On Thursday 24 Jun 2004 00:06, Kenneth Aafl�y wrote: > Hi, > > I've done some more probing on my Typhoon CI today, and I have found that: > > Remark: gpio 0 is hardwired to the saa7113 interface. > > CAM detection can be performed by: > - lovering gpio 1-3 > - setting gpio 3 to input > - small delay > - checking if gpio 3 is high > (actually both gpio 2 and 3 go hi) > > Reading the attribute memory of the CAM is done by: > - setting gpio 2-3 low > - setting gpio 1 high > - read data > > This has the strange effect of repeating the CI attribute memory every > 0x200 bytes, any other gpio setting gives zero, so it makes me think that > they might have used the audio_a interface for addressing, as the audio_b > interface is connected to the frontend, but the audio_a interface is shared > between the saa7113 (not present here) and the ci interface. I have yet to > test this theory, but I will soon, what stumps me however is why the heck > would the attribute space be presented when the audio_a interface is not > yet even programmed? > > Hope you get a clear picture of where I stand as of right now, and that you > trow any ideas however crazy they might be, I'll test!!
Nice to see you're having some success! I think there may be something wrong still though; I _think_ I've seen some CAMs which have > 0x200 bytes of attribute memory - it seems to be the fashion to put the PCMCIA Config Option register round about 0x200. Any chance of you getting a digital multimeter and beeping out the GPIO connections on the card? I found that invaluable for the TT one.
