Comment 41 for bug 497546

Hmm.
There is a quirk in the driver for the original hardware: 1028:02f5
It has been in there since a long time ago, actually even since the support for that codec was first added.
I could try rewriting the model to the way I think it should be, because that model isn't used for any other hardware, but yet that doesn't explain how it could be working before...?