netdev - c/r of seccomp filters via underlying eBPF

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Fri,  4 Sep 2015 10:04:18 -0600
From:	Tycho Andersen <tycho.andersen@...onical.com>
To:	Kees Cook <keescook@...omium.org>,
	Alexei Starovoitov <ast@...nel.org>
Cc:	Will Drewry <wad@...omium.org>, Oleg Nesterov <oleg@...hat.com>,
	Andy Lutomirski <luto@...capital.net>,
	Pavel Emelyanov <xemul@...allels.com>,
	"Serge E. Hallyn" <serge.hallyn@...ntu.com>,
	Daniel Borkmann <daniel@...earbox.net>,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org
Subject: c/r of seccomp filters via underlying eBPF

Hi all,

Here is a set that enables checkpoint restore of the underlying eBPF programs
that power seccomp filters via the API we discussed several months ago. A few
notes:

* We expose prog_id in the ebpf dump as the pointer to the ebpf program in
  kernel memory, since this is unique. I'm not sure if this is safe. The goal
  of exposing prog_id is to enable userspace to figure out how the seccomp
  filters are inherited (i.e. is the filter a result of a fork() or an
  identical filter installed). This is relevant because
  SECCOMP_FILTER_FLAG_TSYNC depends on this ancestry.

* I'm not sure what the fd argument to seccomp() should look like in the
  SECCOMP_FILTER_FLAG_EBPF case: it seems ugly either as &fd or fd; any
  thoughts on this are welcome.

* Please double check the first patch. I'm not too framiliar with eBPF, so this
  is mostly monkey see monkey do.

* The last patch is something I discovered while testing this set (I couldn't
  re-import some filters). I'm not sure if this is indicitave of a wider bug or
  if the fix is even correct but it Works For Me.

Thoughts welcome,

Tycho

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html