[linux-dvb] Re: Status of dvb-kernel vs DVB?

[Andrew de Quincey <adq_dvb@lidskialf.net>]

> (Just like the i2c problems with some frontends -- it's likely an 
> saa7146 i2c subsystem issue, we need to find out...)

Interesting, so thats a known problem is it? I've noticed that the
frontends on my Nova-Ts tend not to get detected on first boot... but
work fine after that (both with the original grundig '401 frontend
driver and my modified one). This was not the case with the "DVB"
driver.

I have tried inserting the drivers with all the debug parameters
enabled, but the flaming thing seems to work fine in debug mode! (timing
problem?)

Basically the i2c just doesn't "work"; can't talk to the frontend device
the first time... but then after I remove and re-add the drivers, all is
well. If behaves as if the frontend chip had never been touched... 

I can tell the grundig '401 did not receive any i2c traffic the first
time because of the "magic" reset command sent to the grundig '401 (and
set up its i2c address) as follows:

1st driver insert:
-- Attempt to read from i2c 0x55. FAILS
   => OK, either the device is not there or it is unconfigured 
-- Send reset command (set up i2c address on grundig)
-- Attempt to read from chip. FAILS

2nd driver insert:
-- Attempt to read from i2c 0x55. FAILS
   => OK, either the device is not there or it is unconfigured 
-- Send reset command (set up i2c address on grundig)
-- Attempt to read from chip. SUCCEEDS... 
-- NORMAL SETUP

3rd driver insert (and thereafter):
-- Attempt to read from i2c 0x55. SUCCEEDS
-- NORMAL SETUP

The 1st driver insert obviously does not send the reset command, because
you should only need to send the reset command once (as in insert #2) to
set up the i2c address.

Feel free to suggest code modifications I can try to help debugging.


-- 
Andrew de Quincey <adq_dvb@lidskialf.net>



-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe linux-dvb" as subject.