Last worked in: 2.6.32
Pressing the dock's "undock" button sends an event, but nothing in the kernel seems to act upon it.
If the button has been pressed, you can undock by:
echo 1 >/sys/devices/platform/dock.0/undock
The exact same event is generated when the undock procedure has completed. This suggests that setting up udev to send the echo is probably not a good idea.
Hardware: IBM Thinkpad T42, Dock II
The event is:
KERNEL[1262830399.911051] change /devices/platform/dock.0 (platform)
This is the same event generated when the button is pressed, or when 1 is sent to /sys/devices/platform/dock.0/undock whether or not the laptop is docked. (This is why trapping it in udev would cause an infinite loop.)
is the behavior changed in recent kernel?
I thought we can check the /sys/devices/platform/dock.0/docked file in udev script and does something according to if the dock is present.
The behaviour has changed. In 2.6.31, the kernel needs no assistance from udev.
can you post the acpidump output? Let's check why there is extra event.
The regression appears to be from 2.6.31.
Also, Paul, please respond or we'll have to close the bug due to the lack of reporter activity.
Created attachment 24719 [details]
ACPI dump of Thinkpad T41
Sorry for the delay. Please note that the laptop is a T41 not a T42. (Why I got it wrong when the model number is printed just under the LCD, I haven't a clue.)
On Monday 01 February 2010, Paul Martin wrote:
> On Mon, Feb 01, 2010 at 01:43:18AM +0100, Rafael J. Wysocki wrote:
> > This message has been generated automatically as a part of a report
> > of regressions introduced between 2.6.31 and 2.6.32.
> > The following bug entry is on the current list of known regressions
> > introduced between 2.6.31 and 2.6.32. Please verify if it still should
> > be listed and let me know (either way).
> > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=15000
> > Subject : Thinkpad dock button no longer works
> > Submitter : Paul Martin <email@example.com>
> > Date : 2010-01-07 02:11 (25 days old)
> Yes, the bug still exists in 2.6.33-rc6.
This patch may be related, but wait until Len or someone who knows what ACPI stands for can verify it as correct.
I'll have a look at that, but in the meantime: Paul, any chance to test if the Chris' patch above helps?
Created attachment 24852 [details]
Chris Mason's patch
I can confirm I've tested this patch, and it does fix this docking bug.
Handled-By : Chris Mason <firstname.lastname@example.org>
Patch : http://patchwork.kernel.org/patch/76110/
Alex Chiang's version of Chris' patch is in the acpi tree.
Alex Chiang's version of Chris' patch is in the acpi tree. http://patchwork.kernel.org/patch/76123/
shipped in linux-2.6.33-rc8-git4
this was a 2.6.33-rc regression, and is thus not needed in 2.6.32.stable