[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <alpine.LFD.2.11.1604122315350.1611@ja.home.ssi.bg>
Date: Tue, 12 Apr 2016 23:27:55 +0300 (EEST)
From: Julian Anastasov <ja@....bg>
To: Daniele Orlandi <daniele@...andi.com>
cc: netdev@...r.kernel.org
Subject: Re: Issue with ping source address display
Hello,
On Tue, 12 Apr 2016, Daniele Orlandi wrote:
> On 12/04/2016 21:38, Julian Anastasov wrote:
> >
> > What is the kernel version?
>
> Much time passed and I don't remember, however the latest test with IPv6
> ping was made on this host:
>
> root@...itor:~# uname -a
> Linux monitor 4.4.0-17-generic #33-Ubuntu SMP Tue Mar 29 17:17:28 UTC
> 2016 x86_64 x86_64 x86_64 GNU/Linux
I asked because I remember for such problem but
at the same time your first report was from Sep 2014, long
before the bad period: 4.0 - 4.1. Can you find and try
this fix from 4.2?:
commit 34b99df4e6256ddafb663c6de0711dceceddfe0e
Author: Julian Anastasov <ja@....bg>
Date: Tue Jun 23 08:34:39 2015 +0300
ip: report the original address of ICMP messages
ICMP messages can trigger ICMP and local errors. In this case
serr->port is 0 and starting from Linux 4.0 we do not return
the original target address to the error queue readers.
Add function to define which errors provide addr_offset.
With this fix my ping command is not silent anymore.
Because I'm not sure if your kernel includes it.
I also remember that the ping utility can use random memory
when such 4.0/4.1 kernel returns msg_namelen=0:
http://marc.info/?l=linux-netdev&m=143509000420707&w=2
Thread:
http://marc.info/?t=143503781500001&r=1&w=2
Regards
--
Julian Anastasov <ja@....bg>
Powered by blists - more mailing lists