netdev - Re: [PATCH net 1/2] ipv6: Dump route exceptions too in rt6_dump

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:   Fri, 7 Jun 2019 03:54:16 +0200
From:   Stefano Brivio <sbrivio@...hat.com>
To:     David Ahern <dsahern@...il.com>
Cc:     Martin Lau <kafai@...com>, David Miller <davem@...emloft.net>,
        Jianlin Shi <jishi@...hat.com>, Wei Wang <weiwan@...gle.com>,
        Eric Dumazet <edumazet@...gle.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH net 1/2] ipv6: Dump route exceptions too in
 rt6_dump_route()

On Thu, 6 Jun 2019 16:48:51 -0600
David Ahern <dsahern@...il.com> wrote:

> On 6/6/19 4:37 PM, Martin Lau wrote:
> >> I don't think that can happen in practice, or at least I haven't found a
> >> way to create enough valid exceptions for the same node.  
> > That I am not sure.  It is not unusual to have many pmtu exceptions in
> > a gateway node.
> >   
> 
> yes.
> 
> Stefano: you could generalize this test script
>    http://patchwork.ozlabs.org/patch/1110802/
> to have N-remote hosts

Right, thanks for the pointer. I ended up doing something like that in
pmtu.sh, and it turns out that, starting from 25 exceptions in the same
node, iproute2 doesn't actually retry with a larger buffer. As Martin
predicted (thanks!) the dump doesn't terminate.

I tested a version that counts the number of routes in a partial dump
and skips them on the next one with 10,000 entries, dump terminates and
entries count is consistent (at some point, the buckets are just full,
and number of entries doesn't increase any longer).

Unfortunately, the setup of the test takes a few minutes, so I wouldn't
include it (at least as it is) in the selftest.

I'll post that as v2 soon.

-- 
Stefano