Fix the return types of dot4add_i8packed and dot4add_u8packed.

Change the definition of the HLSL `dot4add_i8packed` and
`dot4add_u8packed` intrinsics in `utils/hct/gen_intrin_main.txt` to
simply spell out the return types, rather than saying that their
return type is determined by their third argument.

This prevents DXC from trying to give those functions declarations like

    declare i64 @"\01?dot4add_u8packed@hlsl@@YA_JII_J@Z"(i32, i32, i64 signext) #1

which seems to expect a 64-bit third argument and return value.

These functions are not generic, and they have only one overload, so
there is no need to use interesting `uComponentTypeId` values to get
the right effects, and `HLSLExternalSource::MatchArguments` seems to
get confused about how to treat argument types that affect the return
types.

Fixes microsoft#7400.

Author: jimblandy

tools/clang/test/DXC/dot4add_i8_u8_packed-types.hlsl

@@ -0,0 +1,21 @@
+// RUN: %dxc /Od /T cs_6_8 %s | FileCheck %s
+
+// Compiling this HLSL would fail this assertion in TranslateDot4AddPacked:
+//
+//     DXASSERT(
+//         !accTy->isVectorTy() && accTy->isIntegerTy(32),
+//         "otherwise, unexpected vector support in high level intrinsic template");
+//
+// Bug was fixed by changing the declarations of dot4add_i8packed and
+// dot4add_u8packed in utils/hct/gen_intrin_main.txt to simply write
+// out their argument and return types, rather than using the $typeN
+// reference syntax.
+
+// CHECK: define void @main()
+
+[numthreads(1, 1, 1)]
+void main()
+{
+    int c_6_ = dot4add_i8packed(0, 0, 0);
+    int c7 = dot4add_i8packed(0, 0, c_6_);
+}

utils/hct/gen_intrin_main.txt

@@ -336,8 +336,8 @@ float<4,3> [[rn]] ObjectToWorld4x3();
 float<4,3> [[rn]] WorldToObject4x3();
 
 // Packed dot products with accumulate:
-$type3 [[rn]] dot4add_u8packed(in uint a, in $type1 b, in uint c);
-$type3 [[rn]] dot4add_i8packed(in uint a, in $type1 b, in int c);
+uint [[rn]] dot4add_u8packed(in uint a, in $type1 b, in uint c);
+int [[rn]] dot4add_i8packed(in uint a, in $type1 b, in int c);
 $type3 [[rn]] dot2add(in float16_t<2> a, in $type1 b, in float c);
 
 // Unpacking intrinsics