Opened 14 years ago
Closed 14 years ago
#7448 closed bug (fixed)
Touchpad sucking CPU when touched
Reported by: | humdinger | Owned by: | czeidler |
---|---|---|---|
Priority: | normal | Milestone: | R1 |
Component: | Drivers/Mouse/PS2 | Version: | R1/Development |
Keywords: | synaptics, touchpad | Cc: | |
Blocked By: | Blocking: | ||
Platform: | All |
Description
This is hrev41255.
Probably since the recent work on the synaptics/touchpad, input_servers "Ps/2 Touchpad 1" thread consumes about 15% CPU in my Core2Duo when the touchpad is in use. I don't think there was a visible usage-increase before...
Attachments (4)
Change History (11)
comment:1 by , 14 years ago
comment:2 by , 14 years ago
Using hrev41256 (gcc2hybrid) with an old centrino notebook (Pentium M 1,3 GHz, 1280 MB RAM) with Synaptics Touchpad.
Top command in terminal shows a cpu usage of
- ca. 15 % for thread "Ps/2 Touchpad 1 watcher", while using 2-finger-scrolling with touchpad (see "top_touchpad.txt").
- < 1 % for thread "USB Mouse 1 watcher", while using scroll wheel of usb mouse (see "top_usb_mouse.txt").
For comparison, the output of top with hrev37573:
- < 1 % for 2-finger-scrolling with touchpad (see "top_touchpad_old.txt")
- < 1 % for scrolling with usb mouse (see "top_usb_mouse_old.txt")
by , 14 years ago
Attachment: | top_usb_mouse_old.txt added |
---|
by , 14 years ago
Attachment: | top_usb_mouse.txt added |
---|
by , 14 years ago
Attachment: | top_touchpad_old.txt added |
---|
by , 14 years ago
Attachment: | top_touchpad.txt added |
---|
comment:3 by , 14 years ago
I tried the nightly hrev41185, but couldn't get it to finish the bootup, because of that synaptics crash you have fixed with a later commit.
I then went further back to hrev41122, which should be before all your touchpad works. That one boots up and shows a CPU usage of just 0.1%, similar to taos' findings. It seems something of your changes did do the damage. :)
comment:4 by , 14 years ago
When using the touchpad, I have my syslog filled with tons and tons of things like
KERN: ALPS: movement set buttons KERN: ALPS: tap gesture timed out KERN: ALPS: movement event KERN: ALPS: movement set buttons KERN: ALPS: tap gesture timed out KERN: ALPS: no touch event KERN: ALPS: tap drag gesture timed out KERN: ALPS: tap gesture timed out KERN: ALPS: no touch event KERN: ALPS: tap drag gesture timed out KERN: ALPS: tap gesture timed out
May this excessive debugging info be the cause of the CPU suckage?
comment:5 by , 14 years ago
ah jetzt ja! that sounds like the only reasonable explanation. Could you check my last commit? thanks!
comment:6 by , 14 years ago
Perfect! That was it, everything's back to normal, i.e. 0.1% usage. Thanks very much!
Could you please check if it really comes from my changes? Can't imagine where this should could come from. Can't see the usage increase with ALPS...
thx!