Bug 12275 - FN-Fx for volume not working on Eee 901
Summary: FN-Fx for volume not working on Eee 901
Status: CLOSED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Platform (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Corentin Chary
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-12-22 13:49 UTC by Patrick Bregman
Modified: 2010-10-08 18:20 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.28-rc9 (latest from GIT as of 21:39:50 GMT)
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
Contents of /etc/acpi (30.00 KB, application/octet-stream)
2008-12-22 14:15 UTC, Patrick Bregman
Details
Output of acpidump (105.71 KB, text/plain)
2008-12-23 00:47 UTC, Patrick Bregman
Details
eeepc-laptop: use select and not depends on (1.01 KB, patch)
2008-12-23 12:59 UTC, Corentin Chary
Details | Diff

Description Patrick Bregman 2008-12-22 13:49:40 UTC
Latest working kernel version: Don't know, probably 2.6.27
Earliest failing kernel version: Not sure, guess 2.6.28-rc3 (my first RC kernel)
Distribution: Slackware Linux 12.x
Hardware Environment: 
- Asus EEE PC 901, 12GB (Windows version)
- - Atheros 802.11n WiFi card, supported by ath9k
- - New 32GB secondary SSD
Software Environment: Slackware Linux 12.1/12.2, latest updates
Problem Description:
On my Eee PC 901 the keys to change the volume do not work since I switched to the 2.6.28-rc kernel. The first time I noticed this was with kernel 2.6.28-rc3. Before this, I compiled my own 2.6.27 kernel in which it worked just fine. 

When I heard about the 1 second boot time of the 2.6.28 kernel, I switched to compiling this kernel out of the GIT tree. I have been doing so since 2.6.28-rc3. The FN-Fx buttons seemed to work (brightness up/down and sleep), but when I tried the volume keys, it did nothing. I kept using the same acpid (version 1.0.8) and the same ACPI scripts.

Steps to reproduce:
1. Compile a 2.6.28 kernel from GIT
1b. Install/reboot
2. Press FN-F7, FN-F8 or FN-F9 while watching alsamixer

Linked files:
2.6.28-rc9 config: http://www.patrickbregman.eu/config-2.6.28-rc9-901
My kernel: http://www.patrickbregman.eu/vmlinuz-2.6.28-rc9-901
Comment 1 Patrick Bregman 2008-12-22 14:15:45 UTC
Created attachment 19437 [details]
Contents of /etc/acpi
Comment 2 ykzhao 2008-12-22 17:04:31 UTC
Hi, Patrick
    From the problem description it seems that this is a regression. Will you please use git-bisect to identify which commit causes the regression.
    Will you please attach the output of acpidump?
    Thanks.
Comment 3 Patrick Bregman 2008-12-23 00:47:22 UTC
Created attachment 19452 [details]
Output of acpidump

Here is the output of 'acpidump'. I am currently looking how to use git-bisect. Never done that before. Could you give me a hint?
Comment 4 ykzhao 2008-12-23 01:09:38 UTC
Hi, Patrick
    Will you please do the following test on the kernel of 2.6.28-rc3?
    a. kill the process which is using /proc/acpi/event (use the command of "lsof /proc/acpi/event" to get the process using /proc/acpi/event)
    b. cat /proc/acpi/event
    c. press the volumes hotkey and see whether the ACPI event can be reported

    It will be great if you can do the same test on the kernel of 2.6.27 kernel.
    thanks.
    
Comment 5 ykzhao 2008-12-23 01:22:14 UTC
Hi, Patrick
    How to use the git-bisect can be found in 
    >http://www.kernel.org/pub/software/scm/git/docs/user-manual.html
    The good commit is v2.6.27 and the bad commit is v2.6.28-rc3.
    
   Good luck.
Thanks.
Comment 6 Patrick Bregman 2008-12-23 02:07:15 UTC
(In reply to comment #4)
Well, tried it with 2.6.28-rc3 and there is no output when pressing the keys. Only "^[[18~", but this is not reproducable.

When trying this with 2.6.27, the kernel hangs, seems like a bad compile (I did one bad compile before switching to 2.6.28-rc3). I'm going to recompile this in a minute.

I also tried 2.6.26.3 (also self compiled), and here I get some output from /proc/acpi/event. The output is:
hotkey ATKD 00000013 00000000 #First time I hit mute
hotkey ATKD 00000013 00000001 #Second time I hit mute
hotkey ATKD 00000014 00000000 #First time I hit volume down
hotkey ATKD 00000014 00000001 #Second time I hit volume down
hotkey ATKD 00000015 00000000 #First time I hit volume up
hoteky ATKD 00000015 00000001 #Second time I hit volume down

But I think I see the problem. When booting 2.6.27, I see the module "eeepc" loading, I didn't see this with 2.6.28. Also, when booting 2.6.26.3, I see a module eeepc_laptop. 

I'm going to recompile 2.6.27 now, and see if there is a eeepc or eeepc_laptop module for 2.6.28.
Comment 7 Patrick Bregman 2008-12-23 02:20:53 UTC
I just see there IS a file named eeepc_laptop in 2.6.28, it is located in drivers/misc. There just is no way I can select this.

I found it, I did not select HWMON. I'm going to compile the latest 2.6.28 kernel with the eeepc_laptop module, and I'm going to see if that works. If not, I'll try and see what a 2.6.27 kernel gives.
Comment 8 Patrick Bregman 2008-12-23 02:26:09 UTC
Alright...

After compiling the eeepc_laptop module, it works...
Damn, do I feel stupid...

Thanks anyway!

P.S: I guess I should close this bug?
Comment 9 Len Brown 2008-12-23 09:50:27 UTC
Kconfig should offer users features, and not make users
devine their dependencies.

Because of this, EEEPC_LAPTOP should probably select HWMON
rather than depend on HWMON, like THINKPAD_ACPI does.

(ditto for BACKLIGHT_CLASS_DEVICE and RFKILL)
Comment 10 Corentin Chary 2008-12-23 12:59:33 UTC
Created attachment 19459 [details]
eeepc-laptop: use select and not depends on

Here is a patch. I've done a similar patch for asus-laptop. Will post it to the ml soon.
Comment 11 Len Brown 2008-12-23 13:15:29 UTC
patch in comment #10 applied to acpi tree
Comment 12 Len Brown 2009-01-16 11:17:09 UTC
shipped in 2.6.29-rc1

Note You need to log in before you can comment on or make changes to this bug.