From: Benjamin LaHaise <[EMAIL PROTECTED]> Date: Tue, 7 Mar 2006 14:59:10 -0800
> The patch below merges the use of the wait queue lock and socket spinlock > into one. This gains us ~100-150Mbit/s on netperf, mostly due to the fact > that because we know how the spinlock is used, we can avoid the whole irq > save, disable and reenable sequence since the spinlock only needs bh > protection. As a bonus, this also removes one atomic operation per wakeup. We can't do this, unfortunately. We very much might be hitting the socket wait queue from hardware interrupt context in the future. For that reason I think it's best to keep these seperate for now. And hey, we have your TCP hash table RCU stuff to work on meanwhile. That should keep us busy :-) - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
