Hi, > it works on boards with hardware i2c enabled. Changing the I2C_M_NOSTART > to 0, as you did, is the right thing to do here.
Ahh, OK. > Attached is a patch against CVS that makes software I2C work, as well as > implementing the second policy. I think that's a good way to handle the parameters. > requests. The one change that isn't is a deviation from the recommended > settings which should only make a difference if a crystal oscillator in > your tuner is out of spec (the buf[7] = 0x06 instead of 0x05 slightly The value 0x06 is used by the windows-driver, but I think 0x05 is correct. Maybe a rounding error. Wolfram