Binary package hint: bluez
bluetoothd is started via udev here, but it does not want to die on SIGTERM:
root@pluto:~ # ps aux | grep bluetoothd
root 2946 0.1 0.2 3704 1712 ? S<s 21:51 0:00 /usr/sbin/bluetoothd --udev
root@pluto:~ # kill -15 2946
root@pluto:~ # ps aux | grep bluetoothd
root 2946 0.0 0.2 3704 1712 ? S<s 21:51 0:00 /usr/sbin/bluetoothd --udev
It only wants to die on SIGKILL.
This has the negative side effect, that the process is not stopped by /etc/rc[56].d/K74bluetooth stop
and /etc/rc[56].d/sendsigs will try for 10 times to stop the process using SIGTERM, delaying the shutdown process for 10 secs!
ProblemType: Bug
Architecture: i386
Date: Tue Sep 22 21:48:57 2009
DistroRelease: Ubuntu 9.10
Package: bluez 4.51-0ubuntu1
ProcEnviron:
LANGUAGE=de_DE.UTF-8
PATH=(custom, no user)
LANG=de_DE.UTF-8
SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-10.34-generic
SourcePackage: bluez
Uname: Linux 2.6.31-10-generic i686
You're ever-so-slightly ahead of me in filing the bug.
I'd isolated bluetoothd, and that it's blocked SIGTERM out of its mask despite installing a signal handler for it:
SigPnd: 0000000000000000
ShdPnd: 0000000000014000
SigBlk: fffffffe7ffbdeff
SigIgn: 0000000000001000
SigCgt: 0000000180004802
Question is whether this is a bluetoothd bug or a repeat of that udevd bug