Re: [PATCH RFC] ipv6: fix route selection if kernel is not compiled with CONFIG_IPV6_ROUTER_PREF

[Nicolas Dichtel <nicolas.dichtel@6wind.com>]

Le 12/07/2013 10:51, Hannes Frederic Sowa a écrit :
> Hello Nicolas,
>
> On Thu, Jul 11, 2013 at 04:57:42PM +0200, Nicolas Dichtel wrote:
>> Le 11/07/2013 16:46, Hannes Frederic Sowa a écrit :
>>> On Thu, Jul 11, 2013 at 12:24:41PM +0200, Hannes Frederic Sowa wrote:
>>>> I fear, I'll need to do a bit more research.
>>>
>>> My proposal is to take my patch and check for RTF_ADDRCONF plus
>>> RTF_DYNAMIC,
>>> too. The RTF_DYNAMIC check would prevent routes created from icmpv6
>>> redirects
>>> entering an ecmp route set.
>>>
>>> Do you agree?
>> Yes.
>
> There is still some window where things go wrong now, I fear. If we have ecmp
> routes active and we update the pmtu of that rt6_info, we might end up with a
> route in the ecmp set, which might not get recountet if another ecmp route
> joins the set. I will have to think how to deal with this. Do you have an
> idea?
It's possible to add a glue to check this counter when we play with these flags, 
but it's ugly.

Maybe the check against RTF_EXPIRES is fundamentally wrong. Checking 
RTF_ADDRCONF|RTF_DYNAMIC should be enough, what do you think?

In another hand, we can discuss about the initial assumption, that was "only 
static routes are part of ECMP routes". I'm thinking of what are the consequence 
if we accept to accept all routes, without checking any flags.


Regards,
Nicolas