Hi Aaron, and the Linux wacom team,
Following bug 342, dedicated to kernel now, I did what Arron asked today, triggered the EKR stop working bug, and here are the results ...
0/ here are my packages updated this morning :
linux 4.12.8
gnome-desktop 3.24.2
gnome-control-center 3.24.3
libwacom 0.24
xf86-input-wacom 0.35
input-wacom patched for the EKR by Aaron
xorg-server 1.19.3
libinput git with a patch from peter : https://bugs.freedesktop.org/show_bug.cgi?id=102235
1/ xsetwacom list give me a Cintiq 27 with "touch" in lowercase :
[nylnook@nylnook-home ~]$ xsetwacom list Wacom Cintiq 27QHD touch Pen stylus id: 12 type: STYLUS Wacom Express Key Remote Pad pad id: 13 type: PAD Wacom Cintiq 27QHD touch Finger touch id: 14 type: TOUCH Wacom Cintiq 27QHD touch Pen eraser id: 16 type: ERASER
2/ evtest is working after the EKR stopped to respond to krita, and to the Gnome desktop in general... see following attachements... and now finally the EKR is working after the bug if I pair it again and run my xsetwacom script (because of a package update ?).
3/ It seem to me Krita produced an interesting error at the time the EKR stopped working :
(krita:5469): Gdk-CRITICAL **: gdk_device_update_tool: assertion 'GDK_IS_DEVICE (device)' failed
see attachement
4/ look like Mypaint is detecting the device if I run it command line even during the bug occur and I can't use the EKR :
INFO: gui.device: New device 'Wacom Express Key Remote Pad pad' (GDK_SOURCE_TABLET_PAD, axes:6, class=X11DeviceXI2, vendor='056a', product='0331')
see attachement too...
Seem to me we are progressing !
Thanks a lot for you work !
joined evtest and krita (laucnched again) logs after the bug occured
joined mypaint logs, launched after bug occur
Hi,
To clarify about #4 - So after the bug the EKR doesn't work in krita or mypaint.
Also could you get take an xinput log during the bug:
To take a log do (for example) the following:
$ xinput list
⎡ Virtual core pointer id=2 [master pointer (3)]
⎜ ↳ Virtual core XTEST pointer id=4 [slave pointer (2)]
⎜ ↳ Lite-On Technology Corp. USB Multimedia Keyboard id=11 [slave pointer (2)]
⎜ ↳ MOSART Semi. 2.4G Wireless Mouse id=15 [slave pointer (2)]
⎜ ↳ Wacom Express Key Remote Pad pad id=22 [slave pointer (2)]
⎜ ↳ Wacom Intuos Pro M Pad pad id=9 [slave pointer (2)]
⎜ ↳ Wacom Intuos Pro M Pen stylus id=12 [slave pointer (2)]
⎜ ↳ Wacom Intuos Pro M Pen eraser id=13 [slave pointer (2)]
⎜ ↳ Wacom Intuos Pro M Pen cursor id=14 [slave pointer (2)]
⎜ ↳ Wacom Intuos Pro M Finger touch id=16 [slave pointer (2)]
⎣ Virtual core keyboard id=3 [master keyboard (2)]
↳ Virtual core XTEST keyboard id=5 [slave keyboard (3)]
↳ Power Button id=6 [slave keyboard (3)]
↳ Video Bus id=7 [slave keyboard (3)]
↳ Power Button id=8 [slave keyboard (3)]
↳ Lite-On Technology Corp. USB Multimedia Keyboard id=10 [slave keyboard (3)]
$ sudo xinput test-xi2 22
Hi Aaron,
I couldn't reproduce the bug today (sometimes randomness takes time ... ;p) and I'm leaving this evening for work but away from my Cintiq... I'm back on next Tuesday, and I'll continue testing....
Thanks again !
Hi Aaron,
I finally got one interesting log today, joined. I've been doing some 3D related projects last week so I didn't used the EKR much...
Please note that I updated my kernel to 4.12.12 and compiled again your patched input-wacom before this log...
It's hard to tell when the bug started because it was coming back to my screen after leaving it for 15 minutes (and switching it off an on, but the bug occur with or without switching off and on the Cintiq)... So I think somehow the last 25 events occur during the bug, starting at a "EVENT type 16 (RawButtonRelease)".
As you asked, I trigerred every button of the Expresskey during the bug.
An other intersting point is that I think xsetwacom actually stopped during the bug, because the screen become touch sentitive again, and I usually disable it with my script (because Krita doesn't handle it well).
So maybe it's not an EKR bug, but a xsetwacom bug ?
I hope it will bring you some light !
Cheers,
Camille
Hi Camille,
It seems like this is an X or Gnome bug. Maybe Gnome is "grabbing" the device or in some way removing your xsetwacom settings... but since it affects xsetwacom I"m surprised that we haven't heard about this bug from anyone else.
Do you have any other scripts or cronjobs that you are running that might affect the behavior of Gnome or the X server?
Best,
Aaron
Hi ;)
Yes, it seems so... I don't think I have another script or cronjob that can affect Xsetwacom...
My xsetwacom script start from Gnome actually, using Gnome tweak tools, maybe I should try another way to launch it at startup... ?
And I have two cronjobs using rsync to backup my datas on secondary hard drives, one daily and one weekly...
That's all I'm aware of, but there are deamons running installed from packages like gnome ones, dbus, gdm, evolution, gsd, tracker, zeigeist...
Should I wait for Gnome solving the EKR graphical integration (supposed to works today, but bugs are on the way), or should I report a new bug, and where.... ?
Maybe xsetwacom is not much used theses days because Gnome and KDE are providing graphicals ways to set up the ExpressKeys... so that's maybe why I'm the first...
Best regards, and thanks !
Camille
PS: your patched input-wacom works well even with latest kernel, it seems good to go for me ;)
Hi Camille,
Ping suggested to me that you try disabling the gnome-settings-daemon wacom plugin and see if the issue still occurs. I'm not sure if this will affect the rest of your setup or not but please try it out:
gsettings set org.gnome.settings-daemon.plugins.gsdwacom active false
It’s posted at http://linuxwacom.sourceforge.net/wiki/index.php/Tablet_Configuration#gnome-settings-daemon
Best,
Aaron
Hi Aaron, and Ping ;)
Look like a nice idea, but this seem to be somehow outdated, or something is missing on my system :
my gsettings version is 2.52.3 ... I tried to search for "wacom" with d-conf editor, but no results...
Thanks... !
Hi Camille,
Still trying to figure out what's going on with this. Could you do
$ gsettings list-children org.gnome.settings-daemon.plugins
for me?
Best,
Aaron
Hi Aaron,
here it is :
have a nice week-end !
Camille
Maybe Carlos have an idea on what is modified on the Gnome side ?
You are right. Carlos is the expert. There have been a lot of changes in desktop environment recently. xsetwacom will not be supported, sooner or later. Plus, it is not a good idea to run xsetwacom under Gnome environment. Please file a feature request at Gnome bugzilla to make sure they will add an enable/disable Touch button in WACOM TABLET panel to offer people like you the freedom to disable touch feature.
Hi Ping,
and thanks for you answer, I reported the bug here :
https://bugzilla.gnome.org/show_bug.cgi?id=787932
I'll try to see if the bug occur if I launch my Xsetwacom script with the ~/.xinitrc file.
Thanks a lot for your work to both of you !
Last edit: Camille Bissuel 2017-09-20
Hi Ping and Aaron,
I'm sorry to say that after several weeks, starting the script from .xinitrc instaed of Gnome Tweaks do not change anything...
Cheers nevertheless !
Hi,
As explained in the the other bug, the bug is no more occuring at the kernel level after a fresh install :
linux 3.13.7
libwacom 0.24
xf86-input-wacom 0.35
gnome 3.26
xorg-server 1.19.5
Here is a log of xinput-test from today, but I do not notice any difference when the bug occur at the end of the log : simply, no more events are recorded.
My best regards,
Camille
Hi Camille,
Thanks for the update! It is nice to know the issue is not in kernel. Otherwise, Aaron would have to make a special kernel package for you to test.
Since you don't get events any more, I wonder if it is due to the device is disconnected/unpaired. Do you see the issue after you stop using the EKR for a while or while you are using it?
Ping
Hi Ping,
It's always happening when I use it : it's working at startup (without pairing) with the xsetwacom script launched from Gnome Tweaks. If I don't use it nothing happen, and it always works when I start using it.
And now when the bug occur, if I launch the script again it's temporarly fixed until next iteration (random time). Previously, I had to pair again it after the bug (or even reboot), but it's no more needed.
I try to understand since the begining what is trigerring the bug, but I can't say more than it's when I am drawing, using Krita, and pressing an EKR button (no one in particular)... it happen at a random time : can be 5 minutes, usually something like half an hour, or can be 2 or 3 hours.
Thanks for your time !
Ok, that means it was not due to the device was idled for too long that it got automatically disconnected. Your xinput log only tells us xinput doesn't get any events. Did you check kernel output?
Hi Ping... I'm sorry, how do I check kernel output ?
evtest would be one of the tools to display kernel events for you. The whole command would be "sudo evtest /dev/input/event#", where # is the event port that reports EKR events. To get a hint of the event numbers, issue "ls -l /dev/input/by-id"
Hi Ping,
here is a log with evtest during the bug (nothing happen pressing buttons in Krita and the whole Gnome desktop).
Trying to press every button, I first got a evtest crash (line 122), then every buttons where working in evtest (tested all of them several times).
Launching my xestwacom script again fix the bug. But touch support wasn't working during the bug, as set in the script, so I think xsetwacom didn't crash.
Hope this will help !
Cheers,
Camille
Last edit: Camille Bissuel 2017-10-25
Hi Camille,
Are these the steps that produced the attached log?
If evtest gets the "No such device" message we also need the journalctl log
from that moment. In general the journalctl log will be helpful.
Best,
Aaron
On Wed, Oct 25, 2017 at 5:30 AM, Camille Bissuel yagraph@users.sf.net
wrote:
Related
Bugs:
#345Hi,
yes, that's correct, and here is the log for this journalctl boot.
I think the interesting part is at "oct. 25 14:22:08" line 2299 ...
and when the bug occur, there is a lot of "gdk_device_update_tool: assertion 'GDK_IS_DEVICE (device)' failed"...
Thansk again !
Camille
I see in this log file that the EKR is created in the kernel multiple times:
mocaw@mocwa:~/Downloads$ cat EKR\ journalctl\ 25oct2017.log | grep acom |
grep input
oct. 25 13:29:30 nylnook-home kernel: input: Wacom Cintiq 27QHD touch Pen
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.3/3-1.3:1.0/0003:056A:032B.0004/input/input10
oct. 25 13:29:30 nylnook-home kernel: input: Wacom Cintiq 27QHD touch Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.3/3-1.3:1.0/0003:056A:032B.0004/input/input12
oct. 25 13:29:30 nylnook-home kernel: wacom 0003:056A:032B.0004: hidraw3:
USB HID v1.10 Mouse [Wacom Co.,Ltd. Cintiq 27QHDT Tablet] on
usb-0000:04:00.0-1.3/input0
oct. 25 13:29:30 nylnook-home kernel: wacom 0003:056A:0331.0005: hidraw4:
USB HID v1.10 Device [Wacom Co.,Ltd. ExpressKey Remote ] on
usb-0000:04:00.0-1.4/input0
oct. 25 13:29:30 nylnook-home kernel: input: Wacom Cintiq 27QHD touch
Finger as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.6/3-1.6:1.0/0003:056A:032C.0006/input/input17
oct. 25 13:29:30 nylnook-home kernel: wacom 0003:056A:032C.0006: hidraw5:
USB HID v1.11 Device [Wacom Co., Ltd. Cintiq 27QHDT Touch] on
usb-0000:04:00.0-1.6/input0
oct. 25 13:29:30 nylnook-home kernel: input: Wacom Express Key Remote Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.4/3-1.4:1.0/0003:056A:0331.0005/input/input28
oct. 25 14:20:20 nylnook-home kernel: input: Wacom Express Key Remote Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.4/3-1.4:1.0/0003:056A:0331.0005/input/input35
oct. 25 14:22:15 nylnook-home kernel: input: Wacom Express Key Remote Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.4/3-1.4:1.0/0003:056A:0331.0005/input/input36
oct. 25 14:37:29 nylnook-home kernel: input: Wacom Express Key Remote Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.4/3-1.4:1.0/0003:056A:0331.0005/input/input37
oct. 25 14:50:26 nylnook-home kernel: input: Wacom Express Key Remote Pad
as
/devices/pci0000:00/0000:00:1c.0/0000:04:00.0/usb3/3-1/3-1.4/3-1.4:1.0/0003:056A:0331.0005/input/input38
Which looks to me like it is going away and coming back. Though that could
be because you unplugged and re-plugged the USB Dongle or issued the
"unpair_remote" command in sysfs. Did you do either of those things, or
anything else that would cause you to repair the device?
If not, the next time you see this behavior could could also "ls -la
/dev/input" to see if the earlier Remotes were removed by the system.
For my notes, I setup an up to date Antergos system with input-wacom
0.37.1. I then did an evemu recording of some fairly aggressive button
pressing. I then replayed that recording in a loop for a couple hours, I
wasn't able to reproduce the multiple input devices in journalctl seen
above. This still could be a hardware issue.
Best,
Aaron