On Mon, Mar 23, 2009 at 01:37:49AM -0700, David Miller wrote:

From: Eric Dumazet <redacted>
Date: Mon, 23 Mar 2009 09:32:39 +0100

quoted

I dont understand, doesnt it defeat the ticket spinlock thing and fairness ?

Thread doing __qdisc_run() already owns the __QDISC_STATE_RUNNING bit.

Right.

Remember, the way this is designed is that if there is a busy
cpu taking packets out of the queue and putting them into the
device then other cpus will simply add to the queue and immediately
return.

But this "busy cpu" can't take packets out of the queue when it's
waiting on the contended spinlock. Anyway, it's only for testing,
and I didn't say it has to be right.

Jarek P.