netdev - Re: [PATCH net] Panic when tc_lookup_action_n finds a partially initialized action.

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

Date:   Mon, 3 Oct 2016 11:22:33 -0700
From:   Cong Wang <xiyou.wangcong@...il.com>
To:     Krister Johansen <kjlx@...pleofstupid.com>
Cc:     Jamal Hadi Salim <jhs@...atatu.com>,
        Linux Kernel Network Developers <netdev@...r.kernel.org>
Subject: Re: [PATCH net] Panic when tc_lookup_action_n finds a partially
 initialized action.

On Sat, Oct 1, 2016 at 8:13 PM, Krister Johansen
<kjlx@...pleofstupid.com> wrote:
> A tc_action_ops structure is visibile as soon as it is placed in the
> act_base list.  When tcf_regsiter_action adds an item to this list and
> drops act_mod_lock, registration is not complete until
> register_pernet_subsys() finishes.

Hmm, good catch, but does the fix have to be so complicated?

How about moving register_pernet_subsys() under act_mod_lock?
Similar is needed for unregister too of course. This also means
we need to convert act_mod_lock to a mutex which allows blocking.
Fortunately, we don't have to take act_mod_lock in any atomic context.

Please try the attached patch. I also convert the read path to RCU
to avoid a possible deadlock. A quick test shows no lockdep splat.

Thanks!

View attachment "act_api.diff" of type "text/plain" (3031 bytes)