[Bug 105496] Using a single-plane imageview from a multi-plane image is broken

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Fri Mar 16 14:33:20 UTC 2018


--- Comment #5 from Lionel Landwerlin <lionel.g.landwerlin at linux.intel.com> ---
(In reply to atomnuker from comment #3)
> (In reply to Lionel Landwerlin from comment #1)
> > I think the mistake here is that you're using a single image with 2
> > different view. This requires a compatible format between the 2 views. R8 &
> > G8_B8_R8_3PLANE_420 are not compatible.
> Yes they are. Read this. It's right at the top.
> https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.
> html#features-formats-compatible-planes
> > 
> > The feature you want to use is VK_IMAGE_CREATE_ALIAS_BIT. You'll need to
> > create 2 image & 2 views and have the 2 images share the same memory.
> It's a solution, though not the one I want, is recommended by the spec or
> something I'd like to waste time just to find out it doesn't work.

One thing that was also missing on the traces is that no image seem to be
created with VK_IMAGE_USAGE_STORAGE_BIT, which what is needed for descriptors
That seems to hit an assert in my debug build.

> > 
> > This is a bit confusing to be fair :) I had to reread the spec, even I
> > implemented this...
> That would explain why it doesn't work.
> > 
> > Consider running your application with the validation layers, it'll help
> > spotting potential mistakes like this.
> I am running it through validation layers every time and I've reported all
> of the issues on multiplane images, including the one where they don't list
> R8 and G8_B8_R8 as compatible. It'll get fixed.
> Though considering how many assertions have been made in the validation
> layers that no one will ever bother with multiplane images I'm getting
> tired. Though the fact that they're slowly getting fixed one at a time gives
> me hope.

You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20180316/2bff2c0a/attachment.html>

More information about the intel-3d-bugs mailing list