Shader Input Layout Invalid - c++

Question

I am trying to add things to my working code, but it is returning an error [Invalid Arg] when calling CreateInputLayout().
It works before I add the texture stuff, but fails when I add it.
The shader file compiled with no errors.
I think the way I made the layout is bad. Is this correct?
C++ Vertex structure:
struct VertexData{
XMFLOAT3 Pos;
XMFLOAT4 Color;
XMFLOAT2 TexCoord;
};
C++ Layout:
D3D11_INPUT_ELEMENT_DESC inputLayout[]={
{ "POSITION", 0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0, D3D11_INPUT_PER_VERTEX_DATA, 0 },
{ "COLOR", 0, DXGI_FORMAT_R32G32B32A32_FLOAT, 0, 12, D3D11_INPUT_PER_VERTEX_DATA, 0 },
{ "TEXCOORD", 0, DXGI_FORMAT_R32G32_FLOAT, 0, 28, D3D11_INPUT_PER_VERTEX_DATA, 0 },
}; uint lSize=sizeof(inputLayout)/sizeof(inputLayout[0]);
Shader code:
// Stuff from DX Tutorial 7
Texture2D txDiffuse : register( t0 );
SamplerState samLinear : register( s0 );
cbuffer ViewData : register(b0){
matrix World;
matrix View;
matrix Projection;
}
struct VS_INPUT{
float4 pos : SV_POSITION;
float4 color : COLOR0;
float2 tex : TEXCOORD0;
};
struct PS_INPUT{
float4 pos : SV_POSITION;
float4 color : COLOR0;
float2 tex : TEXCOORD0;
};
PS_INPUT VS(VS_INPUT input){
PS_INPUT output=(PS_INPUT)0;
output.pos=mul(input.pos, World);
output.pos=mul(output.pos, View);
output.pos=mul(output.pos, Projection);
output.color = input.color;
output.tex = input.tex;
return output;
}
float4 PS(PS_INPUT input) : SV_Target {
return input.color;
}
Edit:
After a bunch of twiddling with the shader code, I found that if I change the semantics of the...
1st position to: float4 pos : POSITION;
2nd position to: float4 pos : SV_POSITION;
...the input layout successfully creates.
I tried both without the SV_ [system value] prefix, but that also failed. I dont know why I cant have them be the same.
The Semantics seem a bit magical. Not sure if I should answer my own Q with this, or wait for someone smarter to answer better.

Answer 1

Your structs do not match:
struct VertexData
{
XMFLOAT3 Pos; // float3 here
XMFLOAT4 Color;
XMFLOAT2 TexCoord;
};
struct VS_INPUT{
float4 pos : SV_POSITION; // float4 here
float4 color : COLOR0;
float2 tex : TEXCOORD0;
};
For position in 3D space I prefer to use float3. So, as your input layout:
"POSITION", 0, R32G32B32_FLOAT
About semantics magic:
Names in HLSL must match input layout names. For example:
Layout:
{ "POSITION", 0, .. }
Vertex Shader:
float4 pos : POSITION0;
The only exception is that pixel shader must always have float4 pos : SV_POSITION semantics for the input position of pixel (It is a one of the System Value semantics).
And as a bonus, some commonly used stuff I'd like to recommend:
int size = ARRYSIZE(inputayoutDesc);
use D3D11_APPEND_ALIGNED_ELEMENT instead of 12, 28, numbers as alignment offset.

Answer 2

Input semantics shouldn't be SV_POSITION if they are in your input vertices, they should be POSITION I think.
Also you declare this -
struct VS_INPUT{
float4 pos : SV_POSITION;
Shouldn't this be float3 pos : POSITION;
Your input layout only has 3 floats for that item. Not sure if it's the problem but it doesn't look quite right.