#342 ExpressKey Remote kernel bug
Hi dear Linux Wacom mainteners...
We discussed in february about the support of the ExpressKey remote :
https://sourceforge.net/p/linuxwacom/mailman/message/35651898/
The issue continue, meaning now I have a systemd error each time I start my system :
août 14 09:20:19 nylnook-home kernel: wacom 0003:056A:0331.0005: wacom_set_report: ran out of retries (last error = -32)
To start the EKR, I have to pair it (holding the button 5 seconds), then start my xsetxacom script (joined)...
Times to time while using it, the EKR stops working, and I have to pair it again and restart my script again, and it works again.
I'm still using a Cintiq 27 and Antergos Linux with Gnome.
My packages are :
xf86-input-wacom 0.35
libwacom 0.24
linux 4.12
gnome-shell and gnome-control-center 3.24.3
input-wacom is not installed
and I'm drawing with Krita 3.1.4
before pairing the EKR, lsusb give me:
Bus 004 Device 004: ID 056a:0337 Wacom Co., Ltd
Bus 004 Device 003: ID 056a:0337 Wacom Co., Ltd
Bus 004 Device 002: ID 056a:032d Wacom Co., Ltd
Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 003 Device 008: ID 056a:032c Wacom Co., Ltd
Bus 003 Device 007: ID 0403:6014 Future Technology Devices International, Ltd FT232H Single HS USB-UART/FIFO IC
Bus 003 Device 006: ID 056a:0331 Wacom Co., Ltd
Bus 003 Device 005: ID 056a:032b Wacom Co., Ltd
Bus 003 Device 004: ID 056a:0338 Wacom Co., Ltd
Bus 003 Device 003: ID 056a:0338 Wacom Co., Ltd
Bus 003 Device 002: ID 056a:032e Wacom Co., Ltd
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 003: ID 046d:c311 Logitech, Inc. Y-UF49 [Internet Pro Keyboard]
Bus 001 Device 002: ID 0458:003a KYE Systems Corp. (Mouse Systems) NetScroll+ Mini Traveler / Genius NetScroll 120
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
and nothing change after pairing...
and for xinput list:
⎡ Virtual core pointer id=2 [master pointer (3)]
⎜ ↳ Virtual core XTEST pointer id=4 [slave pointer (2)]
⎜ ↳ Genius USB Optical Mouse id=9 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Pen stylus id=12 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Finger touch id=13 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Pen eraser id=15 [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)]
↳ Power Button id=7 [slave keyboard (3)]
↳ Sleep Button id=8 [slave keyboard (3)]
↳ BTC USB Multimedia Keyboard id=10 [slave keyboard (3)]
↳ BTC USB Multimedia Keyboard id=11 [slave keyboard (3)]
↳ Eee PC WMI hotkeys id=14 [slave keyboard (3)]
which change after pairing and starting the xsetwacom script :
⎡ Virtual core pointer id=2 [master pointer (3)]
⎜ ↳ Virtual core XTEST pointer id=4 [slave pointer (2)]
⎜ ↳ Genius USB Optical Mouse id=9 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Pen stylus id=12 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Finger touch id=13 [slave pointer (2)]
⎜ ↳ Wacom Cintiq 27QHD touch Pen eraser id=15 [slave pointer (2)]
⎜ ↳ Wacom Express Key Remote Pad pad 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)]
↳ Power Button id=7 [slave keyboard (3)]
↳ Sleep Button id=8 [slave keyboard (3)]
↳ BTC USB Multimedia Keyboard id=10 [slave keyboard (3)]
↳ BTC USB Multimedia Keyboard id=11 [slave keyboard (3)]
↳ Eee PC WMI hotkeys id=14 [slave keyboard (3)]
Any idea after several months ?
please ask me any other test you need...
Thanks again !
Hi,
I haven't ben able to reproduce this exactly, but I'm just going to respond here to let you know that I started looking at this again.
And I have a quick question: do you have to re-pair the remote every time you restart the machine (that is if you rebooted before you experience the bug)? If so, did you have to re-pair on boot like this before kernel 4.9?
And I'll leave myself a couple breadcrumbs here too.
Best,
Aaron
Actually I changed my mind, that patch I mentioned in #2 could be more important than I thought. Could you test this branch: https://github.com/skomra/input-wacom/tree/EKR ?
If you need help building it let me know back channel.
Thanks,
Aaron
Hi Aaron, and thanks again for your dedicated support !
Sorry we are so far away and have to wait for the night to answer each other because of the jet lag....
Yes, I have to re-pair every time I restart the computer, and that wasn't happening before 4.9... But I can live with that, because now re-pairing and launching the script again immediatly (but temporarly) fix the bug... That's why I didn't bother you before, now I have some time to try to fix this.
Before this situation (before linux 4.9 I suppose), the EKR was working at boot, but when it stoped working I had to reboot or delete a file in /sys as you explained to me.
By the way, I'm still in hope the EKR will receive a proper support in Gnome (corresponding bug here as you know as the author : https://bugzilla.gnome.org/show_bug.cgi?id=750745)
I'll try to build that patch and let you know...
After some errors, I succesfully build and installed your branch, and I confirm: the systemd message disapear at startup, wich is nice, but I still have to re-pair and launch the script to make it work.
I just found out that if the EKR is connected with an USB cable (I was testing that in the discussion with Carlos Ganacho), I don't have to re-pair it again at startup after a reboot !
(using your branch obviously, but I din't find that before, so it may be your branch wich fixed this...)
Last edit: Camille Bissuel 2017-08-15
After some more tests, it seems that's fixed with your patch... USB cable connected or not... I don't know why it didn't work the first time, maybe I had not rebooted... ?
Can I let this bug open for the moment, and confirm you later it's solved once I tested the EKR instensively with your patch ?
Too beautiful to be true : in fact I'm back to the pre-4.9 situation...
The EKR is correctly detected at statup and work, but after a while it stop working, and I can't make it work again (even with your trick in /sys), unless I reboot...
So, there is something in your patch which is right, and something which is wrong ?
Thanks again Aaron !
Hi, thanks for the testing. So the systemd message wasn't directly
related as I had suspected at first. Your issue is still that it
randomly stops working and only a reboot will fix the issue.
Since the systemd patch could affect this issue, I have to check this
again: when it stops working do you see a serial number in the wacom
remote directory? eg 18437 below:
find the remote dir
$ find . -iname "unpair_remote"
./devices/pci0000:00/0000:00:14.0/usb3/3-2/3-2:1.0/0003:056A:0331.0001/wacom_remote/unpair_remote
list the directory that "unpair_remote" is in
$ ls ./devices/pci0000:00/0000:00:14.0/usb3/3-2/3-2:1.0/0003:056A:0331.0001/wacom_remote/
18437 unpair_remote
Best,
Aaron
On Tue, Aug 15, 2017 at 8:32 AM, Camille Bissuel yagraph@users.sf.net wrote:
Hi Aaron,
for some reason, the find command you ask me to enter didn't work, but I managed to find a serial number and the "unpair-remote" (still using your patch) :
Im' not sure, but I have the feeling the bug occur when I press the button 22 (bottom right) of the EKR, mapped to "Ctrl Z" (cancel), right after using button 21 (bottom left), which is mapped to "Ctrl" which allow to zoom in and out in Krita, among other things. Maybe button 21 is not fully raised when I press button 22 (sometimes EKR buttons get a little stuck), so "Ctrl" is triggered twice at the same time. And maybe that cause the bug...
I'll try to check that by testing....
Wrong idea, I have no succes in making the bug occur by triggering several buttons at the same time...
Hi Camille,
It seems like we've fixed the "ran out of retries" error here. That (with my new branch) your remote is still paired at the kernel level and still sends kernel events when you expereience the bug. And that you have to re-run your xsetwacom script for the remote to work in Krita. As I re-read it, I think that was the status of the earlier bug that we discussed in the message thread at https://sourceforge.net/p/linuxwacom/mailman/message/35651898/
If my understanding of that bug and the current situation is correct, could you open a separate bug to track the fact that you have to re-run the script to make the EKR work.
Aaron
Hi Aaron,
yes we have fixed the "ran out of retries" ;)
But the situation now is not exactly as previous : Now after the bug, even with pairing again and re-run the script, the EKR will not respond until reboot....
Let me know if I need to start another bug...
Ok thanks. Wait on the other bug for now.
Before this "ran out of retries" bug, when you have to re-pair and
re-run the script: Has the LED light on the remote always gone off
(have you just turned the EKR back on) or do you sometimes see the bug
when the LED is still lit? My remote turns off after what I'm guessing
is 10 minutes, but I'm not sure how long it takes the production EKRs
to turn off.
On Wed, Aug 16, 2017 at 12:09 PM, Camille Bissuel yagraph@users.sf.net wrote:
Related
Bugs:
#342Ok for not the other bug ;)
I'm not absolutely sure about the led still or blinking, let me check that tomorrow... but when the bug occured, it was while using it, drawing... It always happen while drawing and using it. So I suppose the led is on and still when the bug occur.
And it's also true that if I don't use the EKR for said 10 minutes, it turns off automatically. Never timed that preciselely ;)
But, again and inversely the bug do not occur if I don't use it. I don't know what is triggering it sadly, even if I try to ask myself what was the last pattern or action each time...
So one problem I'm having here is that I can't reproduce this, I set
up a system with Antenegros and a display tablet (I may need to track
down a 27QHD to exactly match your setup).
The second problem is that there are two bugs, and I haven't been
asking the right questions. From your point of view the EKR stops
working and you'll say "EKR will not respond until reboot." and " I
can't make it work again " but I didn't ask if the kernel events
stopped (evtest) or if the EKR just stopped working in Krita.
So let's separate this in to two bugs. This (bug 342) should be the
kernel bug. If the remote stops working and you can no longer see
events in evtest it is a kernel bug and it belongs here.
If (after the EKR stops working in an application like Krita) you do
see events in evtest (this seems related to the problem where you have
to re-run your script) then put that in a separate bug.
Also when you run Krita, run it from the command line. The output in
its logs may be helpful.
Best,
Aaron
On Wed, Aug 16, 2017 at 1:50 PM, Camille Bissuel yagraph@users.sf.net wrote:
Looking through your script, and this is really for the other bug, I'd
prefer to wait for you to open the other bug, but I'm posting it here
because of the time difference. Don't reply to this message here, but
instead on the new bug.
These two lines
"xsetwacom set "Wacom Cintiq 27QHD touch Pen stylus" Area 400 472 119740 67520
xsetwacom set "Wacom Cintiq 27QHD touch Finger touch" Touch off"
don't seem like they should work. The "touch"es that follow 27QHD
should be TOUCH (all caps) and I think the "touch" that precedes "Pen
stylus" should actually be "TABLET". I would expect you should get an
error similar to 'Cannot find device "Wacom Cintiq 27QHD touch Pen
stylus'" for these two lines.
You can double check the device names with "xsetwacom list".
Hi Aaron,
I'm sorry to say I'm leaving today for some (once in a year) holidays, far from my computer, and I'll be back on Monday, August, the 28th... I'm sorry I put you at work and let you wait... But I'm ready to continue to test when I come back. Many thanks for your work !
Triggering the bug takes a random time while using the EKR, so I can't answer your latest questions fully right now, but :
1. I didn't try evetest after the bug with your patch, but the EKR is not working outside Krita neither. I'll try running Krita command line and evtest when I come back.
2. my script is working for me ... here is my xsetwacom list
Again, thanks a lot, I'm sorry for the delay, my best regards from the French Alps !
Hi Aaron,
I'm back, and I continued in bug 345 ...
It evolved : today I can make the EKR work again after a bug If I pair it again and re-run the script...
Seem to me the kernel part is fixed by your patch anyway !
I can't edit the title of this bug, but you can add "kernel" to it if you want and can ...
Cheers, and thanks a lot,
Camille
Hi Aaron,
I'm back to uou beacause today I experience a crash again with linux 4.13.3 patched with your input-wacom-EKR branch.
The patch make the systemd "ran out of retries" error disapear, but apparentelty the bug came back, and it don't work even if I pair the EKR again and relaunch my xsetwacom script
During the crash, I see the EKR in xinput, but the EKR is not reacting to touch with evtest and evemu-record.
Joined is a log with all the logs I was able to think of...
Thanks again, this is a long story...
and $ find . -iname "unpair_remote" doesn't find anything...
On Thu, Oct 5, 2017 at 8:07 AM, Camille Bissuel yagraph@users.sf.net
wrote:
Can you check the output of "modinfo wacom | grep version" to see that the
input-wacom driver you installed is still loaded?
Best,
Aaron
Related
Bugs:
#342On Thu, Oct 5, 2017 at 8:37 AM, Aaron Skomra skomra@users.sf.net wrote:
Related
Bugs:
#342Hi, here is the result:
$ modinfo wacom | grep version
version: v2.00-input-wacom-EKR
srcversion: 36BDD850B4C8E6194E7671D
vermagic: 4.13.3-1-ARCH SMP preempt mod_unload modversions
for some reason I just found there is no more input-wacom package in Archlinux :
https://www.archlinux.org/packages/?sort=&q=input-wacom&maintainer=&flagged=
(no more in AUR)
So I'm compiling the 0.37 version from source waiting for the kernel integration ;)
Cheers !