UNSOLVED

Xps13-9350-user

updated

10 years ago

X9

Xps13-9350-user

1 Rookie

•

11 Posts

1

67346

November 9th, 2015 19:00

xps9350: touchscreen stops working after sleep

Running a 4.3 kernel: touchscreen works great on boot, but fails after sleep/resume. I've tried the following to no avail:

 * rmmod hid_multitouch  && sleep 5 && modprobe hid_multitouch

This does reveal in /var/log/syslog:

Nov 9 22:44:16 morel kernel: [32441.878773] input: ELAN Touchscreen as /devices/pci0000:00/0000:00:14.0/usb1/1-4/1-4:1.0/0003:04F3:20D0.0001/input/input34
Nov 9 22:44:16 morel kernel: [32441.932623] hid-multitouch 0003:04F3:20D0.0001: input,hiddev0,hidraw0: USB HID v1.10 Device [ELAN Touchscreen] on usb-0000:00:14.0-4/input0

-- no errors visible, but the touchscreen simply doesn't work.

Thoughts on how I can debug or report further?

  • agb32

    11 Posts

    3193

    0

    Posted April 15th, 2016 06:00

    Same problem for me - on 16.04.

  • krolaw

    13 Posts

    3193

    0

    Posted April 24th, 2016 14:00

    Apparently (I haven't tested), your fix works with kernel 4.6.  See bbs.archlinux.org/viewtopic.php

  • z_tmr

    2 Posts

    3193

    0

    Posted April 27th, 2016 15:00

    After some experiments, this did work for me:

    $ sudo inputattach --daemon -w8001 /dev/ttyS4

    ...what I don't understand, because my ELAN Touchscreen is definitely not on serial port so there should not be reason to use inputattach at all.

    The problem with this is that inputattach process gets D status (uninterruptible sleep) so it makes the suspend/resume to take about 20 seconds (which is some timeout) and sometimes it does not work at all. Since you cannot kill this proccess, it is quite sad and it might imply any other issues eventually.

    So this is not solution but might be some indice to find out what the real problem is. I also see this:

    $ dmesg | grep -ie elan
    [ 7.637031] usb 1-4: Manufacturer: ELAN
    [ 10.101754] input: ELAN Touchscreen as /devices/pci0000:00/0000:00:14.0/usb1/1-4/1-4:1.0/0003:04F3:20D0.0001/input/input12
    [ 10.101852] hid-multitouch 0003:04F3:20D0.0001: input,hiddev0,hidraw0: USB HID v1.10 Device [ELAN Touchscreen] on usb-0000:00:14.0-4/input0
    [ 698.403796] input: ELAN Touchscreen as /devices/pci0000:00/0000:00:14.0/usb1/1-4/1-4:1.0/0003:04F3:20D0.0001/input/input18
    [ 698.404299] hid-multitouch 0003:04F3:20D0.0001: input,hiddev0,hidraw0: USB HID v1.10 Device [ELAN Touchscreen] on usb-0000:00:14.0-4/input0

    ...I wonder why it is input18 instead of input12 after first resume (might change after another suspend/resume?!). So maybe, there is something waiting for input12 but it has changed to input18 while something expects it on 12? I don't know.

    That's just my two cents, and I am hefeby joining the question how to fix this completely and properly :)

  • agb32

    11 Posts

    3193

    0

    Posted June 17th, 2016 02:00

    Yes, I confirm that works for me too.  With the 4.4 kernel, 16.04.

  • krolaw

    13 Posts

    3193

    0

    Posted June 17th, 2016 03:00

    Awesome find :-)  Works well.

  • z_tmr

    2 Posts

    3193

    0

    Posted June 17th, 2016 17:00

    Wonderful! I did it very quickly so that I didn't have to press the power button and it just locked screen and started to work again. Given this, we may be able to do some kernel debugging now...

  • Sydius

    5 Posts

    3193

    0

    Posted June 29th, 2016 18:00

    Yeah, the quick lid close works for me as well. Weird.

    Any luck debugging this further?

  • 3193

    0

    Posted July 4th, 2016 00:00

    Can also confirm same workaround works. Don't even have to close the lid fully, just move it down a bit then up - the display doesn't even go dark - and touch is working again.

  • Sydius

    5 Posts

    1035

    0

    Posted July 4th, 2016 19:00

    This seems to still be an issue with the latest 1.4.4 firmware. It's not that big of a deal to me given the workaround of closing it and then opening it again right away seems to fix it.

  • cnshenyi

    1 Message

    1036

    0

    Posted September 4th, 2016 11:00

    Did you solve the problem?