Internal error

[LemonParty]

Hello.
I get this error Fatal: Internal error 2009112903 on line

Code: ASM  [Select][+][-]

1. ldr     q1, [x9]

FPC version 3.2.2, architecture AARCH64.
Why is this happening?

[jamie]

Maybe x09
It

[PascalDragon]

I get this error Fatal: Internal error 2009112903 on line

Code: ASM  [Select][+][-]

1. ldr     q1, [x9]

FPC version 3.2.2, architecture AARCH64.
Why is this happening?

Please report a bug with a complete, self contained example.

[d2010]

In old days,year1995, with BP7w, I remember. this trick

Code: [Select]

  PUSH X9
  POP Q1
Eu multumesc ArsenieBoca.

Code: [Select]

Var d1,d2,d3:word;
 Begin asm mov ax,[d1]
                 push ax
                 mov ax,[d2]
                 push ax
                 mov ax,[d3]
                 pop   Rax
   end;
End;

O you make backup of q1

Code: [Select]

Var IsaveQ1:int64;
Begin  IsaveQ1:=asm.Q1;


Maybe x09
It

Code: [Select]

    asm.Q1:=IsaveQ1
End;
               

[WooBean]

Hello.
I get this error Fatal: Internal error 2009112903 on line

Code: ASM  [Select][+][-]

1. ldr     q1, [x9]

FPC version 3.2.2, architecture AARCH64.
Why is this happening?

More Pascalish context looks like below:

Code: Pascal  [Select][+][-]

1. function reg_cgsize(const reg: tregister): tcgsize;
2.       begin
3.         case getregtype(reg) of
4.           R_INTREGISTER:
5.             case getsubreg(reg) of
6.               R_SUBD:
7.                 result:=OS_32
8.               else
9.                 result:=OS_64;
10.             end;
11.           R_MMREGISTER :
12.             begin
13.               case getsubreg(reg) of
14.                 R_SUBMMD:
15.                   result:=OS_F64;
16.                 R_SUBMMS:
17.                   result:=OS_F32;
18.                 R_SUBMMWHOLE:
19.                   result:=OS_M128;
20.                 else
21.                   internalerror(2009112903);
22.               end;
23.             end;
24.           else
25.             internalerror(200303181);
26.           end;
27.         end;
28.  
29.  

The citation above taken from FPC source "\fpcsrc\compiler\aarch64\cpubase.pas".

This suggest that:
- presented asm code for aarch64 architecture is invalid,  or
- FPC was written in a hurry at this point (much less possible).

[PascalDragon]

This suggest that:
- presented asm code for aarch64 architecture is invalid,  or
- FPC was written in a hurry at this point (much less possible).

Especially for the assembly readers its more possible than you might think, considering that one does not have every possible construct in mind that a user might use. (I'm saying this as someone who already fixed a bunch of errors in the assembler readers 😅)

That's why I said to LemonParty to please report a bug.