Date:	Tue, 7 Apr 2015 09:58:45 -0700
From:	Vivek Venkatraman <vivek@...ulusnetworks.com>
To:	"Eric W. Biederman" <ebiederm@...ssion.com>
Cc:	roopa <roopa@...ulusnetworks.com>,
	Andy Gospodarek <gospo@...ulusnetworks.com>,
	Stephen Hemminger <shemming@...cade.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	Robert Shearman <rshearma@...cade.com>
Subject: Re: [PATCH net-next 6/8] iproute2: Add support for the RTA_VIA attribute

On Tue, Apr 7, 2015 at 9:09 AM, Eric W. Biederman <ebiederm@...ssion.com> wrote:
> roopa <roopa@...ulusnetworks.com> writes:
>
>> On 4/6/15, 4:27 PM, Andy Gospodarek wrote:
>>> On Mon, Apr 06, 2015 at 04:04:06PM -0700, roopa wrote:
>>>> On 3/15/15, 12:52 PM, Eric W. Biederman wrote:
>>>>> Add support for the RTA_VIA attribute that specifies an address family
>>>>> as well as an address for the next hop gateway.
>>>>>
>>>>> To make it easy to pass this reorder inet_prefix so that it's tail
>>>>> is a proper RTA_VIA attribute.
>>>>>
>>>>> Signed-off-by: "Eric W. Biederman" <ebiederm@...ssion.com>
>>>>> ---
>>>>>   include/linux/rtnetlink.h |  7 +++++
>>>>>   include/utils.h           |  7 +++--
>>>>>   ip/iproute.c              | 76 +++++++++++++++++++++++++++++++++++++++++------
>>>>>   man/man8/ip-route.8.in    | 18 +++++++----
>>>>>   4 files changed, 90 insertions(+), 18 deletions(-)
>>>>>
>>>>> diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
>>>>> index 3eb78105399b..03e4c8df8e60 100644
>>>>> --- a/include/linux/rtnetlink.h
>>>>> +++ b/include/linux/rtnetlink.h
>>>>> @@ -303,6 +303,7 @@ enum rtattr_type_t {
>>>>>    RTA_TABLE,
>>>>>    RTA_MARK,
>>>>>    RTA_MFC_STATS,
>>>>> +  RTA_VIA,
>>>> eric, if its not too late, what do you think about renaming RTA_VIA
>>>> attribute to
>>>> RTA_NEWGATEWAY (similar to your new RTA_NEWDST attribute to specify a label
>>>> dst) ?. RTA_VIA is fine too.
>>>> This is indeed a new way to specify a gateway (and can/will be used by RFC
>>>> 5549 in the future).
>>>>
>>>> If there is interest in renaming it to RTA_NEWGATEWAY (or any other name,
>>>> cant think of anything better right now),
>>>> I will be happy to submit a follow-on patch.
>>> FWIW, I actually do not mind the name RTA_VIA.  I was planning to
>>> replace use of RTA_GATEWAY in iproute2 and just usa RTA_VIA for all
>>> nexthops regardless of the address family of the dest route or nexthop
>>> and would allow easy creation of the infrastructure needed to support
>>> RFC5549 -- obviously while keeping backwards compatibility in the
>>> kernel.
>> ok, good to know.
>
> To answer the original question.  The new in RTA_NEWDST is not new as in
> a new attribute it is new as in replace the destination address with a
> new destination address.  NAT in other words.  Which is how mpls routing
> works.  Each hop NATs the address before sending the packet on.
>

At the edge, when doing IPoMPLS, we'll be imposing a set of labels on
top of the packet rather than replacing, but the same semantics can be
applied because the destination address is now different and becomes a
label stack.

One thing to note is that the destination address replaced/imposed
could change based on the path selected, when there is ECMP. So, I
propose that the iproute2 syntax of "as [to]" be reconsidered for
MPLS, otherwise we'll end up with something like the following when
this is extended to setup IPoMPLS direct forwarding with ECMP:

ip route add 147.1.1.0/24 nexthop as to 400/2230 via inet 192.168.1.1
dev eth0 nexthop as to 600/2400 via inet 192.168.2.1 dev eth1

Instead, if we use the specifier "label", we'll get:

ip route add 147.1.1.0/24 nexthop via inet 192.168.1.1 dev eth0 label
400/2230 nexthop via inet 192.168.2.1 dev eth1 label 600/2400

The transit case (label swapping) would look like:

ip -f mpls route add 400 via inet 192.168.1.10 dev eth0 label 500

The syntax can then be better extended to specify a label operation
such as "pop" which would be needed when performing ultimate hop pop
(UHP) and then lookup/forward based on underlying label stack or IP
header.

A new application besides MPLS that needs to modify the destination
address would use its own keyword but encode using the RTA_NEWDST
attribute.

>>> This was what my orignal set did (not submitted to netdev, but discussed
>>> with others at netconf) and it was much cleaner code-wise (but not ideal
>>> as I overloaded the use of RTA_GATEWAY and that was not pleasing to me
>>> or others).
>> ok, yeah i remember you had RTA_GATEWAY6 or something like that.
>>
>> just to clarify, i was not suggesting overloading.
>> eric introduced cleaner abstracted attributes for RTA_DST and RTA_GATEWAY.
>> One is called RTA_NEWDST and I was thinking if changing RTA_GATEWAY to
>> RTA_NEWGATEWAY
>> would be less confusing (because, the rest of the structures
>> (ipv4/ipv6) where you will put the
>> RTA_VIA information is still called gw).
>>
>> No worries, RTA_VIA can stay if more people prefer that.
>
> As long as the number and the semantics don't change I don't much care.
>
> However I think via is probably what we should have called the concept
> and the field in the first place, and certainly there are corner cases
> where the machine where we are going via is not actually a gateway but
> the final destination, when you are talking about multiple protocols.
>
> Regardless the name RTA_VIA is my best attempt in that direction.
>
> All of my added support in iproute2 should work for RFC5549.  As well as
> for mpls.
>
> Eric
--
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

Powered by blists - more mailing lists