Re: [PATCH 02/15] security: move LSM xattrnames to xattr.h

From: Casey Schaufler
Date: Sun Jun 27 2010 - 12:54:44 EST


Mimi Zohar wrote:
> On Thu, 2010-06-24 at 20:49 -0700, Casey Schaufler wrote:
>
>> Mimi Zohar wrote:
>>
>>> Make the security extended attributes names global.
>>>
>>> Signed-off-by: Mimi Zohar <zohar@xxxxxxxxxx>
>>> Acked-by: Serge Hallyn <serue@xxxxxxxxxx>
>>> ---
>>> include/linux/capability.h | 3 ---
>>> include/linux/xattr.h | 10 ++++++++++
>>> security/selinux/hooks.c | 3 ---
>>> security/smack/smack.h | 2 --
>>> 4 files changed, 10 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/include/linux/capability.h b/include/linux/capability.h
>>> index 39e5ff5..90012b9 100644
>>> --- a/include/linux/capability.h
>>> +++ b/include/linux/capability.h
>>> @@ -49,9 +49,6 @@ typedef struct __user_cap_data_struct {
>>> } __user *cap_user_data_t;
>>>
>>>
>>> -#define XATTR_CAPS_SUFFIX "capability"
>>> -#define XATTR_NAME_CAPS XATTR_SECURITY_PREFIX XATTR_CAPS_SUFFIX
>>> -
>>> #define VFS_CAP_REVISION_MASK 0xFF000000
>>> #define VFS_CAP_REVISION_SHIFT 24
>>> #define VFS_CAP_FLAGS_MASK ~VFS_CAP_REVISION_MASK
>>> diff --git a/include/linux/xattr.h b/include/linux/xattr.h
>>> index 0cfa1e9..62ca853 100644
>>> --- a/include/linux/xattr.h
>>> +++ b/include/linux/xattr.h
>>> @@ -33,6 +33,16 @@
>>> #define XATTR_USER_PREFIX "user."
>>> #define XATTR_USER_PREFIX_LEN (sizeof (XATTR_USER_PREFIX) - 1)
>>>
>>> +/* Security namespace */
>>> +#define XATTR_SELINUX_SUFFIX "selinux"
>>> +#define XATTR_NAME_SELINUX XATTR_SECURITY_PREFIX XATTR_SELINUX_SUFFIX
>>> +
>>> +#define XATTR_SMACK_SUFFIX "SMACK64"
>>> +#define XATTR_NAME_SMACK XATTR_SECURITY_PREFIX XATTR_SMACK_SUFFIX
>>> +
>>> +#define XATTR_CAPS_SUFFIX "capability"
>>> +#define XATTR_NAME_CAPS XATTR_SECURITY_PREFIX XATTR_CAPS_SUFFIX
>>> +
>>> struct inode;
>>> struct dentry;
>>>
>>> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
>>> index 0f524b7..85338f0 100644
>>> --- a/security/selinux/hooks.c
>>> +++ b/security/selinux/hooks.c
>>> @@ -87,9 +87,6 @@
>>> #include "netlabel.h"
>>> #include "audit.h"
>>>
>>> -#define XATTR_SELINUX_SUFFIX "selinux"
>>> -#define XATTR_NAME_SELINUX XATTR_SECURITY_PREFIX XATTR_SELINUX_SUFFIX
>>> -
>>> #define NUM_SEL_MNT_OPTS 5
>>>
>>> extern int selinux_nlmsg_lookup(u16 sclass, u16 nlmsg_type, u32 *perm);
>>> diff --git a/security/smack/smack.h b/security/smack/smack.h
>>> index c6e9aca..9c773e3 100644
>>> --- a/security/smack/smack.h
>>> +++ b/security/smack/smack.h
>>> @@ -126,10 +126,8 @@ struct smack_known {
>>> /*
>>> * xattr names
>>> */
>>> -#define XATTR_SMACK_SUFFIX "SMACK64"
>>> #define XATTR_SMACK_IPIN "SMACK64IPIN"
>>> #define XATTR_SMACK_IPOUT "SMACK64IPOUT"
>>> -#define XATTR_NAME_SMACK XATTR_SECURITY_PREFIX XATTR_SMACK_SUFFIX
>>> #define XATTR_NAME_SMACKIPIN XATTR_SECURITY_PREFIX XATTR_SMACK_IPIN
>>> #define XATTR_NAME_SMACKIPOUT XATTR_SECURITY_PREFIX XATTR_SMACK_IPOUT
>>>
>>>
>> Why just the SMACK64 attribute name, and not the others? They are
>> manipulated with the same interfaces (well, fsetxattr, fgetxattr)
>> as the SMACK64 attribute. There isn't any conceptual difference and
>> the rationale for moving attribute names really ought to apply to
>> them as well.
>>
>
> Hi Casey,
>
> Moving the other SMACK xattrs is fine, but are they used on persistent
> files and need to be EVM protected?
>

They are not persistent in any configuration available today, but
as they say, tomorrow never knows. The real reason I'm concerned is
that I always dislike having to go look in two places for something
like attr names. Either global or local is fine, but it should be
one or the other.


> Mimi
>
>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/