Debian Bug report logs - #475206
gdb: SIGTRAP interfering with debugging

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Jerry Quinn <jlquinn@optonline.net>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: gdb: SIGTRAP interfering with debugging

Date: Wed, 09 Apr 2008 12:58:37 -0400

Package: gdb
Version: 6.7.1-2
Severity: normal


I'm having trouble with gdb stopping on SIGTRAP a lot.  I can run up to the
first breakpoint, then sometimes I try to print or step and gdb will
claim that the program is responding to SIGTRAP.  I'm debugging C++ code and
I'm doing nothing with signals in this code.

As an example, I've hit a break point and printing values in a vector.  The
first value prints fine, then gdb starts doing the following below.  The
"set unwindonsignal on" doesn't help.  I still get SIGTRAPS after this,
almost as if gdb thinks there's a breakpoint on every statement from then on.

(gdb) p context->h[iwi]
$30 = (const int &) @0x8058194: 3
(gdb) p context->h[2]

Program received signal SIGTRAP, Trace/breakpoint trap.
0xb7ce6913 in std::vector<int, std::allocator<int> >::operator[] (
    this=0x8057d84, __n=2) at /usr/include/c++/4.2/bits/stl_vector.h:489
489	      operator[](size_type __n) const
The program being debugged was signaled while in a function called from GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwindonsignal on"
Evaluation of the expression containing the function (std::vector<int, std::allocator<int> >::operator[](unsigned int) const) will be abandoned.
(gdb) fin
Run till exit from #0  0xb7ce6913 in std::vector<int, std::allocator<int> >::operator[] (this=0x8057d84, __n=2) at /usr/include/c++/4.2/bits/stl_vector.h:489

Program received signal SIGTRAP, Trace/breakpoint trap.
std::vector<int, std::allocator<int> >::operator[] (this=0x8057d84, __n=2)
    at /usr/include/c++/4.2/bits/stl_vector.h:490
490	      { return *(this->_M_impl._M_start + __n); }


The only way I've been able to work around this and shut gdb up is:

handle SIGTRAP noprint nostop

However, once I've hit a break point, I find the code tends to execute
really slowly.  I suspect this is related to the handler change above.  But
I can't confirm because gdb misbehaves without changing the handler.



-- System Information:
Debian Release: lenny/sid
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable'), (500, 'stable')
Architecture: i386 (i686)

Kernel: Linux 2.6.24-1-686 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages gdb depends on:
ii  libc6                     2.7-10         GNU C Library: Shared libraries
ii  libexpat1                 1.95.8-4       XML parsing C library - runtime li
ii  libncurses5               5.6+20080308-1 Shared libraries for terminal hand
ii  libreadline5              5.2-3          GNU readline and history libraries

gdb recommends no packages.

-- no debconf information

Message #10 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Daniel Jacobowitz <drow@false.org>

To: Jerry Quinn <jlquinn@optonline.net>, 475206@bugs.debian.org

Subject: Re: Bug#475206: gdb: SIGTRAP interfering with debugging

Date: Wed, 9 Apr 2008 13:26:45 -0400

On Wed, Apr 09, 2008 at 12:58:37PM -0400, Jerry Quinn wrote:
> As an example, I've hit a break point and printing values in a vector.  The
> first value prints fine, then gdb starts doing the following below.  The
> "set unwindonsignal on" doesn't help.  I still get SIGTRAPS after this,
> almost as if gdb thinks there's a breakpoint on every statement from then on.

This is probably a kernel bug in the management of eflags.  You might
want to check with the kernel maintainers.

I remember this coming up and being fixed, but it may have come back.

-- 
Daniel Jacobowitz
CodeSourcery

Message #15 received at 475206@bugs.debian.org (full text, mbox, reply):

From: "Jerry Quinn (ibm)" <jlquinn@optonline.net>

To: Daniel Jacobowitz <drow@false.org>

Cc: 475206@bugs.debian.org

Subject: Re: Bug#475206: gdb: SIGTRAP interfering with debugging

Date: Wed, 09 Apr 2008 13:32:16 -0400

Daniel Jacobowitz wrote:
> On Wed, Apr 09, 2008 at 12:58:37PM -0400, Jerry Quinn wrote:
>   
>> As an example, I've hit a break point and printing values in a vector.  The
>> first value prints fine, then gdb starts doing the following below.  The
>> "set unwindonsignal on" doesn't help.  I still get SIGTRAPS after this,
>> almost as if gdb thinks there's a breakpoint on every statement from then on.
>>     
>
> This is probably a kernel bug in the management of eflags.  You might
> want to check with the kernel maintainers.
>
> I remember this coming up and being fixed, but it may have come back.
>   
I'm running
linux-image-2.6.24-1-686   2.6.24-5


which is the latest kernel available in unstable.  I don't see any 
reference to eflags in the Debian changelog for the kernel package.

Jerry

Message #20 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Daniel Jacobowitz <drow@false.org>

To: "Jerry Quinn (ibm)" <jlquinn@optonline.net>

Cc: 475206@bugs.debian.org, debian-kernel@lists.debian.org

Subject: Re: Bug#475206: gdb: SIGTRAP interfering with debugging

Date: Wed, 9 Apr 2008 13:36:50 -0400

Kernel maintainers, does this look familiar?  Maybe one of the 2.6.24
patch releases broke this?

On Wed, Apr 09, 2008 at 01:32:16PM -0400, Jerry Quinn (ibm) wrote:
> Daniel Jacobowitz wrote:
>> On Wed, Apr 09, 2008 at 12:58:37PM -0400, Jerry Quinn wrote:
>>   
>>> As an example, I've hit a break point and printing values in a vector.  The
>>> first value prints fine, then gdb starts doing the following below.  The
>>> "set unwindonsignal on" doesn't help.  I still get SIGTRAPS after this,
>>> almost as if gdb thinks there's a breakpoint on every statement from then on.
>>>     
>>
>> This is probably a kernel bug in the management of eflags.  You might
>> want to check with the kernel maintainers.
>>
>> I remember this coming up and being fixed, but it may have come back.
>>   
> I'm running
> linux-image-2.6.24-1-686   2.6.24-5
>
>
> which is the latest kernel available in unstable.  I don't see any  
> reference to eflags in the Debian changelog for the kernel package.
>
> Jerry
>

-- 
Daniel Jacobowitz
CodeSourcery

Message #25 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Richard Guenther <richard.guenther@gmx.net>

To: 475206@bugs.debian.org

Subject: gdb: SIGTRAP interfering with debugging

Date: Fri, 11 Apr 2008 00:12:22 +0200

I see this behavior as well when debugging gcc.

ii  gdb            6.8-1
ii  linux-image-2. 2.6.24-5

it doesn't happen with

ii  linux-image-2. 2.6.22-6

so I suspect the i386/x86_64 arch merge might have broken this.

Richard.

Message #30 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Arthur Meacham <meacham@cs.nyu.edu>

To: 475206@bugs.debian.org

Subject: gdb: SIGTRAP interfering with debugging

Date: Tue, 27 May 2008 13:39:18 -0400

I am having the same problem.  It is making life very difficult.  Is
there something I can do, aside from completely downgrade?


(gdb: 6.8-debian 
 linux: 2.6.24-17-generic)

Message #35 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Daniel Jacobowitz <drow@false.org>

To: Arthur Meacham <meacham@fake.edu>, 475206@bugs.debian.org

Subject: Re: Bug#475206: gdb: SIGTRAP interfering with debugging

Date: Tue, 27 May 2008 13:51:09 -0400

On Tue, May 27, 2008 at 01:39:18PM -0400, Arthur Meacham wrote:
> I am having the same problem.  It is making life very difficult.  Is
> there something I can do, aside from completely downgrade?

Sorry, you've really got to ask the kernel maintainers; it's a kernel
bug, not a GDB bug.

-- 
Daniel Jacobowitz
CodeSourcery

Message #40 received at 475206@bugs.debian.org (full text, mbox, reply):

From: Richard Guenther <richard.guenther@gmx.net>

To: Daniel Jacobowitz <drow@false.org>, 475206@bugs.debian.org

Subject: Re: Bug#475206: gdb: SIGTRAP interfering with debugging

Date: Wed, 28 May 2008 20:05:41 +0200

Daniel Jacobowitz wrote:
> On Tue, May 27, 2008 at 01:39:18PM -0400, Arthur Meacham wrote:
>> I am having the same problem.  It is making life very difficult.  Is
>> there something I can do, aside from completely downgrade?
> 
> Sorry, you've really got to ask the kernel maintainers; it's a kernel
> bug, not a GDB bug.

It seems to be fixed with 2.6.25, at least it didn't reproduce there
for me sofar.

Richard.

Message #45 received at 475206@bugs.debian.org (full text, mbox, reply):

From: DEXTER <mydexterid@gmail.com>

To: 475206@bugs.debian.org

Date: Wed, 30 Jul 2008 15:59:32 +0200

https://bugs.launchpad.net/ubuntu/+source/gdb/+bug/230315/comments/9

Message #50 received at 475206@bugs.debian.org (full text, mbox, reply):

From: "Eugene V. Lyubimkin" <jackyf.devel@gmail.com>

To: 475206@bugs.debian.org

Subject: close this?

Date: Fri, 13 Mar 2009 12:19:11 +0200

[Message part 1 (text/plain, inline)]

Hello, I suspect this bug can be safely closed (came here from the
kernel 2.6.24 bug report for Ubuntu).

-- 
Eugene V. Lyubimkin aka JackYF, JID: jackyf.devel(maildog)gmail.com
C++/Perl developer, Debian Maintainer

[signature.asc (application/pgp-signature, attachment)]

Message #55 received at 475206-done@bugs.debian.org (full text, mbox, reply):

From: Daniel Jacobowitz <drow@false.org>

To: 475206-done@bugs.debian.org

Subject: Fixed in current kernels

Date: Sun, 5 Jul 2009 18:46:13 -0400

This bug no longer occurs with current Debian kernels, as far as I can
tell.

-- 
Daniel Jacobowitz
CodeSourcery

Send a report that this bug log contains spam.