Hi,
Tim Small wrote:
> Jos? Luis Bolos wrote:
>
>> I was the one having problems with the 300i, and I manage to get it
>> running with 2.6.12.5 and 20050627, but I'm having the same problems
>> as you, there's extreme blockyness and artifacts in audio and video. I
>> was starting to think than that may be reception problems, but maybe
>> it isn't. The card is running ok in analog mode. I have tried to get
>> the card running in windows with pinnacle's software (which is utterly
>> cr**, btw) but it refuses to store the detected digital channels.
>>
>> Are they (the artifacts) a known bug? is there a way to get the card
>> running correctly? Any help is appreciated.
>>
>>
>
> If I tune the card under 2.6.12/20050627, and then reboot into 2.6.14.1
> + current CVS, the output looks fine, and works for about 10 seconds, so
> it looks like the artefacts are fixed in the current CVS code. It then
> stutters, and stops. I get this sort of thing in the kernel message log
> (saa7134 i2c_debug=2):
>
Hm, there was a bug in the TS DMA code of the SAA7134. I found and fixed it
after the driver update in 2.6.13. This can explain the artefacts. The bug
was not visible on all PCs.
> kernel: start xfer
> kernel: send address
> kernel: saa7134[0]: i2c data => 0x3c
> kernel: saa7134[0]: i2c stat <= BUSY
> last message repeated 4 times
> kernel: saa7134[0]: i2c stat <= DONE_WRITE
> kernel: saa7134[0]: i2c stat <= DONE_WRITE
> kernel: write bytes
> kernel: saa7134[0]: i2c data => 0x1
> kernel: saa7134[0]: i2c stat <= BUSY
> last message repeated 6 times
> kernel: saa7134[0]: i2c stat <= ARB_LOST
> kernel: saa7134[0]: i2c stat <= ARB_LOST
> kernel: mt352_read_register: readreg error (reg=1, ret==-5)
> kernel: saa7134[0]: i2c stat <= ARB_LOST
> kernel: saa7134[0]: i2c reset
> kernel: saa7134[0]: i2c stat <= ARB_LOST
> kernel: saa7134[0]: i2c stat => ARB_LOST
> kernel: saa7134[0]: i2c stat <= ARB_LOST
> kernel: saa7134[0]: i2c stat <= IDLE
> kernel: saa7134[0]: i2c attr => NOP
> kernel: start xfer
> kernel: send address
> kernel: saa7134[0]: i2c data => 0x3c
> kernel: saa7134[0]: i2c stat <= BUSY
> last message repeated 15 times
> kernel: mt352_read_register: readreg error (reg=0, ret==-5)
> kernel: saa7134[0]: i2c stat <= TO_ARB
>
>
> Which again looks i2c problems with the tuner, I think. Anyone got any
> ideas? Anything at all? I don't really know much about i2c (apart from
> the fact that it is a low speed serial bus, with 2 wires)... If not,
> then I'd guess the next job is to get the MT352, and SAA7134 data sheets
> and/or errata?
>
I know the I2C bus and have saa7134 documentation. But i have no experience
with the MT352. The errata sheet does not mention a I2C bug. But the lost
arbitration is interesting. Assuming there is only one master on the bus,
this indicates that there either is wild oscillation or a protocol problem
with the MT352 resp. the card. Did somebody try to slow down the bus?
This often helps if i.e the pull up resistor values are too high.
Best regards
Hartmut
