Announcement

Collapse
No announcement yet.

Intel's Linux Vulkan Driver Lands Workaround For HITMAN 3

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

  • Intel's Linux Vulkan Driver Lands Workaround For HITMAN 3

    Phoronix: Intel's Linux Vulkan Driver Lands Workaround For HITMAN 3

    Intel Arc Graphics A750 and A770 work on Linux if you are running the very latest Linux kernel and Mesa. The gaming experience is decent aside from occasional driver issues. One of the games that has been pesky with the open-source Intel "ANV" Vulkan driver has been the HITMAN 3 title running under Steam Play but with the newest Mesa 22.3 code should now be fixed up...

    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

  • #2
    Okay, ending up in random memory and crashing the driver is a bug on the driver side, that is clear.

    But I wonder if the game is using some invalid behavior that would be also a misuse of the API? I believe it would be neat if the driver could log some warning if apps are using the API incorrectly or somehow else try to push the developers for not doing such.

    If there was just a workaround to 'handle' this behavior that is transparent to everyone but the deep down driver developer, than it basically becomes part of the accepted API and muddies it.

    I hope this is not the case here.

    Comment


    • #3
      Originally posted by Draget View Post
      Okay, ending up in random memory and crashing the driver is a bug on the driver side, that is clear.

      But I wonder if the game is using some invalid behavior that would be also a misuse of the API? I believe it would be neat if the driver could log some warning if apps are using the API incorrectly or somehow else try to push the developers for not doing such.

      If there was just a workaround to 'handle' this behavior that is transparent to everyone but the deep down driver developer, than it basically becomes part of the accepted API and muddies it.

      I hope this is not the case here.
      Most likely the response here is going to be "It works fine in Windows, we don't support Linux," if they are abusing any given API. Steam Deck & Linux (through Play) doesn't have enough users for most developers to care about unless for some reason just want to do so or getting paid to care by Valve. I don't know what the licensing and sales agreements are like for big name publishers with Steam, but the Hitman 3 website says nothing about supporting Steam Play that I can tell. That means if you want to go this route Valve is the one having to deal with the compatibility layer support and IO Interactive doesn't have to lift a finger because they only support Windows (and the main consoles). All of that boils down to the fact that if Valve wants this to work in Steam Play on Intel GPUs they have to work around what Hitman is doing, because IO Interactive isn't likely to fix it unless it shows up in Windows on the same hardware and it's not a driver bug.
      Last edited by stormcrow; 19 October 2022, 09:42 AM.

      Comment


      • #4
        Originally posted by stormcrow View Post
        Most likely the response here is going to be "It works fine in Windows, we don't support Linux," if they are abusing any given API. Steam Deck & Linux (through Play) doesn't have enough users for most developers to care about unless for some reason just want to do so or getting paid to care by Valve.
        ...
        All of that boils down to the fact that if Valve wants this to work in Steam Play on Intel GPUs they have to work around what Hitman is doing, because IO Interactive isn't likely to fix it unless it shows up in Windows on the same hardware and it's not a driver bug.
        Well, one counter-argument to that is whether other Mesa drivers are affected by the same problem, which as far as I understand, they're not. I also don't think the other Mesa drivers have any workarounds specific to this game (though I didn't bother to check so maybe I'm wrong). So, this to me suggests it's an Intel-specific issue and not an issue with the game developers or the OS.
        Despite everything I just said, it's still possible the game was developed incorrectly. If everyone does things the wrong way but it works, then sometimes doing things the right way can be a problem. Think of it like driving on a highway where everybody is going 20+ above the speed limit, and then there's you going no faster than the speed limit. Even though you're driving at the speed you're supposed to, you're now the danger on the road.

        Comment


        • #5
          Originally posted by Draget View Post
          Okay, ending up in random memory and crashing the driver is a bug on the driver side, that is clear.

          But I wonder if the game is using some invalid behavior that would be also a misuse of the API? I believe it would be neat if the driver could log some warning if apps are using the API incorrectly or somehow else try to push the developers for not doing such.

          If there was just a workaround to 'handle' this behavior that is transparent to everyone but the deep down driver developer, than it basically becomes part of the accepted API and muddies it.

          I hope this is not the case here.
          Vulkan was intentionally designed to allow for crashing when the API is used incorrectly. It removed all the error checking that OpenGL did in order to attempt to increase runtime performance.

          There are validation layers you can run your application through and it should report incorrect uses of the API, or at least the bits it knows to check for. But that's all outside the actual driver.

          Comment

          Working...
          X