Lenovo Yoga 13 touchpad regression

Lenovo Yoga 13 touchpad regression

[H. Peter Anvin]

Hello,

The touchpad on Lenovo Yoga 13 seems to have broken some time between 
3.4 and the current kernel.  The dmesg is full of spewage:

[  256.246481] atkbd serio0: Use 'setkeycodes e03e <keycode>' to make it 
known.
[  256.251473] atkbd serio0: Unknown key released (translated set 2, 
code 0xbe on isa0060/serio0

... about once per second; I am wondering if the touchpad is spewing 
output onto the keyboard channel or if there is something else wrong.

I have tried i8042.nomux and i8042.noloop to no effect.

Unless someone has a better idea I'll try to do a bisect.

	-hpa

Re: Lenovo Yoga 13 touchpad regression

[Kevin Cernekee]

On Mon, May 6, 2013 at 10:25 PM, H. Peter Anvin <hpa@zytor.com> wrote:
> Hello,
>
> The touchpad on Lenovo Yoga 13 seems to have broken some time between 3.4
> and the current kernel.  The dmesg is full of spewage:
>
> [  256.246481] atkbd serio0: Use 'setkeycodes e03e <keycode>' to make it
> known.
> [  256.251473] atkbd serio0: Unknown key released (translated set 2, code
> 0xbe on isa0060/serio0
>
> ... about once per second; I am wondering if the touchpad is spewing output
> onto the keyboard channel or if there is something else wrong.

>From googling around this looks like a Synaptics unit.

Is it being detected by the ALPS driver?  If so, did it use the ALPS
driver in 3.4?

alps.c was changed to identify several new models, but PS/2 touchpad
identification uses "magic knock sequences" which may be prone to
false positives.  This is something we should watch for.

Re: Lenovo Yoga 13 touchpad regression

[Benjamin Tissoires]

On Tue, May 7, 2013 at 7:25 AM, H. Peter Anvin <hpa@zytor.com> wrote:
> Hello,
>
> The touchpad on Lenovo Yoga 13 seems to have broken some time between 3.4
> and the current kernel.  The dmesg is full of spewage:
>
> [  256.246481] atkbd serio0: Use 'setkeycodes e03e <keycode>' to make it
> known.
> [  256.251473] atkbd serio0: Unknown key released (translated set 2, code
> 0xbe on isa0060/serio0
>
> ... about once per second; I am wondering if the touchpad is spewing output
> onto the keyboard channel or if there is something else wrong.
>
> I have tried i8042.nomux and i8042.noloop to no effect.
>
> Unless someone has a better idea I'll try to do a bisect.

On top of Kevin's answer:

I already saw that with Yoga 13 laptops.
The touchpad went into this mode, and Linux was not able to recognize
it anymore.
What "solves" the pb was to stop the machine, then go to the bios (or
UEFI), then the touchpad was back alive when the guy rebooted under
Linux. I bet the bios managed to set the touchpad in the right mode
where Linux can then handle it correctly.

This is not an ideal solution, but at least this may help restoring
your touchpad to a functional state while the correct driver is being
developed.

Anyway, if you are able to do a real bisect (instead of my dirty
trick), that would be great because it would help fixing the problem.

Cheers,
Benjamin

Re: Lenovo Yoga 13 touchpad regression

[H. Peter Anvin]

On 05/06/2013 11:05 PM, Kevin Cernekee wrote:
> On Mon, May 6, 2013 at 10:25 PM, H. Peter Anvin <hpa@zytor.com> wrote:
>> Hello,
>>
>> The touchpad on Lenovo Yoga 13 seems to have broken some time between 3.4
>> and the current kernel.  The dmesg is full of spewage:
>>
>> [  256.246481] atkbd serio0: Use 'setkeycodes e03e <keycode>' to make it
>> known.
>> [  256.251473] atkbd serio0: Unknown key released (translated set 2, code
>> 0xbe on isa0060/serio0
>>
>> ... about once per second; I am wondering if the touchpad is spewing output
>> onto the keyboard channel or if there is something else wrong.
>
>  From googling around this looks like a Synaptics unit.
>
> Is it being detected by the ALPS driver?  If so, did it use the ALPS
> driver in 3.4?
>
> alps.c was changed to identify several new models, but PS/2 touchpad
> identification uses "magic knock sequences" which may be prone to
> false positives.  This is something we should watch for.
>

This is how it is detected:

[    1.843709] psmouse serio1: synaptics: Touchpad model: 1, fw: 8.1, 
id: 0x1e2b1, caps: 0xd00223/0x840300/0x123c00, board id: 1800, fw id: 
1148377
[    1.902569] input: SynPS/2 Synaptics TouchPad as 
/devices/platform/i8042/serio1/input/input3

I think this is the Synaptics driver as opposed to ALPS.

What is really bizarre is that some time today the touchpad 
spontaneously started to work.  However, after a reboot it no longer 
worked again.  This is on 3.8.11.  The keycode spewage continued, however.

	-hpa

Re: Lenovo Yoga 13 touchpad regression

[H. Peter Anvin]

On 05/07/2013 11:43 AM, H. Peter Anvin wrote:
> 
> What is really bizarre is that some time today the touchpad
> spontaneously started to work.  However, after a reboot it no longer
> worked again.  This is on 3.8.11.  The keycode spewage continued, however.
> 

OK, "spontaneously" seems to mean "after going in and out of suspend to
RAM".

	-hpa

Re: Lenovo Yoga 13 touchpad regression

[Dmitry Torokhov]

On Tuesday, May 07, 2013 04:02:40 PM H. Peter Anvin wrote:
> On 05/07/2013 11:43 AM, H. Peter Anvin wrote:
> > What is really bizarre is that some time today the touchpad
> > spontaneously started to work.  However, after a reboot it no longer
> > worked again.  This is on 3.8.11.  The keycode spewage continued, however.
> 
> OK, "spontaneously" seems to mean "after going in and out of suspend to
> RAM".

Does i8042.reset help by any chance?

-- 
Dmitry

Re: Lenovo Yoga 13 touchpad regression

[H. Peter Anvin]

On 05/07/2013 04:19 PM, Dmitry Torokhov wrote:
> On Tuesday, May 07, 2013 04:02:40 PM H. Peter Anvin wrote:
>> On 05/07/2013 11:43 AM, H. Peter Anvin wrote:
>>> What is really bizarre is that some time today the touchpad
>>> spontaneously started to work.  However, after a reboot it no longer
>>> worked again.  This is on 3.8.11.  The keycode spewage continued, however.
>>
>> OK, "spontaneously" seems to mean "after going in and out of suspend to
>> RAM".
> 
> Does i8042.reset help by any chance?
> 

No.

	-hpa