[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87cyqouqfv.fsf@toke.dk>
Date: Wed, 17 Apr 2024 14:25:08 +0200
From: Toke Høiland-Jørgensen <toke@...hat.com>
To: Simon Horman <horms@...nel.org>, Eric Dumazet <edumazet@...gle.com>
Cc: "David S. Miller" <davem@...emloft.net>, Jakub Kicinski
<kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, Jamal Hadi Salim
<jhs@...atatu.com>, Cong Wang <xiyou.wangcong@...il.com>, Jiri Pirko
<jiri@...nulli.us>, netdev@...r.kernel.org, eric.dumazet@...il.com,
cake@...ts.bufferbloat.net
Subject: Re: [PATCH net-next 02/14] net_sched: cake: implement lockless
cake_dump()
Simon Horman <horms@...nel.org> writes:
> + Toke Høiland-Jørgensen <toke@...e.dk>
> cake@...ts.bufferbloat.net
Thanks!
> On Mon, Apr 15, 2024 at 01:20:42PM +0000, Eric Dumazet wrote:
>> Instead of relying on RTNL, cake_dump() can use READ_ONCE()
>> annotations, paired with WRITE_ONCE() ones in cake_change().
>>
>> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
Just to be sure I understand this correctly: the idea is that with
READ/WRITE_ONCE annotations, we can dump the qdisc options without
taking the RTNL lock. This means that a dump not be consistent across a
concurrent reconfig that changes multiple parameters, but each parameter
will be either the new or the old value. Right?
-Toke
Powered by blists - more mailing lists