Announcement

Collapse
No announcement yet.

Radeon UVD Support Merged Into Mesa

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

  • #71
    Originally posted by agd5f View Post
    Also, if it turns out that we can't securely release UVD on those asics we may not be able to. Blame DRM. We never said we would release UVD on all asics, we said we look at the possibility of releasing support if we could do it securely.
    That is basically saying the same: "We deprecated your stuff and we may possibly never give you a feature complete driver. Go fuck yourself, buy something new!"
    No, I won't blame DRM, I blame AMD, because they sell hardware with abysmal support.

    Don't get me wrong, I really appreciate your work, my HD6870 works fine with your drivers, as does my HD3200 (except UVD of course), but this is a very good example for AMD's bad decisions in the last years and it really pisses me (and many others) off. That hardware is still sold in masses and getting sub-par support for it is something that is not making AMD friends. For sure, I could use a 2 years old OS and use the blob, but then I would miss out all the fine stuff that happened in the meantime, especially I had to use an ancient kernel.

    Comment


    • #72
      Originally posted by Vim_User View Post
      That is basically saying the same: "We deprecated your stuff and we may possibly never give you a feature complete driver. Go fuck yourself, buy something new!"
      No, I won't blame DRM, I blame AMD, because they sell hardware with abysmal support.

      Don't get me wrong, I really appreciate your work, my HD6870 works fine with your drivers, as does my HD3200 (except UVD of course), but this is a very good example for AMD's bad decisions in the last years and it really pisses me (and many others) off. That hardware is still sold in masses and getting sub-par support for it is something that is not making AMD friends. For sure, I could use a 2 years old OS and use the blob, but then I would miss out all the fine stuff that happened in the meantime, especially I had to use an ancient kernel.
      We've always made it very clear that we may not able able to support UVD on all asics in the open source driver due to potential security issues. Releasing the code and potentially getting our DRM implementation hacked and losing our ability to sell into a lot of markets isn't going to benefit anyone.

      Comment


      • #73
        Originally posted by agd5f View Post
        We've always made it very clear that we may not able able to support UVD on all asics in the open source driver due to potential security issues. Releasing the code and potentially getting our DRM implementation hacked and losing our ability to sell into a lot of markets isn't going to benefit anyone.
        True, but in the case that the hardware can not be supported by the free driver AMD should not have stopped to release proprietary drivers for it, especially when those chips are still the top of the line integrated GPU for their top of the line CPUs.
        But anyways, this discussion is mood, since it does not lead to anything. Let's just hope that this code gets released and that AMD did not piss off to many customers by this bad move. It is not that they have enough off them that they can afford pissing them off.

        Comment


        • #74
          Differences between UVD, UVD+ and UVD2...?

          Originally posted by agd5f View Post
          Also, if it turns out that we can't securely release UVD on those asics we may not be able to. Blame DRM. We never said we would release UVD on all asics, we said we look at the possibility of releasing support if we could do it securely.
          The OSS driver supports UVD 2.2 onwards, it would seem. And I was always aware that there might be DRM problems supporting UVD; possibly UVD+ too. But DRM problems with UVD 2?

          Comment


          • #75
            Originally posted by chrisr View Post
            The OSS driver supports UVD 2.2 onwards, it would seem. And I was always aware that there might be DRM problems supporting UVD; possibly UVD+ too. But DRM problems with UVD 2?
            Despite the numbering, there are actually bigger hw differences between 2 and 2.2 than 1 and 2.

            Comment


            • #76
              Can you tell us why drm can be a reason when the hdcp master key is already known since 2010? Basically the whole hdcp crap could be disabled (for win) and there would be no diff in security. Nobody cares about it anyway.

              Comment


              • #77
                Originally posted by Kano View Post
                Can you tell us why drm can be a reason when the hdcp master key is already known since 2010? Basically the whole hdcp crap could be disabled (for win) and there would be no diff in security. Nobody cares about it anyway.
                Each link in the playback chain has a different set of requirements/obligations, and the fact that one part is broken doesn't mean we (or any other GPU vendor) can walk away from other parts of the solution. At some point I expect there will be some kind of industry wide "giving up on the whole thing" event where all of the interlocking agreements and standards are cancelled, but until that happens nothing changes for us or any other vendor.

                It's one of those "it doesn't have to make sense, you just have to do it" things...

                Comment


                • #78
                  Originally posted by Kano View Post
                  Can you tell us why drm can be a reason when the hdcp master key is already known since 2010? Basically the whole hdcp crap could be disabled (for win) and there would be no diff in security. Nobody cares about it anyway.
                  Trying to apply logic to the legal system and DRM = FAIL.

                  Comment


                  • #79
                    Originally posted by bridgman View Post
                    Each link in the playback chain has a different set of requirements/obligations, and the fact that one part is broken doesn't mean we (or any other GPU vendor) can walk away from other parts of the solution. At some point I expect there will be some kind of industry wide "giving up on the whole thing" event where all of the interlocking agreements and standards are cancelled, but until that happens nothing changes for us or any other vendor.

                    It's one of those "it doesn't have to make sense, you just have to do it" things...
                    The whole thing is driven by the backwards-thinking control freaks in Hollywood who think that locking down every possible path their content may pass over is somehow going to stop piracy and continue to ignore the vast body of evidence to the contrary.

                    Comment


                    • #80
                      It is multiple media support format high storage capacity, UVD also lunch UVD 3 HD series.

                      Comment

                      Working...
                      X