* We check for errors to try to catch them as soon as possible, and early-out
return even in the middle of functions. The idea being even if this leaves us
in a weird state we won't get called again as the replay controller will
detect the error and substitute out a dummy driver.
* GenerateGLSLShader is reasonably expensive because it uses glslang to
preprocess the shaders. If we can cache the input hash and look up the cache
with that hash then we can skip it entirely.
* This helps catches cases where a discarded image is accidentally used and in
many cases may still have valid data. Particularly on Vulkan this is relevant
for DONT_CARE renderpass load and store ops.
- Separate the pixel history copy pixel shader into two separate
shaders, one for colour copy and one for depth
- Allocate and update descriptor sets on demand
- Add another compute shader for pixel history depth copy
* We still check available features to know which shaders we could create on
replay, to determine whether we should add STORAGE_BIT to MSAA images we want
to store into.
Use a separate compute shader module for MSAA copy, and output
directly into the destination buffer instead of creating
staging resources.
Support case where there is no depth stencil attachment to get post mod
values in per fragment reporting. Previously used the original
framebuffer that might not have had depth/stencil view, so couldn't
count the fragments. Now use the sub image.
To get the post mod color, we need to blend with the premod color, so
we use vkCmdCopyImage to copy from the original image.
- Refactor to use common funcs
- Add per fragment depth data
- Rename pixelhistory.frag to pixelhistory_primid.frag
- Fix how some per fragment pipelines are created
For per fragment data we create 3 pipelines
1) Post modification pipeline
changes the stencil state and scissors around the target pixel
2) Shader output pipeline
in addition to above, disables rasterization discard, depth bounds test,
culling and turns off blending
3) Primitive ID pipe
in addition to above, disables depth test and depth write, and changes
the fragment shader to only output primitive ID
* These features aren't supported by the underlying implementation.
* Triangle size we can feature-detect properly - by checking for geometry shader
support. Quad resolve should be supported but metal doesn't allow atomics on
images so we have to check specifically for a metal backend.
* We don't want to have glslang specific bits mixed in with general SPIR-V bits.
This is also preparing for moving some SPIR-V structs into common code that
can be re-used in new reflection/disassembly as well as editing.
* Split up old uber-cbuffers used for unrelated shaders into shader-specific
cbuffers (DebugPixelCBufferData -> TexDisplayPSCBuffer / CheckerboardCBuffer /
MeshPixelCBuffer).
* Split up HLSL files so not everything is lumped into 'debugdisplay.hlsl' but
has separate files as appropriate.
* Use #include in GLSL and HLSL to better organise shader code together rather
than relying on lumping files together on the C++ side.
* Renamed files like 'debugcbuffers.h' to 'hlsl_cbuffers.h'.
* Trimmed out some extensions/cruft that isn't needed in the GLSL side since we
no longer use separable shaders.
* Combine some shaders like Outline/Checkerboard that were similar into central
place.
Shahbaz Youssefi