Bug 1387005

added as a question on unix.stackexchange
http://unix.stackexchange.com/questions/317692/synaptics-touchpad-in-rawhide

Touchpad last worked in Fedora 24 workstation on kernel 4.4, and on arch with wayland and libinput and kernel 4.7 and on custom kernel 4.8.1 generated by localmodconfig.

Please see the "reporting touchpad bugs" here for the required information:

https://wayland.freedesktop.org/libinput/doc/latest/reporting_bugs.html

The OS broke. During the numerous re-installations I have found that the touchpad starts working after I close the lid (sleep/hibernate), but I have installed "basic desktop" and that doesn't even launch. On further installations I keep getting "The exact message is: [Errno 5] Input/output error", or the screen goes black, or the network setup throws an error. I'll keep trying, but this is ridiculous.

Created attachment 1212782 [details]
evemu recording

udev information for device:

[root@localhost arctiq]# udevadm info /sys/class/input/event3
P: /devices/platform/i8042/serio1/input/input5/event3
N: input/event3
E: DEVNAME=/dev/input/event3
E: DEVPATH=/devices/platform/i8042/serio1/input/input5/event3
E: ID_BUS=i8042
E: ID_INPUT=1
E: ID_INPUT_HEIGHT_MM=67
E: ID_INPUT_TOUCHPAD=1
E: ID_INPUT_TOUCHPAD_INTEGRATION=internal
E: ID_INPUT_WIDTH_MM=104
E: LIBINPUT_DEVICE_GROUP=11/2/7/1b1:isa0060/serio1
E: LIBINPUT_MODEL_SYNAPTICS_SERIAL_TOUCHPAD=1
E: MAJOR=13
E: MINOR=67
E: SUBSYSTEM=input
E: USEC_INITIALIZED=8445240

vendor model number:
Lenovo yoga 13

content of modalias:

[root@localhost arctiq]# cat /sys/class/dmi/id/modalias
dmi:bvnLENOVO:bvr66CN55WW:bd02/28/2013:svnLENOVO:pn20175:pvrLenovoIdeaPadYoga13:rvnLENOVO:rnINVALID:rvr31900003WIN8STDMLT:cvnLENOVO:ct10:cvrLenovoIdeaPadYoga13:

physical dimensions (width*height):
105 mm * 70 mm

Since the buttons aren't working either:
The entire touchpad is clickable.
And I normaly use clickfinger.

It is very simple and consistent to enable the touchpad by suspending. But I see no difference in ps, xinput, libinput, lsmod... Is there anything I can check before/after?

dmesg or the Xorg.log should have have some information. You can get the latter by running journalctl _COMM=gdm-x-session

when you run evemu-record, do you see events coming out of the touchpad when it's not working? If not, then the issue is in the kernel, otherwise it's in libinput somewhere

Like I said neither xev nor evemu are detecting any events.
The evemu recording is actually included as an attachment.
(it's a recording of pressing, sliding, scrolling, but no events)

I've actually switched OS, but I can reinstall and run the command if it would help.Would it?

Oh and I ran dmesg and saw pretty much nothing appart from an enormous spam of atkbd serio0: keypress unknown (or something like that). But on arch the touchpad works but atkbd spam is still there, generated 4 times a second.

right, if evemu doesn't see events, that's definitely a kernel bug.

If suspend/resume fixes the issue, I'll really need the full dmesg from boot containing a suspend/resume.

My guess is that the touchpad gets accessed by an other bus at boot but fails to init through this bus leading to a dead touchpad. Once suspend/resume is triggered, the PS/2 bus re-init the touchpad, and makes it working again.

Sure thing I'll send it tomorrow.
The touchpad did initialize randomly on boot (to installer), roughly every ~30th restart.

Created attachment 1215865 [details]
dmesg output

dmesg output on a fresh install

Created attachment 1215873 [details]
usb related dmesg output

I have the problem that after a fresh reboot the touchpad is working.
After suspend resume the touchpad does not work anymore

I attached the output of: dmesg | grep usb > dmesg-usb-issue.txt

Maybe this is related is a related issue:sometimes after suspend/resume the input via the keyboard sometimes seems to be unreliable: it seems that some strokes to keys are randomly not entering the system.

Thanks for the logs.

I don't think the USB issues are related to the problem (seems that they happen on a different hub/device than the touchscreen one).

I am not entirely sure now on when the problem happens: is it working on boot, but not after suspend/resume, or the other way around?

Dumping some random notes so that when I come back later I'll have them here:

- the laptop is from 2013, the series just before the Thinkpad T440s (one earlier generation of processor).
- the capabilities 0x0c doesn't contain the intertouch bit, so it's unlikely that some other bus triggered an issue.
- so the device is PS/2 only, but an Image Sensor generation
- it might be worth checking if the various settings for PS/2 might actually do something, if not, we need to ask Synaptics if it's a known FW issue
- could be also a HW failure

I'll try to sort out later which PS/2 bits can be tested, depending on the answer of the question earlier in this comment.

For me: the touchpad is working after boot, and not after suspend/resume (that is usually not, but sometimes it is).