Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 364301 - udev wacom rules totally hose system
Summary: udev wacom rules totally hose system
Status: RESOLVED INVALID
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: All Linux
: Normal normal (vote)
Assignee: Gentoo Linux bug wranglers
URL: http://sources.gentoo.org/cgi-bin/vie...
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-04-21 00:45 UTC by Reuben Martin
Modified: 2011-04-21 02:54 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments
wacom udev rules (80-input-wacom.rules,9.40 KB, text/plain)
2011-04-21 02:41 UTC, Reuben Martin
Details
udev db (udev_db_dump.txt,168.24 KB, text/plain)
2011-04-21 02:54 UTC, Reuben Martin
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Reuben Martin 2011-04-21 00:45:30 UTC
There is a nasty bug with the udev rules for wacom devices.

Basically, for some reason udev thinks every device listed under /dev must be a "Serial Wacom Tablet" when X is loading, and X tries to load the wacom driver for every device in /dev.

I've actually noticed this for quite a while, but it was rather harmless, it would just fail on all the devices and be done with it. But a recent change somewhere has changed that. Now it keeps going back to all unmounted block devices and doesn't stop trying to load the driver. The result is excessive device access activity on all unmounted block devices, and ~60% load on all CPUs.

I've been able to stop it for now by commenting out the following lines in /etc/udev/rules.d/80-input-wacom.rules :

# Catch the serial tablets and tell X that's what they are
#ACTION=="add|change", SUBSYSTEM=="pnp", ATTR{id}=="WACf*",      \
#        ENV{NAME}="Serial Wacom Tablet",                        \
#        ENV{ID_INPUT}="1", ENV{ID_INPUT_TABLET}="1"
#
#ACTION=="add|change", SUBSYSTEM=="pnp", ATTR{id}=="FUJ*",       \
#        ENV{NAME}="Serial Wacom Tablet",                        \
#        ENV{ID_INPUT}="1", ENV{ID_INPUT_TABLET}="1"

I'm not very familiar with udev rules, so not totally sure what these are creating so many false positives for Serial Wacom Tablets.

Reproducible: Always
Comment 1 Reuben Martin 2011-04-21 00:47:20 UTC
Also, forgot to mention, the problem may be in xorg's wacom input device definitions rather than udev...
Comment 2 Rafał Mużyło 2011-04-21 01:23:05 UTC
As those rules should have any effect only in X,
could you i.e. restore those rules and check with 'udevadm info --export-db' which devices were actually affected by those rules ?
Also, what's your udev version ?
And what package did provide 80-input-wacom.rules ?
Comment 3 Reuben Martin 2011-04-21 02:40:26 UTC
You know, I didn't even think to check which package the rules belong to, I had assumed it was part of the x11-drivers/xf86-input-wacom package. However, nothing seems to claim it when running an "equery belongs"

It is quite possible that I created that myself in the past from a non-gentoo source. I would think there would need to be a udev rule of some sort though in order to create the wacom input dev nodes.

I'll attach the complete udev rule, but since I can't find any package to claim it you would probably be best off to just mark this as INVALID.
Comment 4 Reuben Martin 2011-04-21 02:41:35 UTC
Created attachment 270735 [details]
wacom udev rules
Comment 5 Jeroen Roovers (RETIRED) gentoo-dev 2011-04-21 02:53:02 UTC
(In reply to comment #4)
> Created attachment 270735 [details]
> wacom udev rules

That's the [URL], left over from the x11-drivers/linuxwacom package. It should have been removed when you uninstalled that, and you should now be using x11-drivers/xf86-input-wacom instead. I guess that makes this bug INVALID.
Comment 6 Reuben Martin 2011-04-21 02:54:40 UTC
Created attachment 270737 [details]
udev db

here's the udev db dump for what it's worth.