[PATCH v8 19/24] ima: Move to LSM infrastructure

[Casey Schaufler]

On 12/27/2023 11:52 AM, Mimi Zohar wrote:
> On Tue, 2023-12-26 at 12:14 -0800, Casey Schaufler wrote:
>> On 12/26/2023 10:14 AM, Mimi Zohar wrote:
>>> On Thu, 2023-12-14 at 18:08 +0100, Roberto Sassu wrote:
>>>> From: Roberto Sassu <roberto.sassu at huawei.com>
>>>>
>>>> Move hardcoded IMA function calls (not appraisal-specific functions) from
>>>> various places in the kernel to the LSM infrastructure, by introducing a
>>>> new LSM named 'ima' (at the end of the LSM list and always enabled like
>>>> 'integrity').
>>>>
>>>> Having IMA before EVM in the Makefile is sufficient to preserve the
>>>> relative order of the new 'ima' LSM in respect to the upcoming 'evm' LSM,
>>>> and thus the order of IMA and EVM function calls as when they were
>>>> hardcoded.
>>>>
>>>> Make moved functions as static (except ima_post_key_create_or_update(),
>>>> which is not in ima_main.c), and register them as implementation of the
>>>> respective hooks in the new function init_ima_lsm().
>>>>
>>>> A slight difference is that IMA and EVM functions registered for the
>>>> inode_post_setattr, inode_post_removexattr, path_post_mknod,
>>>> inode_post_create_tmpfile, inode_post_set_acl and inode_post_remove_acl
>>>> won't be executed for private inodes. Since those inodes are supposed to be
>>>> fs-internal, they should not be of interest of IMA or EVM. The S_PRIVATE
>>>> flag is used for anonymous inodes, hugetlbfs, reiserfs xattrs, XFS scrub
>>>> and kernel-internal tmpfs files.
>>>>
>>>> Conditionally register ima_post_path_mknod() if CONFIG_SECURITY_PATH is
>>>> enabled, otherwise the path_post_mknod hook won't be available.
>>> Up to this point, enabling CONFIG_SECURITY_PATH was not required.  By
>>> making it conditional on CONFIG_SECURITY_PATH, anyone enabling IMA will
>>> also need to enable CONFIG_SECURITY_PATH.  Without it, new files will
>>> not be tagged as a "new" file.
>>>
>>> Casey, Paul, how common is it today not to enable CONFIG_SECURITY_PATH?
>>> Will enabling it just for IMA be a problem?
>> Landlock, AppArmor and TOMOYO require it. Fedora enables Landlock and Ubuntu
>> enables AppArmor. I expect that, except for "minimal" distributions, you
>> won't get any push back. If a distribution is striving for minimal, it's not
>> going to use IMA.
>>
>> It makes me wonder if eliminating CONFIG_SECURITY_PATH might not be a
>> rational alternative.
> Embedded systems were the first to use IMA for file signature
> verification, not distros.               Could they have enabled
> SELinux, lockdown, and IMA?

Yes, they could have. I know some have used Smack and some SELinux.
That's not really relevant, as neither of those use path hooks. My
thought is that CONFIG_SECURITY_PATH adds more aggravation than value,
but I can't quote numbers on either. I don't see a problem with IMA
using path hooks. I also wouldn't see harm in moving the hook(s) you
need for IMA out from that configuration option and into the general
set. With the current rate of new hook additions I can't see moving
an existing hook as a problem.