anv: dEQP-VK.wsi.wayland.<various> failures
The following tests are failing since !12117 (151b65b2)
dEQP-VK.wsi.wayland.incremental_present.scale_none.fifo.identity.opaque.incremental_present
dEQP-VK.wsi.wayland.incremental_present.scale_none.fifo.identity.opaque.reference
dEQP-VK.wsi.wayland.incremental_present.scale_none.mailbox.identity.opaque.incremental_present
dEQP-VK.wsi.wayland.incremental_present.scale_none.mailbox.identity.opaque.reference
dEQP-VK.wsi.wayland.swapchain.create.image_format
dEQP-VK.wsi.wayland.swapchain.private_data.image_format
dEQP-VK.wsi.wayland.swapchain.simulate_oom.image_format
I don't think there's necessarily anything wrong with that patch, other than the fact that it allows more formats to work and that's unmasking the issue.
What I see with Weston is Error during import, and no known conversion for format XB30 in the renderer
which happens because the test is trying to create an opaque window with format AB30, which Weston claims to support just fine, but we go through the "does it have alpha? No? Then go with XB30 instead" path, and I see no mention of XB30 in the list of supported formats.
There is a pair for AR30 and XR30 though, so I'm not sure if this is something that the WSI needs to handle better, or if it's expected that these alpha/non-alpha pairs should always be there.
@emersion I'll appreciate any insight you could throw into this.