Bug 200663 - ELAN 1200 Tochpad looses connection on ASUS GL703GE (rmmod&&insmod hid-multitouch helps)
Summary: ELAN 1200 Tochpad looses connection on ASUS GL703GE (rmmod&&insmod hid-multit...
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Input Devices (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: drivers_input-devices
URL:
Keywords:
Depends on: 199911
Blocks:
  Show dependency tree
 
Reported: 2018-07-26 22:37 UTC by Marc Landolt
Modified: 2018-10-24 23:01 UTC (History)
5 users (show)

See Also:
Kernel Version: 4.19
Tree: Mainline
Regression: No


Attachments
Single Click fires Multiple Clicks on GL703GE (2.13 KB, text/plain)
2018-07-26 22:37 UTC, Marc Landolt
Details
last entries before the Touchpad stops working from Asus GL703GE (1.79 KB, text/plain)
2018-07-26 23:26 UTC, Marc Landolt
Details
attachment-1818-0.html (2.06 KB, text/html)
2018-09-04 20:34 UTC, Andrey Ivanov
Details
attachment-22891-0.html (2.00 KB, text/html)
2018-09-05 19:59 UTC, Andrey Ivanov
Details
attachment-23475-0.html (1.85 KB, text/html)
2018-09-05 20:02 UTC, Andrey Ivanov
Details
attachment-3227-0.html (2.01 KB, text/html)
2018-09-05 20:57 UTC, Andrey Ivanov
Details
attachment-23319-0.html (862 bytes, text/html)
2018-09-05 22:27 UTC, Andrey Ivanov
Details
attachment-23775-0.html (1.37 KB, text/html)
2018-09-05 22:30 UTC, Andrey Ivanov
Details
attachment-26883-0.html (2.59 KB, text/html)
2018-09-06 20:49 UTC, Andrey Ivanov
Details
attachment-4769-0.html (4.16 KB, text/html)
2018-09-06 21:31 UTC, Andrey Ivanov
Details
Benjamin Tissores / ??? patch (make output) (3.42 KB, text/plain)
2018-09-06 21:46 UTC, Marc Landolt
Details
attachment-10237-0.html (1.13 KB, text/html)
2018-09-06 21:55 UTC, Andrey Ivanov
Details
attachment-27498-0.html (2.76 KB, text/html)
2018-09-07 21:17 UTC, Andrey Ivanov
Details
attachment-30864-0.html (1.24 KB, text/html)
2018-09-07 21:25 UTC, Andrey Ivanov
Details
before freeze (10.76 KB, text/plain)
2018-09-10 14:08 UTC, Marc Landolt
Details
after freeze (10.76 KB, text/plain)
2018-09-10 14:08 UTC, Marc Landolt
Details
5 minutes later (10.76 KB, text/plain)
2018-09-10 14:09 UTC, Marc Landolt
Details
dmesg (125.79 KB, text/plain)
2018-09-10 14:09 UTC, Marc Landolt
Details
before freeze 4.19-rc2 (10.76 KB, text/plain)
2018-09-10 14:16 UTC, Marc Landolt
Details
after freeze 4.19-rc2 (10.76 KB, text/plain)
2018-09-10 14:16 UTC, Marc Landolt
Details
5 minutes later 4.19-rc2 (10.76 KB, text/plain)
2018-09-10 14:17 UTC, Marc Landolt
Details
dmesg 4.19-rc2 (78.23 KB, text/plain)
2018-09-10 14:17 UTC, Marc Landolt
Details

Description Marc Landolt 2018-07-26 22:37:25 UTC
Created attachment 277563 [details]
Single Click fires Multiple Clicks on GL703GE

Hello

My ELAN1200:00 04F3:3090 i2c touchpad on my Asus GL703GE clicks multiple with a single click on the button (that is the complete area of the Touchpad) and with two-finger click the same problem.

Workaround is "Tap to click" but maybe this should be fixed, I'll attach the output of a single click of hid-recorder.

Second Problem is that on Asus GL703GE and GL503VS (and maybe other similar devices) the touchpad looses connection and you must rmmod hid-multitouch && insmod hid-multitouch that it works again.

Any help would be nice.

With kind regards
Marc Landolt
Comment 1 Marc Landolt 2018-07-26 23:10:44 UTC
I forgot to say, that there was already a problem with the i2c gpio's on cannon-lake that was fixed here: https://bugzilla.kernel.org/show_bug.cgi?id=199911#c68
Comment 2 Marc Landolt 2018-07-26 23:26:46 UTC
Created attachment 277565 [details]
last entries before the Touchpad stops working from Asus GL703GE

recorded with hid-recorder, the hid-recorder does not record anything until I make a rmmod && modprobe hid-multitouch, also restarting the hid-recorder does not help, so it seems to loose connection
Comment 3 Marc Landolt 2018-07-27 00:54:15 UTC
Anther bit of info:

i can now REPRODUCE THE FREEZE: if you put 5 fingers on the touchpad at once then in freezes always.
Comment 4 Andrey Ivanov 2018-08-27 18:06:16 UTC
You are forget it!

When system is booting, touchpad elan1200 is work perfectly. But in log i can see warnings "Incomplete report 16/2^16".

Via 2 minutes active using touchpad, touchpad lost connections and absolutely not work.

I can monitoring it by xinput --list.

Please solve it!

Btw, i see if i using nouveau touchpad absolutely not work. I poweroff all nouveau modules.
And using intel graphics.

But its not solve crash work of touchpad.

I think problem in ACPI, and IRQ, and in gpio control.. 
Yeah.. please, fix it.. very more peoples say thanks.

My hardware:
Asus strix GL703VD.
Touchpad HID over I2C.
Kernel 4.18.5.
Btw, in kerbel 3.10 touchpad work stable and perfectly, but have very low resolution..
I test it in CENTOS live sd. In centos kernel is 3.10.

My new system is Gentoo, kernel is 4.18.5 from Vanilla-Sources(kernel.org).
Because i can not get work touchpad in Gentoo-Sources (with patches by gentoo team).

So.. please solve lost connections and inconplete report..

Thanks.
Comment 5 Marc Landolt 2018-08-27 20:06:47 UTC
(In reply to Search_Terminal from comment #4)
> You are forget it!
> 
> When system is booting, touchpad elan1200 is work perfectly. But in log i
> can see warnings "Incomplete report 16/2^16".
> 
> Via 2 minutes active using touchpad, touchpad lost connections and
> absolutely not work.
> 
> I can monitoring it by xinput --list.
> 
> Please solve it!
> 
> Btw, i see if i using nouveau touchpad absolutely not work. I poweroff all
> nouveau modules.
> And using intel graphics.
> 
> But its not solve crash work of touchpad.
> 
> I think problem in ACPI, and IRQ, and in gpio control.. 
> Yeah.. please, fix it.. very more peoples say thanks.
> 
> My hardware:
> Asus strix GL703VD.
> Touchpad HID over I2C.
> Kernel 4.18.5.
> Btw, in kerbel 3.10 touchpad work stable and perfectly, but have very low
> resolution..
> I test it in CENTOS live sd. In centos kernel is 3.10.
> 
> My new system is Gentoo, kernel is 4.18.5 from Vanilla-Sources(kernel.org).
> Because i can not get work touchpad in Gentoo-Sources (with patches by
> gentoo team).
> 
> So.. please solve lost connections and inconplete report..
> 
> Thanks.

Hello Search_Terminal

this is not the Bugreport for non-working touchpad this is the bugreport if the touchpad already works but disconnects and fires multiple clicks, to get it running use this thread: https://bugzilla.kernel.org/show_bug.cgi?id=199911.

if you have the same hardware like i have in my GL793GE then you must apply this two patches to get it working: https://bugzilla.kernel.org/show_bug.cgi?id=199911#c68

And otherwise kindly ask Andy Shevchenko in the other thread, he helped me fixing the GPIO stuff

with kind regards
Marc
Comment 6 Andy Shevchenko 2018-09-04 19:38:21 UTC
Can you test ig bug #200899 has a solution for your case?
Comment 7 Andrey Ivanov 2018-09-04 20:34:02 UTC
Created attachment 278295 [details]
attachment-1818-0.html

вт, 4 сент. 2018 г., 22:38 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #6 from Andy Shevchenko (andy.shevchenko@gmail.com) ---
> Can you test ig bug #200899 has a solution for your case?
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.



Please, tell me, how i can test it?

Please, you should to know. My work notebook(laptop) needed for me work and
important.
I can test driver only insmod method, okay? Because for me critical
important have a work laptop.

Please, write for me instructions how to test.


Hmm..

Please, tell developers (cc) about beta-test of driver "for vivien" with my
patch (comment //no_empty_report).
Its solve errors in log of kernel and solve problem of log 2Gb per 3day,
and improve work touchpad.

Please, see last my comments on bugzilla.
Its important, because Asus GL703 will be popular model of hight perf
laptops.

If tests will be good, need to add this code to vanilla-sources of kernel
linux. Its solve problem with elan1200.
Comment 8 Marc Landolt 2018-09-05 19:50:22 UTC
hello @andy shevchenko

i tested with kernel 4.19-rc2 and there is again this GPIO Problem

[   13.957960] cannonlake-pinctrl INT3450:00: pin 225 cannot be used as IRQ
[   13.957962] genirq: Setting trigger mode 8 for irq 138 failed (intel_gpio_irq_type+0x0/0x140)
[   13.957993] i2c_hid i2c-ELAN1200:00: Could not register for ELAN1200:00 interrupt, irq = 138, ret = -1
[   13.958016] i2c_hid: probe of i2c-ELAN1200:00 failed with error -1


and i can not apply the patch you wrote for kernel 4.18-rc6, so i can not test with kernel >= 4.19-rc1

will this patch you wrote for me end up in mainline or not?

and i could try to test the patch from #200899 with 4.18-rc6 where your patch fixed the gpio stuff...

with kind regards
Marc
Comment 9 Andrey Ivanov 2018-09-05 19:59:06 UTC
Created attachment 278319 [details]
attachment-22891-0.html

I absolutely not patch kernel.. i build 4.18.5.. after it i get from github
driver "i2c hid for vivien", improve error, build, and use insmod. Only it.

Before use Benjamin's version driver my processor have 1-2% load for irq.
Its craizy..
Benjamin's version work perfect but i test only yesterday and fast..
I today no have power continue tests..

ср, 5 сент. 2018 г., 22:50 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #8 from Marc Landolt (2009@marclandolt.ch) ---
> hello @andy shevchenko
>
> i tested with kernel 4.19-rc2 and there is again this GPIO Problem
>
> [   13.957960] cannonlake-pinctrl INT3450:00: pin 225 cannot be used as IRQ
> [   13.957962] genirq: Setting trigger mode 8 for irq 138 failed
> (intel_gpio_irq_type+0x0/0x140)
> [   13.957993] i2c_hid i2c-ELAN1200:00: Could not register for ELAN1200:00
> interrupt, irq = 138, ret = -1
> [   13.958016] i2c_hid: probe of i2c-ELAN1200:00 failed with error -1
>
>
> and i can not apply the patch you wrote for kernel 4.18-rc6, so i can not
> test
> with kernel >= 4.19-rc1
>
> will this patch you wrote for me end up in mainline or not?
>
> and i could try to test the patch from #200899 with 4.18-rc6 where your
> patch
> fixed the gpio stuff...
>
> with kind regards
> Marc
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 10 Andrey Ivanov 2018-09-05 20:02:33 UTC
Created attachment 278321 [details]
attachment-23475-0.html

You can not use my patch because i wrote not patch.. i wrote example of
code, and told about error, need just comment one line..
Because without it modules not compile..
Coming soon i has been create github overlay for all (for use benjamin's
version driver).

ср, 5 сент. 2018 г., 22:50 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #8 from Marc Landolt (2009@marclandolt.ch) ---
> hello @andy shevchenko
>
> i tested with kernel 4.19-rc2 and there is again this GPIO Problem
>
> [   13.957960] cannonlake-pinctrl INT3450:00: pin 225 cannot be used as IRQ
> [   13.957962] genirq: Setting trigger mode 8 for irq 138 failed
> (intel_gpio_irq_type+0x0/0x140)
> [   13.957993] i2c_hid i2c-ELAN1200:00: Could not register for ELAN1200:00
> interrupt, irq = 138, ret = -1
> [   13.958016] i2c_hid: probe of i2c-ELAN1200:00 failed with error -1
>
>
> and i can not apply the patch you wrote for kernel 4.18-rc6, so i can not
> test
> with kernel >= 4.19-rc1
>
> will this patch you wrote for me end up in mainline or not?
>
> and i could try to test the patch from #200899 with 4.18-rc6 where your
> patch
> fixed the gpio stuff...
>
> with kind regards
> Marc
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 11 Marc Landolt 2018-09-05 20:08:31 UTC
no i mean the patch andy shevchenko wrote here for me:

https://bugzilla.kernel.org/show_bug.cgi?id=199911#c68

this patch helped that touchpad at least works but with disconnects and multiple clicks when you make a single click

but after that patch the disconnection problem was still there, it just solved the gpio problem...

with kind regards
Marc Landolt
Comment 12 Andrey Ivanov 2018-09-05 20:57:15 UTC
Created attachment 278329 [details]
attachment-3227-0.html

Hmm.. its good, if you solve not correct clicking..
(i only one not understand.. why need 140Mhz freq for toucpad under i2c by
microsoft spec..? )
So.. you have disconnect problem.. its not good.. how are you solve it?
Re-load module?

Can you test centos 3.10 kernel, maybe if your touchpad work well, maybe
developers can to imporove driver by old code.?

Sometimes my touchpad bad recognition events. After click it use sometimes
click left when need to use just right click..

Yeah..
I hope this problems will be solve step by step..
What the your laptop model?

ср, 5 сент. 2018 г., 23:08 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #11 from Marc Landolt (2009@marclandolt.ch) ---
> no i mean the patch andy shevchenko wrote here for me:
>
> https://bugzilla.kernel.org/show_bug.cgi?id=199911#c68
>
> this patch helped that touchpad at least works but with disconnects and
> multiple clicks when you make a single click
>
> but after that patch the disconnection problem was still there, it just
> solved
> the gpio problem...
>
> with kind regards
> Marc Landolt
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 13 Marc Landolt 2018-09-05 21:01:01 UTC
@Search Terminal

Asus GL703GE
Comment 14 Andrey Ivanov 2018-09-05 22:27:31 UTC
Created attachment 278339 [details]
attachment-23319-0.html

Hmmm.. I guess we have absolutely analogy hardware?
Hmm..

чт, 6 сент. 2018 г., 0:01 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #13 from Marc Landolt (2009@marclandolt.ch) ---
> @Search Terminal
>
> Asus GL703GE
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 15 Andrey Ivanov 2018-09-05 22:30:30 UTC
Created attachment 278341 [details]
attachment-23775-0.html

Ouw.. your hardware greater than me  ahah.
Btw, when you test linux, did you see "processor lock, get stuck" on
various distributives of Gnu/Linux?

чт, 6 сент. 2018 г., 1:28 Armv7hf <armv7hf@gmail.com>:

> Hmmm.. I guess we have absolutely analogy hardware?
> Hmm..
>
> чт, 6 сент. 2018 г., 0:01 <bugzilla-daemon@bugzilla.kernel.org>:
>
>> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>>
>> --- Comment #13 from Marc Landolt (2009@marclandolt.ch) ---
>> @Search Terminal
>>
>> Asus GL703GE
>>
>> --
>> You are receiving this mail because:
>> You are on the CC list for the bug.
>
>
Comment 16 Marc Landolt 2018-09-05 23:49:55 UTC
I tested now Kernel 4.19-rc2

+the problem with the multiple clicks seems to be solved
+two finger scroll works
+right click with 2 Fingers and Pressing on the touchpad that is also a button works also

-it still disconnects if you put 5 fingers on the Touchpad and then you must do a sudo rmmod  hid-multitouch && sudo modprobe hid-multitouch that the touchpad works again

For the ASUS GL703GE Model it seems that you need to manually add the lines 1016, 1021 and 1022 from https://bugzilla.kernel.org/attachment.cgi?id=277507&action=diff to /drivers/pinctrl/intel/pinctrl-intel.c from kernel 4.19-rc2

( This is already reported to the Bug https://bugzilla.kernel.org/show_bug.cgi?id=199911#c102 )

with kind regards
Marc Landolt
Comment 17 Marc Landolt 2018-09-06 20:25:46 UTC
Hello Search_Terminal

you sent me this private Mail:

> Can you test this driver??
> https://github.com/EthicalEarth/i2c_hid_driver_elan1200
> Just 
> git clone https://github.com/EthicalEarth/i2c_hid_driver_elan1200
> make
> rmmod i2c_hid
> insmod ./i2c_hid.ko
> -----------------------------------
> This no need install and replace for test.
> Just use insmod. 
> Can you tell me how its work on your PC?
> -----------------------------------

i can not test it i already use kernel 4.19-rc2, and i dont think that the problem is i2c, the touchpad is working so the i2c is working correctly with the current kernel 4.19-rc2 and the modification described here: https://bugzilla.kernel.org/show_bug.cgi?id=199911#c102

the only problem i have is that it looses connection if i put 5 fingers on the touchpad, everything else is fine... and the "140Mhz freq for toucpad" is as far as i know depending on the i2c bus speed, you need at least the doubled speed of the bus clock that you can interpret the signals correctly as far as i understood.

With kind Regards
Marc Landolt
Comment 18 Andrey Ivanov 2018-09-06 20:49:47 UTC
Created attachment 278353 [details]
attachment-26883-0.html

Okay, did you have in dmesg "incomplete report"?
Why you can not test it in your machine?(
I understand, you patched new kernel and work all good.. but can you just
load module or compile?
Or this not work on 4.19 kernel?(

чт, 6 сент. 2018 г., 23:25 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #17 from Marc Landolt (2009@marclandolt.ch) ---
> Hello Search_Terminal
>
> you sent me this private Mail:
>
> > Can you test this driver??
> > https://github.com/EthicalEarth/i2c_hid_driver_elan1200
> > Just
> > git clone https://github.com/EthicalEarth/i2c_hid_driver_elan1200
> > make
> > rmmod i2c_hid
> > insmod ./i2c_hid.ko
> > -----------------------------------
> > This no need install and replace for test.
> > Just use insmod.
> > Can you tell me how its work on your PC?
> > -----------------------------------
>
> i can not test it i already use kernel 4.19-rc2, and i dont think that the
> problem is i2c, the touchpad is working so the i2c is working correctly
> with
> the current kernel 4.19-rc2 and the modification described here:
> https://bugzilla.kernel.org/show_bug.cgi?id=199911#c102
>
> the only problem i have is that it looses connection if i put 5 fingers on
> the
> touchpad, everything else is fine... and the "140Mhz freq for toucpad" is
> as
> far as i know depending on the i2c bus speed, you need at least the doubled
> speed of the bus clock that you can interpret the signals correctly as far
> as i
> understood.
>
> With kind Regards
> Marc Landolt
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 19 Marc Landolt 2018-09-06 21:04:40 UTC
hello Search_terminal

my problem is, if i put 5 fingers on the touchpad i the touchpad looses connection and i must run rmmod hid-multitouch and modprobe hid-multitouch that it works again, so in my opinion the problem is not in i2c-hid because i dont need to reload the i2c-hid...

but i had a look at dmesg, and it also says i2c_hid i2c-ELAN1200:00: ic2_hid_get_input: incomplete report (16/65535). The Laptop runs now since 24h and it occured 696 times since yesterday.

may i ask where you got this files from, have you made this patch or who wrote this? I diff-ed with mainline kernel i2c_hid.c and there are a lot of changes and I would like to stick to the mainline kernel, because if the problem does not get fixed in mainline i (and all other people that has a ASUS GL703GE) have to patch it each time when i setup a new linux on this machine.

With kind regards
Marc Landolt
Comment 20 Andrey Ivanov 2018-09-06 21:31:26 UTC
Created attachment 278355 [details]
attachment-4769-0.html

Its not secret. Autor of this driver (original
https://github.com/bentiss/hid-multitouch/tree/for-vivien) is Benjamin
Tissoires.
Its his variant of driver i2c_hid for more specific hid via i2c technology.

When Benjamin tell about it, i download and try to test. His driiver not
complile because i2c-hid.c have one non correct line. I just see gcc
message and fix.
After it, i try to use.
And work has been mire better. Because its improve error "incomplete
report".
I think this error have negative effect to work of touchpad because needed
addiitioonal time for send in  log more more errors.

Ifter usinng Beennjamin driver (module  i2c_hid my touchpad feels more
well).

I am just load on github my fixed vaiant for personal use..
Maybe in future Benjanin can too close repositoory, but my copy will be in
network.

I want have more info about work and   test of this driver on similiar asus
laptop, maybe  this driver will be work perfect on your pc(i want have
solution not only for me, but for all with this problem, its can to help
more peoples , if its will be help more peoples its can will be reason for
apply this driver to mainlain kernel, i just guess...)
So, please, if you can, test only module i2c_hid and tell me, if its solve
incomplete report error.
Maybe our touchpad will be work well.
Tell me your experience. Just use insmod.
Its not will be change your system.
But you can tell me how its work on your hardware.

Sorry, i know english very bad, its not my root lang, i write it by my self.
I hope you can to understand me..

Hm.. i wil be happy if this driver will be work fine on two Asus laptops
Gl's series.

Best regars, thanks for you attention, Andrey Ivanov aka Search_Terminal.
_________
https://github.com/EthicalEarth/i2c_hid_driver_elan1200

пт, 7 сент. 2018 г., 0:04 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #19 from Marc Landolt (2009@marclandolt.ch) ---
> hello Search_terminal
>
> my problem is, if i put 5 fingers on the touchpad i the touchpad looses
> connection and i must run rmmod hid-multitouch and modprobe hid-multitouch
> that
> it works again, so in my opinion the problem is not in i2c-hid because i
> dont
> need to reload the i2c-hid...
>
> but i had a look at dmesg, and it also says i2c_hid i2c-ELAN1200:00:
> ic2_hid_get_input: incomplete report (16/65535). The Laptop runs now since
> 24h
> and it occured 696 times since yesterday.
>
> may i ask where you got this files from, have you made this patch or who
> wrote
> this? I diff-ed with mainline kernel i2c_hid.c and there are a lot of
> changes
> and I would like to stick to the mainline kernel, because if the problem
> does
> not get fixed in mainline i (and all other people that has a ASUS GL703GE)
> have
> to patch it each time when i setup a new linux on this machine.
>
> With kind regards
> Marc Landolt
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 21 Marc Landolt 2018-09-06 21:46:34 UTC
Created attachment 278357 [details]
Benjamin Tissores / ??? patch (make output)
Comment 22 Marc Landolt 2018-09-06 21:47:22 UTC
Hello Search_Terminal

i can not even build it with kernel 4.19-rc2: https://bugzilla.kernel.org/attachment.cgi?id=278357

With kind regards
Marc Landolt
Comment 23 Andrey Ivanov 2018-09-06 21:55:37 UTC
Created attachment 278359 [details]
attachment-10237-0.html

Ohh i think this driver only for 4.15.x to my 4.18.. i guess..
its not compile on your system((

пт, 7 сент. 2018 г., 0:47 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #22 from Marc Landolt (2009@marclandolt.ch) ---
> Hello Search_Terminal
>
> i can not even build it with kernel 4.19-rc2:
> https://bugzilla.kernel.org/attachment.cgi?id=278357
>
> With kind regards
> Marc Landolt
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 24 Marc Landolt 2018-09-06 22:08:36 UTC
Hello Search Terminal

I booted with Kernel 4.18-rc6 and could build it, i insmod-ed it but the problem with the freeze when you put 5 fingers on the touchpad persists, so it's not the solution for this problem
Comment 25 Dmitry Torokhov 2018-09-07 21:08:22 UTC
Why don't we loop Benjamin in?

Benjamin, it looks like hid-multitouch is unhappy with this device when one places 5 fingers on the surface and stops reporting any data?
Comment 26 Andrey Ivanov 2018-09-07 21:17:56 UTC
Created attachment 278369 [details]
attachment-27498-0.html

Hmm.. Benjamin  write good driver witout "incomplete report error" and..
its solve big log of kernel and disconnect because log is not fast. I think
it.

But in native i2c_hid driver, in Benjamin driver we have error disconnect
when 5 fingers touches.

I last time ask myself, when needed 5 fingers touches in laptop???

Hmm.


So.. disconnect with 5 fing. touches  not solve in all sources of kernel.

After disconnect in native driver elan1200, i have irq load 1%of core..
But touchpad not work..
Its mean touchpad work, but irq bad processing.. i think so.

What is next? Maybe need to debug driver multitouch and i2c hid?

What say Linus? I know, Linux not like patches, but driver work bad. Its
need to improve..

сб, 8 сент. 2018 г., 0:08 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> Dmitry Torokhov (dmitry.torokhov@gmail.com) changed:
>
>            What    |Removed                     |Added
>
> ----------------------------------------------------------------------------
>                  CC|                            |
> benjamin.tissoires@gmail.co
>                    |                            |m,
>                    |                            |dmitry.torokhov@gmail.com
>
> --- Comment #25 from Dmitry Torokhov (dmitry.torokhov@gmail.com) ---
> Why don't we loop Benjamin in?
>
> Benjamin, it looks like hid-multitouch is unhappy with this device when one
> places 5 fingers on the surface and stops reporting any data?
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 27 Marc Landolt 2018-09-07 21:21:49 UTC
Hello Search_termainal

it sometimes happens that you touch the touchpad with your palm, and then it disonnects because the touchpad seems to identify the palm with more then 2 fingers. then it disconnects, so it would be really nice if this would be solved...
Comment 28 Andrey Ivanov 2018-09-07 21:25:32 UTC
Created attachment 278371 [details]
attachment-30864-0.html

Yes yes yes. Go go go. Just i am not good programmer ((( i only little know
about c/pascal(((
Its reason why i write here, try to write Linus about this problems.. and
benjamin.. Vivien is silent..


сб, 8 сент. 2018 г., 0:21 <bugzilla-daemon@bugzilla.kernel.org>:

> https://bugzilla.kernel.org/show_bug.cgi?id=200663
>
> --- Comment #27 from Marc Landolt (2009@marclandolt.ch) ---
> Hello Search_termainal
>
> it sometimes happens that you touch the touchpad with your palm, and then
> it
> disonnects because the touchpad seems to identify the palm with more then 2
> fingers. then it disconnects, so it would be really nice if this would be
> solved...
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 29 Benjamin Tissoires 2018-09-10 09:06:13 UTC
Can we take a step back and provide me the following information (a simple yes/no answer should be good enough).

@Marc:
- plain v4.19-rc2 solves most of the issues but the 5 touches that disconnects
- I am not sure I followed if you still need to patch pinctrl-intel.c to have the proper IRQ mapped
- I am surprised you do not need https://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git/commit/drivers/hid/hid-multitouch.c?h=for-next&id=ec6adef5fbc3f140c70e7499fdad818acb3a46c6 to fix the click issues (this should go in the next PR sent to Linus)
- is the output of /proc/interrupts showing an abnormal usage for the touchpad (interrupts are firing even if no one is touching the touchpad, or it starts storming after the 5 finger touches)?
- if I read correctly, when the touchpad disconnects, hid-recorder doesn't send any more events
- after the touchpad disconnection, are the IRQ still growing? (in /proc/interrupts)

@Andrey
- do you have any more changes in your i2c-hid+hid-multitouch driver than the ones that are on my repo?


Marc, I have a feeling that the `#if` section you commented out is responsible for the disconnect. My theory is that you are still not mapped to the correct GPIO as IRQ, and when you put 5 fingers, the firmware expects you to poll the data, but not too fast or not too slow, and if you have a wrong mapped IRQ, there is a chance the polling simply breaks the I2C chip.

Andrey, the only 'valuable' change I had in my 'for-vivien' branch compared to upstream was https://github.com/bentiss/hid-multitouch/commit/878681f70d6f7deca5aa455619fb404a65495adc This added a magic sequence found by reverse engineering, so it might or not help. It is still a mystery about what this is however. The rest of the driver should be upstream. (Note: next time clone my repo to keep the history, it will solve the issue of not knowing what is in your repo).
Comment 30 Andrey Ivanov 2018-09-10 09:45:47 UTC
В письме от понедельник, 10 сентября 2018 г. 12:06:13 +03 пользователь 
bugzilla-daemon@bugzilla.kernel.org написал:
> https://bugzilla.kernel.org/show_bug.cgi?id=200663
> 
> --- Comment #29 from Benjamin Tissoires (benjamin.tissoires@gmail.com) ---
> Can we take a step back and provide me the following information (a simple
> yes/no answer should be good enough).
> 
> @Marc:
> - plain v4.19-rc2 solves most of the issues but the 5 touches that
> disconnects - I am not sure I followed if you still need to patch
> pinctrl-intel.c to have the proper IRQ mapped
> - I am surprised you do not need
> https://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git/commit/drivers
> /hid/hid-multitouch.c?h=for-next&id=ec6adef5fbc3f140c70e7499fdad818acb3a46c6
> to fix the click issues (this should go in the next PR sent to Linus) - is
> the output of /proc/interrupts showing an abnormal usage for the touchpad
> (interrupts are firing even if no one is touching the touchpad, or it
> starts storming after the 5 finger touches)?
> - if I read correctly, when the touchpad disconnects, hid-recorder doesn't
> send any more events
> - after the touchpad disconnection, are the IRQ still growing? (in
> /proc/interrupts)
> 
> @Andrey
> - do you have any more changes in your i2c-hid+hid-multitouch driver than
> the ones that are on my repo?
> 
> 
> Marc, I have a feeling that the `#if` section you commented out is
> responsible for the disconnect. My theory is that you are still not mapped
> to the correct GPIO as IRQ, and when you put 5 fingers, the firmware
> expects you to poll the data, but not too fast or not too slow, and if you
> have a wrong mapped IRQ, there is a chance the polling simply breaks the
> I2C chip.
> 
> Andrey, the only 'valuable' change I had in my 'for-vivien' branch compared
> to upstream was
> https://github.com/bentiss/hid-multitouch/commit/878681f70d6f7deca5aa455619f
> b404a65495adc This added a magic sequence found by reverse engineering, so
> it might or not help. It is still a mystery about what this is however. The
> rest of the driver should be upstream. (Note: next time clone my repo to
> keep the history, it will solve the issue of not knowing what is in your
> repo).
-------------------------------------------------------------------------------------------------------------------------------

Hello Benjamin. I will be try to answer.

> - do you have any more changes in your i2c-hid+hid-multitouch driver than
> the ones that are on my repo?

YES. 

Driver from you repo not compile under kernel v.4.18.5.

REASON:
git clone https://github.com/bentiss/hid-multitouch/ or zip archive from  
webpage.
make
gcc say error:
 CC [M]  /home/andrey/hid-multitouch-for-vivien/hid-multitouch.o
/home/andrey/hid-multitouch-for-vivien/hid-multitouch.c: in function  
«mt_probe»:
/home/andrey/hid-multitouch-for-vivien/hid-multitouch.c:1534:18: error: 
«HID_QUIRK_NO_EMPTY_INPUT» undeclared (first use in this function); did you 
mean «HID_QUIRK_MULTI_INPUT»?
  hdev->quirks |= HID_QUIRK_NO_EMPTY_INPUT;
                  ^~~~~~~~~~~~~~~~~~~~~~~~
                  HID_QUIRK_MULTI_INPUT
-----
Modules not compiled.
Solution:
open hid-multitouch.c
see 1533 line and 1534:
	hdev->quirks |= HID_QUIRK_MULTI_INPUT;
	hdev->quirks |= HID_QUIRK_NO_EMPTY_INPUT;
----
Change: 
	hdev->quirks |= HID_QUIRK_MULTI_INPUT;
	//hdev->quirks |= HID_QUIRK_NO_EMPTY_INPUT;
save file.

try to compile;
make.
succeseful and work.

The next problem is that this driver could not be compiled under 4.19rc2 even 
with my fix. Mark tried to do this without success.
-------------------------------------------------------------------------------------------------------------------------------
> Andrey, the only 'valuable' change I had in my 'for-vivien' branch compared
> to upstream was
> https://github.com/bentiss/hid-multitouch/commit/878681f70d6f7deca5aa455619f
> b404a65495adc This added a magic sequence found by reverse engineering, so
> it might or not help. It is still a mystery about what this is however. The
> rest of the driver should be upstream. (Note: next time clone my repo to
> keep the history, it will solve the issue of not knowing what is in your
> repo).

Your magical sequence solve error "incomplete report".
for a couple of days I have a huge number of log files with the size of a 
couple of gigabytes due to a mistake in mainline.
Your driver solve it.
-------------------------------------------------------------------------------------------------------------------------------
> Marc, I have a feeling that the `#if` section you commented out is
> responsible for the disconnect. My theory is that you are still not mapped
> to the correct GPIO as IRQ, and when you put 5 fingers, the firmware
> expects you to poll the data, but not too fast or not too slow, and if you
> have a wrong mapped IRQ, there is a chance the polling simply breaks the
> I2C chip.

Can i ask you.. bad IRQ can to broke touchpad i2c chip?
Comment 31 Marc Landolt 2018-09-10 14:07:46 UTC
(In reply to Benjamin Tissoires from comment #29)
> Can we take a step back and provide me the following information (a simple
> yes/no answer should be good enough).
> 
> @Marc:

> - plain v4.19-rc2 solves most of the issues but the 5 touches that
> disconnects

no plain v4.19-rc2 does not work, i need to add line 1016, 1021 and 1022 from https://bugzilla.kernel.org/attachment.cgi?id=277507&action=diff otherwise the touchpad does not work

> - I am not sure I followed if you still need to patch pinctrl-intel.c to
> have the proper IRQ mapped

yes, as described above

> - I am surprised you do not need
> https://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git/commit/drivers/
> hid/hid-multitouch.c?h=for-next&id=ec6adef5fbc3f140c70e7499fdad818acb3a46c6
> to fix the click issues (this should go in the next PR sent to Linus)

multiple clicking issue did not happend again since i use patched 4.19-rc2

> - is the output of /proc/interrupts showing an abnormal usage for the
> touchpad (interrupts are firing even if no one is touching the touchpad, or
> it starts storming after the 5 finger touches)?

i will add before.txt and after.txt, if I put 5 fingers or my palm on the touchpad the number of interrupts does not change (hopefully i looked at the right irq: >   16:    1058607          0          0       4207          0          0    2480853          0          0          0          0          0  IR-IO-APIC   16-fasteoi   idma64.0, i2c_designware.0

> - if I read correctly, when the touchpad disconnects, hid-recorder doesn't
> send any more events

yes, i need to restart hid-multitouch that it records anything again

> - after the touchpad disconnection, are the IRQ still growing? (in
> /proc/interrupts)

im not sure, but if its the 16: ... designware.0 then id does not change at all (before.txt / after.txt / later.txt, dmesg.txt)

Thank you for helping :)

With kind regards
Marc Landolt

> 
> @Andrey
> - do you have any more changes in your i2c-hid+hid-multitouch driver than
> the ones that are on my repo?
> 
> 
> Marc, I have a feeling that the `#if` section you commented out is
> responsible for the disconnect. My theory is that you are still not mapped
> to the correct GPIO as IRQ, and when you put 5 fingers, the firmware expects
> you to poll the data, but not too fast or not too slow, and if you have a
> wrong mapped IRQ, there is a chance the polling simply breaks the I2C chip.
> 
> Andrey, the only 'valuable' change I had in my 'for-vivien' branch compared
> to upstream was
> https://github.com/bentiss/hid-multitouch/commit/
> 878681f70d6f7deca5aa455619fb404a65495adc This added a magic sequence found
> by reverse engineering, so it might or not help. It is still a mystery about
> what this is however. The rest of the driver should be upstream. (Note: next
> time clone my repo to keep the history, it will solve the issue of not
> knowing what is in your repo).
Comment 32 Marc Landolt 2018-09-10 14:08:13 UTC
Created attachment 278405 [details]
before freeze
Comment 33 Marc Landolt 2018-09-10 14:08:40 UTC
Created attachment 278407 [details]
after freeze
Comment 34 Marc Landolt 2018-09-10 14:09:23 UTC
Created attachment 278409 [details]
5 minutes later
Comment 35 Marc Landolt 2018-09-10 14:09:43 UTC
Created attachment 278411 [details]
dmesg
Comment 36 Marc Landolt 2018-09-10 14:11:39 UTC
that were the logs from kernel 4.18-rc6 with the patch from https://bugzilla.kernel.org/show_bug.cgi?id=199911#c68

i will now reboot and put the same files with kernel 4.19-rc2
Comment 37 Marc Landolt 2018-09-10 14:16:19 UTC
Created attachment 278413 [details]
before freeze 4.19-rc2
Comment 38 Marc Landolt 2018-09-10 14:16:47 UTC
Created attachment 278415 [details]
after freeze 4.19-rc2
Comment 39 Marc Landolt 2018-09-10 14:17:21 UTC
Created attachment 278417 [details]
5 minutes later 4.19-rc2
Comment 40 Marc Landolt 2018-09-10 14:17:50 UTC
Created attachment 278419 [details]
dmesg 4.19-rc2
Comment 41 Marc Landolt 2018-09-10 14:21:24 UTC
@Benjamin Tissoires

please tell me if you need more things, or if you like you can write me on skype (marc.landolt.aarau) if you would like to access the device eg. with teamviewer and do your own tests, the machine is not in use yet, so you could do anything you would like to :)
Comment 42 Andrey Ivanov 2018-09-11 10:13:18 UTC
@Benjamin Tissoires
I test latest kernel 4.19_rc3 in gentoo (git-sources), my laptop not fully boot because i915 not correctly loading(fbcon not found but fbcon include in kernel).. 
But i test touchpad in console with GPM (General Puporse Mouse) in console. Disconnect when i touch 5 fingers.
i try to solve buut my laptop fully but i think i will be back to 4.18.5... hmm.. maybe need to waiting stable release of 4.19..
Because X not start( i ca not see sddm login screen)
Comment 43 Andrey Ivanov 2018-09-11 10:42:48 UTC
@Benjamin Tissoires
I create new test, rebuild kernel but intel 915 not work( hmm. 
In 4.19.rc3 not solved incomplete report error..
Comment 44 Hieu Cao 2018-10-17 04:31:40 UTC
Here is any news about this bug?
This bug appear usually so right now, I just come back to windows and use virtual machime for a stable. Hope some good ideas come to you guy and can fix it faster.
Thanks
Comment 45 Andrey Ivanov 2018-10-24 21:55:07 UTC
(In reply to Hieu Cao from comment #44)
> Here is any news about this bug?
> This bug appear usually so right now, I just come back to windows and use
> virtual machime for a stable. Hope some good ideas come to you guy and can
> fix it faster.
> Thanks

Build new kernel 4.19.0.

Touchpad elan1200 not added when loading modules. 4.19.0 not support my touchpad.
Comment 46 Marc Landolt 2018-10-24 21:57:08 UTC
(In reply to Andrey Ivanov from comment #45)
> (In reply to Hieu Cao from comment #44)
> > Here is any news about this bug?
> > This bug appear usually so right now, I just come back to windows and use
> > virtual machime for a stable. Hope some good ideas come to you guy and can
> > fix it faster.
> > Thanks
> 
> Build new kernel 4.19.0.
> 
> Touchpad elan1200 not added when loading modules. 4.19.0 not support my
> touchpad.

you still need to add the 3 lines described in comment 16 for kernel 4.19: https://bugzilla.kernel.org/show_bug.cgi?id=200663#c16

with kind regards
marc
Comment 47 Andrey Ivanov 2018-10-24 21:58:26 UTC
(In reply to Hieu Cao from comment #44)
> Here is any news about this bug?
> This bug appear usually so right now, I just come back to windows and use
> virtual machime for a stable. Hope some good ideas come to you guy and can
> fix it faster.
> Thanks

Hardware Asus GL703VD
Comment 48 Andrey Ivanov 2018-10-24 21:58:51 UTC
(In reply to Hieu Cao from comment #44)
> Here is any news about this bug?
> This bug appear usually so right now, I just come back to windows and use
> virtual machime for a stable. Hope some good ideas come to you guy and can
> fix it faster.
> Thanks

Hardware Asus GL703VD
Comment 49 Andrey Ivanov 2018-10-24 22:16:36 UTC
(In reply to Marc Landolt from comment #46)
> (In reply to Andrey Ivanov from comment #45)
> > (In reply to Hieu Cao from comment #44)
> > > Here is any news about this bug?
> > > This bug appear usually so right now, I just come back to windows and use
> > > virtual machime for a stable. Hope some good ideas come to you guy and
> can
> > > fix it faster.
> > > Thanks
> > 
> > Build new kernel 4.19.0.
> > 
> > Touchpad elan1200 not added when loading modules. 4.19.0 not support my
> > touchpad.
> 
> you still need to add the 3 lines described in comment 16 for kernel 4.19:
> https://bugzilla.kernel.org/show_bug.cgi?id=200663#c16
> 
> with kind regards
> marc

Dear Marc, i little missundderstood how to add(((can you add patch file for me?
Comment 50 Marc Landolt 2018-10-24 22:58:17 UTC
You must edit the File manually:

you need line 1016, 1021 and 1022 from

https://bugzilla.kernel.org/attachment.cgi?id=277507&action=diff 

this is an older patch, most of this things are already in the 4.19 Kernal but not line 1016, 1021 and 1022, so you need to add them manually with an editor in file /drivers/pinctrl/intel/pinctrl-intel.c

in the file from kernel 4.19 it are the lines 1000, 1005, 1006

with kind regards
marc

PS: let the community know if it worked since there is nobody with GL703VD, only me with GL703GE...

Note You need to log in before you can comment on or make changes to this bug.