[rt2x00-users] [PATCH RFC] rt2800usb: TX_STA_FIFO read timer

Johannes Stezenbach js at sig21.net
Tue Jan 11 06:41:42 EST 2011


Hi Ivo,

On Mon, Jan 10, 2011 at 05:40:46PM +0100, Ivo Van Doorn wrote:
> I'm a bit worried about the complexity which is being added.
> So far this patch will start allocating and freeing objects for
> each frame which has been send (in the register_read_async),
> apply timers and all kind of logics to read the registers.

If you have suggestions how to simplify I'm all ears...

> However, I still can't see what we are fixing here, nor do I see
> any improvements in the link quality or my association stresstest.

In the introductory message for the "rt2800usb: read TX_STA_FIFO
asynchronously" patch (Dec 14) I gave some performance numbers:

Performance figures (on 330MHz ARM926EJ-S board):
  current rt2x00/master:
    - RX load (iperf -s): 14.6Mbit/s
    - TX load (iperf -c): 9.5Mbit/s
  with patch:
    - RX load (iperf -s): 16.5Mbit/s
    - TX load (iperf -c): 9.9Mbit/s

You reported that this patch introduced a regression in your
association stresstest, and this is what I'm trying to fix.
But as I wrote in my previous mail I missed to fix the
watchdog issue (actually the code should deal with missing
TX_STA_FIFO entries without needing the watchdog).
I'll try to fix.

> As far as I can see, most problems are currently caused by the
> usage of workqueues itself and the way we flush() before a channel
> switch.

OK, but one step at a time.  I'm trying to finish first what
I started last year.


Johannes



More information about the users mailing list