[Bug 1013647] New: dell 7470 - trackpoint issues
http://bugzilla.suse.com/show_bug.cgi?id=1013647 Bug ID: 1013647 Summary: dell 7470 - trackpoint issues Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-maintainers@forge.provo.novell.com Reporter: tchvatal@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- With kernel 4.8.11 or 4.9.0rc7 the trackpoint still behave weirdly. a) its sensitivity is too high, small nudge moves the screen all around the display -> maybe reducing default speed reported by the device could be enough b) trackpoint buttons sometimes on press report doubleclick on the machine, ie instead of moving cursor in app you mark the text/word -> discussed with jikos and the evtest reports the events correctly, so this might be some libinput issue c) trackpoint profile seems kinda linear, regardless of a pressure it moves too quickly, either small nudge or full press make the same movement. All of the above observable on Gnome in X and Wayland -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c1 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tiwai@suse.com Component|Kernel |X.Org Assignee|kernel-maintainers@forge.pr |xorg-maintainer-bugs@forge. |ovo.novell.com |provo.novell.com QA Contact|qa-bugs@suse.de |xorg-maintainer-bugs@forge. | |provo.novell.com --- Comment #1 from Takashi Iwai <tiwai@suse.com> --- Most of issues sound more like a configuration or libinput issue. Reassigned to X team at first. Feel free to reassign back to kernel once when it's figured out to be a kernel bug. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c2 Stefan Dirsch <sndirsch@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tchvatal@suse.com Flags| |needinfo?(tchvatal@suse.com | |) --- Comment #2 from Stefan Dirsch <sndirsch@suse.com> --- Please check this first: https://wiki.microfocus.net/index.php?title=SUSE/Labs_Publications/Linux_Sup... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c3 Tomáš Chvátal <tchvatal@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(tchvatal@suse.com | |) | --- Comment #3 from Tomáš Chvátal <tchvatal@suse.com> --- (In reply to Stefan Dirsch from comment #2)
Please check this first:
https://wiki.microfocus.net/index.php?title=SUSE/Labs_Publications/ Linux_Supportability_For_Corporate_Notebooks#DELL_Latitude_E7470
As I stated I am on 4.8.11 kernel with latest tumbleweed where the trackpoint is properly identified as a trackpoint. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c4 Max Staudt <mstaudt@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mstaudt@suse.com --- Comment #4 from Max Staudt <mstaudt@suse.com> --- (In reply to Tomáš Chvátal from comment #0)
With kernel 4.8.11 or 4.9.0rc7 the trackpoint still behave weirdly.
Why do you say "still"? If there is a precursor to this bug, could you please link to it? As for the trackpoint, we can't check this without the hardware. I know that ThinkPads have sysfs settings for their trackpoints - maybe your Dell trackpoint has them too and you can use them to turn down the sensitivity? If yes, then we can come up with a udev rule for your stick. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c5 --- Comment #5 from Tomáš Chvátal <tchvatal@suse.com> --- (In reply to Max Staudt from comment #4)
(In reply to Tomáš Chvátal from comment #0)
With kernel 4.8.11 or 4.9.0rc7 the trackpoint still behave weirdly.
Why do you say "still"?
If there is a precursor to this bug, could you please link to it?
As for the trackpoint, we can't check this without the hardware. I know that ThinkPads have sysfs settings for their trackpoints - maybe your Dell trackpoint has them too and you can use them to turn down the sensitivity? If yes, then we can come up with a udev rule for your stick.
Well still because of bsc#987703 which is supposed to be fixed starting 4.8.3. On xinput I can of course slow the trackpoint down, but still it is linear, either full speed or 0, even if it should have pressure control. Ie. xinput set-prop 13 'libinput Accel Speed' -1.0 More problematic is the trackpoint buttons pressing. Really the evtest reports just one click but quite often the result is simply marked text. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c6 --- Comment #6 from Max Staudt <mstaudt@suse.com> --- (In reply to Tomáš Chvátal from comment #5)
Well still because of bsc#987703 which is supposed to be fixed starting 4.8.3.
Thanks! I wasn't aware of that.
On xinput I can of course slow the trackpoint down, but still it is linear, either full speed or 0, even if it should have pressure control. Ie. xinput set-prop 13 'libinput Accel Speed' -1.0
AFAIK, Accel Speed doesn't like negative values. Correct me if I'm wrong though.
More problematic is the trackpoint buttons pressing. Really the evtest reports just one click but quite often the result is simply marked text.
Huh, this keeps getting weirder and weirder. Do you know whether we have spare hardware to test this with? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c7 --- Comment #7 from Tomáš Chvátal <tchvatal@suse.com> --- About the accel values I think so too, it behaves slightly weirdly in neg numbers, but simply put with 0 it is acting like on speed :) FWIW as the 7470 is the currentgen dell IS&T hw we all get MF admins should have some for lending or I think kernel lads should've received some for testing. If neither is true I am more than willing just pack this one and send it to where needed for debugging and fixing :) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 Christian Deckelmann <christian.deckelmann@microfocus.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |christian.deckelmann@microf | |ocus.com -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 Michal Marek <mmarek@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mmarek@suse.com -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c8 Max Staudt <mstaudt@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |CONFIRMED CC| |sknorr@suse.com --- Comment #8 from Max Staudt <mstaudt@suse.com> --- Ok, Stefan Knorr (CC'd) lent me a machine and I tried my best to configure this thing. Unfortunately there is no sysfs interface to configure the stick, unlike with the ThinkPad's TrackPoint. To understand what I mean, please refer to this: https://wiki.archlinux.org/index.php/TrackPoint So what's left is to check userspace. There used to be a tool called "tpconfig", but that doesn't have an option concerning speed. I played around with libinput, and a negative acceleration value of -1 indeed helps. Wow. The best results I've gotten so far are with xf86-input-evdev. It provides more options through xinput, including "Device Accel Constant Deceleration", which I set to 3. It also feels much more natural. Feel free to play around with the other settings as well. However this still doesn't solve the (IMHO) annoying issue of having to push the stick quite a bit before it reacts at all, and when it does, it jumps around quite a bit. The above evdev option softens the blow a lot though. As a workaround, I suggest creating a setting in /etc/X11/xorg.conf.d/99-evdev-dell-latitude-e7470.conf to select the other driver: Section "InputClass" # openSUSE bug #1013647 Identifier "DELL TrackStick by ALPS" MatchProduct "AlpsPS/2 ALPS DualPoint Stick" MatchIsPointer "on" Driver "evdev" Option "EmulateWheel" "on" Option "EmulateWheelButton" "2" # Since middle button is in use now these button events need to be emulated (bnc#869097) Option "Emulate3Buttons" "on" EndSection -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c9 Max Staudt <mstaudt@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jkosina@suse.com Flags| |needinfo?(jkosina@suse.com) --- Comment #9 from Max Staudt <mstaudt@suse.com> --- Jiri, can you please have a look at the kernel driver? The driver does not generate events for light movements of the stick, and floods the stack with events once the tipping point has been crossed. It's basically impossible to get events with very small movements out of this stick. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c10 Jiri Kosina <jkosina@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |brian.coons@microfocus.com Flags|needinfo?(jkosina@suse.com) | --- Comment #10 from Jiri Kosina <jkosina@suse.com> --- (In reply to Max Staudt from comment #9)
Jiri, can you please have a look at the kernel driver?
I unfortunately can't. IS&T never provided me with a machine to develop / debug the driver on.
The driver does not generate events for light movements of the stick, and floods the stack with events once the tipping point has been crossed. It's basically impossible to get events with very small movements out of this stick.
-- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c11 Max Staudt <mstaudt@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|X.Org |Kernel Assignee|xorg-maintainer-bugs@forge. |kernel-maintainers@forge.pr |provo.novell.com |ovo.novell.com QA Contact|xorg-maintainer-bugs@forge. |qa-bugs@suse.de |provo.novell.com | --- Comment #11 from Max Staudt <mstaudt@suse.com> --- Reassigning this to Kernel team, so they can have a look should they ever receive such a machine. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c18 --- Comment #18 from Martin Wilck <martin.wilck@suse.com> --- This recent mail from the Research ML is worthwhile: "More and more people here have the Dell E7470 with the ALPS DualStick Trackpoint. Unfortunatly, it is so sensitive on a normal Linux system, that it is unuseable. But: if I boot the installer, the Stick works perfect! So I tried to find out what the differnce is: in the normal system, the Stick is detected correct as ALPS DualStick, while in the installer, it is only detected as ImPS mouse. I wasn't able to find out why this is the case." And later:
Please try:
psmouse.proto=imps
On your kernel cmdline.
Thanks, that works! The stick works with the imps protocol but not with alps/2. Perhaps we could approach upstream again with this information. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 Stephan Lahl <stephan.lahl@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |stephan.lahl@suse.com -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c19 --- Comment #19 from Max Staudt <mstaudt@suse.com> --- What you're describing is the fallback mode where the trackpad emulates a PS/2 mouse. This is used when no ALPS driver (with extended features) has been loaded. If this mode works fine for you, blacklisting the driver should help. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c20 --- Comment #20 from Martin Wilck <martin.wilck@suse.com> --- So you're saying the HW is operating in a different mode, depending on the driver used? I was hoping that comment #18 might provide some insight what's going wrong in the alps driver (interpretation of HW readings or whatnot). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c21 --- Comment #21 from Max Staudt <mstaudt@suse.com> --- (In reply to Martin Wilck from comment #20)
So you're saying the HW is operating in a different mode, depending on the driver used?
Exactly. That's a very common thing in PC laptops. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 Martin Pluskal <mpluskal@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mpluskal@suse.com -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c24 --- Comment #24 from Martin Wilck <martin.wilck@suse.com> --- 4.11rc4 contains some patches that might make a difference here. I was notified about them by Jörg's git-fixes script: 47e6fb4212d0 Input: ALPS - fix trackstick button handling on V8 devices e7348396c6d5 Input: ALPS - fix V8+ protocol handling (73 03 28) In particular the 2nd one may be interesting, it applies to ALPS devices with device ID 'E7="73 03 28"', not sure if that applies to E7470. I wonder if some of the touchstick users could give 4.11rc4 a try. I could do it myself, but I fear that I'm not qualified to assess whether or not the stick is working as smoothly as it would be expected. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c25 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|CONFIRMED |RESOLVED CC| |jslaby@suse.com Resolution|--- |FIXED --- Comment #25 from Jiri Slaby <jslaby@suse.com> --- Given nobody complained after that, consider this RESOLVED as a ping. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c26 Johannes Segitz <jsegitz@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jsegitz@suse.com --- Comment #26 from Johannes Segitz <jsegitz@suse.com> --- (In reply to Martin Wilck from comment #18) For the record, the psmouse.proto=imps workaround seems to be better since it allows you to speed up the cursor movement by applying more pressure -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c27 --- Comment #27 from Takashi Iwai <tiwai@suse.com> --- (In reply to Johannes Segitz from comment #26)
(In reply to Martin Wilck from comment #18) For the record, the psmouse.proto=imps workaround seems to be better since it allows you to speed up the cursor movement by applying more pressure
Could you check 4.11 or 4.12-rc kernel as Martin suggested in comment 24? At least, due to the recent changes of alps driver, the behavior of the touchpad on E7270 changed slightly (there is a known regression by that, too, unfortunately), so I guess E7470 trackpoint may be affected somehow. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c28 --- Comment #28 from Martin Wilck <martin.wilck@suse.com> --- I find it disconcerting that the generic legacy driver works better than the specific driver "with extended features" (comment 19). But I lack deeper knowledge of these devices, I'm only on this bug because I have an E7470 laptop, too. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c29 Martin Wilck <martin.wilck@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- URL| |https://bugs.freedesktop.or | |g/show_bug.cgi?id=106323 --- Comment #29 from Martin Wilck <martin.wilck@suse.com> --- The POINTINGSTICK_CONST_ACCEL trick ceased to work with libinput 1.9. See https://bugs.freedesktop.org/show_bug.cgi?id=106323 for technical background. Once this bug is solved, there'll probably be a better solution for recent libinput versions, too. Update on possible workarounds, other than falling back from the ALPS driver to the generic PS/2 driver: Solution 1: Use evdev Under X, use the evdev input driver, and set the ConstantDeceleration property to 8 or more. # cat >/etc/X11/xorg.conf.d/60-alps.conf <<EOF Section "InputClass" Identifier "DualPoint Stick" MatchIsPointer "true" MatchProduct "DualPoint Stick" MatchDevicePath "/dev/input/event*" Driver "evdev" Option "AccelerationProfile" "2" Option "ConstantDeceleration" "8" EndSection EOF Solution 2: libinput < 1.9 With libinput < 1.9 (SLE12, Leap 42.x): set POINTINGSTICK_CONST_ACCEL to a low value: # cat >/etc/udev/rules.d/80-libinput.rules <<EOF ACTION=="add|change", KERNEL=="event[0-9]*", ENV{ID_INPUT_POINTINGSTICK}=="?*", \ ENV{POINTINGSTICK_CONST_ACCEL}="0.1" EOF Solution 3: libinput >= 1.9 With libinput 1.9 and higher (SLE15, Tumbleweed), set LIBINPUT_ATTR_TRACKPOINT_RANGE to the maximum value 100: # cat >/etc/udev/rules.d/80-libinput.rules <<EOF ACTION=="add|change", KERNEL=="event[0-9]*", ENV{ID_INPUT_POINTINGSTICK}=="?*", \ ENV{LIBINPUT_ATTR_TRACKPOINT_RANGE}="100" EOF Furthermore, set the acceleration profile in libinput to "flat" and the acceleration factor to a low value such as -0.9, using xinput commands. Note: The results of this are much worse than using evdev, see above. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c30 --- Comment #30 from Martin Wilck <martin.wilck@suse.com> --- The libinput issues are more or less settled now. I'm providing updated packages of latest libinput (1.11 pre) with the relevant patches here:
https://download.opensuse.org/repositories/home:/mwilck:/branches:/X11:/Wayl...
This should make the stick on the DELL E7470 usable with libinput. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c31 --- Comment #31 from Stefan Dirsch <sndirsch@suse.com> --- on sle15 we meanwhile have libinput 1.10.5. Is this not sufficient? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c32 --- Comment #32 from Martin Wilck <martin.wilck@suse.com> --- (In reply to Stefan Dirsch from comment #31)
on sle15 we meanwhile have libinput 1.10.5. Is this not sufficient?
No it's not. We're talking 1.10.901 + a few patches on top here. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c33 --- Comment #33 from Martin Wilck <martin.wilck@suse.com> --- ... but a backport of the ALPS fixes to a stable release shouldn't be too hard. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c34 --- Comment #34 from Martin Wilck <martin.wilck@suse.com> --- I can only encourage everyone to try the package from home:mwilck:branches:X11:Wayland. It works like a charm! I've unconciously started to use the stick which I've been avoiding for about 2 years. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c35 --- Comment #35 from Martin Wilck <martin.wilck@suse.com> --- http://download.opensuse.org/repositories/home:/mwilck:/branches:/X11:/Wayla... now contains libinput 1.10.7 with the backported patches. Works nicely, too. I'll submit this to factory. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c36 --- Comment #36 from Stefan Dirsch <sndirsch@suse.com> --- (In reply to Martin Wilck from comment #35)
http://download.opensuse.org/repositories/home:/mwilck:/branches:/X11:/ Wayland/ now contains libinput 1.10.7 with the backported patches. Works nicely, too. I'll submit this to factory.
Great! Thanks a lot! -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1013647 http://bugzilla.suse.com/show_bug.cgi?id=1013647#c37 --- Comment #37 from Martin Wilck <martin.wilck@suse.com> --- While this was fixed in upstream libinput 1.11, 1.12 managed to break it again. https://gitlab.freedesktop.org/libinput/libinput/issues/148#note_36487 For now, the workaround is to comment out the "Latitude E7470 Trackpoint" section in /usr/share/libinput-1.12.0/libinput/50-system-dell.quirks, or to set the "AttrTrackpointMultiplier" value to a lower value (e.g. 0.125) in that section. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com