Announcement

Collapse
No announcement yet.

HDMI Audio for Evergreen

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • HDMI Audio for Evergreen

    I just switched to the gallium open source driver on my Evergreen card (HD 5770), and I lost HDMI audio. On the xorg wiki I see that it's supposed to work for R600/700 cards, but not for Evergreen, even though they share the same driver.
    Is anyone working on adding support for HDMI audio on Evergreen? How are the two cards different in that respect so that HDMI audio works on one but not on the other?
    Thanks in advance for any reply.

  • #2
    I am also interested.

    Comment


    • #3
      Originally posted by klown72 View Post
      Is anyone working on adding support for HDMI audio on Evergreen? How are the two cards different in that respect so that HDMI audio works on one but not on the other?
      As far as I understood it they use a third party chip for sound and the supplier changed between R700 and Evergreen, so the whole interface changed. I've not heard any progress on the matter in a long, long time so it has probably not made it to the top of anyone's priorities, if it's a third party chip with third party code it might be more lawyers than developer time required.

      Comment


      • #4
        Originally posted by Kjella View Post
        I've not heard any progress on the matter in a long, long time so it has probably not made it to the top of anyone's priorities,
        Actually it's near the top; when you don't hear anything for a while that usually means there's a lot of work to be done with no obvious milestones we can report on.

        Originally posted by Kjella View Post
        if it's a third party chip with third party code it might be more lawyers than developer time required.
        All of the above... it's not actually a third party chip, just some third party IP mixed into our IP so we need to cut very carefully along the dotted line (after we figure out where the dotted line is).

        Comment


        • #5
          Well, wtf is AMD doing? 1 year and still no HDMI audio. You have some really bad lawyers...

          I'm sure code is sitting there somewhere and is getting dusty. If i knew how to reverse i would already started...

          Comment


          • #6
            Originally posted by admiral0 View Post
            Well, wtf is AMD doing? 1 year and still no HDMI audio. You have some really bad lawyers...

            I'm sure code is sitting there somewhere and is getting dusty. If i knew how to reverse i would already started...
            Perhaps we have a disconnect here. The lawyers are doing their job - determining if a proposed plan / release is safe and legal. Sometimes that means saying "no", unfortunately. Do you think their job is just to say "yes" and that they are somehow being "really bad" by doing their job ?

            Comment


            • #7
              Well, it's not like it's no one's fault that HDMI audio is not yet out. If it were a dev issue code would have been already out, it's always a legal issue with that (see synaptics gestures).

              That's why i say: either you've got bad lawyers (they should define what you can open source and what not), either you don't want to open up that code (this is imo quite impossible, because AMD wouldn't have opened up docs and collaborate actively for r600+).

              It's not that "omg the sky is falling", but already a year has passed and AMD is still stuck in legal issues. That's really sad.

              No offence here for AMD, really.

              Comment


              • #8
                Originally posted by admiral0 View Post
                That's why i say: either you've got bad lawyers (they should define what you can open source and what not), either you don't want to open up that code (this is imo quite impossible, because AMD wouldn't have opened up docs and collaborate actively for r600+).
                Sorry, I still don't understand your point. The answer to open sourcing anything starts as "no" and what we do is painstakingly make a case for changing that to "yes" in specific areas.

                We have been working through the different areas in order of importance, starting in 2007 with modesetting and command submission, working our way through 3D acceleration and memory management over the last few years, and now we are getting to more complex (from an IP perspective) areas such as audio and UVD.

                Nobody thought this was going to be easy. I'm pleased that we committed the time, money and resources to make it happen.

                Comment


                • #9
                  Are you perhaps imagining a scenario with hundreds or thousands of lawyers and engineers working on open source documentation (which is what it would take for the kind of timeline you seem to expect) ?

                  If so, I have some bad news for you.

                  Comment


                  • #10
                    Ah i got it. So you still get lots of "no"?

                    I don't understand what's happening in AMD, that's why i made suppositions (from what i read from your last post , i was wrong).

                    What i thought was that you got ok to publish code for HDMI audio, but you are sorting out what to publish.

                    Comment


                    • #11
                      Originally posted by admiral0 View Post
                      Ah i got it. So you still get lots of "no"?

                      I don't understand what's happening in AMD, that's why i made suppositions (from what i read from your last post , i was wrong).

                      What i thought was that you got ok to publish code for HDMI audio, but you are sorting out what to publish.
                      Not so much "no" as "here are a list of specific concerns". Sometimes we find solutions for all of them, sometimes we have to go back and change our plans.

                      We try to look ahead and shape the plans to steer around likely problems but because of the learning curve that comes with each new domain that isn't a 100% solution either.

                      Comment


                      • #12
                        Originally posted by bridgman View Post
                        All of the above... it's not actually a third party chip, just some third party IP mixed into our IP so we need to cut very carefully along the dotted line (after we figure out where the dotted line is).
                        Please don't tell me it's related to DRM and Protected Audio Path. If I hear one more thing which that f#%ing bunch of c#&%&%ers make hopeless, I might have to nuke MPAA from orbit. It's the only way to be sure.

                        Comment


                        • #13
                          Originally posted by bridgman View Post
                          now we are getting to more complex (from an IP perspective) areas such as audio and UVD.
                          So... how soon until there's a timeline for UVD?

                          Comment


                          • #14
                            There is a quite simple and fast way to get HDMI Audio working on Evergreen hardware:

                            Sit down and reverse engineer how it works, and then implement this into the DRM kernel module.
                            If you get stuck, you can always ask on the maillinglist for help.

                            Christian.

                            Comment


                            • #15
                              Originally posted by Deathsimple View Post
                              There is a quite simple and fast way to get HDMI Audio working on Evergreen hardware:

                              Sit down and reverse engineer how it works, and then implement this into the DRM kernel module.
                              If you get stuck, you can always ask on the maillinglist for help.

                              Christian.
                              No offense, but I can't see how reverse engineering a recent graphics card is "quite simple and fast". Any info on this?

                              Comment

                              Working...
                              X