Re: Refcount mismatch when unregistering netdevice from kernel

[David Ahern <dsahern@gmail.com>]

On 1/4/21 8:05 PM, stranche@codeaurora.org wrote:
> 
> We're able to reproduce the refcount mismatch after some experimentation
> as well.
> Essentially, it consists of
> 1) adding a default route (ip -6 route add dev XXX default)
> 2) forcing the creation of an exception route via manually injecting an
> ICMPv6
> Packet Too Big into the device.
> 3) Replace the default route (ip -6 route change dev XXX default)
> 4) Delete the device. (ip link del XXX)
> 
> After adding a call to flush out the exception cache for the route, the
> mismatch
> is no longer seen:
> diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
> index 7a0c877..95e4310 100644
> --- a/net/ipv6/ip6_fib.c
> +++ b/net/ipv6/ip6_fib.c
> @@ -1215,6 +1215,7 @@ static int fib6_add_rt2node(struct fib6_node *fn,
> struct fib6_info *rt,
>                 }
>                 nsiblings = iter->fib6_nsiblings;
>                 iter->fib6_node = NULL;
> +               rt6_flush_exceptions(iter);
>                 fib6_purge_rt(iter, fn, info->nl_net);
>                 if (rcu_access_pointer(fn->rr_ptr) == iter)
>                         fn->rr_ptr = NULL;

Ah, I see now. rt6_flush_exceptions is called by fib6_del_route, but
that won't handle replace.

If you look at fib6_purge_rt it already has a call to remove pcpu
entries. This call to flush exceptions should go there and the existing
one in fib6_del_route can be removed.

Also, can you add the reproducer as another test case to
tools/testing/selftests/net/pmtu.sh? We definitely need one for this
sequence (route, exceptions, replace route).