[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_u8packed` intrinsic (PR #115068)

[Greg Roth via cfe-commits]

================
@@ -942,7 +942,13 @@ uint64_t dot(uint64_t4, uint64_t4);
 
 _HLSL_AVAILABILITY(shadermodel, 6.4)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot4add_i8packed)
-int dot4add_i8packed(unsigned int, unsigned int, int);
+int dot4add_i8packed(uint, uint, int);
----------------
pow2clk wrote:

It's not really a bug fix. It's really minor and we do have some tolerance for incidental clean ups even if we moved the select instruction return fixes to their own PR.

https://github.com/llvm/llvm-project/pull/115068