GDB infinite loop when stepping on "while (1) i++;"

GDB infinite loop when stepping on "while (1) i++;"

[qinwei]

Dears, problems are here:

gcc -o t -gdwarf-2 tp_while1.c
gdb t
(gdb) b main
Breakpoint 1 at 0x8048350: file tp_while1.c, line 3.
(gdb) r
Starting program: /home/qinwei/GJ283/code/build_linux/debug-gdb/t 

Breakpoint 1, main () at tp_while1.c:3
3         int i = 10;
(gdb) n
4         while (1) i++;
(gdb) n

When program stops at "while (1) i++;", and printing "next" command will 
cause gdb infinite loop. 
The target are x86, score, or other? (gdb-6.6 release version)

Is the reason that the following dwarf-2 info:

  "Special opcode 104: advance Address by 7 to 0x8048357 and Line by 1 to 
4"
  "Advance PC by 7 to 804835e"

which makes "while (1)" and "i++;" the same line number? And gdb will get
"stop_pc" from the range [0x8048357, 0x804835e] and will not stop?
How to solve it?  Thanks!

int main (void)
{
  int i = 10; 
  while (1) i++;
  return 0;
}

 Line Number Statements:
  Extended opcode 2: set Address to 0x8048334
  Special opcode 6: advance Address by 0 to 0x8048334 and Line by 1 to 2
  Advance PC by constant 17 to 0x8048345
  Special opcode 160: advance Address by 11 to 0x8048350 and Line by 1 to 
3
  Special opcode 104: advance Address by 7 to 0x8048357 and Line by 1 to 4
  Advance PC by 7 to 804835e 
  Extended opcode 1: End of Sequence

08048334 <main>:
 8048334:   55                      push   %ebp
 8048335:   89 e5                   mov    %esp,%ebp
 8048337:   83 ec 08                sub    $0x8,%esp
 804833a:   83 e4 f0                and    $0xfffffff0,%esp
 804833d:   b8 00 00 00 00          mov    $0x0,%eax
 8048342:   83 c0 0f                add    $0xf,%eax
 8048345:   83 c0 0f                add    $0xf,%eax
 8048348:   c1 e8 04                shr    $0x4,%eax
 804834b:   c1 e0 04                shl    $0x4,%eax
 804834e:   29 c4                   sub    %eax,%esp
 8048350:   c7 45 fc 0a 00 00 00    movl   $0xa,0xfffffffc(%ebp)
 8048357:   8d 45 fc                lea    0xfffffffc(%ebp),%eax
 804835a:   ff 00                   incl   (%eax)
 804835c:   eb f9                   jmp    8048357 <main+0x23>
 804835e:   90                      nop
 804835f:   90                      nop

Best regards,
Qinwei
Mail  qinwei@sunnorth.com.cn
Phone +86-010-62981668-2708
Fax   +86-010-62985972

Re: GDB infinite loop when stepping on "while (1) i++;"

[Kunal Parmar]

Hi,

On Thu, 2007-03-22 at 15:44 +0800, qinwei wrote:
> Dears, problems are here:
> 
> gcc -o t -gdwarf-2 tp_while1.c
> gdb t
> (gdb) b main
> Breakpoint 1 at 0x8048350: file tp_while1.c, line 3.
> (gdb) r
> Starting program: /home/qinwei/GJ283/code/build_linux/debug-gdb/t 
> 
> Breakpoint 1, main () at tp_while1.c:3
> 3         int i = 10;
> (gdb) n
> 4         while (1) i++;
> (gdb) n
> 
> When program stops at "while (1) i++;", and printing "next" command will 
> cause gdb infinite loop. 
> The target are x86, score, or other? (gdb-6.6 release version)
> 
> Is the reason that the following dwarf-2 info:
> 
>   "Special opcode 104: advance Address by 7 to 0x8048357 and Line by 1 to 
> 4"
>   "Advance PC by 7 to 804835e"
> 
> which makes "while (1)" and "i++;" the same line number? And gdb will get
> "stop_pc" from the range [0x8048357, 0x804835e] and will not stop?

Yes. The command `next' is used to move to advance execution to the next
line of the current function. 

> How to solve it?  Thanks!

Try using `nexti' of `stepi'.

Cheers,
Kunal

> 
> int main (void)
> {
>   int i = 10; 
>   while (1) i++;
>   return 0;
> }
> 
>  Line Number Statements:
>   Extended opcode 2: set Address to 0x8048334
>   Special opcode 6: advance Address by 0 to 0x8048334 and Line by 1 to 2
>   Advance PC by constant 17 to 0x8048345
>   Special opcode 160: advance Address by 11 to 0x8048350 and Line by 1 to 
> 3
>   Special opcode 104: advance Address by 7 to 0x8048357 and Line by 1 to 4
>   Advance PC by 7 to 804835e 
>   Extended opcode 1: End of Sequence
> 
> 08048334 <main>:
>  8048334:   55                      push   %ebp
>  8048335:   89 e5                   mov    %esp,%ebp
>  8048337:   83 ec 08                sub    $0x8,%esp
>  804833a:   83 e4 f0                and    $0xfffffff0,%esp
>  804833d:   b8 00 00 00 00          mov    $0x0,%eax
>  8048342:   83 c0 0f                add    $0xf,%eax
>  8048345:   83 c0 0f                add    $0xf,%eax
>  8048348:   c1 e8 04                shr    $0x4,%eax
>  804834b:   c1 e0 04                shl    $0x4,%eax
>  804834e:   29 c4                   sub    %eax,%esp
>  8048350:   c7 45 fc 0a 00 00 00    movl   $0xa,0xfffffffc(%ebp)
>  8048357:   8d 45 fc                lea    0xfffffffc(%ebp),%eax
>  804835a:   ff 00                   incl   (%eax)
>  804835c:   eb f9                   jmp    8048357 <main+0x23>
>  804835e:   90                      nop
>  804835f:   90                      nop
> 
> Best regards,
> Qinwei
> Mail  qinwei@sunnorth.com.cn
> Phone +86-010-62981668-2708
> Fax   +86-010-62985972
-- 
Kunal Parmar
Tools Team,
Celunite Inc. (www.celunite.com)

Re: GDB infinite loop when stepping on "while (1) i++;"

[qinwei]

> > When program stops at "while (1) i++;", and printing "next" command 
will 
> > cause gdb infinite loop. 
> > The target are x86, score, or other? (gdb-6.6 release version)
> > 
> > Is the reason that the following dwarf-2 info:
> > 
> >   "Special opcode 104: advance Address by 7 to 0x8048357 and Line by 1 
to 
> > 4"
> >   "Advance PC by 7 to 804835e"
> > 
> > which makes "while (1)" and "i++;" the same line number? And gdb will 
get
> > "stop_pc" from the range [0x8048357, 0x804835e] and will not stop?
> 
> Yes. The command `next' is used to move to advance execution to the next
> line of the current function. 

Thanks!

Re: GDB infinite loop when stepping on "while (1) i++;"

[Wenbo Yang]

>> (gdb) n
>> 4         while (1) i++;
>> (gdb) n
>
> I think there is nothing wrong. You let GDB next, it next. And your 
> line is a infinite loop, so it never stops.
>
> If "while (1) i++;" in you source code is put in a line. I think the 
> debugging information should be a line. If not, maybe you compiler has 
> problems. It is different with a single line function, i.e. "int a() {}".
>
> If you want to  step into, why not use "step"?
I'm sorry. Step should not work too. I mean "stepi". :-)

Wenbo

--
Wenbo Yang
SimpLight Nanoelectronics Ltd.