[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f4bcf5fd-b1b0-47a8-9eb3-5aae2c5171b7@kernel.org>
Date: Wed, 6 Mar 2024 16:38:29 -0700
From: David Ahern <dsahern@...nel.org>
To: Eric Dumazet <edumazet@...gle.com>, "David S . Miller"
<davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>
Cc: netdev@...r.kernel.org, eric.dumazet@...il.com
Subject: Re: [PATCH net-next 1/4] ipv6: make inet6_fill_ifaddr() lockless
On 3/6/24 8:51 AM, Eric Dumazet wrote:
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index 2f84e6ecf19f48602cadb47bc378c9b5a1cdbf65..480a1f9492b590bb13008cede5ea7dc9c422af67 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -2730,7 +2730,7 @@ int addrconf_prefix_rcv_add_addr(struct net *net, struct net_device *dev,
> if (update_lft) {
> ifp->valid_lft = valid_lft;
> ifp->prefered_lft = prefered_lft;
> - ifp->tstamp = now;
> + WRITE_ONCE(ifp->tstamp, now);
There are a lot of instances of ifp->tstamp not annotated with READ_ONCE
or WRITE_ONCE. Most of them before this function seem to be updated or
read under rtnl. What's the general mode of operation for this patch?
e.g., there are tstamp references just above this one in this function
not modified. Commit message does not describe why some are updated and
others not.
Powered by blists - more mailing lists