support for some types of Asus+Elantech touchpads

support for some types of Asus+Elantech touchpads

[Wladimir Mutel]

Dear Dmitry & linux-input members,

some new Asus (and probably non-Asus) laptops are shipped with a new 
kind of Elantech (not sure if only Elantech) touchpads which are not 
adequately handled by Linux kernel module (psmouse). First problem is 
that these touchpads are dying off (losing their sensitivity and 
response) when the system is put into pm-suspend and then waken up. 
Another problem is that on reload (modprobe -r ... && modprobe ...) of 
psmouse module they are dying off too, in 'dmesg' coming with some 
weird-numbered indexes (like, input5 on initial load and input292 on 
reload of psmouse).

There are some bugs reported in Ubuntu Launchpad on this topic 
(https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519034 is my own, 
also
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1490130 , 
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1506363 , 
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1504445 and some 
others).
Also there is a proposed patch at
https://marc.info/?l=linux-input&m=144312209020616
and one more report at
http://www.spinics.net/lists/linux-input/msg41634.html

For me, these problems look related. I have Asus X302LA on which I am 
testing Ubuntu Xenial and whose touchpad exhibits this problem. If you 
need I could provide my assistance in testing psmouse module on this 
hardware. Thank you in advance for your response.

Re: support for some types of Asus+Elantech touchpads

[Dmitry Torokhov]

Hi Wladimir,

On Tue, Dec 01, 2015 at 09:59:28PM +0200, Wladimir Mutel wrote:
> Dear Dmitry & linux-input members,
> 
> some new Asus (and probably non-Asus) laptops are shipped with a new
> kind of Elantech (not sure if only Elantech) touchpads which are not
> adequately handled by Linux kernel module (psmouse). First problem
> is that these touchpads are dying off (losing their sensitivity and
> response) when the system is put into pm-suspend and then waken up.
> Another problem is that on reload (modprobe -r ... && modprobe ...)
> of psmouse module they are dying off too, in 'dmesg' coming with
> some weird-numbered indexes (like, input5 on initial load and
> input292 on reload of psmouse).
> 
> There are some bugs reported in Ubuntu Launchpad on this topic
> (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519034 is my
> own, also
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1490130 ,
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1506363 ,
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1504445 and
> some others).
> Also there is a proposed patch at
> https://marc.info/?l=linux-input&m=144312209020616
> and one more report at
> http://www.spinics.net/lists/linux-input/msg41634.html
> 
> For me, these problems look related. I have Asus X302LA on which I
> am testing Ubuntu Xenial and whose touchpad exhibits this problem.
> If you need I could provide my assistance in testing psmouse module
> on this hardware. Thank you in advance for your response.

Does the noselftest patch help on your box (you'd have to adjust
i8042_dmi_noselftest_table with the DMI data for your laptop)?

Thanks.

-- 
Dmitry

Re: support for some types of Asus+Elantech touchpads

[Wladimir Mutel]

On Tue, Dec 01, 2015 at 01:08:52PM -0800, Dmitry Torokhov wrote:

> > Also there is a proposed patch at
> > https://marc.info/?l=linux-input&m=144312209020616

> > For me, these problems look related. I have Asus X302LA on which I
> > am testing Ubuntu Xenial and whose touchpad exhibits this problem.
> > If you need I could provide my assistance in testing psmouse module
> > on this hardware. Thank you in advance for your response.

> Does the noselftest patch help on your box (you'd have to adjust
> i8042_dmi_noselftest_table with the DMI data for your laptop)?

I applied this patch to Ubuntu kernel sources v 4.2.0-19, then put X302LA model code into DMI list,
then rebuilt generic kernel and reinstalled it over Ubuntu packages
After rebooting with this patch, the touchpad now survives pm-suspend well. So it is effective in my case.
It still does not survive reload of psmouse module, coming as .../input5 in initial load, as .../input15 on reload, and then dying off.
But now I don't need this workaround so I can live without that. 
Not sure if any other hardware is supposed to survive psmouse module reload at all.