Re: Help ADA and Assembly !!! :-(

Re: Help ADA and Assembly !!! :-(

[Martin C. Carlisle]

In article <68omkm$orf$1@news2.isdnet.net>, jphe <jphe@hol.fr> wrote:
>Hi everybody i have a project that compare ada subprocedure and assembly
>subprocedure but i can't get in ADA the assembly .obj sub-procedure that i
>have made withe TASM could you help me about the use of Pragma import
>because i read the doc of Gnat and ADAGIDE but it doesn't recognize my .obj,
>i send my little source thanks..... JP  ( JPHE@HOL.fr )
>
>Exemple is a test of pragma import that must be works with calcul.obj but
>doesn't. It says udefinied reference to "calcul" ???

I believe you want something like the following:

procedure Test2;
pragma Import(Convention => Ada, Entity => Test2, Link_Name =>
  "calcul"); -- use whatever name is in .obj in place of "calcul"
             -- use Ada's name in place of Test2.
pragma Linker_Options("myobjfile.obj"); -- rename myobjfile to your name

--Martin
   
-- 
Martin C. Carlisle, Computer Science, US Air Force Academy
mcc@cs.usafa.af.mil, http://www.usafa.af.mil/dfcs/bios/carlisle.html
DISCLAIMER:  This content in no way reflects the opinions, standard or 
policy of the US Air Force Academy or the United States Government.

Re: Help ADA and Assembly !!! :-(

[Dmitriy Anisimkov]

Hello.

> Exemple is a test of pragma import that must be works with calcul.obj but
> doesn't. It says udefinied reference to "calcul" ???
 
Gnat don't understand the *.obj files, but you can write in GNU assembler.
GCC compile assembler files *.s and make *.o files. Also you can get
assembler source from ADA source by -S compiler option.

gcc -S -O2 essaicalc.adb

try
-O1 without optimization
-O2 optimization
-O3 strong optimization

from 
---------essaicalc.ads
package essaicalc is
Procedure essaicalcul (n1,n2:integer; r1,r2 : out integer);
end essaicalc;

---------essaicalc.adb
package body essaicalc is
Procedure essaicalcul (n1,n2:integer; r1,r2 : out integer) is
begin
r1 := n1+n2;
r2 := n1*n2;
end;
end essaicalc;

I have get assembler code, this is a main fragment :

_essaicalc__essaicalcul:
.stabn 68,0,2,LM1-_essaicalc__essaicalcul
LM1:
LBB2:
LBE2:
	pushl %ebp
	movl %esp,%ebp
	pushl %ebx
	movl 8(%ebp),%ebx
	movl 12(%ebp),%eax
	movl 16(%ebp),%edx
.stabn 68,0,5,LM2-_essaicalc__essaicalcul
LM2:
	movl %eax,%ecx
	addl %edx,%eax
	imull %edx,%ecx
	movl %eax,(%ebx)
	movl %ecx,4(%ebx)
	movl %ebx,%eax
	movl -4(%ebp),%ebx
	movl %ebp,%esp
	popl %ebp
	ret $4

I'm using Gnat 3.10 for Win32. I think you can get same results in GNAT for
MSDOS.
IMHO Gnat for MSDOS is 32bit compiler, not 16 :

>	MOV	AX,[BP+6]
>	ADD	AX,[BP+8]

Best regards,

Dimitri Anisimkov.

Re: Help ADA and Assembly !!! :-(

[Samuel Tardieu]

>>>>> "Dmitriy" == Dmitriy Anisimkov <ts@quadrat.omsk.su> writes:

Dmitriy> try -O1 without optimization

You mean probably -O0. -O1 turns on some optimizations such as
putting variables into registers even when the "register" keyword is
not present (when compiling C code), it turns on -fthread-jumps and
-fdefer-pop, etc. See "man gcc" for details.

  Sam
-- 
Samuel Tardieu -- sam@ada.eu.org

Re: Help ADA and Assembly !!! :-(

[Robert Dewar]

Samuel said

<<You mean probably -O0. -O1 turns on some optimizations such as
putting variables into registers even when the "register" keyword is
not present (when compiling C code), it turns on -fthread-jumps and
-fdefer-pop, etc. See "man gcc" for details.
>>

Indeed, -O1 does substantial optimization, a typical behavior for many
programs is that -O1 saves a lot over -O0, but -O2 does not save that
much more over -O1.