Implement the unpack_s8s16 HLSL Function

[farzonl]

• Implement unpack_s8s16 clang builtin,
• Link unpack_s8s16 clang builtin with hlsl_intrinsics.h
• Add sema checks for unpack_s8s16 to CheckHLSLBuiltinFunctionCall in SemaChecking.cpp
• Add codegen for unpack_s8s16 to EmitHLSLBuiltinExpr in CGBuiltin.cpp
• Add codegen tests to clang/test/CodeGenHLSL/builtins/unpack_s8s16.hlsl
• Add sema tests to clang/test/SemaHLSL/BuiltIns/unpack_s8s16-errors.hlsl
• Create the int_dx_unpack_s8s16 intrinsic in IntrinsicsDirectX.td
• Create the DXILOpMapping of int_dx_unpack_s8s16 to 219 in DXIL.td
• Create the unpack_s8s16.ll and unpack_s8s16_errors.ll tests in llvm/test/CodeGen/DirectX/
• Create the int_spv_unpack_s8s16 intrinsic in IntrinsicsSPIRV.td
• In SPIRVInstructionSelector.cpp create the unpack_s8s16 lowering and map it to int_spv_unpack_s8s16 in SPIRVInstructionSelector::selectIntrinsic.
• Create SPIR-V backend test case in llvm/test/CodeGen/SPIRV/hlsl-intrinsics/unpack_s8s16.ll

DirectX

DXIL Opcode	DXIL OpName	Shader Model	Shader Stages
219	Unpack4x8	6.6	()

SPIR-V

OpSConvert:

Description:

Convert signed width. This is either a truncate or a sign extend.

Result Type must be a scalar or vector of integer
type.

Signed Value must be a scalar or vector of integer type.
It must have the same number of components as Result Type. The
component width must not equal the component width in Result Type.

Results are computed per component.

Word Count	Opcode	Results	Operands
4	114	<id> Result Type	Result <id>	<id> Signed Value

Test Case(s)

Example 1

//dxc unpack_s8s16_test.hlsl -T lib_6_8 -enable-16bit-types -O0

export int16_t4 fn(int8_t4_packed p1) {
    return unpack_s8s16(p1);
}

HLSL:

Syntax

int16_t<4> unpack_s8s16(p32i8 pk);

Type Description

Name	Template Type	Component Type	Size
ret	vector	int16_t	4
pk	scalar	int8_t4_packed	1

Minimum Shader Model

This function is supported in the following shader models.

Shader Model	Supported
Shader Model 6.6 and higher shader models	yes

Shader Stages

See also

• Intrinsic Functions (DirectX HLSL)
• See Unpack Intrinsics