netdev - Re: missing icmp redirects

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] [day] [month] [year] [list]

Message-Id: <20140611190425.6f5075756e726b743ed3f53c@qrator.net>
Date:	Wed, 11 Jun 2014 19:04:25 +0400
From:	Dmitry Popov <ixaphire@...tor.net>
To:	Per Jessen <per@...puter.org>
Cc:	"David S. Miller" <davem@...emloft.net>,
	Lukas Tribus <luky-37@...mail.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: missing icmp redirects

On Wed, 11 Jun 2014 09:19:22 +0200
Per Jessen <per@...puter.org> wrote:

> Dmitry Popov wrote:
> > On Tue, 10 Jun 2014 20:10:23 +0200
> > Per Jessen <per@...puter.org> wrote:
> > 
> >> Lukas Tribus wrote:
> >>>> See complete description here:
> >>>>
> >>>> https://bugzilla.kernel.org/show_bug.cgi?id=72561
> >>>>
> >>>> Summary - some ICMP redirects seem to be missing.
> >>>>
> > It's probably not a bug, but intended behavior. See ip_rt_send_redirect in 
> > net/ipv4/route.c. The kernel has a built-in per-source-address limiter for icmp
> > redirects. 
> Any possibility that this limiter may have been inactive/faulty in e.g. 2.6.34? 

Hi!

This limiter worked per-destination-address before kernel v3.6. More precisely,
commit 1d861aa4b3(inet: Minimize use of cached route inetpeer) changed that
behavior. I don't know the reasons behind it (both approaches have pros and 
cons), I think you should ask David if you're curious.

* upd: it's not a limiter for icmp redirects only, it is used for other icmp
errors too. 

> > You can disable it for example with 
> > echo -1 > /proc/sys/net/ipv4/route/redirect_silence
> > Though I am not sure if it is ok in wild networks.
> 
> Okay, that worked on the first attempt.
> Should I also be looking at the other redirect settings?
> 
> /proc/sys/net/ipv4/route/redirect_load
> /proc/sys/net/ipv4/route/redirect_number

redirect_silence is enough, you can look at the algorithm (it's simple enough)
in net/ipv4/route.c:ip_rt_send_redirect if you're curious how it works and what 
those settings mean.
--
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