netdev - Re: [PATCH net] soreuseport: use RCU_INIT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 12 Dec 2018 05:08:45 -0800
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     Xiaozhou Liu <liuxiaozhou@...edance.com>, netdev@...r.kernel.org
Cc:     "David S . Miller" <davem@...emloft.net>
Subject: Re: [PATCH net] soreuseport: use RCU_INIT_POINTER() when NULLing



On 12/12/2018 01:58 AM, Xiaozhou Liu wrote:
> Use RCU_INIT_POINTER() instead of rcu_assign_pointer() to
> NULL out the pointer because it is a bit faster.
> 
> Signed-off-by: Xiaozhou Liu <liuxiaozhou@...edance.com>
> ---
>  net/core/sock_reuseport.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/core/sock_reuseport.c b/net/core/sock_reuseport.c
> index d8fe3e549373..25e71355f0a6 100644
> --- a/net/core/sock_reuseport.c
> +++ b/net/core/sock_reuseport.c
> @@ -205,7 +205,7 @@ void reuseport_detach_sock(struct sock *sk)
>  	if (reuse->reuseport_id)
>  		bpf_sk_reuseport_detach(sk);
>  
> -	rcu_assign_pointer(sk->sk_reuseport_cb, NULL);
> +	RCU_INIT_POINTER(sk->sk_reuseport_cb, NULL);
>  
>  	for (i = 0; i < reuse->num_socks; i++) {
>  		if (reuse->socks[i] == sk) {
> 

It is _not_ faster.

Just different ways to do the exact same thing.

For some reason you missed this two years old commit :


commit 3a37f7275cda5ad25c1fe9be8f20c76c60d175fa
Author: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>
Date:   Sun May 1 18:46:54 2016 -0700

    rcu: No ordering for rcu_assign_pointer() of NULL
    
    This commit does a compile-time check for rcu_assign_pointer() of NULL,
    and uses WRITE_ONCE() rather than smp_store_release() in that case.
    
    Reported-by: Christoph Hellwig <hch@...radead.org>
    Signed-off-by: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>