[linux-dvb] S2API - Status - Thu Sep 11th

Steven Toth stoth at linuxtv.org
Mon Sep 15 05:17:02 CEST 2008

Christophe Thommeret wrote:
> Le Sunday 14 September 2008 20:04:06 Christophe Thommeret, vous avez écrit :
>> Le Sunday 14 September 2008 18:47:00 Steven Toth, vous avez écrit :
>>> Steven Toth wrote:
>>>> Johannes Stezenbach wrote:
>>>>> On Fri, Sep 12, 2008, Steven Toth wrote:
>>>>>> Christophe Thommeret wrote:
>>>>>>> As far  as i understand, the cinergyT2 driver is a bit unusual, e.g.
>>>>>>> dvb_register_frontend is never called (hence no dtv_* log messages).
>>>>>>> I don't know if there is others drivers like this, but this has to
>>>>>>> be investigated cause rewritting all drivers for S2API could be a
>>>>>>> bit of work :)
>>>>>>> P.S.
>>>>>>> I think there is an alternate driver for cinergyT2 actually in
>>>>>>> developement but idon't remember where it's located neither its
>>>>>>> state.
>>>>>> Good to know. (I also saw your followup email). I have zero
>>>>>> experience with the cinergyT2 but the old api should still be working
>>>>>> reliably. I plan to investigate this, sounds like a bug! :)
>>>>> Holger was of the opinion that having the demux in dvb-core
>>>>> was stupid for devices which have no hw demux, so he
>>>>> programmed around dvb-core. His plan was to add a
>>>>> mmap-dma-buffers kind of API to the frontend device,
>>>>> but it never got implemented.
>>>>> Anyway, it's bad if one driver is different than all the others.
>>>> Hmm, I didn't realize this, good to know.
>>>> Now it's peaked my interest, I'll have to look at the code.
>>>> The existing API should still work at a bare minimum, if it's not - it
>>>> needs to.
>>> So I looked the the cinergyT2 code, that's a complete eye-opener. It has
>>> it's own ioctl handler, outside of dvb-core.
>>> It's a good news / bad news thing.
>>> The good news is that this driver will not be effected by the S2API
>>> changes, so nothing can break.
>>> The bad news is that this driver will not be effected by the S2API
>>> changes, so it doesn't get the benefit.
>>> Regardless of S2API or multiproto, I see no reason why we shouldn't
>>> bring this driver back into dvb-core.
>>> I don't have a device to test, but here's a patch (0% tested, with bugs
>>> probably) that converts the module back to a regular dvb-core compatible
>>> device, so the S2API would work with this. If anyone wants to test this,
>>> and finds bugs - I won't get back to this driver for a couple of weeks -
>>> so your patches would be welcome. :)
>>> Frankly, is S2API is selected for merge and we have enough users of the
>>> current non-dvb-core driver, I'll probably re-write it from the spec.
>>> So much to do, so little time.
>>> - Steve
>> Steve,
>> as you expected, it's abit buggy and crashes at load ;)
>> i've searched for the alternative cinergyT2 driver.
>> it's available at: http://linuxtv.org/hg/~tmerle/cinergyT2/
>> haven't tried it yet but will do (at least several people seems to have it
>> working).
> Have it working, not as good as Holger' one (often fails to lock) but it 
> works, with both old and new api.
> So, one problem less :)

Good to know Christophe, thanks.

- Steve

More information about the linux-dvb mailing list