43222bd309
This pull request aims to remove any dependency on OpenCL/SPIR-V type information in LLVM IR metadata. While, using metadata might simplify and prettify the resulting SPIR-V output (and restore some of the information missed in the transformation to opaque pointers), the overall methodology for resolving kernel parameter types is highly inefficient. The high-level strategy is to assign kernel parameter types in this order: 1. Resolving the types using builtin function calls as mangled names must contain type information or by looking up builtin definition in SPIRVBuiltins.td. Then: - Assigning the type temporarily using an intrinsic and later setting the right SPIR-V type in SPIRVGlobalRegistry after IRTranslation - Inserting a bitcast 2. Defaulting to LLVM IR types (in case of pointers the generic i8* type or types from byval/byref attributes) In case of type incompatibility (e.g. parameter defined initially as sampler_t and later used as image_t) the error will be found early on before IRTranslation (in the SPIRVEmitIntrinsics pass).
53 lines
2.6 KiB
LLVM
53 lines
2.6 KiB
LLVM
; RUN: llc -O0 -mtriple=spirv32-unknown-unknown %s -o - | FileCheck %s --check-prefix=CHECK-SPIRV
|
|
|
|
;; This test checks that the backend is capable to correctly translate
|
|
;; atomic_cmpxchg OpenCL C 1.2 built-in function [1] into corresponding SPIR-V
|
|
;; instruction.
|
|
|
|
;; __kernel void test_atomic_cmpxchg(__global int *p, int cmp, int val) {
|
|
;; atomic_cmpxchg(p, cmp, val);
|
|
;;
|
|
;; __global unsigned int *up = (__global unsigned int *)p;
|
|
;; unsigned int ucmp = (unsigned int)cmp;
|
|
;; unsigned int uval = (unsigned int)val;
|
|
;; atomic_cmpxchg(up, ucmp, uval);
|
|
;; }
|
|
|
|
; CHECK-SPIRV: OpName %[[#TEST:]] "test_atomic_cmpxchg"
|
|
; CHECK-SPIRV-DAG: %[[#UINT:]] = OpTypeInt 32 0
|
|
; CHECK-SPIRV-DAG: %[[#UINT_PTR:]] = OpTypePointer CrossWorkgroup %[[#UINT]]
|
|
|
|
;; In SPIR-V, atomic_cmpxchg is represented as OpAtomicCompareExchange [2],
|
|
;; which also includes memory scope and two memory semantic arguments. The
|
|
;; backend applies some default memory order for it and therefore, constants
|
|
;; below include a bit more information than original source
|
|
|
|
;; 0x2 Workgroup
|
|
; CHECK-SPIRV-DAG: %[[#WORKGROUP_SCOPE:]] = OpConstant %[[#UINT]] 2
|
|
|
|
;; 0x0 Relaxed
|
|
;; TODO: do we need CrossWorkgroupMemory here as well?
|
|
; CHECK-SPIRV-DAG: %[[#RELAXED:]] = OpConstant %[[#UINT]] 0
|
|
|
|
; CHECK-SPIRV: %[[#TEST]] = OpFunction %[[#]]
|
|
; CHECK-SPIRV: %[[#PTR:]] = OpFunctionParameter %[[#UINT_PTR]]
|
|
; CHECK-SPIRV: %[[#CMP:]] = OpFunctionParameter %[[#UINT]]
|
|
; CHECK-SPIRV: %[[#VAL:]] = OpFunctionParameter %[[#UINT]]
|
|
; CHECK-SPIRV: %[[#]] = OpAtomicCompareExchange %[[#UINT]] %[[#PTR]] %[[#WORKGROUP_SCOPE]] %[[#RELAXED]] %[[#RELAXED]] %[[#VAL]] %[[#CMP]]
|
|
; CHECK-SPIRV: %[[#]] = OpAtomicCompareExchange %[[#UINT]] %[[#PTR]] %[[#WORKGROUP_SCOPE]] %[[#RELAXED]] %[[#RELAXED]] %[[#VAL]] %[[#CMP]]
|
|
|
|
define dso_local spir_kernel void @test_atomic_cmpxchg(i32 addrspace(1)* noundef %p, i32 noundef %cmp, i32 noundef %val) local_unnamed_addr {
|
|
entry:
|
|
%call = tail call spir_func i32 @_Z14atomic_cmpxchgPU3AS1Viii(i32 addrspace(1)* noundef %p, i32 noundef %cmp, i32 noundef %val)
|
|
%call1 = tail call spir_func i32 @_Z14atomic_cmpxchgPU3AS1Vjjj(i32 addrspace(1)* noundef %p, i32 noundef %cmp, i32 noundef %val)
|
|
ret void
|
|
}
|
|
|
|
declare spir_func i32 @_Z14atomic_cmpxchgPU3AS1Viii(i32 addrspace(1)* noundef, i32 noundef, i32 noundef) local_unnamed_addr
|
|
|
|
declare spir_func i32 @_Z14atomic_cmpxchgPU3AS1Vjjj(i32 addrspace(1)* noundef, i32 noundef, i32 noundef) local_unnamed_addr
|
|
|
|
;; References:
|
|
;; [1]: https://www.khronos.org/registry/OpenCL/sdk/2.0/docs/man/xhtml/atomic_cmpxchg.html
|
|
;; [2]: https://www.khronos.org/registry/spir-v/specs/unified1/SPIRV.html#OpAtomicCompareExchange
|