openwrt/staging/blogic.git
14 years agoAppArmor: mediation of non file objects
John Johansen [Thu, 29 Jul 2010 21:48:05 +0000 (14:48 -0700)]
AppArmor: mediation of non file objects

ipc:
AppArmor ipc is currently limited to mediation done by file mediation
and basic ptrace tests.  Improved mediation is a wip.

rlimits:
AppArmor provides basic abilities to set and control rlimits at
a per profile level.  Only resources specified in a profile are controled
or set.  AppArmor rules set the hard limit to a value <= to the current
hard limit (ie. they can not currently raise hard limits), and if
necessary will lower the soft limit to the new hard limit value.

AppArmor does not track resource limits to reset them when a profile
is left so that children processes inherit the limits set by the
parent even if they are not confined by the same profile.

Capabilities:  AppArmor provides a per profile mask of capabilities,
that will further restrict.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: LSM interface, and security module initialization
John Johansen [Thu, 29 Jul 2010 21:48:07 +0000 (14:48 -0700)]
AppArmor: LSM interface, and security module initialization

AppArmor hooks to interface with the LSM, module parameters and module
initialization.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: Enable configuring and building of the AppArmor security module
John Johansen [Thu, 29 Jul 2010 21:48:08 +0000 (14:48 -0700)]
AppArmor: Enable configuring and building of the AppArmor security module

Kconfig and Makefiles to enable configuration and building of AppArmor.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: update Maintainer and Documentation
John Johansen [Thu, 29 Jul 2010 21:48:09 +0000 (14:48 -0700)]
AppArmor: update Maintainer and Documentation

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: functions for domain transitions
John Johansen [Thu, 29 Jul 2010 21:48:06 +0000 (14:48 -0700)]
AppArmor: functions for domain transitions

AppArmor routines for controling domain transitions, which can occur at
exec or through self directed change_profile/change_hat calls.

Unconfined tasks are checked at exec against the profiles in the confining
profile namespace to determine if a profile should be attached to the task.

Confined tasks execs are controlled by the profile which provides rules
determining which execs are allowed and if so which profiles should be
transitioned to.

Self directed domain transitions allow a task to request transition
to a given profile.  If the transition is allowed then the profile will
be applied, either immeditately or at exec time depending on the request.
Immeditate self directed transitions have several security limitations
but have uses in setting up stub transition profiles and other limited
cases.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: file enforcement routines
John Johansen [Thu, 29 Jul 2010 21:48:04 +0000 (14:48 -0700)]
AppArmor: file enforcement routines

AppArmor does files enforcement via pathname matching.  Matching is done
at file open using a dfa match engine.  Permission is against the final
file object not parent directories, ie. the traversal of directories
as part of the file match is implicitly allowed.  In the case of nonexistant
files (creation) permissions are checked against the target file not the
directory.  eg. In case of creating the file /dir/new, permissions are
checked against the match /dir/new not against /dir/.

The permissions for matches are currently stored in the dfa accept table,
but this will change to allow for dfa reuse and also to allow for sharing
of wider accept states.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: userspace interfaces
John Johansen [Thu, 29 Jul 2010 21:48:03 +0000 (14:48 -0700)]
AppArmor: userspace interfaces

The /proc/<pid>/attr/* interface is used for process introspection and
commands.  While the apparmorfs interface is used for global introspection
and loading and removing policy.

The interface currently only contains the files necessary for loading
policy, and will be extended in the future to include sysfs style
single per file introspection inteface.

The old AppArmor 2.4 interface files have been removed into a compatibility
patch, that distros can use to maintain backwards compatibility.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: dfa match engine
John Johansen [Thu, 29 Jul 2010 21:48:01 +0000 (14:48 -0700)]
AppArmor: dfa match engine

A basic dfa matching engine based off the dfa engine in the Dragon
Book.  It uses simple row comb compression with a check field.

This allows AppArmor to do pattern matching in linear time, and also
avoids stack issues that an nfa based engine may have.  The dfa
engine uses a byte based comparison, with all values being valid.
Any potential character encoding are handled user side when the dfa
tables are created.  By convention AppArmor uses \0 to separate two
dependent path matches since \0 is not a valid path character
(this is done in the link permission check).

The dfa tables are generated in user space and are verified at load
time to be internally consistent.

There are several future improvements planned for the dfa engine:
* The dfa engine may be converted to a hybrid nfa-dfa engine, with
  a fixed size limited stack.  This would allow for size time
  tradeoffs, by inserting limited nfa states to help control
  state explosion that can occur with dfas.
* The dfa engine may pickup the ability to do limited dynamic
  variable matching, instead of fixing all variables at policy
  load time.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: contexts used in attaching policy to system objects
John Johansen [Thu, 29 Jul 2010 21:47:59 +0000 (14:47 -0700)]
AppArmor: contexts used in attaching policy to system objects

AppArmor contexts attach profiles and state to tasks, files, etc. when
a direct profile reference is not sufficient.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: basic auditing infrastructure.
John Johansen [Thu, 29 Jul 2010 21:47:58 +0000 (14:47 -0700)]
AppArmor: basic auditing infrastructure.

Update lsm_audit for AppArmor specific data, and add the core routines for
AppArmor uses for auditing.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoAppArmor: misc. base functions and defines
John Johansen [Thu, 29 Jul 2010 21:47:57 +0000 (14:47 -0700)]
AppArmor: misc. base functions and defines

Miscellaneous functions and defines needed by AppArmor, including
the base path resolution routines.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Update version to 2.3.0
Tetsuo Handa [Tue, 27 Jul 2010 08:17:06 +0000 (17:17 +0900)]
TOMOYO: Update version to 2.3.0

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Fix quota check.
Tetsuo Handa [Tue, 27 Jul 2010 01:08:29 +0000 (10:08 +0900)]
TOMOYO: Fix quota check.

Commit d74725b9 "TOMOYO: Use callback for updating entries." broke
tomoyo_domain_quota_is_ok() by counting deleted entries. It needs to
count non-deleted entries.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSELinux: Move execmod to the common perms
Eric Paris [Fri, 23 Jul 2010 15:44:15 +0000 (11:44 -0400)]
SELinux: Move execmod to the common perms

execmod "could" show up on non regular files and non chr files.  The current
implementation would actually make these checks against non-existant bits
since the code assumes the execmod permission is same for all file types.
To make this line up for chr files we had to define execute_no_trans and
entrypoint permissions.  These permissions are unreachable and only existed
to to make FILE__EXECMOD and CHR_FILE__EXECMOD the same.  This patch drops
those needless perms as well.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: place open in the common file perms
Eric Paris [Fri, 23 Jul 2010 15:44:09 +0000 (11:44 -0400)]
selinux: place open in the common file perms

kernel can dynamically remap perms.  Drop the open lookup table and put open
in the common file perms.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSELinux: special dontaudit for access checks
Eric Paris [Fri, 23 Jul 2010 15:44:03 +0000 (11:44 -0400)]
SELinux: special dontaudit for access checks

Currently there are a number of applications (nautilus being the main one) which
calls access() on files in order to determine how they should be displayed.  It
is normal and expected that nautilus will want to see if files are executable
or if they are really read/write-able.  access() should return the real
permission.  SELinux policy checks are done in access() and can result in lots
of AVC denials as policy denies RWX on files which DAC allows.  Currently
SELinux must dontaudit actual attempts to read/write/execute a file in
order to silence these messages (and not flood the logs.)  But dontaudit rules
like that can hide real attacks.  This patch addes a new common file
permission audit_access.  This permission is special in that it is meaningless
and should never show up in an allow rule.  Instead the only place this
permission has meaning is in a dontaudit rule like so:

dontaudit nautilus_t sbin_t:file audit_access

With such a rule if nautilus just checks access() we will still get denied and
thus userspace will still get the correct answer but we will not log the denial.
If nautilus attempted to actually perform one of the forbidden actions
(rather than just querying access(2) about it) we would still log a denial.
This type of dontaudit rule should be used sparingly, as it could be a
method for an attacker to probe the system permissions without detection.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agosecurity: make LSMs explicitly mask off permissions
Eric Paris [Fri, 23 Jul 2010 15:43:57 +0000 (11:43 -0400)]
security: make LSMs explicitly mask off permissions

SELinux needs to pass the MAY_ACCESS flag so it can handle auditting
correctly.  Presently the masking of MAY_* flags is done in the VFS.  In
order to allow LSMs to decide what flags they care about and what flags
they don't just pass them all and the each LSM mask off what they don't
need.  This patch should contain no functional changes to either the VFS or
any LSM.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agovfs: re-introduce MAY_CHDIR
Eric Paris [Fri, 23 Jul 2010 15:43:51 +0000 (11:43 -0400)]
vfs: re-introduce MAY_CHDIR

Currently MAY_ACCESS means that filesystems must check the permissions
right then and not rely on cached results or the results of future
operations on the object.  This can be because of a call to sys_access() or
because of a call to chdir() which needs to check search without relying on
any future operations inside that dir.  I plan to use MAY_ACCESS for other
purposes in the security system, so I split the MAY_ACCESS and the
MAY_CHDIR cases.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSELinux: break ocontext reading into a separate function
Eric Paris [Wed, 21 Jul 2010 16:51:03 +0000 (12:51 -0400)]
SELinux: break ocontext reading into a separate function

Move the reading of ocontext type data out of policydb_read() in a separate
function ocontext_read()

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSELinux: move genfs read to a separate function
Eric Paris [Wed, 21 Jul 2010 16:50:57 +0000 (12:50 -0400)]
SELinux: move genfs read to a separate function

move genfs read functionality out of policydb_read() and into a new
function called genfs_read()

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: fix error codes in symtab_init()
Dan Carpenter [Sat, 12 Jun 2010 18:57:39 +0000 (20:57 +0200)]
selinux: fix error codes in symtab_init()

hashtab_create() only returns NULL on allocation failures to -ENOMEM is
appropriate here.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: fix error codes in cond_read_bool()
Dan Carpenter [Sat, 12 Jun 2010 18:56:01 +0000 (20:56 +0200)]
selinux: fix error codes in cond_read_bool()

The original code always returned -1 (-EPERM) on error.  The new code
returns either -ENOMEM, or -EINVAL or it propagates the error codes from
lower level functions next_entry() or hashtab_insert().

next_entry() returns -EINVAL.
hashtab_insert() returns -EINVAL, -EEXIST, or -ENOMEM.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: fix error codes in cond_policydb_init()
Dan Carpenter [Sat, 12 Jun 2010 18:55:01 +0000 (20:55 +0200)]
selinux: fix error codes in cond_policydb_init()

It's better to propagate the error code from avtab_init() instead of
returning -1 (-EPERM).  It turns out that avtab_init() never fails so
this patch doesn't change how the code runs but it's still a clean up.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: fix error codes in cond_read_node()
Dan Carpenter [Sat, 12 Jun 2010 18:53:46 +0000 (20:53 +0200)]
selinux: fix error codes in cond_read_node()

Originally cond_read_node() returned -1 (-EPERM) on errors which was
incorrect.  Now it either propagates the error codes from lower level
functions next_entry() or cond_read_av_list() or it returns -ENOMEM or
-EINVAL.

next_entry() returns -EINVAL.
cond_read_av_list() returns -EINVAL or -ENOMEM.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: fix error codes in cond_read_av_list()
Dan Carpenter [Sat, 12 Jun 2010 18:52:19 +0000 (20:52 +0200)]
selinux: fix error codes in cond_read_av_list()

After this patch cond_read_av_list() no longer returns -1 for any
errors.  It just propagates error code back from lower levels.  Those can
either be -EINVAL or -ENOMEM.

I also modified cond_insertf() since cond_read_av_list() passes that as a
function pointer to avtab_read_item().  It isn't used anywhere else.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: propagate error codes in cond_read_list()
Dan Carpenter [Sat, 12 Jun 2010 18:51:40 +0000 (20:51 +0200)]
selinux: propagate error codes in cond_read_list()

These are passed back when the security module gets loaded.

The original code always returned -1 (-EPERM) on error but after this
patch it can return -EINVAL, or -ENOMEM or propagate the error code from
cond_read_node().  cond_read_node() still returns -1 all the time, but I
fix that in a later patch.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: cleanup return codes in avtab_read_item()
Dan Carpenter [Sat, 12 Jun 2010 18:50:35 +0000 (20:50 +0200)]
selinux: cleanup return codes in avtab_read_item()

The avtab_read_item() function tends to return -1 as a default error
code which is wrong (-1 means -EPERM).  I modified it to return
appropriate error codes which is -EINVAL or the error code from
next_entry() or insertf().

next_entry() returns -EINVAL.
insertf() is a function pointer to either avtab_insert() or
cond_insertf().
avtab_insert() returns -EINVAL, -ENOMEM, and -EEXIST.
cond_insertf() currently returns -1, but I will fix it in a later patch.

There is code in avtab_read() which translates the -1 returns from
avtab_read_item() to -EINVAL. The translation is no longer needed, so I
removed it.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Stephen D. Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSecurity: capability: code style issue
Chihau Chau [Fri, 16 Jul 2010 16:11:54 +0000 (12:11 -0400)]
Security: capability: code style issue

This fix a little code style issue deleting a space between a function
name and a open parenthesis.

Signed-off-by: Chihau Chau <chihau@gmail.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agosecurityfs: Drop dentry reference count when mknod fails
Tvrtko Ursulin [Thu, 15 Jul 2010 12:25:06 +0000 (13:25 +0100)]
securityfs: Drop dentry reference count when mknod fails

lookup_one_len increments dentry reference count which is not decremented
when the create operation fails. This can cause a kernel BUG at
fs/dcache.c:676 at unmount time. Also error code returned when new_inode()
fails was replaced with more appropriate -ENOMEM.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@sophos.com>
Acked-by: Serge E. Hallyn <serge@hallyn.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: use generic_file_llseek
Arnd Bergmann [Wed, 7 Jul 2010 21:40:10 +0000 (23:40 +0200)]
selinux: use generic_file_llseek

The default for llseek will change to no_llseek,
so selinuxfs needs to add explicit .llseek
assignments. Since we're dealing with regular
files from a VFS perspective, use generic_file_llseek.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: Eric Paris <eparis@parisplace.org>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoima: use generic_file_llseek for securityfs
Arnd Bergmann [Wed, 7 Jul 2010 21:40:15 +0000 (23:40 +0200)]
ima: use generic_file_llseek for securityfs

The default for llseek will change to no_llseek,
so securityfs users need to add explicit .llseek
assignments. Since we're dealing with regular
files from a VFS perspective, use generic_file_llseek.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Explicitly set file_operations->llseek pointer.
Tetsuo Handa [Thu, 8 Jul 2010 12:57:41 +0000 (21:57 +0900)]
TOMOYO: Explicitly set file_operations->llseek pointer.

TOMOYO does not deal offset pointer. Thus seek operation makes
no sense. Changing default seek operation from default_llseek()
to no_llseek() might break some applications. Thus, explicitly
set noop_llseek().

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agosecurity: move LSM xattrnames to xattr.h
Mimi Zohar [Thu, 1 Jul 2010 19:07:43 +0000 (15:07 -0400)]
security: move LSM xattrnames to xattr.h

Make the security extended attributes names global. Updated to move
the remaining Smack xattrs.

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>
14 years agoKEYS: Reinstate lost passing of process keyring ID in call_sbin_request_key()
Justin P. Mattock [Wed, 30 Jun 2010 09:39:11 +0000 (10:39 +0100)]
KEYS: Reinstate lost passing of process keyring ID in call_sbin_request_key()

In commit bb952bb98a7e479262c7eb25d5592545a3af147d there was the accidental
deletion of a statement from call_sbin_request_key() to render the process
keyring ID to a text string so that it can be passed to /sbin/request-key.

With gcc 4.6.0 this causes the following warning:

  CC      security/keys/request_key.o
security/keys/request_key.c: In function 'call_sbin_request_key':
security/keys/request_key.c:102:15: warning: variable 'prkey' set but not used

This patch reinstates that statement.

Without this statement, /sbin/request-key will get some random rubbish from the
stack as that parameter.

Signed-off-by: Justin P. Mattock <justinmattock@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: Use the variable 'key' in keyctl_describe_key()
David Howells [Mon, 28 Jun 2010 13:05:04 +0000 (14:05 +0100)]
KEYS: Use the variable 'key' in keyctl_describe_key()

keyctl_describe_key() turns the key reference it gets into a usable key pointer
and assigns that to a variable called 'key', which it then ignores in favour of
recomputing the key pointer each time it needs it.  Make it use the precomputed
pointer instead.

Without this patch, gcc 4.6 reports that the variable key is set but not used:

building with gcc 4.6 I'm getting a warning message:
 CC      security/keys/keyctl.o
security/keys/keyctl.c: In function 'keyctl_describe_key':
security/keys/keyctl.c:472:14: warning: variable 'key' set but not used

Reported-by: Justin P. Mattock <justinmattock@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Add missing poll() hook.
Tetsuo Handa [Fri, 25 Jun 2010 03:22:09 +0000 (12:22 +0900)]
TOMOYO: Add missing poll() hook.

Commit 1dae08c "TOMOYO: Add interactive enforcing mode." forgot to register
poll() hook. As a result, /usr/sbin/tomoyo-queryd was doing busy loop.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Rename symbols.
Tetsuo Handa [Fri, 25 Jun 2010 02:16:00 +0000 (11:16 +0900)]
TOMOYO: Rename symbols.

Use shorter name in order to make it easier to fit 80 columns limit.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Small cleanup.
Tetsuo Handa [Fri, 25 Jun 2010 00:30:09 +0000 (09:30 +0900)]
TOMOYO: Small cleanup.

Split tomoyo_write_profile() into several functions.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Copy directly to userspace buffer.
Tetsuo Handa [Thu, 24 Jun 2010 05:57:16 +0000 (14:57 +0900)]
TOMOYO: Copy directly to userspace buffer.

When userspace program reads policy from /sys/kernel/security/tomoyo/
interface, TOMOYO uses line buffered mode. A line has at least one word.

Commit 006dacc "TOMOYO: Support longer pathname." changed a word's max length
from 4000 bytes to max kmalloc()able bytes. By that commit, a line's max length
changed from 8192 bytes to more than max kmalloc()able bytes.

Max number of words in a line remains finite. This patch changes the way of
buffering so that all words in a line are firstly directly copied to userspace
buffer as much as possible and are secondly queued for next read request.
Words queued are guaranteed to be valid until /sys/kernel/security/tomoyo/
interface is close()d.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use common code for policy reading.
Tetsuo Handa [Thu, 24 Jun 2010 03:24:19 +0000 (12:24 +0900)]
TOMOYO: Use common code for policy reading.

tomoyo_print_..._acl() are similar. Merge them.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Allow reading only execute permission.
Tetsuo Handa [Thu, 24 Jun 2010 03:00:25 +0000 (12:00 +0900)]
TOMOYO: Allow reading only execute permission.

Policy editor needs to know allow_execute entries in order to build domain
transition tree. Reading all entries is slow. Thus, allow reading only
allow_execute entries.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Change list iterator.
Tetsuo Handa [Thu, 24 Jun 2010 02:28:14 +0000 (11:28 +0900)]
TOMOYO: Change list iterator.

Change list_for_each_cookie to

(1) start from current position rather than next position
(2) remove temporary cursor
(3) check that srcu_read_lock() is held

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use common code for domain transition control.
Tetsuo Handa [Mon, 21 Jun 2010 02:14:39 +0000 (11:14 +0900)]
TOMOYO: Use common code for domain transition control.

Use common code for "initialize_domain"/"no_initialize_domain"/"keep_domain"/
"no_keep_domain" keywords.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Remove alias keyword.
Tetsuo Handa [Mon, 21 Jun 2010 00:58:53 +0000 (09:58 +0900)]
TOMOYO: Remove alias keyword.

Some programs behave differently depending on argv[0] passed to execve().
TOMOYO has "alias" keyword in order to allow administrators to define different
domains if requested pathname passed to execve() is a symlink. But "alias"
keyword is incomplete because this keyword assumes that requested pathname and
argv[0] are identical. Thus, remove "alias" keyword (by this patch) and add
syntax for checking argv[0] (by future patches).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Merge path_group and number_group.
Tetsuo Handa [Thu, 17 Jun 2010 07:55:58 +0000 (16:55 +0900)]
TOMOYO: Merge path_group and number_group.

Use common code for "path_group" and "number_group".

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Aggregate reader functions.
Tetsuo Handa [Thu, 17 Jun 2010 07:54:33 +0000 (16:54 +0900)]
TOMOYO: Aggregate reader functions.

Now lists are accessible via array index. Aggregate reader functions using index.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use array of "struct list_head".
Tetsuo Handa [Thu, 17 Jun 2010 07:53:24 +0000 (16:53 +0900)]
TOMOYO: Use array of "struct list_head".

Assign list id and make the lists as array of "struct list_head".

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Merge tomoyo_path_group and tomoyo_number_group
Tetsuo Handa [Thu, 17 Jun 2010 07:52:29 +0000 (16:52 +0900)]
TOMOYO: Merge tomoyo_path_group and tomoyo_number_group

"struct tomoyo_path_group" and "struct tomoyo_number_group" are identical.
Rename tomoyo_path_group/tomoyo_number_group to tomoyo_group and
tomoyo_path_group_member to tomoyo_path_group and
tomoyo_number_group_member to tomoyo_unmber_group.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: Use current_security() when possible
Paul Moore [Thu, 22 Apr 2010 18:46:19 +0000 (14:46 -0400)]
selinux: Use current_security() when possible

There were a number of places using the following code pattern:

  struct cred *cred = current_cred();
  struct task_security_struct *tsec = cred->security;

... which were simplified to the following:

  struct task_security_struct *tsec = current_security();

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: Convert socket related access controls to use socket labels
Paul Moore [Thu, 22 Apr 2010 18:46:19 +0000 (14:46 -0400)]
selinux: Convert socket related access controls to use socket labels

At present, the socket related access controls use a mix of inode and
socket labels; while there should be no practical difference (they
_should_ always be the same), it makes the code more confusing.  This
patch attempts to convert all of the socket related access control
points (with the exception of some of the inode/fd based controls) to
use the socket's own label.  In the process, I also converted the
socket_has_perm() function to take a 'sock' argument instead of a
'socket' since that was adding a bit more overhead in some cases.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: Shuffle the sk_security_struct alloc and free routines
Paul Moore [Thu, 22 Apr 2010 18:46:18 +0000 (14:46 -0400)]
selinux: Shuffle the sk_security_struct alloc and free routines

The sk_alloc_security() and sk_free_security() functions were only being
called by the selinux_sk_alloc_security() and selinux_sk_free_security()
functions so we just move the guts of the alloc/free routines to the
callers and eliminate a layer of indirection.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: Consolidate sockcreate_sid logic
Paul Moore [Thu, 22 Apr 2010 18:46:18 +0000 (14:46 -0400)]
selinux: Consolidate sockcreate_sid logic

Consolidate the basic sockcreate_sid logic into a single helper function
which allows us to do some cleanups in the related code.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: Set the peer label correctly on connected UNIX domain sockets
Paul Moore [Thu, 22 Apr 2010 18:46:18 +0000 (14:46 -0400)]
selinux: Set the peer label correctly on connected UNIX domain sockets

Correct a problem where we weren't setting the peer label correctly on
the client end of a pair of connected UNIX sockets.

Signed-off-by: Paul Moore <paul.moore@hp.com>
Acked-by: Eric Paris <eparis@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Pass "struct list_head" rather than "void *".
Tetsuo Handa [Wed, 16 Jun 2010 07:31:50 +0000 (16:31 +0900)]
TOMOYO: Pass "struct list_head" rather than "void *".

Pass "struct list_head" to tomoyo_add_to_gc() and bring
list_del_rcu() to tomoyo_add_to_gc().

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Make read function to void.
Tetsuo Handa [Wed, 16 Jun 2010 07:29:59 +0000 (16:29 +0900)]
TOMOYO: Make read function to void.

Read functions do not fail. Make them from int to void.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Merge functions.
Tetsuo Handa [Wed, 16 Jun 2010 07:28:21 +0000 (16:28 +0900)]
TOMOYO: Merge functions.

Embed tomoyo_path_number_perm2() into tomoyo_path_number_perm().

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Remove wrapper function for reading keyword.
Tetsuo Handa [Wed, 16 Jun 2010 07:26:38 +0000 (16:26 +0900)]
TOMOYO: Remove wrapper function for reading keyword.

Keyword strings are read-only. We can directly access them to reduce code size.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Loosen parameter check for mount operation.
Tetsuo Handa [Wed, 16 Jun 2010 07:24:58 +0000 (16:24 +0900)]
TOMOYO: Loosen parameter check for mount operation.

If invalid combination of mount flags are given, it will be rejected later.
Thus, no need for TOMOYO to reject invalid combination of mount flags.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Rename symbols.
Tetsuo Handa [Wed, 16 Jun 2010 07:23:55 +0000 (16:23 +0900)]
TOMOYO: Rename symbols.

Use shorter name in order to make it easier to fix 80 columns limit.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use callback for permission check.
Tetsuo Handa [Wed, 16 Jun 2010 07:22:51 +0000 (16:22 +0900)]
TOMOYO: Use callback for permission check.

We can use callback function since parameters are passed via
"const struct tomoyo_request_info".

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Pass parameters via structure.
Tetsuo Handa [Wed, 16 Jun 2010 07:21:36 +0000 (16:21 +0900)]
TOMOYO: Pass parameters via structure.

To make it possible to use callback function, pass parameters via
"struct tomoyo_request_info".

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use common code for open and mkdir etc.
Tetsuo Handa [Wed, 16 Jun 2010 07:20:24 +0000 (16:20 +0900)]
TOMOYO: Use common code for open and mkdir etc.

tomoyo_file_perm() and tomoyo_path_permission() are similar.
We can embed tomoyo_file_perm() into tomoyo_path_permission().

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoSELinux: seperate range transition rules to a seperate function
Eric Paris [Fri, 11 Jun 2010 16:37:05 +0000 (12:37 -0400)]
SELinux: seperate range transition rules to a seperate function

Move the range transition rule to a separate function, range_read(), rather
than doing it all in policydb_read()

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use common code for garbage collection.
Tetsuo Handa [Tue, 15 Jun 2010 01:10:37 +0000 (10:10 +0900)]
TOMOYO: Use common code for garbage collection.

Use common code for elements using "struct list_head" + "bool" structure.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use callback for updating entries.
Tetsuo Handa [Tue, 15 Jun 2010 00:23:26 +0000 (09:23 +0900)]
TOMOYO: Use callback for updating entries.

Use common code for elements using "struct list_head" + "bool" structure.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use common structure for list element.
Tetsuo Handa [Tue, 15 Jun 2010 00:22:42 +0000 (09:22 +0900)]
TOMOYO: Use common structure for list element.

Use common "struct list_head" + "bool" structure.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use callback for updating entries.
Tetsuo Handa [Sat, 12 Jun 2010 11:46:22 +0000 (20:46 +0900)]
TOMOYO: Use callback for updating entries.

Use common "struct list_head" + "bool" + "u8" structure and
use common code for elements using that structure.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: Make /proc/keys check to see if a key is possessed before security check
David Howells [Fri, 11 Jun 2010 16:31:10 +0000 (17:31 +0100)]
KEYS: Make /proc/keys check to see if a key is possessed before security check

Make /proc/keys check to see if the calling process possesses each key before
performing the security check.  The possession check can be skipped if the key
doesn't have the possessor-view permission bit set.

This causes the keys a process possesses to show up in /proc/keys, even if they
don't have matching user/group/other view permissions.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: Authorise keyctl_set_timeout() on a key if we have its authorisation key
David Howells [Fri, 11 Jun 2010 16:31:05 +0000 (17:31 +0100)]
KEYS: Authorise keyctl_set_timeout() on a key if we have its authorisation key

Authorise a process to perform keyctl_set_timeout() on an uninstantiated key if
that process has the authorisation key for it.

This allows the instantiator to set the timeout on a key it is instantiating -
provided it does it before instantiating the key.

For instance, the test upcall script provided with the keyutils package could
be modified to set the expiry to an hour hence before instantiating the key:

[/usr/share/keyutils/request-key-debug.sh]
 if [ "$3" != "neg" ]
 then
+    keyctl timeout $1 3600
     keyctl instantiate $1 "Debug $3" $4 || exit 1
 else

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Update profile structure.
Tetsuo Handa [Thu, 3 Jun 2010 11:38:44 +0000 (20:38 +0900)]
TOMOYO: Update profile structure.

This patch allows users to change access control mode for per-operation basis.
This feature comes from non LSM version of TOMOYO which is designed for
permitting users to use SELinux and TOMOYO at the same time.

SELinux does not care filename in a directory whereas TOMOYO does. Change of
filename can change how the file is used. For example, renaming index.txt to
.htaccess will change how the file is used. Thus, letting SELinux to enforce
read()/write()/mmap() etc. restriction and letting TOMOYO to enforce rename()
restriction is an example usage of this feature.

What is unfortunate for me is that currently LSM does not allow users to use
SELinux and LSM version of TOMOYO at the same time...

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Add pathname aggregation support.
Tetsuo Handa [Thu, 3 Jun 2010 11:38:03 +0000 (20:38 +0900)]
TOMOYO: Add pathname aggregation support.

This patch allows users to aggregate programs which provide similar
functionality (e.g. /usr/bin/vi and /usr/bin/emacs ).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Allow wildcard for execute permission.
Tetsuo Handa [Thu, 3 Jun 2010 11:37:26 +0000 (20:37 +0900)]
TOMOYO: Allow wildcard for execute permission.

Some applications create and execute programs dynamically. We need to accept
wildcard for execute permission because such programs contain random suffix
in their filenames. This patch loosens up regulation of string parameters.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Support longer pathname.
Tetsuo Handa [Thu, 3 Jun 2010 11:36:43 +0000 (20:36 +0900)]
TOMOYO: Support longer pathname.

Allow pathnames longer than 4000 bytes.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Several fixes for TOMOYO's management programs.
Tetsuo Handa [Thu, 3 Jun 2010 11:35:53 +0000 (20:35 +0900)]
TOMOYO: Several fixes for TOMOYO's management programs.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoLSM: Remove unused arguments from security_path_truncate().
Tetsuo Handa [Wed, 2 Jun 2010 04:24:43 +0000 (13:24 +0900)]
LSM: Remove unused arguments from security_path_truncate().

When commit be6d3e56a6b9b3a4ee44a0685e39e595073c6f0d "introduce new LSM hooks
where vfsmount is available." was proposed, regarding security_path_truncate(),
only "struct file *" argument (which AppArmor wanted to use) was removed.
But length and time_attrs arguments are not used by TOMOYO nor AppArmor.
Thus, let's remove these arguments.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Acked-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agosmack: opt_dentry is never null in in smack_d_instantiate()
Dan Carpenter [Tue, 1 Jun 2010 07:14:04 +0000 (09:14 +0200)]
smack: opt_dentry is never null in in smack_d_instantiate()

This patch removes some unneeded code for if opt_dentry is null because
that can never happen.

The function dereferences "opt_dentry" earlier when it checks
"if (opt_dentry->d_parent == opt_dentry) {".  That code was added in
2008.

This function called from security_d_instantiate().  I checked all the
places which call security_d_instantiate() and dentry is always non-null.
I also checked the selinux version of this hook and there is a comment
which says that dentry should be non-null if called from
d_instantiate().

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Split files into some pieces.
Tetsuo Handa [Mon, 17 May 2010 01:12:46 +0000 (10:12 +0900)]
TOMOYO: Split files into some pieces.

security/tomoyo/common.c became too large to read.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Add interactive enforcing mode.
Tetsuo Handa [Mon, 17 May 2010 01:11:36 +0000 (10:11 +0900)]
TOMOYO: Add interactive enforcing mode.

Since the behavior of the system is restricted by policy, we may need to update
policy when you update packages.

We need to update policy in the following cases.

    * The pathname of files has changed.
    * The dependency of files has changed.
    * The access permissions required has increased.

The ideal way to update policy is to rebuild from the scratch using learning
mode. But it is not desirable to change from enforcing mode to other mode if
the system has once entered in production state. Suppose MAC could support
per-application enforcing mode, the MAC becomes useless if an application that
is not running in enforcing mode was cracked. For example, the whole system
becomes vulnerable if only HTTP server application is running in learning mode
to rebuild policy for the application. So, in TOMOYO Linux, updating policy is
done while the system is running in enforcing mode.

This patch implements "interactive enforcing mode" which allows administrators
to judge whether to accept policy violation in enforcing mode or not.
A demo movie is available at http://www.youtube.com/watch?v=b9q1Jo25LPA .

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Add mount restriction.
Tetsuo Handa [Mon, 17 May 2010 01:10:31 +0000 (10:10 +0900)]
TOMOYO: Add mount restriction.

mount(2) has three string and one numeric parameters.
Split mount restriction code from security/tomoyo/file.c .

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Split file access control functions by type of parameters.
Tetsuo Handa [Mon, 17 May 2010 01:09:15 +0000 (10:09 +0900)]
TOMOYO: Split file access control functions by type of parameters.

Check numeric parameters for operations that deal them
(e.g. chmod/chown/ioctl).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Use structure for passing common arguments.
Tetsuo Handa [Mon, 17 May 2010 01:08:05 +0000 (10:08 +0900)]
TOMOYO: Use structure for passing common arguments.

Use "struct tomoyo_request_info" instead of passing individual arguments.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoTOMOYO: Add numeric values grouping support.
Tetsuo Handa [Mon, 17 May 2010 01:06:58 +0000 (10:06 +0900)]
TOMOYO: Add numeric values grouping support.

This patch adds numeric values grouping support, which is useful for grouping
numeric values such as file's UID, DAC's mode, ioctl()'s cmd number.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoselinux: remove all rcu head initializations
Paul E. McKenney [Tue, 18 May 2010 19:11:25 +0000 (12:11 -0700)]
selinux: remove all rcu head initializations

Remove all rcu head inits. We don't care about the RCU head state before passing
it to call_rcu() anyway. Only leave the "on_stack" variants so debugobjects can
keep track of objects on stack.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Reviewed-by: James Morris <jmorris@namei.org>
Cc: Eric Paris <eparis@parisplace.org>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoLinux 2.6.35
Linus Torvalds [Sun, 1 Aug 2010 22:11:14 +0000 (15:11 -0700)]
Linux 2.6.35

14 years agoNFS: Fix a typo in include/linux/nfs_fs.h
Trond Myklebust [Sun, 1 Aug 2010 17:40:40 +0000 (13:40 -0400)]
NFS: Fix a typo in include/linux/nfs_fs.h

nfs_commit_inode() needs to be defined irrespectively of whether or not
we are supporting NFSv3 and NFSv4.

Allow the compiler to optimise away code in the NFSv2-only case by
converting it into an inlined stub function.

Reported-and-tested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Sat, 31 Jul 2010 02:02:51 +0000 (19:02 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  cyber2000fb: fix console in truecolor modes
  cyber2000fb: fix machine hang on module load
  SA1111: Eliminate use after free
  ARM: Fix Versatile/Realview/VExpress MMC card detection sense
  ARM: 6279/1: highmem: fix SMP preemption bug in kmap_high_l1_vipt
  ARM: Add barriers to io{read,write}{8,16,32} accessors as well
  ARM: 6273/1: Add barriers to the I/O accessors if ARM_DMA_MEM_BUFFERABLE
  ARM: 6272/1: Convert L2x0 to use the IO relaxed operations
  ARM: 6271/1: Introduce *_relaxed() I/O accessors
  ARM: 6275/1: ux500: don't use writeb() in uncompress.h
  ARM: 6270/1: clean files in arch/arm/boot/compressed/
  ARM: Fix csum_partial_copy_from_user()

14 years agoMerge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Sat, 31 Jul 2010 02:02:21 +0000 (19:02 -0700)]
Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
  NFS: Ensure that writepage respects the nonblock flag
  NFS: kswapd must not block in nfs_release_page
  nfs: include space for the NUL in root path

14 years agoMerge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Sat, 31 Jul 2010 02:01:11 +0000 (19:01 -0700)]
Merge branch 'drm-fixes' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/edid: Fix the HDTV hack sync adjustment
  drm/radeon/kms: fix radeon mid power profile reporting

14 years agomm: fix ia64 crash when gcore reads gate area
Hugh Dickins [Fri, 30 Jul 2010 17:58:26 +0000 (10:58 -0700)]
mm: fix ia64 crash when gcore reads gate area

Debian's ia64 autobuilders have been seeing kernel freeze or reboot
when running the gdb testsuite (Debian bug 588574): dannf bisected to
2.6.32 62eede62dafb4a6633eae7ffbeb34c60dba5e7b1 "mm: ZERO_PAGE without
PTE_SPECIAL"; and reproduced it with gdb's gcore on a simple target.

I'd missed updating the gate_vma handling in __get_user_pages(): that
happens to use vm_normal_page() (nowadays failing on the zero page),
yet reported success even when it failed to get a page - boom when
access_process_vm() tried to copy that to its intermediate buffer.

Fix this, resisting cleanups: in particular, leave it for now reporting
success when not asked to get any pages - very probably safe to change,
but let's not risk it without testing exposure.

Why did ia64 crash with 16kB pages, but succeed with 64kB pages?
Because setup_gate() pads each 64kB of its gate area with zero pages.

Reported-by: Andreas Barth <aba@not.so.argh.org>
Bisected-by: dann frazier <dannf@debian.org>
Signed-off-by: Hugh Dickins <hughd@google.com>
Tested-by: dann frazier <dannf@dannf.org>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoCIFS: Remove __exit mark from cifs_exit_dns_resolver()
David Howells [Fri, 30 Jul 2010 14:25:19 +0000 (15:25 +0100)]
CIFS: Remove __exit mark from cifs_exit_dns_resolver()

Remove the __exit mark from cifs_exit_dns_resolver() as it's called by the
module init routine in case of error, and so may have been discarded during
linkage.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agocyber2000fb: fix console in truecolor modes
Ondrej Zary [Thu, 29 Jul 2010 20:40:54 +0000 (22:40 +0200)]
cyber2000fb: fix console in truecolor modes

Return value was not set to 0 in setcolreg() with truecolor modes. This causes
fb_set_cmap() to abort after first color, resulting in blank palette - and
blank console in 24bpp and 32bpp modes.

Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agocyber2000fb: fix machine hang on module load
Ondrej Zary [Thu, 29 Jul 2010 20:32:20 +0000 (22:32 +0200)]
cyber2000fb: fix machine hang on module load

I was testing two CyberPro 2000 based PCI cards on x86 and the machine always
hanged completely when the cyber2000fb module was loaded. It seems that the
card hangs when some registers are accessed too quickly after writing RAMDAC
control register. With this patch, both card work.

Add delay after RAMDAC control register write to prevent hangs on module load.

Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoSA1111: Eliminate use after free
Julia Lawall [Fri, 30 Jul 2010 15:17:28 +0000 (17:17 +0200)]
SA1111: Eliminate use after free

__sa1111_remove always frees its argument, so the subsequent reference to
sachip->saved_state represents a use after free.  __sa1111_remove does not
appear to use the saved_state field, so the patch simply frees it first.

A simplified version of the semantic patch that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@@
expression E,E2;
@@

__sa1111_remove(E)
...
(
  E = E2
|
* E
)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoARM: Fix Versatile/Realview/VExpress MMC card detection sense
Russell King [Thu, 29 Jul 2010 14:58:59 +0000 (15:58 +0100)]
ARM: Fix Versatile/Realview/VExpress MMC card detection sense

The MMC card detection sense has become really confused with negations
at various levels, leading to some platforms not detecting inserted
cards.  Fix this by converting everything to positive logic throughout,
thereby getting rid of these negations.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoARM: 6279/1: highmem: fix SMP preemption bug in kmap_high_l1_vipt
Gary King [Thu, 29 Jul 2010 16:37:20 +0000 (17:37 +0100)]
ARM: 6279/1: highmem: fix SMP preemption bug in kmap_high_l1_vipt

smp_processor_id() must not be called from a preemptible context (this
is checked by CONFIG_DEBUG_PREEMPT).  kmap_high_l1_vipt() was doing so.
This lead to a problem where the wrong per_cpu kmap_high_l1_vipt_depth
could be incremented, causing a BUG_ON(*depth <= 0); in
kunmap_high_l1_vipt().

The solution is to move the call to smp_processor_id() after the call
to preempt_disable().

Originally by: Andrew Howe <ahowe@nvidia.com>

Signed-off-by: Gary King <gking@nvidia.com>
Acked-by: Nicolas Pitre <nico.as.pitre@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoNFS: Ensure that writepage respects the nonblock flag
Trond Myklebust [Fri, 30 Jul 2010 19:31:57 +0000 (15:31 -0400)]
NFS: Ensure that writepage respects the nonblock flag

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
14 years agoNFS: kswapd must not block in nfs_release_page
Trond Myklebust [Fri, 30 Jul 2010 19:31:54 +0000 (15:31 -0400)]
NFS: kswapd must not block in nfs_release_page

See https://bugzilla.kernel.org/show_bug.cgi?id=16056

If other processes are blocked waiting for kswapd to free up some memory so
that they can make progress, then we cannot allow kswapd to block on those
processes.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
14 years agonfs: include space for the NUL in root path
Dan Carpenter [Tue, 13 Jul 2010 11:34:59 +0000 (13:34 +0200)]
nfs: include space for the NUL in root path

In root_nfs_name() it does the following:

        if (strlen(buf) + strlen(cp) > NFS_MAXPATHLEN) {
                printk(KERN_ERR "Root-NFS: Pathname for remote directory too long.\n");
                return -1;
        }
        sprintf(nfs_export_path, buf, cp);

In the original code if (strlen(buf) + strlen(cp) == NFS_MAXPATHLEN)
then the sprintf() would lead to an overflow.  Generally the rest of the
code assumes that the path can have NFS_MAXPATHLEN (1024) characters and
a NUL terminator so the fix is to add space to the nfs_export_path[]
buffer.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
14 years agoMerge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Thu, 29 Jul 2010 22:23:28 +0000 (15:23 -0700)]
Merge branch 'for-linus' of git://git390.marist.edu/linux-2.6

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6:
  [S390] etr: fix clock synchronization race
  [S390] Fix IRQ tracing in case of PER

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Thu, 29 Jul 2010 22:21:44 +0000 (15:21 -0700)]
Merge git://git./linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  watchdog: update MAINTAINERS entry