Re: xfrm_state locking regression...

[David Miller <davem@davemloft.net>]

From: Herbert Xu <herbert@gondor.apana.org.au>
Date: Wed, 3 Sep 2008 15:50:41 +1000

> ipsec: Move state dump list unlinking into GC
> 
> This patch moves the unlinking of x->all (for dumping) into the
> GC so that we know for sure what locks we hold when we unlink the
> entry.
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

After your patch I just applied, things are a bit different now.

Are you going to respin this or is it not relevant any longer?

> diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
> index 89551b6..0244325 100644
> --- a/net/xfrm/xfrm_state.c
> +++ b/net/xfrm/xfrm_state.c
> @@ -380,6 +380,10 @@ static void xfrm_put_mode(struct xfrm_mode *mode)
>  
>  static void xfrm_state_gc_destroy(struct xfrm_state *x)
>  {
> +	spin_lock_bh(&xfrm_state_lock);
> +	list_del(&x->all);
> +	spin_unlock_bh(&xfrm_state_lock);
> +
>  	del_timer_sync(&x->timer);
>  	del_timer_sync(&x->rtimer);
>  	kfree(x->aalg);
> @@ -540,10 +544,6 @@ void __xfrm_state_destroy(struct xfrm_state *x)
>  {
>  	WARN_ON(x->km.state != XFRM_STATE_DEAD);
>  
> -	spin_lock_bh(&xfrm_state_lock);
> -	list_del(&x->all);
> -	spin_unlock_bh(&xfrm_state_lock);
> -
>  	spin_lock_bh(&xfrm_state_gc_lock);
>  	hlist_add_head(&x->bydst, &xfrm_state_gc_list);
>  	spin_unlock_bh(&xfrm_state_gc_lock);
> 
> Thanks,
> -- 
> Visit Openswan at http://www.openswan.org/
> Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt