Re: [PATCH net-next v2] ipv6: fix multipath route replace error recovery

[roopa <roopa@cumulusnetworks.com>]

On 9/7/15, 5:03 AM, Nicolas Dichtel wrote:
> The bug you're trying to fix has been introduced by the commit 
> 51ebd3181572
> ("ipv6: add support of equal cost multipath (ECMP)").
> Commit 27596472473a ("ipv6: fix ECMP route replacement") didn't try to 
> fix this
> problem.
The reason I say "27596472473a ("ipv6: fix ECMP route replacement") ", 
has introduced the problem
  i am trying to fix is because of the below change. The part where it 
deletes all siblings of an existing route

diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 96dbfff..bde57b1 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
[..snip]
@@ -835,8 +851,27 @@ add:
info->nl_net->ipv6.rt6_stats->fib_route_nodes++;
                         fn->fn_flags |= RTN_RTINFO;
                 }
+               nsiblings = iter->rt6i_nsiblings;
                 fib6_purge_rt(iter, fn, info->nl_net);
                 rt6_release(iter);
+
+               if (nsiblings) {
+                       /* Replacing an ECMP route, remove all siblings */
+                       ins = &rt->dst.rt6_next;
+                       iter = *ins;
+                       while (iter) {
+                               if (rt6_qualify_for_ecmp(iter)) {
+                                       *ins = iter->dst.rt6_next;
+                                       fib6_purge_rt(iter, fn, 
info->nl_net);
+                                       rt6_release(iter);
+                                       nsiblings--;
+                               } else {
+                                       ins = &iter->dst.rt6_next;
+                               }
+                               iter = *ins;
+                       }
+                       WARN_ON(nsiblings != 0);
+               }
         }

         return 0;

>
> Note that the first patch you submitted to fix this pb (cf
> http://patchwork.ozlabs.org/patch/362296/) was in June 2014, ie one 
> year before
> the commit 27596472473a.
>
yes, i had submitted the patch you mention above to fix a slightly 
different problem that existed then..which
was introduced by "51ebd3181572 ("ipv6: add support of equal cost 
multipath (ECMP)")".
Commit "35f1b4e96b9258a3668872b1139c51e5a23eb876 ipv6: do not delete 
previously existing ECMP routes if add fails"
subsequently fixed it.

Thanks,
Roopa