Ok, here is my super monster shader. It's a Normal Diffuse Specular Shader which handles 6 Lights in 3 passes with shader target version 2.o. Thanks to xXxGuitar for the view trick, I made the corrections. Now the shader could use even more lights, but I think it would be to slow (and you need the beta for the new light managerment). As additional feature I made a 1.1 fallback (the shader JCL postet).

Again, I'm not really skilled with HLSL, even after studying the shaderx2 series.

Code:

//--------------------------------------------------------------
// Diffuse and specular shader
// -------------------------------------------------------------

float4x4 matWorldViewProj;
float4x4 matWorld;
float4 vecLightPos[8]; //light position
float4 vecLightColor[8]; //light position
float4 vecViewPos;
float4 vecFog;

texture entSkin1; //this is the color map
texture entSkin2; //this is the normal map

sampler ColorMapSampler = sampler_state
{
Texture = <entSkin1>;
MinFilter = Linear;
MagFilter = Linear;
MipFilter = Linear;
AddressU = wrap;
AddressV = wrap;
};


sampler BumpMapSampler = sampler_state
{
Texture = <entSkin2>;
MinFilter = Linear;
MagFilter = Linear;
MipFilter = Linear;
AddressU = wrap;
AddressV = wrap;
};


// -------------------------------------------------------------
// 2.0
// -------------------------------------------------------------


struct VS_OUTPUT0
{
float4 Pos : POSITION;
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light1 : TEXCOORD2;
float Att1 : TEXCOORD3;

float3 Light2 : TEXCOORD4;
float Att2 : TEXCOORD5;

float Fog : FOG;
};


VS_OUTPUT0 VS_PASS0(float4 Pos : POSITION, float2 texcoord0 : TEXCOORD0, float3 Normal : NORMAL, float3 Tangent : TEXCOORD0 )
{
VS_OUTPUT0 Out = (VS_OUTPUT0)0;
Out.Pos = mul(Pos, matWorldViewProj); // transform Position

// compute the 3x3 tranform matrix
// to transform from world space to tangent space
float3x3 worldToTangentSpace;
worldToTangentSpace[0] = mul(Tangent, matWorld);
worldToTangentSpace[1] = mul(cross(Tangent, Normal), matWorld);
worldToTangentSpace[2] = mul(Normal, matWorld);

Out.Tex = texcoord0.xy;

float3 PosWorld = mul(Pos, matWorld);

float3 Viewer = PosWorld - vecViewPos;
Out.View = mul(worldToTangentSpace, - Viewer); // V

//light 1
float3 Light1 = PosWorld - vecLightPos[0] ;
Out.Light1.xyz = mul(worldToTangentSpace, -Light1); // L

Out.Att1 = distance(PosWorld,vecLightPos[0])/vecLightPos[0].w; // Point light

//light 2
float3 Light2 = PosWorld - vecLightPos[1] ;
Out.Light2.xyz = mul(worldToTangentSpace, -Light2); // L

Out.Att2 = distance(PosWorld,vecLightPos[1])/vecLightPos[1].w; // Point light

float ofog = 1 - (distance(PosWorld, vecViewPos) - vecFog.x) * (vecFog.z);
Out.Fog = ofog;

return Out;
}


struct PS_INPUT0
{
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light1 : TEXCOORD2;
float Att1 : TEXCOORD3;

float3 Light2 : TEXCOORD4;
float Att2 : TEXCOORD5;
};


float4 PS_PASS0( PS_INPUT0 psInStruct ):COLOR
{

float4 color = tex2D(ColorMapSampler, psInStruct.Tex); // fetch color map
float3 bumpNormal = 2 * (tex2D(BumpMapSampler, psInStruct.Tex) - 0.5); // fetch bump map
float4 gloss = tex2D( BumpMapSampler, psInStruct.Tex );

float3 ViewDir = normalize(psInStruct.View);

//light1
float3 LightDir1 = normalize(psInStruct.Light1);
float4 diff1 = saturate(dot(bumpNormal, LightDir1)); // diffuse component
float shadow1 = saturate(4 * diff1);
float3 Reflect1 = normalize(2 * diff1 * bumpNormal - LightDir1); // R
float4 spec1 = pow(saturate(dot(Reflect1, ViewDir)), 15);
float4 Attenuation1 = saturate(dot(psInStruct.Att1, psInStruct.Att1));

//light2
float3 LightDir2 = normalize(psInStruct.Light2);
float4 diff2 = saturate(dot(bumpNormal, LightDir2)); // diffuse component
float shadow2 = saturate(4 * diff2);
float3 Reflect2 = normalize(2 * diff2 * bumpNormal - LightDir2); // R
float4 spec2 = pow(saturate(dot(Reflect2, ViewDir)), 15);
float4 Attenuation2 = saturate(dot(psInStruct.Att2, psInStruct.Att2));

return
(
(0.3 * color) + //ambient
((shadow1 * (color * diff1 + (spec1*gloss.w)) * (1 -Attenuation1))*vecLightColor[0])+
((shadow2 * (color * diff2 + (spec2*gloss.w)) * (1 -Attenuation2))*vecLightColor[1])
);
}


struct VS_OUTPUT1
{
float4 Pos : POSITION;
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light3 : TEXCOORD2;
float Att3 : TEXCOORD3;

float3 Light4 : TEXCOORD4;
float Att4 : TEXCOORD5;

float Fog : FOG;
};


VS_OUTPUT1 VS_PASS1(float4 Pos : POSITION, float2 texcoord0 : TEXCOORD0, float3 Normal : NORMAL, float3 Tangent : TEXCOORD0 )
{
VS_OUTPUT1 Out = (VS_OUTPUT1)0;
Out.Pos = mul(Pos, matWorldViewProj); // transform Position

// compute the 3x3 tranform matrix
// to transform from world space to tangent space
float3x3 worldToTangentSpace;
worldToTangentSpace[0] = mul(Tangent, matWorld);
worldToTangentSpace[1] = mul(cross(Tangent, Normal), matWorld);
worldToTangentSpace[2] = mul(Normal, matWorld);

Out.Tex = texcoord0.xy;

float3 PosWorld = mul(Pos, matWorld);

float3 Viewer = PosWorld - vecViewPos;
Out.View = mul(worldToTangentSpace, -Viewer); // V

//light 3
float3 Light3 = PosWorld - vecLightPos[2] ;
Out.Light3.xyz = mul(worldToTangentSpace, -Light3); // L

Out.Att3 = distance(PosWorld,vecLightPos[2])/vecLightPos[2].w; // Point light


//light 4
float3 Light4 = PosWorld - vecLightPos[3] ;
Out.Light4.xyz = mul(worldToTangentSpace, -Light4); // L

Out.Att4 = distance(PosWorld,vecLightPos[3])/vecLightPos[3].w; // Point light

float ofog = 1 - (distance(PosWorld, vecViewPos) - vecFog.x) * (vecFog.z);
Out.Fog = ofog;

return Out;
}


struct PS_INPUT1
{
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light3 : TEXCOORD2;
float Att3 : TEXCOORD3;

float3 Light4 : TEXCOORD4;
float Att4 : TEXCOORD5;
};


float4 PS_PASS1( PS_INPUT1 psInStruct ):COLOR
{

float4 color = tex2D(ColorMapSampler, psInStruct.Tex); // fetch color map
float3 bumpNormal = 2 * (tex2D(BumpMapSampler, psInStruct.Tex) - 0.5); // fetch bump map
float4 gloss = tex2D( BumpMapSampler, psInStruct.Tex );

float3 ViewDir = normalize(psInStruct.View);

//light3
float3 LightDir3 = normalize(psInStruct.Light3);
float4 diff3 = saturate(dot(bumpNormal, LightDir3)); // diffuse component
float shadow3 = saturate(4 * diff3);
float3 Reflect3 = normalize(2 * diff3 * bumpNormal - LightDir3); // R
float4 spec3 = pow(saturate(dot(Reflect3, ViewDir)), 15);
float4 Attenuation3 = saturate(dot(psInStruct.Att3, psInStruct.Att3));

//light4
float3 LightDir4 = normalize(psInStruct.Light4);
float4 diff4 = saturate(dot(bumpNormal, LightDir4)); // diffuse component
float shadow4 = saturate(4 * diff4);
float3 Reflect4 = normalize(2 * diff4 * bumpNormal - LightDir4); // R
float4 spec4 = pow(saturate(dot(Reflect4, ViewDir)), 15);
float4 Attenuation4 = saturate(dot(psInStruct.Att4, psInStruct.Att4));

return
(
((shadow3 * (color * diff3 + (spec3*gloss.w)) * (1 -Attenuation3))*vecLightColor[2])+
((shadow4 * (color * diff4 + (spec4*gloss.w)) * (1 -Attenuation4))*vecLightColor[3])

);
}



struct VS_OUTPUT2
{
float4 Pos : POSITION;
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light5 : TEXCOORD2;
float Att5 : TEXCOORD3;

float3 Light6 : TEXCOORD4;
float Att6 : TEXCOORD5;

float Fog : FOG;
};


VS_OUTPUT2 VS_PASS2(float4 Pos : POSITION, float2 texcoord0 : TEXCOORD0, float3 Normal : NORMAL, float3 Tangent : TEXCOORD0 )
{
VS_OUTPUT2 Out = (VS_OUTPUT2)0;
Out.Pos = mul(Pos, matWorldViewProj); // transform Position

// compute the 3x3 tranform matrix
// to transform from world space to tangent space
float3x3 worldToTangentSpace;
worldToTangentSpace[0] = mul(Tangent, matWorld);
worldToTangentSpace[1] = mul(cross(Tangent, Normal), matWorld);
worldToTangentSpace[2] = mul(Normal, matWorld);

Out.Tex = texcoord0.xy;

float3 PosWorld = mul(Pos, matWorld);

float3 Viewer = PosWorld - vecViewPos;
Out.View = mul(worldToTangentSpace, -Viewer); // V

//light 5
float3 Light5 = PosWorld - vecLightPos[4] ;
Out.Light5.xyz = mul(worldToTangentSpace, -Light5); // L

Out.Att5 = distance(PosWorld,vecLightPos[4])/vecLightPos[4].w; // Point light

//light 6
float3 Light6 = PosWorld - vecLightPos[5] ;
Out.Light6.xyz = mul(worldToTangentSpace, -Light6); // L

Out.Att6 = distance(PosWorld,vecLightPos[5])/vecLightPos[5].w; // Point light

float ofog = 1 - (distance(PosWorld, vecViewPos) - vecFog.x) * (vecFog.z);
Out.Fog = ofog;

return Out;
}



struct PS_INPUT2
{
float2 Tex : TEXCOORD0;
float3 View : TEXCOORD1;

float3 Light5 : TEXCOORD2;
float Att5 : TEXCOORD3;

float3 Light6 : TEXCOORD4;
float Att6 : TEXCOORD5;
};


float4 PS_PASS2( PS_INPUT2 psInStruct ):COLOR
{

float4 color = tex2D(ColorMapSampler, psInStruct.Tex); // fetch color map
float3 bumpNormal = 2 * (tex2D(BumpMapSampler, psInStruct.Tex) - 0.5); // fetch bump map
float4 gloss = tex2D( BumpMapSampler, psInStruct.Tex );

float3 ViewDir = normalize(psInStruct.View);

//light5
float3 LightDir5 = normalize(psInStruct.Light5);
float4 diff5 = saturate(dot(bumpNormal, LightDir5)); // diffuse component
float shadow5 = saturate(4 * diff5);
float3 Reflect5 = normalize(2 * diff5 * bumpNormal - LightDir5); // R
float4 spec5 = pow(saturate(dot(Reflect5, ViewDir)), 15);
float4 Attenuation5 = saturate(dot(psInStruct.Att5, psInStruct.Att5));

//light2
float3 LightDir6 = normalize(psInStruct.Light6);
float4 diff6 = saturate(dot(bumpNormal, LightDir6)); // diffuse component
float shadow6 = saturate(4 * diff6);
float3 Reflect6 = normalize(2 * diff6 * bumpNormal - LightDir6); // R
float4 spec6 = pow(saturate(dot(Reflect6, ViewDir)), 15);
float4 Attenuation6 = saturate(dot(psInStruct.Att6, psInStruct.Att6));

return
(
((shadow5 * (color * diff5 + (spec5*gloss.w)) * (1 -Attenuation5))*vecLightColor[4])+
((shadow6 * (color * diff6 + (spec6*gloss.w)) * (1 -Attenuation6))*vecLightColor[5])

);
}



// -------------------------------------------------------------
// 1.1
// -------------------------------------------------------------

float3x3 matTangent;
float4x4 matWorldView;
float4 vecSunDir;

float4 DoTransform(float4 Pos)
{
return mul(Pos,matWorldViewProj);
}

float DoFog(float4 Pos)
{
float3 P = mul(Pos,matWorldView); // convert vector to view space to get it's depth (.z)
return saturate((vecFog.y-P.z) * vecFog.z); // apply the linear fog formula
}

float4 DoPos(float4 Pos)
{
return (mul(Pos,matWorld));
}
float3 DoPos(float3 Pos)
{
return (mul(Pos,matWorld));
}

void CreateTangents(float3 inNormal,float3 inTangent)
{
matTangent[0] = DoPos(inTangent);
matTangent[1] = DoPos(cross(inTangent,inNormal)); // binormal
matTangent[2] = DoPos(inNormal);
}

float3 DoTangent(float3 inVector)
{
return normalize(mul(matTangent,inVector));
}

struct out_bump
{
float4 Pos: POSITION;
float Fog: FOG;
float4 Color: COLOR;
float2 Tex: TEXCOORD0;
float2 Bump: TEXCOORD1;
float3 Normal: TEXCOORD2;
float3 Light: TEXCOORD3;
};

out_bump vs_bump( in float4 inPos: POSITION, in float3 inNormal: NORMAL, in float2 inTex: TEXCOORD0, in float3 inTangent: TEXCOORD2)
{
out_bump Out;

Out.Pos = DoTransform(inPos);
Out.Tex = inTex;
Out.Bump = inTex; // different coordinates required for ps_1_1
Out.Color = float4(1.0,1.0,1.0,1.0);
Out.Fog = DoFog(inPos);

CreateTangents(inNormal,inTangent);
float3 N = matTangent[2];

// transform the output values into the 0..1 range
Out.Light = DoTangent(-vecSunDir) * 0.5 + 0.5;
Out.Normal = DoTangent(N) * 0.5 + 0.5;

return Out;
}


float4 ps_bump(out_bump In): COLOR
{
float4 base = tex2D(ColorMapSampler,In.Tex);
float3 bumpNormal = tex2D(BumpMapSampler,In.Bump);
float diffuse = saturate(dot(In.Light*2 - 1,bumpNormal*2 - 1));
diffuse *= saturate(4 * dot(In.Light*2 - 1,In.Normal*2 - 1));

return base * diffuse;
}


// -------------------------------------------------------------
// techniques//
// -------------------------------------------------------------

// 2.0
technique SpecularNormalMapping_20
{
pass P0
{
alphablendenable=false;
srcblend=zero;
// compile shaders
VertexShader = compile vs_2_0 VS_PASS0();
PixelShader = compile ps_2_0 PS_PASS0();
}

pass P1
{
//blend second pass additively with first
alphablendenable=true;
srcblend=one;
destblend=one;

// compile shaders
VertexShader = compile vs_2_0 VS_PASS1();
PixelShader = compile ps_2_0 PS_PASS1();
}

pass P2
{
//blend second pass additively with first and second
alphablendenable=true;
srcblend=one;
destblend=one;

// compile shaders
VertexShader = compile vs_2_0 VS_PASS2();
PixelShader = compile ps_2_0 PS_PASS2();
}
}

// 1.1
technique SpecularNormalMapping_11
{
pass P0
{
VertexShader = compile vs_1_1 vs_bump();
PixelShader = compile ps_1_1 ps_bump();
}
}

// Fallback; If nothing works
technique fallback { pass one { } }