CityK wrote:
maillist wrote:
I think I was going down the wrong path on the IR Remote. What I
originally noticed is that the driver in windows is polling the GPIO
register. I figured that must be how it reads the bits coming in. In
the saa7134 code in the saa7134-ir.c (or was it saa7134-i2c.c) you see
that is supports polling. I went down that route but it wasn't making
any sense. So I went back and studied the logs a bit harder and now I
see that it sometimes reads from a device through the I2C port. I
don't have my notes with me at work so I can't tell you the address,
but I see it reads from an I2C device and gets a series of 0xFF data
followed by 0x10 and then 0xFF again. This didn't at first seem
helpful because I was pressing two different keys, but later in the
log I found an instance of 0xFF, 0x11, 0xFF. So maybe it does mean
something.
It is quite possible that the unknown part on the board, the KS007
part is a microcontroller with code that handles the IR. I am going to
try to capture more tonight with my bus analyzer and push more buttons
on the remote than two.
I was thinking about this earlier on and then remembered Henry Wong's
recent efforts with the MSI TV@nywhere. See:
- this thread:
https://www.redhat.com/mailman/private/video4linux-list/2005-November/msg00756.html
... note that there is one post (about the Pinicle 110) that breaks the
thread in the middle, but just click next in thread and it will pick up
again for 4 or 5 more posts
- this post:
https://www.redhat.com/mailman/private/video4linux-list/2005-December/msg00435.html
- and this post:
https://www.redhat.com/mailman/private/video4linux-list/2006-January/msg00280.html
This previous post almost exactly describes the problems I am now having.
It looks like the KS007 device is the IR decoder. Probably a generic
micro programmed to decode the KWorld brand of remote control. This is
hooked up to GPIO as 0x100 to generate a low pulse when you press a key.
Then you have to read back from the i2c device to see what the code
result is.
This IR decoding device is address 0xC2 (or 0x61 on the i2c bus).
The problem I have is two fold.
1. I now know why the tuner (tda9887) is not first setup correctly. The
i2c probing initiated by the tuner module finds something at 61 and
things this is the tda9887 but it isn't. This is the IR decoder. The
tuner is at 0x43. It seems that the tda9887 probe routine needs to read
something from the i2c device to correctly identify it as such device.
2. The way that the tuner module probes it will get a response from the
IR decoder but the ir-kbd module just tries to read 1 byte and it
doesn't detect anything from the IR decoder. The device seems to require
a valid read transaction to respond.
I never saw a resolution to the "MSI TV Anywhere" thread referenced
above. Can anyone comment on what happened to that problem and how I
might be able to solve a similar problem with this Kworld card?
Curt
_______________________________________________
linux-dvb@xxxxxxxxxxx
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
