netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* forwarding of ipv4 link local addresses
@ 2015-12-01 21:35 David Ahern
  2015-12-03 19:35 ` David Miller
  0 siblings, 1 reply; 5+ messages in thread
From: David Ahern @ 2015-12-01 21:35 UTC (permalink / raw)
  To: netdev@vger.kernel.org, David Miller

RFC 3927 states that packets from/to IPv4 link-local addresses 
(169.254/16) should not be forwarded, yet the Linux networking stack 
happily forwards them. Before sending in a patch I wanted to inquire if 
this behavior is intentional.

Thanks,
David

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: forwarding of ipv4 link local addresses
  2015-12-01 21:35 forwarding of ipv4 link local addresses David Ahern
@ 2015-12-03 19:35 ` David Miller
  2015-12-09 13:44   ` Thomas Graf
  0 siblings, 1 reply; 5+ messages in thread
From: David Miller @ 2015-12-03 19:35 UTC (permalink / raw)
  To: dsa; +Cc: netdev

From: David Ahern <dsa@cumulusnetworks.com>
Date: Tue, 1 Dec 2015 13:35:58 -0800

> RFC 3927 states that packets from/to IPv4 link-local addresses
> (169.254/16) should not be forwarded, yet the Linux networking stack
> happily forwards them. Before sending in a patch I wanted to inquire
> if this behavior is intentional.

It probably won't break anything if we prohibit this, so sure send
a patch.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: forwarding of ipv4 link local addresses
  2015-12-03 19:35 ` David Miller
@ 2015-12-09 13:44   ` Thomas Graf
  2015-12-09 13:46     ` Thomas Graf
  2015-12-09 19:04     ` David Ahern
  0 siblings, 2 replies; 5+ messages in thread
From: Thomas Graf @ 2015-12-09 13:44 UTC (permalink / raw)
  To: David Miller; +Cc: dsa, netdev

On 12/03/15 at 02:35pm, David Miller wrote:
> From: David Ahern <dsa@cumulusnetworks.com>
> Date: Tue, 1 Dec 2015 13:35:58 -0800
> 
> > RFC 3927 states that packets from/to IPv4 link-local addresses
> > (169.254/16) should not be forwarded, yet the Linux networking stack
> > happily forwards them. Before sending in a patch I wanted to inquire
> > if this behavior is intentional.
> 
> It probably won't break anything if we prohibit this, so sure send
> a patch.

I don't have the full email context so apologies if this is not
relevant. The RFC states that such addresses should not be forwarded
_beyond the local link_. So as long as you are not breaking forwarding
of these addresses on the local host, I'm perfectly fine.

I bring this up specifically because of:

commit d0daebc3d622f95db181601cb0c4a0781f74f758
Author: Thomas Graf <tgraf@suug.ch>
Date:   Tue Jun 12 00:44:01 2012 +0000

    ipv4: Add interface option to enable routing of 127.0.0.0/8
    
    Routing of 127/8 is tradtionally forbidden, we consider
    packets from that address block martian when routing and do
    not process corresponding ARP requests.

    [...]

This feature is being used by a popular PaaS which leverages the
127/8 address space locally without polluting an entire routeable
address space.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: forwarding of ipv4 link local addresses
  2015-12-09 13:44   ` Thomas Graf
@ 2015-12-09 13:46     ` Thomas Graf
  2015-12-09 19:04     ` David Ahern
  1 sibling, 0 replies; 5+ messages in thread
From: Thomas Graf @ 2015-12-09 13:46 UTC (permalink / raw)
  To: David Miller; +Cc: dsa, netdev

On 12/09/15 at 02:44pm, Thomas Graf wrote:
> This feature is being used by a popular PaaS which leverages the
> 127/8 address space locally without polluting an entire routeable
> address space.

Kubernetes also wants to rely on this eventually, so please let's
not break it ;-)

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: forwarding of ipv4 link local addresses
  2015-12-09 13:44   ` Thomas Graf
  2015-12-09 13:46     ` Thomas Graf
@ 2015-12-09 19:04     ` David Ahern
  1 sibling, 0 replies; 5+ messages in thread
From: David Ahern @ 2015-12-09 19:04 UTC (permalink / raw)
  To: Thomas Graf, David Miller; +Cc: netdev

On 12/9/15 6:44 AM, Thomas Graf wrote:
> On 12/03/15 at 02:35pm, David Miller wrote:
>> From: David Ahern <dsa@cumulusnetworks.com>
>> Date: Tue, 1 Dec 2015 13:35:58 -0800
>>
>>> RFC 3927 states that packets from/to IPv4 link-local addresses
>>> (169.254/16) should not be forwarded, yet the Linux networking stack
>>> happily forwards them. Before sending in a patch I wanted to inquire
>>> if this behavior is intentional.
>>
>> It probably won't break anything if we prohibit this, so sure send
>> a patch.
>
> I don't have the full email context so apologies if this is not
> relevant. The RFC states that such addresses should not be forwarded
> _beyond the local link_. So as long as you are not breaking forwarding
> of these addresses on the local host, I'm perfectly fine.

Hi Thomas:

The above is the full email context.

The behavior that one of our testers tripped over is packets sent to 
169.254 addresses received on link A are forwarded out link B. That's 
the behavior that was surprising and seems to violate the RFC.

>
> I bring this up specifically because of:
>
> commit d0daebc3d622f95db181601cb0c4a0781f74f758
> Author: Thomas Graf <tgraf@suug.ch>
> Date:   Tue Jun 12 00:44:01 2012 +0000
>
>      ipv4: Add interface option to enable routing of 127.0.0.0/8
>
>      Routing of 127/8 is tradtionally forbidden, we consider
>      packets from that address block martian when routing and do
>      not process corresponding ARP requests.
>
>      [...]
>
> This feature is being used by a popular PaaS which leverages the
> 127/8 address space locally without polluting an entire routeable
> address space.
>

Daniel pointed out this commit as well. I am referring strictly to 
169.254/16 addresses.

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2015-12-09 19:04 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-12-01 21:35 forwarding of ipv4 link local addresses David Ahern
2015-12-03 19:35 ` David Miller
2015-12-09 13:44   ` Thomas Graf
2015-12-09 13:46     ` Thomas Graf
2015-12-09 19:04     ` David Ahern

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).