Announcement

Collapse
No announcement yet.

Coreboot's Flashrom Working On Radeon GPU Flashing Support

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

  • Coreboot's Flashrom Working On Radeon GPU Flashing Support

    Phoronix: Coreboot's Flashrom Working On Radeon GPU Flashing Support

    Former RadeonHD driver developer Luc Verhaegen is back at the AMD Radeon GPU reverse-engineering game. He's now pursuing Radeon firmware flashing with the Coreboot Flashrom utility...

    http://www.phoronix.com/scan.php?pag...MD-SPI-Patches

  • #2
    His blog reveals some pretty interesting stuff and is well written. Can only recommend reading

    Comment


    • #3
      This past fight he talks about between ATI and AMD's driver efforts are way over my head to judge, but his blog post reads like a vendetta to me: "So much for history, as giving ATI the two fingered salute is actually more of an added bonus for me."

      Comment


      • #4
        While interesting, it read like a bridgman smear campaign.

        Comment


        • #5
          So there is a chance there will be FOSS firmware for modern AMD GPUs too? No DRM and all?

          Comment


          • #6
            Pretty cool - Linux desperately needs tools like this. When I OC my GPUs, I like to modify the BIOS of them so the changes apply across reboots without the need of any other 3rd party applications running in the background.

            Originally posted by shmerl View Post
            So there is a chance there will be FOSS firmware for modern AMD GPUs too? No DRM and all?
            Not necessarily. As far as I'm concerned, you can still upload an open-source firmware to your GPU via atiflash (a closed-source Windows utility).
            Last edited by schmidtbag; 17 October 2018, 11:11 AM.

            Comment


            • #7
              Sounds like he's not a big fan of bridgman ... But besides that the VEGA part is wrong or at least only partialy right. Flashing VEGA is one side of the medal. The other side is you can't flash modified BIOSes to VEGA hardware. Only with a signed BIOS the GPU will start. So even if it is nice to have an open-source flashtool you can't get modified BIOSes to work on VEGA hardware.

              Comment


              • #8
                Originally posted by shmerl View Post
                So there is a chance there will be FOSS firmware for modern AMD GPUs too? No DRM and all?
                There very likely would have been. And then AMD lost it, between bad products, the credit crunch, and the ATI rubbleheap, openness died. And there are a few X.org community members who helped make it so.

                But no, open firmware is not very likely today. There is little useful information made public any more. And i do not have the resources to do more than scratch the surface and expose the rot, and even there i seem to be the only doing so.

                Comment


                • #9
                  Originally posted by Yoshi View Post
                  Sounds like he's not a big fan of bridgman ... But besides that the VEGA part is wrong or at least only partialy right. Flashing VEGA is one side of the medal. The other side is you can't flash modified BIOSes to VEGA hardware. Only with a signed BIOS the GPU will start. So even if it is nice to have an open-source flashtool you can't get modified BIOSes to work on VEGA hardware.
                  As stated, i do not have vega hardware. I had read some rumours, but saw nothing solid, and kept focussed on the hardware i could test.

                  I am wondering how they would achieve such lock-down though. Given that no-one before me has gone and REed the trivial standard SPI interface for earlier cards, i would not depend on rumours here.

                  The latest version of ATIFlash contains code for flashing vega, and some valuable info, and some functions which look remarkably like what i have seen in the linux version. So if the windows version can flash vega BIOSes, is it perhaps atiwinflash that checks the signed binary before flashing? How would a trivial hw spi interface figure out whether this binary is correctly signed? Is there even an external SPI chip still? How does ATI or the board makers put the initial BIOS on these cards? How does ATI execute atomBIOS's ASICInit(), and would that still bring up the whole hardware, and would that really needed to be loaded from the SPI rom?

                  There something weird going on there, and i think there is a chance that either ATI has severely crippled their hw, or are just using smoke and mirrors, or, in case of a magic knock, we should be able to reproduce the knock.

                  Comment


                  • #10
                    So, I'd like to congratulate Luc on his return to development. It's been a rough road back for you, with a lot of personal struggle. It's great that you feel ok with working on development and GPU projects again.

                    That said, can't you contribute without having to personally attack people? You attack and you attack and you attack, but comments like this:

                    While the level of support Bridgman had from his own ATI management is unclear to me, to AMD management he claimed that Alex was only there to help out the AMD sponsored project
                    show the crux of the problem in your attitude: you don't know what was really going on at ATI/AMD at the time. You're just guessing, and all of your logic is based on your bitter feelings about how things went. The decisions made don't seem to make sense to you, so you suggest foul play. Which is possible. Anything is possible. But what's far more likely is that someone higher up made decisions and the mid-level representatives were forced to adjust without tossing blame around. That's how corporations work. Someone high up the ladder makes a call based more on instinct or "experience" than logic, and then everyone else lower on the chain has to communicate and execute it. They can't talk about it, because that'll just get them kicked off the project or fired. They've just gotta swallow the bitter synergy pill and deal with it. There's no point in getting angry at them, because they weren't the ones who made the call. And what's clear is that you don't know who was making the calls about all this back then.

                    You attacking people who are still involved and contributing to open source development doesn't help. And, honestly, it makes your claims of being blackballed kind of suspicious. Because insisting on maintaining a one-sided public feud with other figures in the same industry for years kind of shows up during a google search. Only incompetent recruiters would miss it. Hell, David Airlie is the guy sending pull requests to Linus for the DRM subsystem. There doesn't have to be any attempt to blackball. Potential employers can just look at your livejournal and speculate as to whether you'd be able to interact with him on DRM mailing lists in a civil fashion.

                    Bridgeman, Airlie and Deucher seem more than willing to bury the hatchet and not deal with this nonsense from a goddamn decade ago. They can be professional about it, despite probably being irritated by being constantly dragged through the mud. Can you? Can't we all just write some goddamn code without devolving into an episode of Jersey Shore? Do we really have to perpetuate the emotionally immature autistic manchild stereotype that plagues Linux developers?

                    By all means, write a history of the facts if you can't let it go. Say what happened and when. But drop all the bullshit speculation about what people intended to do and talk about what made it to market, what got shelved, and how much effort was put in. You undermine the objectivity of your narrative just by being so ridiculously, publicly biased.

                    Honestly, though, you should just let it go. Zen, therapy, whatever. Just stop trying to drag dirty laundry back into the public eye. Stop dwelling on old feuds. Delete the angsty shit, hit the gym. Move the fuck on.

                    Comment

                    Working...
                    X