Bug 1092246 - Tumbleweed: hid2hci called from a udev rule is constantly failing leading to a busy system
Summary: Tumbleweed: hid2hci called from a udev rule is constantly failing leading to...
Status: IN_PROGRESS
Alias: None
Product: openSUSE Tumbleweed
Classification: openSUSE
Component: Basesystem (show other bugs)
Version: Current
Hardware: x86-64 openSUSE Factory
: P5 - None : Normal (vote)
Target Milestone: ---
Assignee: Vladimir Botka
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-05-07 21:39 UTC by Neil Rickert
Modified: 2018-05-22 15:32 UTC (History)
2 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
A transcript showing the output of "ps -ef | grep udevd" (3.79 KB, text/plain)
2018-05-07 21:40 UTC, Neil Rickert
Details
debug logs (155.98 KB, application/gzip)
2018-05-17 20:58 UTC, Neil Rickert
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Neil Rickert 2018-05-07 21:39:31 UTC
User-Agent:       Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Firefox/52.0
Build Identifier: 

I am only seeing this on one Tumbleweed system.  The "top" command show one "systemd-udevd" process at near 100% processor time, and another at a bit under 50% processor time.  I will attach the output of "ps -ef | grep udevd".

This started in March or April -- I'm not sure which Tumbleweed snapshot.  The system is currently up-to-date, and the problem continues.

Reproducible: Always
Comment 1 Neil Rickert 2018-05-07 21:40:36 UTC
Created attachment 769287 [details]
A transcript showing the output of "ps -ef | grep udevd"
Comment 2 Franck Bui 2018-05-17 15:45:01 UTC
Could you try to attach to one of the udev processes with gdb and show the backtrace ?

Otherwise can you try to reproduce with the udev debug logs enabled (see /etc/udev/udev.conf) ?
Comment 3 Neil Rickert 2018-05-17 20:58:31 UTC
Created attachment 770648 [details]
debug logs

I set debug loggin in "udev.conf".  I am attaching the compressed output from
 journalctl -b | grep udev

(second try -- I failed to compress on the first try, and exceeded max attachment size)
Comment 4 Franck Bui 2018-05-18 09:00:35 UTC
Apparently there's something broken with 'hid2hci':

  Process 'hid2hci --method=dell --devpath=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.6/2-1.6.2/2-1.6.2:1.0' failed with exit code 1

Perhaps something goes wrong with your HW and the tool is not able to handle gracefully...

Where does this tool come from ? I can't find it in the regular repos...
Comment 5 Neil Rickert 2018-05-18 13:09:12 UTC
>Where does this tool come from ? I can't find it in the regular repos...

  /usr/lib/udev/hid2hci

Part of the "bluez" package, version 5.49-2.1-x86_64 from vendor openSUSE

I'll note that I am not seeing the same problem in Leap 15.0 (on the same computer) with "bluez" version 5.48-lp150.3.1-x86_64

Also, I am not actually using any bluetooth devices.
Comment 6 Franck Bui 2018-05-18 13:29:11 UTC
OK let's re-assign this issue to the Bluez package maintainer since the problem happens in the rules shipped by this package.
Comment 7 Franck Bui 2018-05-22 09:19:57 UTC
This might be related: https://lkml.org/lkml/2018/5/7/326
Comment 8 Neil Rickert 2018-05-22 15:32:39 UTC
Yes, that link from lkml.org does look related.

I am not able to find that file on my system, so I suppose it has to do with rules that are built into the software.

For the present, I am disabling bluetooth in the BIOS, which seems to avoid the problem.