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: <44b504ca-722c-552c-6813-5fad799941fd@gmail.com>
Date:   Sun, 12 Aug 2018 16:34:26 -0600
From:   David Ahern <dsahern@...il.com>
To:     Damir Mansurov <dnman@...etlabs.ru>, netdev@...r.kernel.org
Cc:     "Alexandra N. Kossovsky" <Alexandra.Kossovsky@...etlabs.ru>
Subject: Re: The recvmsg() with IP_PKTINFO for local addresses returns various
 ipi_ifindex

On 8/9/18 2:13 AM, Damir Mansurov wrote:
> Greetings,
> 
> I use the IP_PKTINFO to detect ipi_ifindex from which the packet was
> arrived, it used to work for local addresses also.
> 
> For local addresses ipi_ifindex always returned 1, but starting from
> Linux 4.14 ip_ifindex began to return various values.
> 
> Example host configuration:
> 1:  lo:   127.0.0.1
> 2:  eth0: 192.168.3.45
> 3:  eth1: 192.168.4.45
> 
> I use sendto() with addresses {127.0.0.1, 192.168.3.45 and
> 192.168.4.45}, call recvmsg() and than use standard procedure to get
> ipi_ifindex, it shows results:
> 
>                  |      ipi_ifindex     |      ipi_ifindex
>  sendto(address) |   Linux ver < 4.14   |   Linux ver >= 4.14
> ---------------------------------------------------------------
>   127.0.0.1      |          1           |           1
> ---------------------------------------------------------------
>   192.168.3.45   |          1           |           2
> ---------------------------------------------------------------
>   192.168.4.45   |          1           |           3
> 
> 
> It seems that this behavior depends from commit:
> net: ipv4: set orig_oif based on fib result for local traffic
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-4.14.y&id=839da4d98960bcc68e6b7b945b33ad3916ec1e92
> 
> 
> I believe that ipi_ifindex for local addresses should be 1.
> Is there a Bug for Linux >= 4.14 or is this a valid behavior?

The index should be the device with the address. It was reporting
loopback prior to the mentioned commit because loopback is used for
packet Tx which is an implementation detail for sending packets locally.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ