This is the mail archive of the
gdb-prs@sourceware.org
mailing list for the GDB project.
Re: gdb/2054: can't start X server under gdb
- From: Adam Jackson <ajackson at redhat dot com>
- To: nobody at sources dot redhat dot com
- Cc: gdb-prs at sources dot redhat dot com,
- Date: 20 Dec 2005 00:28:01 -0000
- Subject: Re: gdb/2054: can't start X server under gdb
- Reply-to: Adam Jackson <ajackson at redhat dot com>
The following reply was made to PR gdb/2054; it has been noted by GNATS.
From: Adam Jackson <ajackson@redhat.com>
To: Daniel Jacobowitz <drow@false.org>
Cc: gdb-gnats@sources.redhat.com
Subject: Re: gdb/2054: can't start X server under gdb
Date: Mon, 19 Dec 2005 19:26:02 -0500
On Dec 19, 2005, at 7:12 PM, Daniel Jacobowitz wrote:
> On Mon, Dec 19, 2005 at 06:59:13PM -0500, Adam Jackson wrote:
>>
>> On Dec 19, 2005, at 5:05 PM, Daniel Jacobowitz wrote:
>>
>>> Huh - that (0x4057f) is SIGTRAP | (PTRACE_EVENT_EXEC << 16). GDB
>>> shouldn't
>>> be receiving PTRACE_EVENT_EXEC notification. Is this an unmodified
>>> kernel.org kernel? Did X exec?
>>
>> Yes, totally stock kernel. X does exec xkbcomp(1) as part of startup,
>> and in fact that's where this is failing (bt after the
>> internal-error):
>>
>> (gdb) bt
>> #0 0xffffe410 in __kernel_vsyscall ()
>> #1 0xb7e01aa6 in __libc_fork () at fork.c:127
>> #2 0x081affdf in Popen (
>> command=0x8218370 "\"/opt/debrix/bin/xkbcomp\" -w 1
>> \"-R/opt/debrix/share/X11/xkb\" -xkm \"-\" -em1 \"The XKEYBOARD keymap
>> compiler (xkbcomp) reports:\" -emp \"> \" -eml \"Errors from xkbcomp
>> are not fatal to the X server\" \"/opt/d"..., type=0x81be16f "w") at
>> utils.c:1766
>> #3 0x08191b2d in XkbDDXCompileKeymapByNames (xkb=0x78,
>> names=0xb7a36980,
>> want=120, need=120, nameRtrn=0xbf8a3d40 "T=\212?|=\212??\001",
>> nameRtrnLen=4096) at ddxLoad.c:379
>> ...
>>
>> maybe a bad glibc?
>
> We're in __libc_fork. We've, presumably, just forked. Your kernel has
> somehow reported event 4, when it ought to have reported event 1
> (fork).
>
> I don't have the faintest idea how this could happen. It's possible
> the event is somehow coming from the forked child; could you please
> strace gdb for me? That'll show more clearly what processes we're
> tracing.
That turned into about 50M of log data. The truncated version is at:
http://people.freedesktop.org/~ajax/gdb-bug-2054/strace-of-gdb.txt
The rt_sigaction at the end repeated forever until i sigkilled strace
(4.5.1). I never got to the internal-error message from gdb.
- ajax