A referencing problem

It’s possible that you can place any data type even though it is not advisable because “struct” belongs to a value type. But the actual problem arises when we place a type like array or some thing as a buffer and call some unmanaged external API to fill it. In this case when a garbage collection occurs there is a chance that the memory of the buffer can get reallocated as a result of GC’s memory compaction process and thus a page fault will result when the API function try to access this buffer. A kind of work around can be done using the “Fixed” keyword which will pin that data type to memory. But this is not the exact one. For a long time I was searching for a solution and I came across a code snippet in one web site which is as follows

using System.Runtime.InteropServices;
[StructLayout(LayoutKind.Explicit)]
struct tstruct
{
[FieldOffset(0)]
public byte service1;
[FieldOffset(1)]
public byte service2;
[FieldOffset(2)]
public byte data1;
[FieldOffset(12)]
public byte data2;
[FieldOffset(22)]
public byte max_data1;
[FieldOffset(23)]
public byte max_data2;
}

To access the data1/2 arrays,use:
tstruct t;
...
byte* b1=&t.data1;
byte* b2=&t.data2;

Then just access b1/2 like a normal array. The thing is that there is will be no index-checking or other thing done by CLR because it directly considers it as a perfect value type and will get completely allocated into stack there by avoiding all operations of GC