Hello.

SD wrote:
> Why is it difficult to filter IP requests on kernel higher 2.6.30?
> I am quite happy that TOMOYO 1.6.8 can filter network requests.
> Why can't we do the same on kernel after 2.6.30?

TOMOYO filters network requests based on TOMOYO's domainname which the process
calling socket()/bind()/listen()/connect()/sendmsg()/accept()/recvmsg() belongs
to rather than the process called socket() belongs to.

However, LSM hooks for filtering incoming requests upon dequeue time (accept()
and recvmsg()) based on TOMOYO's domainname which receiver process belongs to
are missing.

LSM hooks for filtering _before_ dequeue time (security_socket_accept() and
security_socket_recvmsg()) are provided, but I won't use them because they are
racy and trigger almost-denial-of-service state by spending CPU resource for
accept() and recvmsg() which won't success.
I want to use LSM hooks for filtering _after_ dequeue time
(security_socket_post_accept() and security_socket_post_recvmsg()).

Regarding security_socket_post_accept(), it will be acceptable for mainline
because "snet: Security for NETwork syscalls" is proposing it and no objections
hearing regarding it.

Regarding security_socket_post_recvmsg(), I have to propose it.
TOMOYO 1.x is doing it in a way which is not acceptable for mainline.
Proposing it in a way which is acceptable for mainline is in my to-do list.
But I'm planning to complete file access restrictions before adding network
access restrictions.

Regards.

_______________________________________________
tomoyo-users-en mailing list
[email protected]
http://lists.sourceforge.jp/mailman/listinfo/tomoyo-users-en

Reply via email to