Bug 242816

Summary:	Defunct hcid processes after resume from S3
Product:	[Fedora] Fedora	Reporter:	Matthias Saou <matthias>
Component:	bluez-utils	Assignee:	David Woodhouse <dwmw2>
Status:	CLOSED NEXTRELEASE	QA Contact:
Severity:	low	Docs Contact:
Priority:	low
Version:	7	CC:	bnocera
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2008-01-17 13:20:29 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Matthias Saou 2007-06-05 22:35:48 UTC

On a Dell Latitude X1 laptop, with integrated bluetooth which works fine, I get
defunct hcid processes, one after each resume from suspend (S3).

 3844 ?        Ss     0:00 /usr/sbin/hcid
 3856 ?        S      0:00  \_ /usr/lib/bluetooth/bluetoothd-service-input
 4302 ?        Z      0:00  \_ [hcid] <defunct>
 4303 ?        Z      0:00  \_ [hcid] <defunct>

If I restart the "bluetooth" service, then they go away. Here are extract from
/var/log/messages which might be useful :

Jun  6 00:31:21 twister hcid[3844]: HCI dev 0 down
Jun  6 00:31:21 twister hcid[3844]: Stopping security manager 0
Jun  6 00:31:21 twister hcid[3844]: Device hci0 has been disabled
Jun  6 00:31:21 twister hcid[3844]: HCI dev 0 unregistered
Jun  6 00:31:21 twister hcid[3844]: Unregister path: /org/bluez/hci0
Jun  6 00:31:21 twister hcid[3844]: Device hci0 has been removed
Jun  6 00:31:21 twister hcid[3844]: HCI dev 0 registered
Jun  6 00:31:21 twister hcid[3844]: HCI dev 0 up
Jun  6 00:31:21 twister hcid[3844]: Device hci0 has been added
Jun  6 00:31:21 twister hcid[3844]: Starting security manager 0
Jun  6 00:31:21 twister hcid[3844]: Device hci0 has been activated

This is with F7 and all current updates and updates-testing applied (as of
today). Note that bluetooth still works fine, it's just that a new defunct
process is there after each and every suspend.

Comment 1 Matthias Saou 2007-06-05 22:38:04 UTC

Oh, it's actually two new defunct processes for each suspend, as can be seen by
the nearby pids (3 suspends for this result) :

 3844 ?        Ss     0:00 /usr/sbin/hcid
 4302 ?        Z      0:00  \_ [hcid] <defunct>
 4303 ?        Z      0:00  \_ [hcid] <defunct>
 4838 ?        Z      0:00  \_ [hcid] <defunct>
 4847 ?        Z      0:00  \_ [hcid] <defunct>
 5193 ?        Z      0:00  \_ [hcid] <defunct>
 5194 ?        Z      0:00  \_ [hcid] <defunct>

Comment 2 Bastien Nocera 2007-10-05 11:43:35 UTC

There's been 10 releases of bluez-utils since then, and I can't find any
relevant changes for that particular problem. Would you mind trying upgrading
bluez-utils and bluez-libs to their current rawhide versions and seeing whether
you see the same problem?

Comment 3 Matthias Saou 2007-10-10 20:30:16 UTC

I updated the 3 following packages on F7 i386 :
bluez-gnome-0.14-7.fc8.i386.rpm
bluez-libs-3.20-1.fc8.i386.rpm
bluez-utils-3.20-2.fc8.i386.rpm

# /usr/sbin/hcid -s -n
hcid[8544]: Bluetooth HCI daemon
/usr/sbin/hcid: symbol lookup error: /usr/sbin/hcid: undefined symbol:
dbus_watch_get_unix_fd

So it looks like the bluez-gnome package could use an explicit dbus dependency
on some higher version... or dbus could have used some soname bump at some point.

Comment 4 Matthias Saou 2007-10-10 20:32:05 UTC

...and since the F8 dbus requires a newer libexpat, the packages can't be
updated as-is on F7. I'll try a rebuild when I have some time.

Comment 5 Matthias Saou 2008-01-17 13:20:29 UTC

This is no longer happening with Fedora 8, so closing.