jamal wrote:
On Wed, 2004-12-15 at 10:03, Thomas Spatzier wrote:
jamal <hadi@xxxxxxxxxx> wrote on 15.12.2004 14:50:27:
When you netif_stop_queue you should never receive packets anymore
at the device level. If you receive any its a bug and you should drop
them and bitch violently. In other words i think what you have at the
moment is bandaid not the solution.
When we do a netif_stop_queue, we do not get any more packets.
So this behaviour is ok. The problem is that the socket write
queues fill up then and get blocked as soon as they are full.
This is the strange part. Anyone still willing to provide a sample
program that hangs?
I haven't tried this myself, but surely it can happen when the
socket send-buffer is smaller than what can be queued up for
transmission: i.e. in the qdisc queue and device DMA ring.
And even more so, when sending to multiple devices from one socket.
Look at e1000 they do it right there.
On link up:
netif_carrier_on(netdev);
netif_wake_queue(netdev);
On Link Down:
netif_carrier_off(netdev); // wonder if these need swapping
netif_stop_queue(netdev);
Well, this is the same as what started this whole thread.
I believe that stopping the queue on link-down events is simply
bad behavior of the driver.
-Tommy
|