Hi Werner,

thank you a lot for an explanation. Now I see what's the problem,
but in my opinion your approach is at variance with main problem.
The problem you found shows imperfect of current rx/tx races
implementation (with which we've faced before), so I'd like to
think about global solution instead of applying a large sets of patches
in attempt to cork up all the races holes.

I plan to rework rx/tx mechanism and remove workers according Dmitry
suggestions,
but it will be a little bit later, currently I have no free time :-(

Thank you,
Alexander

2011/7/7 Werner Almesberger <wer...@almesberger.net>

> Alexander Smirnov wrote:
> > Currently I don't clearly imagine a situation, when in BUSY_RX condition
> > we will request for RX_ON.
>
> Sorry, my description may have been a bit fuzzy. This happens after
> at86rf230_state writes STATE_RX_ON to SR_TRX_CMD. It then tries to
> verify that SR_TRX_STATUS has really changed to STATE_RX_ON.
>
> However, if the host is slow enough and someone is sending at just
> the right moment, we may transition -> STATE_TRANSITION_IN_PROGRESS
> -> STATE_RX_ON -> STATE_BUSY_RX before reading SR_TRX_STATUS. And
> then at86rf230_state complains.
>
> Nothing bad happens besides the complaint, because the error code
> trickles up to mac802154_xmit_worker and is simply ignored there.
>
> You can see this happen quite often if your hardware is just slow
> enough :-) (We're working on the USB-to-SPI infrastructure for
> atusb and it's still quite sluggish at the moment. Which is of
> course great for spotting races.)
>
> - Werner
>
------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
Linux-zigbee-devel mailing list
Linux-zigbee-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-zigbee-devel

Reply via email to