16f9c5da45
This PR resolves the issue of the SPIR-V specification, requiring Shader-coupled capabilities to read/write images in the OpenCL SPIR-V environment, from the perspective of the LLVM SPIR-V backend. See https://github.com/KhronosGroup/SPIRV-Headers/issues/487 for details and discussion. Current implementation correctly reproduces requirements of the SPIR-V specification, however, since the requirements are problematic, out current implementation blocks generation of valid SPIR-V code for compute environments. This PR is to implement a solution discussed at the SPIR-V WG to allow proceeding with generation of valid SPIR-V code for the OpenCL environment and do not impact Vulkan environment at the same time.
58 lines
3.0 KiB
LLVM
58 lines
3.0 KiB
LLVM
; RUN: llc -O0 -mtriple=spirv64-unknown-unknown %s -o - | FileCheck %s --check-prefix=CHECK-SPIRV
|
|
|
|
; StorageImageReadWithoutFormat/StorageImageWriteWithoutFormat implicitly
|
|
; declare Shader, causing a SPIR-V module to be rejected by the OpenCL
|
|
; run-time. See https://github.com/KhronosGroup/SPIRV-Headers/issues/487
|
|
; De-facto, OpImageRead and OpImageWrite are allowed to use Unknown Image
|
|
; Formats when the Kernel capability is declared. We reflect this behavior
|
|
; in the test case, and leave the check under CHECK-SPIRV-NOT to track
|
|
; the issue and follow-up its final resolution when ready.
|
|
; CHECK-SPIRV-NOT: OpCapability StorageImageReadWithoutFormat
|
|
|
|
; CHECK-SPIRV: %[[#IntTy:]] = OpTypeInt
|
|
; CHECK-SPIRV: %[[#IVecTy:]] = OpTypeVector %[[#IntTy]]
|
|
; CHECK-SPIRV: %[[#FloatTy:]] = OpTypeFloat
|
|
; CHECK-SPIRV: %[[#FVecTy:]] = OpTypeVector %[[#FloatTy]]
|
|
; CHECK-SPIRV: OpImageRead %[[#IVecTy]]
|
|
; CHECK-SPIRV: OpImageRead %[[#FVecTy]]
|
|
|
|
;; __kernel void kernelA(__read_only image3d_t input) {
|
|
;; uint4 c = read_imageui(input, (int4)(0, 0, 0, 0));
|
|
;; }
|
|
;;
|
|
;; __kernel void kernelB(__read_only image3d_t input) {
|
|
;; float4 f = read_imagef(input, (int4)(0, 0, 0, 0));
|
|
;; }
|
|
|
|
define dso_local spir_kernel void @kernelA(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %input) {
|
|
entry:
|
|
%input.addr = alloca target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0), align 8
|
|
%c = alloca <4 x i32>, align 16
|
|
%.compoundliteral = alloca <4 x i32>, align 16
|
|
store target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %input, ptr %input.addr, align 8
|
|
%0 = load target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0), ptr %input.addr, align 8
|
|
store <4 x i32> zeroinitializer, ptr %.compoundliteral, align 16
|
|
%1 = load <4 x i32>, ptr %.compoundliteral, align 16
|
|
%call = call spir_func <4 x i32> @_Z12read_imageui14ocl_image3d_roDv4_i(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %0, <4 x i32> noundef %1)
|
|
store <4 x i32> %call, ptr %c, align 16
|
|
ret void
|
|
}
|
|
|
|
declare spir_func <4 x i32> @_Z12read_imageui14ocl_image3d_roDv4_i(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %0, <4 x i32> noundef %1)
|
|
|
|
define dso_local spir_kernel void @kernelB(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %input) {
|
|
entry:
|
|
%input.addr = alloca target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0), align 8
|
|
%f = alloca <4 x float>, align 16
|
|
%.compoundliteral = alloca <4 x i32>, align 16
|
|
store target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %input, ptr %input.addr, align 8
|
|
%0 = load target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0), ptr %input.addr, align 8
|
|
store <4 x i32> zeroinitializer, ptr %.compoundliteral, align 16
|
|
%1 = load <4 x i32>, ptr %.compoundliteral, align 16
|
|
%call = call spir_func <4 x float> @_Z11read_imagef14ocl_image3d_roDv4_i(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %0, <4 x i32> noundef %1)
|
|
store <4 x float> %call, ptr %f, align 16
|
|
ret void
|
|
}
|
|
|
|
declare spir_func <4 x float> @_Z11read_imagef14ocl_image3d_roDv4_i(target("spirv.Image", void, 2, 0, 0, 0, 0, 0, 0) %0, <4 x i32> noundef %1)
|