Description of problem: Version-Release number of selected component (if applicable): Fedora 26 Fedora 27 Fedora Rawide How reproducible: Boot into Fedora on a GL702ZC laptop Actual results: Touchpad not detected Expected results: Touchpad detected Additional info: evemu-record does not show the touchpad at all. There is a kernel backtrace in dmesg. I've attached the output
Created attachment 1349138 [details] lspci output
Created attachment 1349139 [details] dmesg output
Created attachment 1349141 [details] dmidecode output
Created attachment 1349142 [details] lscpu output
I have a little bit more information: I built linux 4.13.11 from https://github.com/M-Bab/linux-kernel-amdgpu/tree/mbab_4.13 and enabled i2c debugging. I see what appears to be normal output and the touchpad works but occasionally i2c output from i2c_designware AMDI0010:03 stops completely and the touchpad stops working. If I rmmod and modprobe i2c_hid the touchpad starts working again for a while. I see no errors leading up to the touchpad stopping working.
Created attachment 1349305 [details] i2c debugging output leading up to the problem
Created attachment 1349318 [details] i2c rmmod/modprobe failure
I've added dmesg output for the i2c debug output as well as an instance in which rmmod/modprobe of the i2c_hid module did not result in a working touchpad again. When the touchpad/i2c/whichever component this is gets in this state only a hard reboot seems to fix it.
Thanks for the report. Hans had a similar issue with an other cheap i2c-hid chip. See https://patchwork.kernel.org/patch/10046575/. I wonder if you apply the same quirk, your device may just work.
I just built 4.14.0-rc8 and the touchpad is now also detected and works. I haven't had a random stop yet. I'll update the ticket if it hasn't happened for a day or two. Or sooner if it does happen. I still have i2c debugging enabled.
So I have some results. The problem didn't reappear with the config I cobbled together from various sources. I decided to do a config more similar to the default Fedora .config and the problem reappeared. I haven't yet figured out what settings cause the problem to appear. For now I'll attach both the working and non-working .config, perhaps someone can do a diff and have an a-ha moment. I'll go through the ones that seem relevant to me and try it though.
Created attachment 1350251 [details] kernel .config not working
Created attachment 1350252 [details] kernel .config working
I've heard from another user of this same laptop that 4.14.0 with the default Ubuntu configs don't suffer from this problem. I'm still trying to work out what the exact issue is. There are a lot of differences :)
I spent a little more time on this issue now. I've found some things: * Disabling CONFIG_X86_AMD_PLATFORM_DEVICE in my working .config results in a dmesg message that is the same as with the default Fedora .config. * Enabling CONFIG_X86_AMD_PLATFORM_DEVICE in the Fedora .config by itself makes the dmesg error about i2c_dw_clk_rate+0x1b/0x20 listed above go away but the touchpad is still not moving. There is no mention of the ELAN driver in dmesg, however: $ ls -l /sys/bus/i2c/devices/i2c-0/ total 0 --w-------. 1 root root 4096 Dec 5 23:18 delete_device lrwxrwxrwx. 1 root root 0 Dec 5 23:18 device -> ../../AMDI0010:03 lrwxrwxrwx. 1 root root 0 Dec 5 23:18 firmware_node -> ../../../LNXSYSTM:00/LNXSYBUS:00/AMDI0010:03 drwxr-xr-x. 3 root root 0 Dec 6 2017 i2c-ELAN1200:00 -r--r--r--. 1 root root 4096 Dec 5 23:18 name --w-------. 1 root root 4096 Dec 5 23:18 new_device drwxr-xr-x. 2 root root 0 Dec 5 23:18 power lrwxrwxrwx. 1 root root 0 Dec 6 2017 subsystem -> ../../../../bus/i2c -rw-r--r--. 1 root root 4096 Dec 5 23:18 uevent So there is something being detected now. There are many more configuration differences to check though.
Anyone made any progress?
Sorry, I haven't been spending huge amounts of time on this. I haven't yet located the exact configuration change between the .config I use and the one used by Fedora which causes the touchpad to work besides CONFIG_X86_AMD_PLATFORM_DEVICE being required.
Yesterday I downloaded both configs and put them besides each other using meld. In meld you can use the search tool. Searching for "I2C" revealed some major differences. I will do a diff search for "ELAN", too. Maybe I can get a difftool to print out a patch.
I'm pretty sure it's not a config issue. I think they have fiddled with the sources over at canonical. My notebook has been pretty active for a few days now, I have tried a few dozens of configs based on configs merged from vanilla and the working one you uploaded. Unfortunately for me the working one does not work. I do not get any graphics support. Not even plain tty, so it's just uninitialized stuff on screen. Anyway no luck here and I'm positive it's not a config related issue, since I copied over anything I can think of affecting bus systems and hid drivers related to I2C and ELAN input. Modprobing doesn't change anything either. If I find the time I will go to a LUG and put up a challenge :D
I finally got it to work. Today I stumbled upon this git repo: https://notabug.org/hp/linux-gl702zc Whoever this guy is: I owe you a beer or two. It's as easy as compile, install reboot. It comes complete with working config. Very well done. I posted this in case someone stumbled upon this and for the kernel devs to check out a running version. Just to be clear: this is not an ubuntu based kernel. This kernel pretty much behaves just like a RHEL kernel. Including selinux. The only downside was that I had to put in the original Realtek wifi adpter, since the kernel in the repo had no drivers for my newly bought intel chip on board. This would be a job for the future once I'm done celebrating hat my notebook is a self contained unit at last :D
Yeah, that's mine :) No patches other than the k10temp stuff. Still have no idea what is causing this. I'm beginning to suspect some of the more arcane base address or alignment values.
This bug is still present in the Fedora 28 stock kernel. I still don't know what combination of settings makes it work and what doesn't. I'm almost certainly missing something rather obvious. I'll try to spend some more time on this soon. I haven't had much incentive since my custom kernels seem to work well, but that's not really helpful for the wider Fedora community :)
Depending on where you're located we could meet up in May and work on that together for a weekend or so. This month I'm occupied with writing my thesis.
Just a quick note to say that I, too have the same laptop, HAD the same issues until I compiled the kernel provided by Mr. Braam at https://notabug.org/hp/linux-gl702zc and everything works great now.. Wireless and touchpad being the major issues. Likewise for brightness control; that works too.
Happy anniversary for this bug! Can it be fixed before the laptop goes out of fashion? PS: I'm a linux user and a C programmer. Can I assist in fixing it? Please let me know how to proceed..
(In reply to Turgut Kalfaoglu from comment #25) > Happy anniversary for this bug! > Can it be fixed before the laptop goes out of fashion? > > PS: I'm a linux user and a C programmer. Can I assist in fixing it? > Please let me know how to proceed.. You can help by comparing the .config file which defines all the CONFIG_KERNEL_OPTION_FOO= values for building the kernel between the Fedora kernels and Mr. Braam's kernels and then if you see something which is "suspect" which is enabled in Fedora and not in Mr. Braam's kernels, or the other way around change that in the Fedora .config, build a new kernel with that config and test, until you've found the culprit Kconfig option.
Thank you! There are lots of changes unfortunately.. The best thing I found was the "beyond compare" to compare these two. I'm attaching the result and Mr. Braam's config as well..
Created attachment 1438488 [details] differences between fedora 4.16.8-300.fc28.x86_64 and braams kernels
Created attachment 1438489 [details] Mr. Braam's kernel config
Ok, I think I've found the issue, Fedora kernels are not selecting CONFIG_X86_AMD_PLATFORM_DEVICE which is necessary for the i2c-controllers on recent AMD platforms to get enumerated. I've started a test kernel build with this enabled here: https://koji.fedoraproject.org/koji/taskinfo?taskID=27041524 (note this will take a couple of hours to finish). Once finish, please follow these generic instructions for testing kernels directly from koji: https://fedorapeople.org/~jwrdegoede/kernel-test-instructions.txt And let me know if the test kernel fixes things.
I noted that in https://bugzilla.redhat.com/show_bug.cgi?id=1510649#c15. Only enabling the AMD platform device is not by itself sufficient to make the touchpad function.
Wonderful, thank you very much, you are a genious :)
(In reply to Hein-Pieter van Braam from comment #31) > I noted that in https://bugzilla.redhat.com/show_bug.cgi?id=1510649#c15. > Only enabling the AMD platform device is not by itself sufficient to make > the touchpad function. Ah I missed that, still we have a bit newer base kernel now, so might be worth a shot, also dmesg output with the test kernel I build would be appreciated / useful.
The touchpad still cannot be enabled unfortunately.. I'm enclosing a dmesg..
Created attachment 1438768 [details] dmesg of the test kernel
(In reply to Turgut Kalfaoglu from comment #34) > The touchpad still cannot be enabled unfortunately.. Bummer.
Ok, so I had some inspiration about this but this morning. I realized that besides the i2c bus the touchpad also needs an interrupt and that is connected to a GPIO, which means we may need a pinctrl and/or gpio driver for this. Which led to me finding that CONFIG_GPIO_AMDPT and CONFIG_PINCTRL_AMD are not set in the Fedora test kernels, so here is a new test build with these enabled: https://koji.fedoraproject.org/koji/taskinfo?taskID=27060582 As usual this will take a couple of hours to finish and testing instructions are here: https://fedorapeople.org/~jwrdegoede/kernel-test-instructions.txt Regardless of if this fixes things, can you please run alsa-info.sh let it save to a file and attach the file here? I know this is not a sound related bug, but that is the easiest way to get a log of which ACPI devices are present on your systems and what their status is. Also if things still do not work, please run "ls -l /sys/bus/i2c/devices" and copy and paste the output here.
Created attachment 1438927 [details] output of alsa-info.sh
Many thanks.. I'm checking it now.. I haven't installed the latest kernel yet.. [root@tklaptop ~]# ls -l /sys/bus/i2c/devices total 0 lrwxrwxrwx 1 root root 0 May 19 2018 i2c-0 -> ../../../devices/platform/AMDI0010:03/i2c-0 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-1 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-1 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-10 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/drm/card0/card0-DP-1/i2c-10 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-11 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/drm/card0/card0-DP-2/i2c-11 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-12 -> ../../../devices/pci0000:00/0000:00:14.0/i2c-12 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-13 -> ../../../devices/pci0000:00/0000:00:14.0/i2c-13 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-14 -> ../../../devices/pci0000:00/0000:00:14.0/i2c-14 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-15 -> ../../../devices/pci0000:00/0000:00:14.0/i2c-15 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-2 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-2 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-3 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-3 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-4 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-4 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-5 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-5 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-6 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-6 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-7 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-7 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-8 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/i2c-8 lrwxrwxrwx 1 root root 0 May 19 18:27 i2c-9 -> ../../../devices/pci0000:00/0000:00:03.1/0000:0c:00.0/drm/card0/card0-eDP-1/i2c-9 lrwxrwxrwx 1 root root 0 May 19 2018 i2c-ELAN1200:00 -> ../../../devices/platform/AMDI0010:03/i2c-0/i2c-ELAN1200:00 [root@tklaptop ~]#
THANK YOU! It is solved; and the touchpad now works :) Do we need to do something special so that these changes are reflected to the mainstream Fedora kernel? I'd love these changes to stay if possible..
PS: Right click results in a left click; but that's probably a gnome issue..
(In reply to Turgut Kalfaoglu from comment #40) > THANK YOU! > It is solved; and the touchpad now works :) > > Do we need to do something special so that these changes are reflected to > the mainstream Fedora kernel? I'd love these changes to stay if possible.. I will push these changes to the Fedora kernel, to be picked up by the next official Fedora kernel build, so there is nothing which you need to do.
(In reply to Turgut Kalfaoglu from comment #41) > PS: Right click results in a left click; but that's probably a gnome issue.. Right, GNOME3 now defaults to clickfinger, which means that clicking anywhere with 2 fingers down results in a right click, to get a right click area at the bottom right of the touchpad as you probable expect, install gnome-tweaks, start it, go to "Keyboard & Mouse" and change the "Mouse Click Emulation" from "Fingers" to "Area".
Thank you very much for all your help!
kernel-4.16.10-300.fc28 has been submitted as an update to Fedora 28. https://bodhi.fedoraproject.org/updates/FEDORA-2018-8ce90c8b24
kernel-4.16.10-200.fc27 has been submitted as an update to Fedora 27. https://bodhi.fedoraproject.org/updates/FEDORA-2018-42898720ed
kernel-4.16.10-200.fc27 has been pushed to the Fedora 27 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-42898720ed
kernel-4.16.10-300.fc28 has been pushed to the Fedora 28 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-8ce90c8b24
kernel-4.16.11-200.fc27 has been submitted as an update to Fedora 27. https://bodhi.fedoraproject.org/updates/FEDORA-2018-93c2e74446
*** Bug 1573133 has been marked as a duplicate of this bug. ***
kernel-4.16.10-300.fc28 has been pushed to the Fedora 28 stable repository. If problems still persist, please make note of it in this bug report.
kernel-4.16.11-200.fc27 has been pushed to the Fedora 27 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-93c2e74446
kernel-4.16.11-200.fc27 has been pushed to the Fedora 27 stable repository. If problems still persist, please make note of it in this bug report.
4.16.12-300.fc28.x86_64 fixed the issue for me. Awesome!
The issue is back.. GL702ZC touchpad no longer works.. $ uname -a Linux .... 4.20.10-200.fc29.x86_64 #1 SMP Fri Feb 15 18:33:14 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux ]$ more /etc/sysconfig/grub GRUB_CMDLINE_LINUX="pti=off"
I just checked, the 2 kernel config options which were missing before are still present in the current Fedora kernels, so that is not the problem.