Simple read/write CS shader with errors when compiled to GLSL

[pixelnerve]

Hello there,
I'm having a weird error when compiling a simple CS shader to GLSL.
Below is the HLSL shader i'm using and what i found to be weird on the generated GLSL code.

RWBuffer<float4>   BufferOutput            : register( u0 );
RWBuffer<float4>     PositionBuffer         	: register( u1 );
RWBuffer<float4>     RotationBuffer          : register( u2 );

[numthreads(32, 32, 1)]
void main( uint3 DispatchThreadID : SV_DispatchThreadID )
{
    uint globalIndex = DispatchThreadID.y * NUM_WORK_GROUPS_X + DispatchThreadID.x;

    if( globalIndex >= NUM_INSTANCES )
        return;

    int3 outputIndex = (uint3)( globalIndex ) * (uint3)( INSTANCE_STRIDE ) + int3( 0, 1, 2 );
    
    float4 pos = PositionBuffer[ globalIndex ];
    float4 Q = RotationBuffer[ globalIndex ];

    float4 scale = float4( pos.www, 1.0f );

    // Pos, Quat, Scale, float4
    BufferOutput[ outputIndex.x ] = pos;
    BufferOutput[ outputIndex.y ] = Q;
    BufferOutput[ outputIndex.z ] = scale;
}   

It reads PositionBuffer as:

    //LD_UAV_TYPED
    u_xlati1 = floatBitsToInt(imageLoad(PositionBuffer, int(u_xlatu0.x)));

and writes as:

    //STORE_UAV_TYPED
    imageStore(BufferOutput, u_xlati0.x, intBitsToFloat(u_xlati1).yzwx);

Another thing is , scale writes looks wrong. it writes .xxxw

    //MOV
    u_xlati1.w = 0;
    //Instruction 3057
    //STORE_UAV_TYPED
    imageStore(BufferOutput, u_xlati0.w, intBitsToFloat(u_xlati1).xxxw);

This is the full generated GLSL code:

 #version 440 core
#extension GL_ARB_explicit_attrib_location : require
#extension GL_ARB_explicit_uniform_location : require
#extension GL_ARB_shading_language_420pack : require

layout(binding = 3, std140) uniform InstanceData {
    vec4 CustomInputParams0;
    vec4 InstanceDataReserved[15];
};
writeonly layout(binding=0) uniform imageBuffer BufferOutput;
readonly layout(binding=1, rgba32f) highp uniform imageBuffer PositionBuffer;
readonly layout(binding=2, rgba32f) highp uniform imageBuffer RotationBuffer;
ivec4 u_xlati0;
uvec2 u_xlatu0;
ivec4 u_xlati1;
ivec4 u_xlati2;
int u_xlati3;
bool u_xlatb3;
layout(local_size_x = 32, local_size_y = 32, local_size_z = 1) in;
void main()
{
    //Instruction 3044
    //FTOU
    u_xlatu0.xy = uvec2(CustomInputParams0.yx);
    //Instruction 3045
    //IMAD
    u_xlatu0.x = gl_GlobalInvocationID.y * u_xlatu0.x + gl_GlobalInvocationID.x;
    //Instruction 3046
    //UGE
    u_xlatb3 = u_xlatu0.x>=u_xlatu0.y;
    //Instruction 3047
    //IF
    if(u_xlatb3){
        //Instruction 3048
        //RET
        return;
        //Instruction 3049
    //ENDIF
    }
    //Instruction 3050
    //LD_UAV_TYPED
    u_xlati1 = floatBitsToInt(imageLoad(PositionBuffer, int(u_xlatu0.x)));
    //Instruction 3051
    //LD_UAV_TYPED
    u_xlati2 = floatBitsToInt(imageLoad(RotationBuffer, int(u_xlatu0.x)));
    //Instruction 3052
    //IMUL
    u_xlati3 = int(u_xlatu0.x) * 3;
    //Instruction 3053
    //IMAD
    u_xlati0.xzw = ivec3(u_xlatu0.xxx) * ivec3(3, 3, 3) + ivec3(0, 1, 2);
    //Instruction 3054
    //STORE_UAV_TYPED
    imageStore(BufferOutput, u_xlati0.x, intBitsToFloat(u_xlati1).yzwx);
    //Instruction 3055
    //STORE_UAV_TYPED
    imageStore(BufferOutput, u_xlati0.z, intBitsToFloat(u_xlati2));
    //Instruction 3056
    //MOV
    u_xlati1.w = 0;
    //Instruction 3057
    //STORE_UAV_TYPED
    imageStore(BufferOutput, u_xlati0.w, intBitsToFloat(u_xlati1).xxxw);
    //Instruction 3058
    //RET
    return;
}

On a final note, i was able to get proper GLSL code by wrapping the pos.www with a abs(). so:

float4 scale = float4( abs(pos.www), 1.0f );

That worked