[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <56FDBA6E.5030508@stressinduktion.org>
Date: Fri, 1 Apr 2016 02:01:50 +0200
From: Hannes Frederic Sowa <hannes@...essinduktion.org>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: davem@...emloft.net, netdev@...r.kernel.org,
sasha.levin@...cle.com, daniel@...earbox.net,
alexei.starovoitov@...il.com, mkubecek@...e.cz
Subject: Re: [PATCH net 4/4] tcp: various missing rcu_read_lock around
__sk_dst_get
Hi Eric,
On 01.04.2016 01:39, Eric Dumazet wrote:
> On Fri, 2016-04-01 at 01:29 +0200, Hannes Frederic Sowa wrote:
>> Various fixes which were discovered by this changeset. More probably
>> to come...
>
>
> Really ?
>
> Again, TCP stack locks the socket most of the time.
>
> The fact that lockdep does not understand this is not a reason to add
> this overhead.
I don't see how lockdep does not understand this? I think I added the
appropriate helper to exactly verify if we have the socket lock with
lockdep, please have a look at lockdep_sock_is_held in patch #2.
Some of the patches also just reorder the rcu_read_lock, which is anyway
mostly free.
Bye,
Hannes
Powered by blists - more mailing lists