Re: [PATCH v2 net-next 00/15] ipv6: Drop RTNL from mcast.c and anycast.c

On 6/27/25 2:49 AM, Kuniyuki Iwashima wrote:

On Thu, Jun 26, 2025 at 6:27 AM Paolo Abeni [off-list ref] wrote:

quoted

On 6/24/25 10:24 PM, Kuniyuki Iwashima wrote:

quoted

From: Kuniyuki Iwashima <kuniyu@google.com>

This is a prep series for RCU conversion of RTM_NEWNEIGH, which needs
RTNL during neigh_table.{pconstructor,pdestructor}() touching IPv6
multicast code.

Currently, IPv6 multicast code is protected by lock_sock() and
inet6_dev->mc_lock, and RTNL is not actually needed.

In addition, anycast code is also in the same situation and does not
need RTNL at all.

This series removes RTNL from net/ipv6/{mcast.c,anycast.c} and finally
removes setsockopt_needs_rtnl() from do_ipv6_setsockopt().

I went through the whole series I could not find any obvious bug.

Still this is not trivial matter and I recently missed bugs in similar
changes, so let me keep the series in PW for a little longer, just in
case some other pair of eyes would go over it ;)

Thank you Paolo!

quoted

BTW @Kuniyuki: do you have a somewhat public todo list that others could
peek at to join this effort?

I  don't have a public one now, but I can create a public repo on GitHub
and fill the Issues tab as the todo list.  Do you have any ideas ?

Not really, that is way I asked ;) Hopefully someone ~here could help.

Quickly skimming over the codebase I suspect/hope mroute{4,6} should be
doable (to be converted to own lock instead of rtnl).

/P