Elan touchpad stopped working, I fixed it but...

[Zucca]

Long story short: My Huawei MateBook D touchpad stopped working while I was using my laptop. I then thought it was a driver issue and rebooted. It didn't help. At this point I was sure this was a hardware failure. I haven't made any updates etc because we just moved to a new house and I have tons of other things to do. So hardware issue right? Maybe... BUT unloading and loading i2c_hid kernel module brought the touchpad back to life.

I can't think of any reason why this is like it is. Anyone have any guesses?

I'll be sure to save all the kernel messages if this happens again.

[roccobaroccoSC]

I disable the touchpad sometimes by accident when pressing the wrong keys (Fn+F7 on my machine). Check if you disabled it unintentionally.

[Zucca]

There's no quick key to disable touchpad on my laptop, afaik. Fn+F7 disables microphone.

But the touchpad really gave some error messages before dying.

[Zucca]

It happened again.

• These messages started to appear on dmesg:

  Code: Select all

  i2c_hid i2c-ELAN2204:00: i2c_hid_get_input: incomplete report (14/65535)

  And then i did remove/unload i2c_hid and then loaded it again, but then:

  Code: Select all

  i2c_hid i2c-ELAN2204:00: supply vdd not found, using dummy regulator
  i2c_hid i2c-ELAN2204:00: supply vddl not found, using dummy regulator
  i2c_hid i2c-ELAN2204:00: failed to reset device.
  i2c_hid i2c-ELAN2204:00: failed to reset device.
  i2c_hid i2c-ELAN2204:00: failed to reset device.
  i2c_hid i2c-ELAN2204:00: failed to reset device.
  i2c_hid i2c-ELAN2204:00: can't add hid device: -61
  i2c_hid: probe of i2c-ELAN2204:00 failed with error -61

I guess compiling i2c_hid into kernel would only worsen things.
But also I just realized this "hangup" happens only short time after resuming from suspend. So the very ugly solution would be to unload the module before going to suspend...

[allexj]

I have the SAME issue and the SAME laptop.

It happened again.
So the very ugly solution would be to unload the module before going to suspend...

So you mean to create a suspend hook (to unload i2c_hid) and resume hook(to reload i2c_hid)?
Also, modprobe doesn't let me unload i2c_hid... I have to unload i2c_hid_acpi before, and then I can remove i2c_hid. Do you have the same scenario?

[Zucca]

Hi, allexj!

So you mean to create a suspend hook (to unload i2c_hid) and resume hook(to reload i2c_hid)?

Yes, that could solve the issue.

Also, modprobe doesn't let me unload i2c_hid... I have to unload i2c_hid_acpi before, and then I can remove i2c_hid. Do you have the same scenario?

I don't have i2c_hid_acpi, however.
Which kernel version you run?

Code: Select all

# uname -rp
5.10.61-huawei-matebook-d-amd AMD Ryzen 5 3500U with Radeon Vega Mobile Gfx

I bet we're not alone and this indicates that this isn't a hardware problem.

Oh, and welcome to Gentoo!

[allexj]

Which kernel version you run?

Oh, and welcome to Gentoo!

Hi! Thanks for welcome, but to be honest I am running ArchLinux. I searched for this issue and I noticed this thread so I commented.

Code: Select all

$ uname -rp
5.14.16-arch1-1 unknown

Doing as you suggested (creating sleep and resume hooks) seem to have fixed the issue!

If you want more resources about the bug, here there are some links:
https://www.reddit.com/r/linuxhardware/ ... bernation/
https://bugzilla.kernel.org/show_bug.cgi?id=214749

Thanks for the suggestion, I'll let you know if the bug happens again

[Zucca]

So let's wait for the fix, I guess.

Darn. I'm not very eager to test out bleeding edge kernels, but this time, I guess, I have to when the time comes.