Bug 101300

Bad news here, the jump isn't really detectable. so I think you'll have to live with the warnings in the log. 

In your evemu output, this one is the one that causes the jump:
  E: 3.805534 0003 0035 2627      # EV_ABS / ABS_MT_POSITION_X    2627
  E: 3.805534 0003 0036 4463      # EV_ABS / ABS_MT_POSITION_Y    4463
  E: 3.805534 0003 003a 0066      # EV_ABS / ABS_MT_PRESSURE      66
  E: 3.805534 0003 0000 2627      # EV_ABS / ABS_X                2627
  E: 3.805534 0003 0001 4463      # EV_ABS / ABS_Y                4463
  E: 3.805534 0003 0018 0066      # EV_ABS / ABS_PRESSURE         66
  E: 3.805534 0000 0000 0000      # ------------ SYN_REPORT (0) ---------- +14ms

This is 3 events after one at 4330/2314. The only possible indication about the jump is the presence of three pressure-only events, but I'm not sure that's reliable enough. Unless there are other side-effects, living with the warning in the log is the simplest solution here.

Please do attach the output of the touchpad-edge-detector though, this device doesn't have axis ranges set so we need to override it.

Sure :)
# touchpad-edge-detector 105x70 /dev/input/event7 
Touchpad SynPS/2 Synaptics TouchPad on /dev/input/event7
Move one finger around the touchpad to detect the actual edges
Kernel says:	x [1278..5664], y [1206..4648]
Touchpad sends:	x [1285..5665], y [1207..4649] /^C[F

Touchpad has no resolution, size unknown
User-specified touchpad size: 105x70mm
Calculated ranges: 4380/3442

Suggested udev rule:
# <Laptop model description goes here>
evdev:name:SynPS/2 Synaptics TouchPad:dmi:bvnDellInc.:bvr01.00.05:bd03/01/2017:svnDellInc.:pnInspiron157000Gaming:pvr:rvnDellInc.:rn065C71:rvrX02:cvnDellInc.:ct10:cvr:*
 EVDEV_ABS_00=1285:5665:42
 EVDEV_ABS_01=1207:4649:49
 EVDEV_ABS_35=1285:5665:42
 EVDEV_ABS_36=1207:4649:49

thanks, I'll get that into systemd upstream. Please also attach your dmesg so benjamin can have a look why this is needed in the first place

Created attachment 131765 [details]
dmesg output

Attached

(In reply to Pacho Ramos from comment #4)
> Created attachment 131765 [details]
> dmesg output
> 
> Attached

Thanks. However, I do not see what could be wrong besides the firmware not providing the correct resolution :(

Given that you are on gentoo, could you give a shot at a kernel v4.12-rcX and boot it with "psmouse.synaptics_intertouch=1" appended to the kernel boot line to see if your device works with RMI4 over SMBus, and if this solves the issue.

Are you sure it's the dmesg output the want you need? I mean, the errors are shown in "journalctl -e" output (coming from X I think)

I can try, of course :)

(In reply to Pacho Ramos from comment #6)
> Are you sure it's the dmesg output the want you need? I mean, the errors are
> shown in "journalctl -e" output (coming from X I think)

Yes, I am interested in the dmesg. We know why there are errors in X or Wayland: the kernel sends garbage (or at least an unconfigured device), so I am interested in what the kernel exports. And for that, dmesg is the place to look for errors or debug information.

> I can try, of course :)

Thanks!

Created attachment 131778 [details]
updated dmesg with psmouse tweak

This is the updated dmesg output... for now I don't see any errors related with the jumping in logs :/

Please also enable MOUSE_PS2_SYNAPTICS_SMBUS and RMI4_SMB in your .config (and probably any RMI4_* option).

Created attachment 131791 [details]
updated dmesg with psmouse tweak

this is the updated one

Created attachment 131792 [details]
/proc/config.gz

I was confused why the touchpad doesn't appear on the SMBus.

And then I realized that you have CONFIG_I2C_HID not set. Please turn this on, and the touchpad should be handled by HID now.

Then, you should at least enable CONFIG_HID_MULTITOUCH, and CONFIG_HID_RMI for this touchpad to be handled properly.

Then, you should probably review trhe config of most HID modules (especially CONFIG_HID_LOGITECH_DJ and CONFIG_HID_LOGITECH_HIDPP) if you ever want to plug an input device that we support in a special way in the kernel. I suspect you might also be interested in CONFIG_HID_SENSOR_HUB in case your laptop exports some accelerometer/light sensor.

Created attachment 131793 [details]
/proc/config.gz

updated config

Created attachment 131794 [details]
updated dmesg with psmouse tweak

updated one

...and you probably also need CONFIG_I2C_DESIGNWARE_PLATFORM, CONFIG_I2C_DESIGNWARE_PCI to have a driver for the I2C controller that can handle the touchpad.

Created attachment 131799 [details]
updated dmesg with psmouse tweak

Created attachment 131800 [details]
/proc/config.gz

I feel a littlbe bit like playing whack a mole here.

Could you just to be sure download a livecd of Fedora 25 and see if your touchpad behave the same?
If it is handled by hid-multitouch, you probably want to compare the loaded modules from the livecd to yours under gentoo.

Also, just to be sure that it's a HID touchpad, could you attach the output of acpidump (as root)?

Anyway, since running kernel 4.12-rc4 with this boot options I don't see the "jump" error in logs :/

Created attachment 131888 [details]
acpidump

(In reply to Pacho Ramos from comment #19)
> Anyway, since running kernel 4.12-rc4 with this boot options I don't see the
> "jump" error in logs :/

that's a good sign, unless you noticed a pointer jump?

But without that boot options, I still see the messages... anyway, the messages don't show any issues with a normal usage, I only reported because I saw the errors in journalctl output and wanted to let you know :/

for the 4.12 kernel, the boot option is the default for your device. so once you update to 4.12, everything should work fine :)

OK, thanks a lot :)