RE: baffling assembly-level weirdness

["Pierre Muller" <muller@ics.u-strasbg.fr>]

Hi,

maybe you are more used to intel disassembler
this is att disassembler
mov 0x805e3c0,%edx
means put the content of address 0x805e3c0 into register edx
putting the constant 0x805e3c0 into edx
would be written as
mov $0x805e3c0,%edx
if I am not mistaken.

If you are more used to intel assembly
you can use
'set disassembly-flavor intel'

In your case, this means that the generated code is faulty...


Pierre Muller
Pascal language support maintainer for GDB



> -----Message d'origine-----
> De : gdb-owner@sourceware.org [mailto:gdb-owner@sourceware.org] De la
> part de Mathieu Lacage
> Envoyé : Monday, January 26, 2009 4:25 PM
> À : gdb@sourceware.org
> Objet : baffling assembly-level weirdness
> 
> hi,
> 
> The following gdb session baffles me completely: %edx is reset to zero
> by the mov at address 0x0804ad62 instead of being set to the constant
> 0x804ad62. Of course, this code segfaults at $pc = 0x804ad68 when zero
> is dereferenced...
> 
> Version: GNU gdb 6.8
> 
> (gdb) disas $pc $pc+10
> Dump of assembler code from 0x804ad62 to 0x804ad6c:
> 0x0804ad62 <indent+50>:	mov    0x805e3c0,%edx
> 0x0804ad68 <indent+56>:	mov    0x14(%edx),%eax
> 0x0804ad6b <indent+59>:	cmp    0x18(%edx),%eax
> End of assembler dump.
> (gdb) p $pc
> $3 = (void (*)()) 0x804ad62 <indent+50>
> (gdb) p $edx
> $4 = 1
> (gdb) ni
> 0x0804ad68	108	  return _IO_putc_unlocked (__c, stdout);
> (gdb) p $edx
> $5 = 0
> (gdb) disas $pc $pc+10
> Dump of assembler code from 0x804ad68 to 0x804ad72:
> 0x0804ad68 <indent+56>:	mov    0x14(%edx),%eax
> 0x0804ad6b <indent+59>:	cmp    0x18(%edx),%eax
> 0x0804ad6e <indent+62>:	jae    0x804adbe <indent+142>
> 0x0804ad70 <indent+64>:	movb   $0x20,(%eax)
> End of assembler dump.
> (gdb) x/20bx 0x0804ad62
> 0x804ad62 <indent+50>:	0x8b	0x15	0xc0	0xe3	0x05	0x08	0x8b	0x42
> 0x804ad6a <indent+58>:	0x14	0x3b	0x42	0x18	0x73	0x4e	0xc6	0x00
> 0x804ad72 <indent+66>:	0x20	0x83	0xc0	0x01
> 
> Anyone would have even a remote idea of what could have gone wrong here
> ?
> 
> Mathieu
> --
> Mathieu Lacage <mathieu.lacage@gmail.com>