Re: [net-next-2.6 PATCH v6 2/7 RFC] TCPCT part 1b: generate Responder Cookie

[William Allen Simpson <william.allen.simpson@gmail.com>]

Eric Dumazet wrote:
> Small point :
> 
> +	if (unlikely(time_after_eq(jiffy, tcp_secret_generating->expires))) {
> +		spin_lock_bh(&tcp_secret_locker);
> +		if (!time_after_eq(jiffy, tcp_secret_generating->expires)) {
> +			/* refreshed by another */
> +			spin_unlock_bh(&tcp_secret_locker);
> +			memcpy(bakery,
> +			       &tcp_secret_generating->secrets[0],
> +			       sizeof(tcp_secret_generating->secrets));
> 
> Technically speaking, you should perform the memcpy() before spin_unlock_bh()
> 
> +	if (unlikely(time_after_eq(jiffy, tcp_secret_generating->expires))) {
> +		spin_lock_bh(&tcp_secret_locker);
> +		if (!time_after_eq(jiffy, tcp_secret_generating->expires)) {
> +			/* refreshed by another */
> +			memcpy(bakery,
> +			       &tcp_secret_generating->secrets[0],
> +			       sizeof(tcp_secret_generating->secrets));
> +			spin_unlock_bh(&tcp_secret_locker);
> 
Technically speaking is exactly what I need!  I'd thought from your
earlier description of memory barriers that it should be after the unlock.
Done.