On 6/11/19 12:05 PM, Józef Kucia wrote: > On Tue, Jun 11, 2019 at 11:57 AM Samuel Iglesias Gonsálvez > <sigles...@igalia.com> wrote: >> According to the Vulkan spec, uniform blocks are not allowed to be >> updated through vkCmdPushDescriptorSetKHR(). >> >> There are these spec quotes from "13.2.1. Descriptor Set Layout" >> that are relevant for this case: >> >> "VK_DESCRIPTOR_SET_LAYOUT_CREATE_PUSH_DESCRIPTOR_BIT_KHR specifies >> that descriptor sets must not be allocated using this layout, and >> descriptors are instead pushed by vkCmdPushDescriptorSetKHR." >> >> "If flags contains >> VK_DESCRIPTOR_SET_LAYOUT_CREATE_PUSH_DESCRIPTOR_BIT_KHR, then all >> elements of pBindings must not have a descriptorType of >> VK_DESCRIPTOR_TYPE_INLINE_UNIFORM_BLOCK_EXT". >> >> There is no explicit mention in vkCmdPushDescriptorSetKHR() to forbid >> this case but it is implied in the creation of the descriptor set >> layout as aforementioned. >> >> Signed-off-by: Samuel Iglesias Gonsálvez <sigles...@igalia.com> >> --- >> >> My only doubt is I did well in the case of >> radv_meta_push_descriptor_set(), let me know if you prefer to change >> it to false. >> > Perhaps I'm missing something, but what is the point to add the > additional checks for invalid usage? A correct program must not use > inline uniform blocks with push descriptors. > Right, it should be detected by the Validation Layers. However it is arguable what to do in the driver's side. We can just keep it as it is now, ignore inline uniform block updates (this patch) or even add an assert if it affects stability of the HW (it is not the case here, we tested it). I think ignoring the updates it's the best option, but I am OK with what RADV developers prefer to do.
Sam
signature.asc
Description: OpenPGP digital signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
