netdev - Re: [PATCH net-next 2/2] net: sched: protect against stack overflow in TC act

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:   Tue, 25 Jun 2019 10:42:51 +0100
From:   John Hurley <john.hurley@...ronome.com>
To:     Florian Westphal <fw@...len.de>
Cc:     Eyal Birger <eyal.birger@...il.com>,
        Linux Netdev List <netdev@...r.kernel.org>,
        David Miller <davem@...emloft.net>,
        Jamal Hadi Salim <jhs@...atatu.com>,
        Simon Horman <simon.horman@...ronome.com>,
        Jakub Kicinski <jakub.kicinski@...ronome.com>,
        oss-drivers@...ronome.com, shmulik@...anetworks.com
Subject: Re: [PATCH net-next 2/2] net: sched: protect against stack overflow
 in TC act_mirred

On Tue, Jun 25, 2019 at 10:15 AM Florian Westphal <fw@...len.de> wrote:
>
> John Hurley <john.hurley@...ronome.com> wrote:
> > Hi Eyal,
> > The value of 4 is basically a revert to what it was on older kernels
> > when TC had a TTL value in the skb:
> > https://elixir.bootlin.com/linux/v3.19.8/source/include/uapi/linux/pkt_cls.h#L97
>
> IIRC this TTL value was not used ever.

It was used to carry out this looping check on ingress redirects:
https://elixir.bootlin.com/linux/v3.19.8/source/net/core/dev.c#L3468
It appears this was removed/unused after changes in 4.2


>
> > I also found with my testing that a value greater than 4 was sailing
> > close to the edge.
> > With a larger value (on my system anyway), I could still trigger a
> > stack overflow here.
> > I'm not sure on the history of why a value of 4 was selected here but
> > it seems to fall into line with my findings.
> > Is there a hard requirement for >4 recursive calls here?
>
> One alternative would be to (instead of dropping the skb), to
> decrement the ttl and use netif_rx() instead.

Yes, this seems like something worth investigating.
Thanks