git/viro/audit-current.git;a=summary
and uses the inotify kernel API:
http://lkml.org/lkml/2006/6/1/145Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Don't return -ENOMEM when callers of these functions are checking for
a NULL return.  Bug noticed by Serge Hallyn.
Signed-off-by: NAmy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

We should not send a pile of replies while holding audit_netlink_mutex
since we hold the same mutex when we receive commands.  As the result,
we can get blocked while sending and sit there holding the mutex while
auditctl is unable to send the next command and get around to receiving
what we'd sent.

Solution: create skb and put them into a queue instead of sending;
once we are done, send what we've got on the list.  The former can
be done synchronously while we are handling AUDIT_LIST or AUDIT_LIST_RULES;
we are holding audit_netlink_mutex at that point.  The latter is done
asynchronously and without messing with audit_netlink_mutex.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Hi,

The patch below builds upon the patch sent earlier and adds subject label to
all audit events generated via the netlink interface. It also cleans up a few
other minor things.
Signed-off-by: NSteve Grubb <sgrubb@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

This patch provides the ability to filter audit messages based on the
elements of the process' SELinux context (user, role, type, mls sensitivity,
and mls clearance).  It uses the new interfaces from selinux to opaquely
store information related to the selinux context and to filter based on that
information.  It also uses the callback mechanism provided by selinux to
refresh the information when a new policy is loaded.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Semaphore to mutex conversion.

The conversion was generated via scripts, and the result was validated
automatically via a script as well.
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Darrel Goeddel initiated a discussion on IRC regarding the possibility
of audit_comparator() returning -EINVAL signaling an invalid operator.

It is possible when creating the rule to assure that the operator is one
of the 6 sane values.  Here's a snip from include/linux/audit.h  Note
that 0 (nonsense) and 7 (all operators) are not valid values for an
operator.

...

/* These are the supported operators.
 *      4  2  1
 *      =  >  <
 *      -------
 *      0  0  0         0       nonsense
 *      0  0  1         1       <
 *      0  1  0         2       >
 *      0  1  1         3       !=
 *      1  0  0         4       =
 *      1  0  1         5       <=
 *      1  1  0         6       >=
 *      1  1  1         7       all operators
 */
...

Furthermore, prior to adding these extended operators, flagging the
AUDIT_NEGATE bit implied !=, and otherwise == was assumed.

The following code forces the operator to be != if the AUDIT_NEGATE bit
was flipped on.  And if no operator was specified, == is assumed.  The
only invalid condition is if the AUDIT_NEGATE bit is off and all of the
AUDIT_EQUAL, AUDIT_LESS_THAN, and AUDIT_GREATER_THAN bits are
on--clearly a nonsensical operator.

Now that this is handled at rule insertion time, the default -EINVAL
return of audit_comparator() is eliminated such that the function can
only return 1 or 0.

If this is acceptable, let's get this applied to the current tree.

:-Dustin

--
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
(cherry picked from 9bf0a8e137040f87d1b563336d4194e38fb2ba1a commit)

Hi,

The following patch adds a little more information to the add/remove rule message emitted
by the kernel.
Signed-off-by: NSteve Grubb <sgrubb@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Updated patch to dynamically allocate audit rule fields in kernel's
internal representation.  Added unlikely() calls for testing memory
allocation result.

Amy Griffis wrote:     [Wed Jan 11 2006, 02:02:31PM EST]
> Modify audit's kernel-userspace interface to allow the specification
> of string fields in audit rules.
>
> Signed-off-by: Amy Griffis <amy.griffis@hp.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
(cherry picked from 5ffc4a863f92351b720fe3e9c5cd647accff9e03 commit)

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

This fixes the per-user and per-message-type filtering when syscall
auditing isn't enabled.

[AV: folded followup fix from the same author]
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>