linux-toradex.git/security/integrity, branch v2.6.33.2

security: fix error return path in ima_inode_alloc

2010-03-15T16:05:50+00:00

commit baac35c4155a8aa826c70acee6553368ca5243a2 upstream.

If radix_tree_preload is failed in ima_inode_alloc, we don't need
radix_tree_preload_end because kernel is alread preempt enabled

Signed-off-by: Xiaotian Feng 
Signed-off-by: Mimi Zohar 
Signed-off-by: James Morris 
Signed-off-by: Greg Kroah-Hartman

ima: rename PATH_CHECK to FILE_CHECK

2010-02-07T08:06:23+00:00

With the movement of the ima hooks functions were renamed from *path* to
*file* since they always deal with struct file.  This patch renames some of
the ima internal flags to make them consistent with the rest of the code.

Signed-off-by: Mimi Zohar 
Signed-off-by: Eric Paris 
Signed-off-by: Al Viro

ima: rename ima_path_check to ima_file_check

2010-02-07T08:06:22+00:00

ima_path_check actually deals with files!  call it ima_file_check instead.

Signed-off-by: Eric Paris 
Acked-by: Mimi Zohar 
Signed-off-by: Al Viro

ima: initialize ima before inodes can be allocated

2010-02-07T08:06:22+00:00

ima wants to create an inode information struct (iint) when inodes are
allocated.  This means that at least the part of ima which does this
allocation (the allocation is filled with information later) should
before any inodes are created.  To accomplish this we split the ima
initialization routine placing the kmem cache allocator inside a
security_initcall() function.  Since this makes use of radix trees we also
need to make sure that is initialized before security_initcall().

Signed-off-by: Eric Paris 
Acked-by: Mimi Zohar 
Signed-off-by: Al Viro

fix ima breakage

2010-02-07T08:06:22+00:00

The "Untangling ima mess, part 2 with counters" patch messed
up the counters.  Based on conversations with Al Viro, this patch
streamlines ima_path_check() by removing the counter maintaince.
The counters are now updated independently, from measuring the file,
in __dentry_open() and alloc_file() by calling ima_counts_get().
ima_path_check() is called from nfsd and do_filp_open().
It also did not measure all files that should have been measured.
Reason: ima_path_check() got bogus value passed as mask.
[AV: mea culpa]
[AV: add missing nfsd bits]

Signed-off-by: Mimi Zohar 
Signed-off-by: Al Viro

ima: limit imbalance msg

2009-12-16T17:16:48+00:00

Limit the number of imbalance messages to once per filesystem type instead of
once per system boot.  (it's actually slightly racy and could give you a
couple per fs, but this isn't a real issue)

Signed-off-by: Mimi Zohar 
Signed-off-by: Al Viro

Untangling ima mess, part 3: kill dead code in ima

2009-12-16T17:16:47+00:00

Kill the 'update' argument of ima_path_check(), kill
dead code in ima.

Current rules: ima counters are bumped at the same time
when the file switches from put_filp() fodder to fput()
one.  Which happens exactly in two places - alloc_file()
and __dentry_open().  Nothing else needs to do that at
all.

Signed-off-by: Al Viro

ima: call ima_inode_free ima_inode_free

2009-12-16T17:16:46+00:00

ima_inode_free() has some funky #define just to confuse the crap out of me.

void ima_iint_delete(struct inode *inode)

and then things actually call ima_inode_free() and nothing calls
ima_iint_delete().

Signed-off-by: Eric Paris 
Signed-off-by: Al Viro

IMA: clean up the IMA counts updating code

2009-12-16T17:16:46+00:00

We currently have a lot of duplicated code around ima file counts.  Clean
that all up.

Signed-off-by: Eric Paris 
Acked-by: Serge Hallyn 
Signed-off-by: Al Viro

ima: only insert at inode creation time

2009-12-16T17:16:46+00:00

iints are supposed to be allocated when an inode is allocated (during
security_inode_alloc())  But we have code which will attempt to allocate
an iint during measurement calls.  If we couldn't allocate the iint and we
cared, we should have died during security_inode_alloc().  Not make the
code more complex and less efficient.

Signed-off-by: Eric Paris 
Signed-off-by: Al Viro