Announcement

Collapse
No announcement yet.

No, AMD Will Not Be Opening Up Its Firmware/Microcode

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

  • #21
    Originally posted by libv View Post
    Brent: the original specification of AtomBIOS was that the function tables would only hold ASIC specific things, and not board specific information. Board specific information was to be provided in data tables. This was gradually and increasingly broken in r6xx.

    If we had been able to continue this project like that, there would have been a high probability that we ended up in a situation where we would also be able to open source the firmware of the 3d engine/CP. But that of course did not happen.
    I wouldn't really call xf86-video-radeon a fork of xf86-video-radeonhd seeing as it existed first

    Comment


    • #22
      Oh, i forgot the one AtomBIOS document we once got, which, by the time we got it was not that useful anymore. I think it described atombios for DCE-2 or something. Iirc, nothing like that was available, not even at ATI internal, when we started. We just had register level docs, hw, and the first version of the atombios interpreter, and a lot of clue and motivation.

      Comment


      • #23
        Originally posted by starshipeleven View Post
        Not gonna work, people isn't going to violate NDAs publicly just for lulz.

        He cannot make such project unless AMD actually publishes stuff first.
        As far as I'm aware a clean room wall is still legal.

        Comment


        • #24
          Luc, are you talking about the big r600 programming guide ? If so then we did end up publishing cleaned-up versions of that information albeit in other docs. We did not publish the microcode emulation of 2D packets because that was dead-ended after r600 and didn't even make it into the 6xx derivative chips (610, 630) and some other bits got sanitized out, but I think we covered all of the useful bits.

          Not sure what the second doc you are referring to is, pls let me know.
          Test signature

          Comment


          • #25
            Fireburn: nope. There was radeonhd code bolted into -ati afterwards, and then partially rewritten, and some copyrights removed. All of that happened in November 2007.

            Comment


            • #26
              bridgman: Nah, something later, with display engine stuff. That was like Q3-Q4 2008, shortly before AMD ran out of cash, so i think there was also a time factor involved in why these never were released.

              Comment


              • #27
                Originally posted by libv View Post
                bridgman: Nah, something later, with display engine stuff. That was like Q3-Q4 2008, shortly before AMD ran out of cash, so i think there was also a time factor involved in why these never were released.
                Ah yes... I was doing a presentation about the open source driver effort to some of our architects; while talking about documentation I got an "OMG you can't publish that" about one specific (but essential) aspect of the display documentation.

                It took a couple of years of discussion with Legal and eventually the conclusion was that it was OK to publish that information but by that time the display logic had become so complicated (largely thanks to displayport) that we stayed away from display hardware documentation. Probably more of a time and money thing as you say than a legal thing AFAIK.
                Test signature

                Comment


                • #28
                  Originally posted by starshipeleven View Post
                  you know what microcode means? https://en.wikipedia.org/wiki/Microcode
                  well right, but it can still be something between "real" microcode or just assembled against some ISA. The firmware files used within nouveau are actually written in a falcon assmebly code, which gets simply assembled to the falcon ISA.

                  Comment


                  • #29
                    Originally posted by karolherbst View Post
                    well right, but it can still be something between "real" microcode or just assembled against some ISA. The firmware files used within nouveau are actually written in a falcon assmebly code, which gets simply assembled to the falcon ISA.
                    We have some of that (UVD, VCE, MEC) but also a bunch of lower level HW microcode (memory controller, SDMA controller, ME etc..).
                    Test signature

                    Comment


                    • #30
                      Originally posted by libv View Post
                      For display/chip init, we had it all with RadeonHD, a native C driver (C code was remarketed as legacy by the fork later on), register level documentation being made available (which stopped as soon as RadeonHD stopped), we just needed some enterprising individuals to provide a corebooted version of ASICInit() (about 200-400loc). But Redhat and a few "community members", just had to side with ATI, create a competing fork, and then implement everything the ATI way.
                      I followed RadeonHD since 2007, but unfortunately I had a Nvidia at the time. When I finally managed to get an ATI radeonhd was already forked and later I had to switch to radeon because the project got abandoned. Noone will prevent anybody to restore the work on a more free driver once the times get mature: sooner or later firmware blobs will become more and more invasive/buggy and peoples will start to get tired. Funny fact: the vast majority of issues I've never been able to solve even after years were firmware related.
                      ## VGA ##
                      AMD: X1950XTX, HD3870, HD5870
                      Intel: GMA45, HD3000 (Core i5 2500K)

                      Comment

                      Working...
                      X