two problems in dev/e1000/if_lem.c::lem_handle_rxtx()

[John Baldwin]

On Saturday, January 19, 2013 8:19:19 pm Adrian Chadd wrote:
> On 19 January 2013 08:14, John Baldwin <jhb at freebsd.org> wrote:
> 
> > However, I did describe an alternate setup where you can fix this.  Part of
> > the key is to get various NICs to share a single logical queue of tasks.  You
> > could simulate this now by having all the deferred tasks share a single
> > taskqueue with a pool of tasks, but that will still not fully cooperate with
> > ithreads.  To do that you have to get the interrupt handlers themselves into
> > the shared taskqueue.  Some changes I have in a p4 branch allow you to do that
> > by letting interrupt handlers reschedule themselves (avoiding the need for a
> > separate task and preventing the task from running concurrently with the
> > interrupt handler) and providing some (but not yet all) of the framework to
> > allow multiple devices to share a single work queue backed by a shared pool of
> > threads.
> 
> How would that work when I want to pin devices to specific cores?

Note that the setup allows you to bind things however you want.  By default it
uses the current model (each IRQ uses a dedicated queue with a single thread).
The idea is to provide the flexbility so that you can glue things together in
whatever way makes the most sense.  In a router that tends to get into livelock
using a shared queue may make more sense.  However, you are not forced to use that
for other workloads where it does not.

-- 
John Baldwin