Bug 31927

Summary:	Segfault when starting httpd.socket
Product:	systemd	Reporter:	Matthias Kuhn <matthias.kuhn>
Component:	general	Assignee:	Lennart Poettering <lennart>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	medium
Version:	unspecified
Hardware:	All
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description Matthias Kuhn 2010-11-25 20:53:33 UTC

I want to get my httpd launched when somebody connects to port 80. So I put the following file into /etc/systemd/system/ and started it (sudo systemctl start httpd.socket)
What happened then was, that I was thrown to tty1 (still, gnome kept running) and I was unable to interact with systemd. i.e. any command just times out. syslog states the following:

kernel: systemd[1]: segfault at 0 ip 000000000043adea sp 00007fff0172daf0 error 4 in systemd[400000+d2000]
kernel: systemd[1]: Caught <SEGV>, dumped core as pid 1237.

My file might be not how a socket file should look, but it should definitely not kill systemd.

Comment 1 Lennart Poettering 2011-01-20 15:34:45 UTC

Hmm, does this problem still exist? If so, can you extract a backtrace from that core dump?

Comment 2 Matthias Kuhn 2011-01-21 12:18:46 UTC

Just tried with a git-build from today, same effect.

Hope this information helps:

<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /bin/systemd...done.
[New Thread 15786]
Cannot access memory at address 0xfffff888858b48ff
(gdb) where
#0  0x00007f65a0c5510b in ?? ()
#1  0x0000000000407821 in crash (sig=11) at src/main.c:118
#2  0x00007f65a0c55230 in ?? ()
#3  0x0000000000000001 in ?? ()
#4  0x0000000000000000 in ?? ()
(gdb) 

For future requests: is there a nice way of shutting down the computer under such circumstances? Right now I do it by hitting the power button for 10 secs...

Comment 3 Lennart Poettering 2011-01-21 13:57:34 UTC

(In reply to comment #2)
> Just tried with a git-build from today, same effect.
> 
> Hope this information helps:
> 
> <http://www.gnu.org/software/gdb/bugs/>...
> Reading symbols from /bin/systemd...done.
> [New Thread 15786]
> Cannot access memory at address 0xfffff888858b48ff
> (gdb) where
> #0  0x00007f65a0c5510b in ?? ()
> #1  0x0000000000407821 in crash (sig=11) at src/main.c:118
> #2  0x00007f65a0c55230 in ?? ()
> #3  0x0000000000000001 in ?? ()
> #4  0x0000000000000000 in ?? ()
> (gdb) 
> 
> For future requests: is there a nice way of shutting down the computer under
> such circumstances? Right now I do it by hitting the power button for 10
> secs...

Uh, are you sure you properly rebooted after installing today's snapshot, can't really believe that this is the same issue. Hmm, the backtrace isn't really useful, could you redo it with debug symbols installed?

There's always "reboot -f" to reboot the machine forcibly.

Comment 4 Matthias Kuhn 2011-01-22 06:31:30 UTC

Nice, it works now.
Sorry for the wrong information. You're right, I didn't reboot.
So it looks like the only task left is getting the deamon to start properly, as at least something is listening on port 80 now.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.