mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Eric Dumazet <dada1@cosmosbay.com>
To: minyard@acm.org
Cc: Linux Kernel <linux-kernel@vger.kernel.org>,
	netdev@vger.kernel.org, shemminger@vyatta.com,
	paulmck@linux.vnet.ibm.com
Subject: Re: [PATCH 3/3] Convert the UDP hash lock to RCU
Date: Mon, 06 Oct 2008 23:22:31 +0200	[thread overview]
Message-ID: <48EA8197.6080502@cosmosbay.com> (raw)
In-Reply-To: <20081006185026.GA10383@minyard.local>

Corey Minyard a écrit :
> Change the UDP hash lock from an rwlock to RCU.
> 
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> ---
>  include/net/udp.h |    9 +++++----
>  net/ipv4/udp.c    |   47 +++++++++++++++++++++++++++--------------------
>  net/ipv6/udp.c    |   17 +++++++++--------
>  3 files changed, 41 insertions(+), 32 deletions(-)
> 
> diff --git a/include/net/udp.h b/include/net/udp.h
> index addcdc6..35aa104 100644
> --- a/include/net/udp.h
> +++ b/include/net/udp.h
> @@ -51,7 +51,7 @@ struct udp_skb_cb {
>  #define UDP_SKB_CB(__skb)	((struct udp_skb_cb *)((__skb)->cb))
>  
>  extern struct hlist_head udp_hash[UDP_HTABLE_SIZE];
> -extern rwlock_t udp_hash_lock;
> +extern spinlock_t udp_hash_wlock;
>  
>  
>  /* Note: this must match 'valbool' in sock_setsockopt */
> @@ -112,12 +112,13 @@ static inline void udp_lib_hash(struct sock *sk)
>  
>  static inline void udp_lib_unhash(struct sock *sk)
>  {
> -	write_lock_bh(&udp_hash_lock);
> -	if (sk_del_node_init(sk)) {
> +	spin_lock_bh(&udp_hash_wlock);
> +	if (sk_del_node_init_rcu(sk)) {
>  		inet_sk(sk)->num = 0;
>  		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
>  	}
> -	write_unlock_bh(&udp_hash_lock);
> +	spin_unlock_bh(&udp_hash_wlock);
> +	synchronize_rcu();

UDP central rwlock can hurt performance, because of cache line ping pong,
so your patch really makes sense.

Me wondering what impact this synchronize_rcu() can have on mono-threaded
VOIP applications using lot of UDP sockets. What is the maximum delay of
this function ?

For "struct file" freeing, we chose call_rcu() instead of synchronize_rcu()

Maybe we could add a generic rcu head to struct sock, and use call_rcu() in
sk_prot_free() for sockets needing RCU (udp after your patch is applied, maybe
tcp on future patches, while I believe previous work on the subject concluded
RCU was not giving good results for short lived http sessions) ?

Or just add SLAB_DESTROY_BY_RCU to slab creation in proto_register()
for "struct proto udp_prot/udpv6_prot" so that kmem_cache_free() 
done in sk_prot_free() can defer freeing to RCU...





  reply	other threads:[~2008-10-06 21:22 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-10-06 18:50 Corey Minyard
2008-10-06 21:22 ` Eric Dumazet [this message]
2008-10-06 21:40   ` David Miller
2008-10-06 23:08     ` Corey Minyard
2008-10-07  8:37       ` Evgeniy Polyakov
2008-10-07 14:16         ` Christoph Lameter
2008-10-07 14:29           ` Evgeniy Polyakov
2008-10-07 14:38             ` Christoph Lameter
2008-10-07 14:33           ` Paul E. McKenney
2008-10-07 14:45             ` Christoph Lameter
2008-10-07 15:07               ` Eric Dumazet
2008-10-07 15:07               ` Paul E. McKenney
2008-10-07  5:24     ` Eric Dumazet
2008-10-07  8:54       ` Benny Amorsen
2008-10-07 12:59         ` Eric Dumazet
2008-10-07 14:07           ` Stephen Hemminger
2008-10-07 20:55             ` David Miller
2008-10-07 21:20               ` Stephen Hemminger
2008-10-08 13:55               ` Eric Dumazet
2008-10-08 18:45                 ` David Miller
2008-10-07 16:43           ` Corey Minyard
2008-10-07 18:26       ` David Miller
2008-10-08  8:35         ` Eric Dumazet
2008-10-08 16:38           ` David Miller
2008-10-07  8:31     ` Peter Zijlstra
2008-10-07 14:36       ` Paul E. McKenney
2008-10-07 18:29       ` David Miller
2008-10-06 22:07   ` Corey Minyard
2008-10-07  8:17   ` Peter Zijlstra
2008-10-07  9:24     ` Eric Dumazet
2008-10-07 14:15       ` Christoph Lameter
2008-10-07 14:38         ` Paul E. McKenney
2008-10-07 14:50         ` Eric Dumazet
2008-10-07 15:05           ` Paul E. McKenney
2008-10-07 15:09           ` Peter Zijlstra
2008-10-07 15:23           ` Christoph Lameter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=48EA8197.6080502@cosmosbay.com \
    --to=dada1@cosmosbay.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=minyard@acm.org \
    --cc=netdev@vger.kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=shemminger@vyatta.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome