Program terminated with signal 5, Trace/breakpoint trap

Program terminated with signal 5, Trace/breakpoint trap

[Mathew Yeates]

Hi
With gdb-7.2.50.20100713 I get this error while loading a corefile.

uname
 x86_64 x86_64 x86_64 GNU/Linux

gcc -v
gcc-4.2.4

the first potential error I see, prior to "Program terminated with
signal 5, Trace/breakpoint trap" is
"warning: Can't read pathname for load map: Input/output error."

Help?

-Mathew

Re: Program terminated with signal 5, Trace/breakpoint trap

[Jan Kratochvil]

On Fri, 16 Jul 2010 02:55:47 +0200, Mathew Yeates wrote:
[...]
> With gdb-7.2.50.20100713 I get this error while loading a corefile.
[...]
>  x86_64 x86_64 x86_64 GNU/Linux
[...]
> "warning: Can't read pathname for load map: Input/output error."

I think you face this one:
	[patch] Fix GNU/Linux core open: Can't read pathname for load map: Input/output error.
	http://sourceware.org/ml/gdb-patches/2009-10/msg00142.html

OTOH I still believe the right fix would be in glibc (referenced there).


Regards,
Jan

Re: Program terminated with signal 5, Trace/breakpoint trap

[Mathew Yeates]

After looking at the patch, it looks like the warning I was seeing was
harmless and probably not responsible for my program terminating.
Right?

On Fri, Jul 16, 2010 at 7:54 AM, Jan Kratochvil
<jan.kratochvil@redhat.com> wrote:
> On Fri, 16 Jul 2010 02:55:47 +0200, Mathew Yeates wrote:
> [...]
>> With gdb-7.2.50.20100713 I get this error while loading a corefile.
> [...]
>>  x86_64 x86_64 x86_64 GNU/Linux
> [...]
>> "warning: Can't read pathname for load map: Input/output error."
>
> I think you face this one:
>        [patch] Fix GNU/Linux core open: Can't read pathname for load map: Input/output error.
>        http://sourceware.org/ml/gdb-patches/2009-10/msg00142.html
>
> OTOH I still believe the right fix would be in glibc (referenced there).
>
>
> Regards,
> Jan
>

Re: Program terminated with signal 5, Trace/breakpoint trap

[Mathew Yeates]

I just noticed another warning
"warning: no loadable sections found in added symbol-file
system-supplied DSO at 0x2aaaaaaab000"

On Fri, Jul 16, 2010 at 8:52 AM, Mathew Yeates <mat.yeates@gmail.com> wrote:
> After looking at the patch, it looks like the warning I was seeing was
> harmless and probably not responsible for my program terminating.
> Right?
>
> On Fri, Jul 16, 2010 at 7:54 AM, Jan Kratochvil
> <jan.kratochvil@redhat.com> wrote:
>> On Fri, 16 Jul 2010 02:55:47 +0200, Mathew Yeates wrote:
>> [...]
>>> With gdb-7.2.50.20100713 I get this error while loading a corefile.
>> [...]
>>>  x86_64 x86_64 x86_64 GNU/Linux
>> [...]
>>> "warning: Can't read pathname for load map: Input/output error."
>>
>> I think you face this one:
>>        [patch] Fix GNU/Linux core open: Can't read pathname for load map: Input/output error.
>>        http://sourceware.org/ml/gdb-patches/2009-10/msg00142.html
>>
>> OTOH I still believe the right fix would be in glibc (referenced there).
>>
>>
>> Regards,
>> Jan
>>
>

Re: Program terminated with signal 5, Trace/breakpoint trap

[Jan Kratochvil]

On Fri, 16 Jul 2010 17:59:54 +0200, Mathew Yeates wrote:
> I just noticed another warning
> "warning: no loadable sections found in added symbol-file
> system-supplied DSO at 0x2aaaaaaab000"

This was an issue on some x86_64 Linux kernels with 8KB vDSO (instead of
common 4KB).  There was Fedora GDB workaround for it but recent Linux kernels
do not need the workaround as they have 4KB vDSO again.
	https://bugzilla.redhat.com/show_bug.cgi?id=312011


Regards,
Jan

Re: Program terminated with signal 5, Trace/breakpoint trap

[Jan Kratochvil]

On Fri, 16 Jul 2010 17:52:40 +0200, Mathew Yeates wrote:
> After looking at the patch, it looks like the warning I was seeing was
> harmless and probably not responsible for my program terminating.
> Right?

Yes.  There is either some forgotten breakpoint in the inferior or some
un-waitpid()ed SIGTRAP signal left in the inferior, just guessing.


Regards,
Jan