[linux-dvb] AnyTV AUTV002 USB ATSC/QAM Tuner Stick

Michael Krufky mkrufky at linuxtv.org
Thu Nov 27 20:28:49 CET 2008


On Thu, Nov 27, 2008 at 11:11 AM, CityK <cityk at rogers.com> wrote:
> Bob Cunningham wrote:
>> What are the next steps?
>> 1. How much effort will be needed to make this driver work?
>> 2. Do all of the pieces already exist?  (All the chips seem to be
>> mentioned somewhere in the DVB tree.)
>> 3. What new code is needed?
>> 4. How much reverse-engineering needs to be done?  Will a full
>> schematic be needed?
>>
>> Most importantly:
>> 5. How can I help?
>>
>> I'm a real-time embedded systems programmer with 25 years experience,
>> though most of what I've written runs on "bare metal", often without
>> an OS.  I know nothing about Linux device drivers.  However, I am very
>> good at getting local hardware to "play nice", first by poking it with
>> a debugger, then generally by scripting through /dev/port.
>> I've never worked with hardware across the USB bus, though I have
>> brought up USB interface hardware from the CPU side, and have had to
>> snoop USB traffic to diagnose problems.  Unfortunately, I don't have a
>> Windows system available to use to snoop the USB traffic from the
>> driver provided by the vendor, though I haven't yet tried to get
>> anything to work via Wine.
>>
>> I'll be taking lots of time off in December, and should have some time
>> to put toward this project.
>>
>> If anyone else is curious, or wants to help, the product is here:
>> http://www.dealextreme.com/details.dx/sku.15569
>>
>> I also found a brief description of a reference design here:
>> http://www.auvitek.com/AU8522%20MT%20USB%20TV%20Stick%20Design%20Brief_R1.0.pdf
>
> 1 - probably not a lot
> 2 - sounds like it, but I do not know how far developed the MT2131
> driver is
> 3 - probably just the "glue code" to tie all the pieces together
> 4 - likely none (there may be a chance the the components are wired up
> slightly differently, such as in the case with GPIO pins, in which case
> you'd have to try to discover the true configuration....schematics would
> obviously help, but probably 99.99% of cases are resolved without such
> aide )
> 5 - you can add the support for the device !  :P    Because of the
> advanced state (component drivers already exist) you won't need any in
> depth knowledge about the developing device drivers.  At this point, all
> you will need to do is figure out where to insert the necessary glue
> code in the existing modules - have a look in the source code of the
> respective components (eg. /linux/drivers/media/video/au0828/ ;
> /linux/drivers/media/dvb/frontends/ ; .... ).  Note that the AU0828 does
> not currently support analog
> (http://marc.info/?l=linux-video&m=122459807631633&w=2).
>
> I suspect that solving the EEPROM issue is just a matter of adjusting
> the code so that it is not specifically expecting a Hauppauge
> signature.  Getting the device to attach the correct tuner should be, I
> imagine, fairly straight forward too .  It would appear that this device
> is highly similar to the Woodbury  (see:
> http://marc.info/?l=linux-dvb&m=122617795121243&w=2); I do not know
> whether there is significant difference between the respective MT parts
> -- might be trivial, then again it might not be

I am hosting experimental support for these devices in the following
mercurial repository:

http://linuxtv.org/hg/~mkrufky/teledongle

Please generate any patches, if any, against the above tree.

If you read the comments that I wrote in the changesets in that tree,
it explains what I know about these devices.

Ignore the "syntek teledongle" name for now -- there are so many of
these different devices floating around with the same USB ID that the
name really doesnt matter.

If this works for you, please let me know.  So far, the tda18271
version works, but the mt2131 version does not.

-Mike



More information about the linux-dvb mailing list