On Wednesday 28 April 2004 20:00, Allen Briggs wrote:
> On Thu, Apr 29, 2004 at 04:54:25AM +0900, Izumi Tsutsui wrote:
> > Isn't it better to put RU suspend state in the last RFA
> > rather than ABORT command on RNR?
>
> That might be better.  If we do that, we will need to send CU_RESUME
> instead of CU_START when recovering from RNR.
>
> I wish there was a way to ask the chip what state it's in (IDLE,
> ACTIVE, etc.).  I haven't seen one, though.

See FXP_CSR_SCB_RUSCUS and FXP_SCB_RUS_*.  However, there are race conditions 
here...