Re: [FFADO-user] Cycle timer apparently moving backwards

[Stefan R. <st...@s5...>]

On Mar 27 Jonathan Woithe wrote:
> Where does the OHCI driver obtain the cycle number from?  It's not in the
> packet header quadlet, so I'm guessing that it either listens for the "cycle
> start" packet and bases it on that, or reads the cycle timer on reception of
> each packet.

The OHCI link-layer device takes the cycle number from the last cycle
start packet and writes it into the DMA buffer together with the packet.
(OHCI spec section 10.6)

According to Stanislovas' ffado-test ListDevices logs, the VT6315 is cycle
master.  I wonder whether the VT6315's cycle master part and IR part just
disagree, or whether it generates wrong cycle packets in the first place.
-- 
Stefan Richter
-=====-===-= --== ==-==
http://arcgraph.de/sr/