[regression] [hardy] sony eject button doesn't work anymore

Hi unggnu,

Care to give a reference to the patch you mention? Thanks.

Sorry, I don't know the patch exactly but I guess there is one since the eject button only works with Feisty and Gutsy and the Ubuntu Kernel since my own compiled 2.6.22-2.6.24 vanilla kernel doesn't support the eject function but all other sony-laptop features.

Per the kernel team's bug policy, can you please attach the following information. Please be sure to attach each file as a separate attachment.

* uname -a > uname-a.log
* cat /proc/version_signature > version.log
* dmesg > dmesg.log
* sudo lspci -vvnn > lspci-vvnn.log

For more information regarding the kernel team bug policy, please refer to https://wiki.ubuntu.com/KernelTeamBugPolicies . Thanks again.

the eject button acpi event under Hardy (doesn't work):
[Sat Feb 2 12:52:31 2008] received event "sony/hotkey SPIC 00000001 0000001b"
[Sat Feb 2 12:52:31 2008] notifying client 4983[0:0]
[Sat Feb 2 12:52:31 2008] notifying client 5253[108:116]
[Sat Feb 2 12:52:31 2008] executing action "/etc/acpi/ejectbtn.sh"
[Sat Feb 2 12:52:31 2008] BEGIN HANDLER MESSAGES
[Sat Feb 2 12:52:31 2008] END HANDLER MESSAGES
[Sat Feb 2 12:52:31 2008] action exited with status 0
[Sat Feb 2 12:52:31 2008] completed event "sony/hotkey SPIC 00000001 0000001b"

the eject button acpi event under Gutsy (does work):
[Sun Feb 10 12:02:16 2008] received event "sony/hotkey SPIC 00000001 0000001b"
[Sun Feb 10 12:02:16 2008] notifying client 5049[107:114]
[Sun Feb 10 12:02:16 2008] notifying client 5112[0:0]
[Sun Feb 10 12:02:16 2008] executing action "/etc/acpi/ejectbtn.sh"
[Sun Feb 10 12:02:16 2008] BEGIN HANDLER MESSAGES
[Sun Feb 10 12:02:16 2008] END HANDLER MESSAGES
[Sun Feb 10 12:02:16 2008] action exited with status 0
[Sun Feb 10 12:02:16 2008] completed event "sony/hotkey SPIC 00000001 0000001b"

Maybe it is just a device issue so the script isn't able to open it with Vanilla and Hardy kernel cd device path.

I have forgot to mention that I see a gnome eject applet for some seconds if I press the button under Gutsy but not under Hardy.

Hi unggnu,

I just want to make sure that the acpi_fakekey is working on the correct device, can you send me the output of:

head -n 1 /sys/class/input/input*/name

and also the output from:

strace acpi_fakekey 161

Many thanks.

This command gives me an error output only.

Hi,

just to do a quick sanity check, can you eject using the command line "eject" command?

thanks

Yes, it works fine and as I said before gnome doesn't show the eject applet after pressing the key.

Hi,

A couple of questions just to clarify some things for me:

>>sudo strace acpi_fakekey 161
> This command gives me an error output only.

1, What error output output are you referring to?
2, does "acpi_fakekey 161" do an eject?

Colin

1. the strace command of comment https://bugs.launchpad.net/ubuntu/+source/linux/+bug/187980/comments/12 only outputs error ( 2> ) - every output should be in the attached file
2. No it doesn't eject.

It appears that difference between Feisty and Hardy is the appearance of the sony-laptop module in Hardy - this is overriding the sonypi module that handles the hotkeys. I suggest as an interim workaround to black list the sony-laptop module:

sudo vim /etc/modprobe.d/blacklist

and add the following line:

blacklist sony-laptop

reboot and let me know if this interim workaround helps.

I will investigate why the sony-laptop module isn't working correctly

...or if vim is not you favorite editor do:

sudo gedit /etc/modprobe.d/blacklist

Alternatively, if you do not want to blacklist the sony-laptop module then one may need to add the following lines to your /etc/X11/xorg.conf file to use the Sony Vaio hotkeys with the sony-laptop module:

Section "InputDevice"
        Identifier "Vaio keys"
        Driver "evdev"
        Option "Name" "Sony Vaio Keys"
        Option "XkbLayout" "jp"
        Option "XkbModel" "jp106"
EndSection

Section "ServerLayout"
        ...
        InputDevice "Vaio keys" "SendCoreEvents"
EndSection

c.f.: http://www.linux.it/~malattia/wiki/index.php/Sony-laptop

Thanks for your help.

First work around doesn't work because the module is still loaded. I guess it is a initrd problem but it would break gnome brightness and other things so I guess it is no real option and if I remove sony-laptop it still doesn't work (removing of sonypi and adding it again doesn't help too). Afaik Gutsy has sony-laptop module too.

Second work around doesn't seem to have an effect.

Hi,

since I don't have Sony Vaio TX2 to hand, can you tell me which sequence of keys you are pressing to do the eject with? Hopefully I can examine the driver and also see how it matches with /usr/share/hal/fdi/information/10freedesktop/30-keymap-module-sony-laptop.fdi

Thanks.

I am not sure what you mean. I don't need an FN combination. There is an eject key beside the power button and the multimedia keys.

Hi,

Can you edit /etc/modprobe.d/options:

sudo gedit /etc/modprobe.d/options

and add:

options sony-laptop debug

and reboot. Then press the CD eject key and then send a copy of the dmesg log. This may give me more information to work on.

Thank you.

I am sorry but I got the error "sony_laptop: `' invalid for parameter `debug'".
What might interesting is if the eject buttons of the other Sony laptops still work.

Apologies, my mistake, it should have been:

options sony-laptop debug=1

Hi there,

I've looked at the key code sequence and it appears that the current sony-laptop driver in Hardy does not support it for your model type. What is more curious is that I cannot see how it worked for feisty or gutsy kernels in the sonypi.c driver. Since this is reported as a regression, if you have an older system still installed that works with the eject key, can you do the following on that working system:

sudo gedit /etc/modprobe.d/options

and add

options sonypi verbose=2

reboot and press the key several times and send the dmesg output.

This way I can see where the sonypi and the newer sony-laptop drivers differ in their handling of the eject key.

Thank you.

Output from Gutsy Live CD with verbose=2. It is really interesting that eject even works with a LiveCD :).
Btw. you are right that sony-laptop was used the first time in Hardy.

eject key output:
[ 3316.064000] sonypi: event port1=0x32,port2=0x21
[ 3348.564000] sonypi: event port1=0x32,port2=0x21
[ 3469.364000] sonypi: event port1=0x32,port2=0x21

After looking at the debug sofar, my conclusion is that this is a problem not in the kernel driver sony-laptop, but at a higher level between acpi_fakekey and the desktop.

The facts are that Eject key events are being received by acpid:

1. We see the acpi log message:
 [Sat Feb 2 12:52:31 2008] received event "sony/hotkey SPIC 00000001 0000001b"

..which means that sony-laptop is managing to detect the key and push the correct keycode (SONYPI_EVENT_FNKEY_E (which is 0x1b)) via kernel routine acpi_bus_generate_proc_event().

2. and that acpi_fakekey is pushing out the eject cd code 161 to a higher level.

3. The legacy driver sonypi detects the same port values (0x32,0x21) which the sony-laptop module handles in the same manner, meaning there is not a regression between drivers sonypi and sony-laptop.

I get the bug to the appropriate team that can handle this bug at the appropriate level.

I never saw this kind of hardware either, but maybe you can compare the hal behaviour on a gutsy and hardy live CD? Please boot it, then do

  sudo killall hald
  sudo hald --verbose=yes --daemon=no 2>&1|tee /tmp/hal.log

then let the output settle. Press return a few times, so that you can visually separate the new output. Then press the eject key. Does this generate any new output?

Please press Control-C and attach /tmp/hal.log here (saying which is gutsy and which is hardy). Would be interesting to compare both, I think.

Thanks for your replies.

I have used my current Hardy and Gutsy installation to generate the logs. I hope that they are fine too.

This not a hal thing. The "problem" is in the kernel. I've posted a patch in the bug 217504, which is a duplicate of this one.