Announcement

Collapse
No announcement yet.

DMA-BUF Explicit Sync Improvements Being Discussed To Help Vulkan, Other Modern Users

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

  • DMA-BUF Explicit Sync Improvements Being Discussed To Help Vulkan, Other Modern Users

    Phoronix: DMA-BUF Explicit Sync Improvements Being Discussed To Help Vulkan, Other Modern Users

    Intel's lead developer of the ANV Vulkan Linux driver started a discussion last week about adding an API to DMA-BUF for importing/exporting of sync files as part of allowing for explicit synchronization capabilities in better handling of modern APIs from user-space...

    http://www.phoronix.com/scan.php?pag...licit-Sync-RFC

  • #2
    History repeating:

    Intel's lead developer of the ANV Vulkan Linux driver started a discussion last week about adding an API to DMA-BUF for importing/exporting of sync files as part of allowing for explicit synchronization capabilities in better handling of modern APIs from user-space.
    Intel: It's fun! It's fast!

    AMD open-source driver developer Christian K├Ânig expressed interest in the concept. For exporting of sync files he referred to it as an "absolutely great idea" though pointed out in the current design the importing could allow security issues. Christian noted, "See we can't allow userspace to mess with DMA-buf fences in that way because it rips open a security hole you can push an elephant through."
    Others: Well, security.

    Comment


    • #3
      Originally posted by Hibbelharry View Post
      History repeating:



      Intel: It's fun! It's fast!



      Others: Well, security.
      Or, someone who mostly doesn't work in the kernel writing up a proposed patch (marked as RFC, so for early feedback and not submission) with a bug in the implementation, and the bug gets found during review. (AKA process working as intended)

      Comment


      • #4
        Originally posted by BNieuwenhuizen View Post

        Or, someone who mostly doesn't work in the kernel writing up a proposed patch (marked as RFC, so for early feedback and not submission) with a bug in the implementation, and the bug gets found during review. (AKA process working as intended)
        seems like most developers (myself included) and development teams

        Comment

        Working...
        X