[TIP] Catalyst / fglrx and DVI to HDMI adapters (audio)

Christian König deathsimple at vodafone.de
Mon Oct 7 10:51:01 CEST 2013

Am 07.10.2013 10:06, schrieb Rafał Miłecki:
> For the last few days it was I was testing fglrx on various cards to
> trace HDMI setup. One thing that was killing me was no HDMI audio when
> using some generic/cheap DVI to HDMI adapter with my HD 4850.
> I started digging and it has appeared to be common problem when using
> Catalyst / fglrx. Windows users were also experiencing this problem,
> without any known workaround.
> I've found one explanation by droidhacker:
> http://phoronix.com/forums/showthread.php?23451-HDMI-Audio-with-fglrx&p=124891#post124891
>> If that is the case, you need to be aware that AMD sells a couple of PROPRIETARY adapter plugs, which fglrx can DETECT in order to decide whether or not to enable HDMI audio.
> That sounded weird, I didn't believe that. There wasn't any
> explanation how it's handled, I couldn't imagine how fglrx can "talk"
> to the simple adapter. Overall it should just map the PINs, not
> include any logic!
> But today I've found another discussion:
> http://www.avsforum.com/t/1080627/atis-magic-dvi-to-hdmi-dongle
> That finally explains what is happening. So it appears that ATI's
> proprietary adapter include some tiny I2C protocol that allows
> identifying them! It seems that Catalyst / fglrx uses some simple I2C
> talk to check if the adapter is made by ATI and allows or refuses to
> use HDMI mode.
> That also explains why radeon driver never had any problems with audio
> over DVI, no matter what adapter was used.
> Another ATI/fgllx mystery explained ;)

Why didn't you just asked me?

I've figured this out over five years ago by applying brute force to one 
of the "magic" DVI->HDMI adapters that came with my original RV635. And 
it indeed contains an extra EEPROM on the I2C bus ;)


More information about the dri-devel mailing list