Re: [PATCH] net: af_unix should disable preempt around sock_prot_inuse_add()

[Eric Dumazet <dada1@cosmosbay.com>]

Alexander Beregalov a écrit :
> 
> BUG: using smp_processor_id() in preemptible [00000000] code: ini
> 
> caller is sock_prot_inuse_add+0x24/0x42
> Pid: 738, comm: init-early.sh Tainted: G        W  2.6.28-rc6-nex
> 
> Call Trace:
>  [<ffffffff8037fb72>] debug_smp_processor_id+0xca/0xe0
>  [<ffffffff8046afc9>] sock_prot_inuse_add+0x24/0x42
>  [<ffffffff804baf2c>] unix_create1+0x161/0x176
>  [<ffffffff804bb03d>] unix_stream_connect+0x94/0x3b0
>  [<ffffffff80468a84>] sys_connect+0x68/0x8e
>  [<ffffffff8020b75b>] system_call_fastpath+0x16/0x1b
> 
> Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
> ---
> 
>  net/unix/af_unix.c |    8 ++++++--
>  1 files changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
> index a9b3c90..f7a1d46 100644
> --- a/net/unix/af_unix.c
> +++ b/net/unix/af_unix.c
> @@ -361,7 +361,9 @@ static void unix_sock_destructor(struct sock *sk)
>  		unix_release_addr(u->addr);
>  
>  	atomic_dec(&unix_nr_socks);
> +	preempt_disable();
>  	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
> +	preempt_enable();
>  #ifdef UNIX_REFCNT_DEBUG
>  	printk(KERN_DEBUG "UNIX %p is destroyed, %d are still alive.\n", sk,
>  		atomic_read(&unix_nr_socks));
> @@ -612,9 +614,11 @@ static struct sock *unix_create1(struct net *net, struct socket *sock)
>  out:
>  	if (sk == NULL)
>  		atomic_dec(&unix_nr_socks);
> -	else
> +	else {
> +		preempt_disable();
>  		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
> -
> +		preempt_enable();
> +	}
>  	return sk;
>  }
>  
> 
> 

Thanks Alexander, but this problem (and others) was fixed few
hours ago by two commits on net-next-2.6

commit 6f756a8c36bf54d0afb1d457082b3e3033d951a7

net: Make sure BHs are disabled in sock_prot_inuse_add()

The rule of calling sock_prot_inuse_add() is that BHs must
be disabled.  Some new calls were added where this was not
true and this tiggers warnings as reported by Ilpo.

Fix this by adding explicit BH disabling around those call sites.

Signed-off-by: David S. Miller <davem@davemloft.net>

commit 920de804bca61f88643bc9171bcd06f1a56c6258

net: Make sure BHs are disabled in sock_prot_inuse_add()

The rule of calling sock_prot_inuse_add() is that BHs must
be disabled.  Some new calls were added where this was not
true and this tiggers warnings as reported by Ilpo.

Fix this by adding explicit BH disabling around those call sites,
or moving sock_prot_inuse_add() call inside an existing BH disabled
section.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>