On Mon, Jun 03, 2019 at 05:18:11PM -0600, David Ahern wrote:
> On 6/3/19 5:05 PM, Wei Wang wrote:
> > On Mon, Jun 3, 2019 at 3:35 PM David Ahern <dsah...@gmail.com> wrote:
> >>
> >> On 6/3/19 3:58 PM, Wei Wang wrote:
> >>> Hmm... I am still a bit concerned with the ip6_create_rt_rcu() call.
> >>> If we have a blackholed nexthop, the lookup code here always tries to
> >>> create an rt cache entry for every lookup.
> >>> Maybe we could reuse the pcpu cache logic for this? So we only create
> >>> new dst cache on the CPU if there is no cache created before.
> >>
> >> I'll take a look.
> >>
>
> BTW, I am only updating ip6_pol_route to use pcpu routes for blackhole
> nexthops.
>
> ip6_pol_route_lookup will continue as is. That function does not use
> pcpu routes and will stay as is.
>
I have concern on calling ip6_create_rt_rcu() in general which seems
to trace back to this commit
dec9b0e295f6 ("net/ipv6: Add rt6_info create function for ip6_pol_route_lookup")
This rt is not tracked in pcpu_rt, rt6_uncached_list or exception bucket.
In particular, how to react to NETDEV_UNREGISTER/DOWN like
the rt6_uncached_list_flush_dev() does and calls dev_put()?
The existing callers seem to do dst_release() immediately without
caching it, but still concerning.
