Re: Any way of persuading GNAT/GCC to implement a true overlay and not a pointer?

["Doobs" <doobs@doobs.com>]

">> I was under the impression that code of the following form :
>>
>> X : <Some Type>;
>> Y : <Some Type>;
>> for Y'Address use X'Address;
>>
>> would result in an overlay in the resulting code.

>>
>
> Try applying pragma volatile to both declarations. Also for completeness 
> you should add a pragma import to Y to avoid any default initialization.
>
> There is an AI that talks (to some exent) to this issue.
>
> I am not entirely sure that the volatile will fix your problem because I 
> don't quite get what the resulting behaviour is that you are seeing/not 
> seeing.

Tried Pragma volatile - had no effect.....

What I am seeing is that in the map file X and Y are located at different 
addresses.  Even though Y is a large array in my implementation, the linker 
has only reserved four bytes for it.  When examining the contents of Y at 
run time it contains the address of X - i.e. it is a pointer to X

What I actually want is that Y and X are physically located at the same 
address (i.e. a true overlay) so that test equipment that uses the symbol 
table to resolve data can look at X and resolve the contents as one data 
type or look at Y and resolve the contents as another - but they are 
physically reading the same memory.  I dont want the run-time overhead of 
copying the contents via unchecked conversion to another data type.