Re: [linux-dvb] [PATCH] cx23885 analog TV and audio support for HVR-1500

[Steven Toth <stoth@linuxtv.org>]

Michael Krufky wrote:
> 2008/8/29 Tim Lucas <lucastim@gmail.com>:
>> Mijhail Moreyra wrote:
>>> Steven Toth wrote:
>>>> Mijhail Moreyra wrote:
>>>>> Steven Toth wrote:
>>>>>> Mijhail,
>>>>>>
>>>>>> http://linuxtv.org/hg/~stoth/cx23885-audio
>>>>>>
>>>>>> This tree contains your patch with some minor whitespace cleanups
>>>>>> and fixes for HUNK related merge issues due to the patch wrapping at
>>>>>> 80 cols.
>>>>>>
>>>>>> Please build this tree and retest in your environment to ensure I
>>>>>> did not break anything. Does this tree still work OK for you?
>>>>>>
>>>>>> After this I will apply some other minor cleanups then invite a few
>>>>>> other HVR1500 owners to begin testing.
>>>>>>
>>>>>> Thanks again.
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> Steve
>>>>> Hi, sorry for the delay.
>>>>>
>>>>> I've tested the http://linuxtv.org/hg/~stoth/cx23885-audio tree and
>>>>> it doesn't work well.
>>>>>
>>>>> You seem to have removed a piece from my patch that avoids some register
>>>>> modification in cx25840-core.c:cx23885_
>> initialize()
>>>>> -       cx25840_write(client, 0x2, 0x76);
>>>>> +       if (state->rev != 0x0000) /* FIXME: How to detect the bridge
>>>>> type ??? */
>>>>> +               /* This causes image distortion on a true cx23885
>>>>> board */
>>>>> +               cx25840_write(client, 0x2, 0x76);
>>>>>
>>>>> As the patch says that register write causes a horrible image distortion
>>>>> on my HVR-1500 which has a real cx23885 (not 23887, 23888, etc) board.
>>>>>
>>>>> I don't know if it's really required for any bridge as everything seems
>>>>> to be auto-configured by default, maybe it can be simply dropped.
>>>>>
>>>>> Other than that the cx23885-audio tree works well.
>>>>>
>>>>> WRT the whitespaces, 80 cols, etc; most are also in the sources I took
>>>>> as basis, so I didn't think they were a problem.
>>>> That's a mistake, I'll add that later tonight, thanks for finding
>>>> this. I must of missed it when I had to tear apart your email because
>>>> of HUNK issues caused by patch line wrapping.
>>>>
>>>> Apart from this, is everything working as you expect?
>>>>
>>>> Regards,
>>>>
>>>> Steve
>>>>
>>>>
>>> OK.
>>>
>>> And sorry about the patch, I didn't know it was going to be broken that
>>> way by being sent by email.
>>>
>>>  >> Other than that the cx23885-audio tree works well.
>>>
>>> Great, thanks for confirming.
>>> Regards,
>>> Steve
>> I'll try asking again since my replies in gmail were not including the
>> correct subject heading.
>> Can this code for cx23885 analog support be adapted for the DViCO Fusion
>> HDTV7 Dual Express which also uses the cx23885?  Currently the driver for
>> that card is digital only and I am stuck with a free antiquated large
>> satellite system that is analog only in my apartment. I am willing to put in
>> the work if someone can point me in the right direction.  Thank you,
> 
> Tim,
> 
> The patch currently being tested only enables the analog video path on
> the HVR1500, but this does lay down the ground work to bring up analog
> on all of the other cards.
> 
> If the HVR1500 is working properly, then it will be easy to add analog
> support for the HVR1500Q ... Once that is done, the exact same code
> (for analog) will be reused for the FusionHDTV7 Dual.
> 
> Keep in mind, however, that you will only get ONE analog video device
> on the F7 Dual, and that you must not try to use the first DVB adapter
> while using the analog on that board.  You can always use the 2nd
> adapter , though.
> 
> Actually, I think it might be a good idea to reverse the registration
> order of the DVB adapters in the cx23885 driver, but I'll have to talk
> to Steve about that.
> 
> I think it makes more sense to register VIDC first, since VIDC is
> always going to be a DTV device, where there *might* be an encoder on
> VIDB.
> 
> VIDB may or may not share a tuner with VIDA, but VIDC will always be
> independant.
> 
> What do you think about that, Steve?

I think we should indicate the load order in the cards struct, the 
default should remain as-is, but each specific card could chose to do C, 
then B, or B then C.

I don't think it's safe to assume that VIDC will not share a tuner with 
VID_A.

I'll put some patches together in the cx23885-audio tree when I do some 
testing this weekend, allowing the creation order to be changed.

- Steve



_______________________________________________
linux-dvb mailing list
linux-dvb@linuxtv.org
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb