Announcement

Collapse
No announcement yet.

NVIDIA Pushes 62MB Of GSP Binary Firmware Blobs Into Linux-Firmware.Git

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

  • #71
    Originally posted by cl333r View Post

    The biggest place where such money concentrates is London.

    Most "hackers" are simply workers for rich people, usually a proxy of MI6 and CIA. Most rich assholes in the world who steal 8+ figures in their countries flee to London when they're about to be persecuted for their crimes, then in London they follow the typical procedure - they share part of their stolen money with Britain, often agree to lie and fight the govts of their countries and Britain in return gives them asylum and portrays them as innocent victims of X regime (recall all the Russian oligarchs who fled?).
    I bet you never read this in western history books or in western news. Instead all you hear that Britain is doing is "fighting for freedom, democracy and human rights" - which is the modus operandi of the strongest and worst terrorist states in the world.

    Which is one of the big reasons why Britain (the oldest still acting big terrorist state, not as active as the US) is having such economic woes - the world's worst thieves don't flee to it as often as before because it's not safe anymore (because of sanctions).

    I'm still surprised how in every little country the most corrupt assholes who cripple their little countries have very strong relations to London and US in general, often have children there.
    you speak the truth you did enlighten me. and i have something to add and at the same time england threat heroes like Julian Assange like a terrorist.

    this means they only thread evil people good and good people are threated like shit.
    Phantom circuit Sequence Reducer Dyslexia

    Comment


    • #72
      Originally posted by cl333r View Post
      this is the most blatantly cringe crap I have ever heard of. Countries (their govts) are the biggest criminals in the world by definition. The fact that you're clueless about history and geopolitics is no wonder. The worst ones are the UK and US because both follow the policy of the British Empire - colonialism that portrays itself as fighting against all world's Evil.
      i know you are 100% right... but do not expect any truth from Quackdoc​ he resisted many facts for example about intel gpus to the point that he himself in the end discovered that he can not really use his intel gpu he bought because essencial parts to run steam games in proton and vulkan do not exist and so one.
      then if you point him to these contradicted points he himself produce he becomes aggressive. and blames me for toxicity and character assassination​.

      problem with that is humans in generall who think in a spezific way do this on a meta level means Quackdoc thinks about intel cpus similar he thinks about politics of US and UK,

      if someone tells himt he truth about intel GPUs he reacts in the same way as in your case you tell him the truth about UK and USA...

      but to be honest other forum members like avis​ are much worst than quackdoc...
      Phantom circuit Sequence Reducer Dyslexia

      Comment


      • #73
        Originally posted by avis View Post
        I'm well aware of that. I'll tell you something: even when nouveau/NVK start supporting newer NVIDIA GPUs well, most NVIDIA Linux users will likely still continue using their proprietary drivers because they have been rock solid. The only reason most AMD users are not using AMDPro is because it's not been hassle free at all and then AFAIK (I can be wrong, I don't remember it well now) it had its own kernel module with which people struggled with and AMD did a very poor job maintaining it.
        yet another lie you claim AMD PRO driver is not hassle free and this you claim is the reason why people don't use it.

        the true reason why people do not use it is that the AMDGPU-kerneldriver+RADV-vulkandriver+RadeonSI-OpenGLDriver+ACO compiler means the MESA stack is superior to what you get from the AMD PRO driver if you select the closed source option.

        why should i go to AMD.COM and download the AMD PRO driver if the MESA-Stack opensource driver RADV/RadeonSI/ACO is superior and has zero problems in my case ??? i am on Fedora 39 by the way...

        your thoghts and ideas are corrupted you believe that AMD PRO driver from AMD.com has the same meaning as the Nvidia closed source driver.
        but this is clearly not the case.

        for users of bleeding edge linux distros like Fedora 39 this AMD PRO driver has no meaning at all

        this driver is for outdated distros like Debian stable or Redhat enterprise linux and remember with the AMD PRO driver you can just install the opensource driver this means some people use it to update the outdated driver stack of these debian stable and redhat enterprise linux
        Phantom circuit Sequence Reducer Dyslexia

        Comment


        • #74
          Originally posted by Quackdoc View Post
          VM VFIO emulates the actual pcie bus, you can passthrough pcie cards that are unsupported on linux to windows and have them work in windows​
          Your point is totally irrelevant. PCIe passthrough doesn't change the fact that one GPU card can run one firmware only. So each VM must be in sync in driver update if firmware ABI compatibility isn't there. This problem is orthogonal to which OS the host or the VMs are using. In fact, if there isn't PCIe passthrough, things might have been more manageable as the VM can fake a virtual GPU and expose only a stable userspace API to each VM.

          Comment


          • #75
            We talked about code sharing between operating systems, I proved my point. An AMD engineer proved it as well.

            Your ability to change the topic completely and accuse me of the tangential things I didn't talk about is simply staggering.

            You could become an excellent spin writer, qarium

            And you continue to baselessly throw "you're a liar". I'm not interested in this kind of dialogue. Continue to insult someone else please. I'll ignore your posts from now on. I did it before but these forums keep notifying me about your replies despite you being on my ignore list.

            Comment


            • #76
              Originally posted by qarium View Post

              i know you are 100% right... but do not expect any truth from Quackdoc​ he resisted many facts for example about intel gpus to the point that he himself in the end discovered that he can not really use his intel gpu he bought because essencial parts to run steam games in proton and vulkan do not exist and so one.
              then if you point him to these contradicted points he himself produce he becomes aggressive. and blames me for toxicity and character assassination​.

              problem with that is humans in generall who think in a spezific way do this on a meta level means Quackdoc thinks about intel cpus similar he thinks about politics of US and UK,

              if someone tells himt he truth about intel GPUs he reacts in the same way as in your case you tell him the truth about UK and USA...

              but to be honest other forum members like avis​ are much worst than quackdoc...
              Im literally using it now you absolute mongoloid and it is working perfectly fine for me, intel is addressing the sparse residency support and I think parts of it have even landed? not too sure on that one. However maybe if you had less of a hate boner for me, you could post something that actually makes sense for once.

              no shit I get aggressive when you constantly actively lie about me

              Comment


              • #77
                Originally posted by billyswong View Post

                Your point is totally irrelevant. PCIe passthrough doesn't change the fact that one GPU card can run one firmware only. So each VM must be in sync in driver update if firmware ABI compatibility isn't there. This problem is orthogonal to which OS the host or the VMs are using. In fact, if there isn't PCIe passthrough, things might have been more manageable as the VM can fake a virtual GPU and expose only a stable userspace API to each VM.
                as long as you don't need to reflash the card you should be fine

                Comment


                • #78
                  Sorry guys, I didn't have much choice. I would have been okay with leaving the firmware on Nvidia's github, but some distros insist on grabbing firmware only from linux-firmware.git.

                  The good news is that these images compress really well with zstd, so the actual images on your initrd should be a lot smaller than 62MB. Also, you only need one of the 30MB images depending on which GPU you have. Hopefully your distro will be smart enough to install only what you need.

                  Comment


                  • #79
                    Originally posted by tabicat View Post
                    Sorry guys, I didn't have much choice. I would have been okay with leaving the firmware on Nvidia's github, but some distros insist on grabbing firmware only from linux-firmware.git.

                    The good news is that these images compress really well with zstd, so the actual images on your initrd should be a lot smaller than 62MB. Also, you only need one of the 30MB images depending on which GPU you have. Hopefully your distro will be smart enough to install only what you need.
                    The lack of ABI stablitity is a problem so is the way distributions have been doing things..

                    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite


                    Yes it possible for distributions to support multi initrd files that merge at boot but most distributions don't support this feature. Yes not supporting this feature the firmware can turn into x number by kernels installed. Lack of ABI stability also is trouble when you attempt to optimize this as well.

                    Also there is another consideration distributions like having kernels for rescue cases that that basically have almost every bit of possible firmware need. Yes you don't have X GPU but you might have a GPU failure and that X GPU might be the one you have on hand to drop into the system.... So only what you need is really does depend on the case system rescue kernels and livecd kernels and cases like it equal a lot more firmware required.

                    Comment


                    • #80
                      Originally posted by avis View Post
                      We talked about code sharing between operating systems, I proved my point. An AMD engineer proved it as well.
                      bridgman did told you the exact opposit that the AMD PRO driver just spins the AMDVLK driver who is shared source code with their windows driver.
                      additional to that he told you that FGLRX the driver we all used in 2007/2008 on AMD hardware had shared all parts of the driver with windows and the people me included hated it completely.
                      then he told you they did abolish this foolish nightmare and started to write linux spezific kernel driver who is not shared with windows sourcecode.
                      this is the AMDGPU kernel driver and AMDVLK and also RADV use that kernel driver.

                      so what exactly did you proof ? people really involved in driver development tell you that the Nvidia modell of sharing all driver parts between windows and linux is a complete bullshit idea and results in nasty complete shit driver experience.

                      AMD just did end this madness step by step and i am 100% sure they will abolish AMDVLK in the future then of course we will not have any bullshit code from microsoft windows in our drivers.

                      Originally posted by avis View Post
                      Your ability to change the topic completely and accuse me of the tangential things I didn't talk about is simply staggering.
                      no its simple you who claim what nvidia does is a good idea and will give you good results.
                      i abolished my last nvidia card in the end of 2007 and from 2007 to 2023 Nvidia did plain and simple never exited this madness.
                      this "62mb-of-gsp-binary-firmware" is yet another madness.

                      Originally posted by avis View Post

                      You could become an excellent spin writer, qarium
                      And you continue to baselessly throw "you're a liar". I'm not interested in this kind of dialogue. Continue to insult someone else please. I'll ignore your posts from now on. I did it before but these forums keep notifying me about your replies despite you being on my ignore list.
                      dude you told me 1000 and over 1000 times that you will ignore my posts and you did never ignore my posts

                      that you always say you will ignore me and in the end you do not ignore me proofs that you are "insane"
                      Phantom circuit Sequence Reducer Dyslexia

                      Comment

                      Working...
                      X