Opened 2 years ago
Closed 2 years ago
#17841 closed bug (fixed)
PS2 Touchpad Disconnects
Reported by: | navjack27 | Owned by: | waddlesplash |
---|---|---|---|
Priority: | normal | Milestone: | R1/beta4 |
Component: | Drivers/Input/PS2/Mouse | Version: | R1/Development |
Keywords: | Cc: | ||
Blocked By: | Blocking: | ||
Platform: | All |
Description
On my Gateway laptop model MS2274 I have Haiku's nighty installed. Sorry for being vague but somewhere between hrev56272 and hrev56297 a change was done that causes the PS2 touchpad to disconnect after being touched. I'll attach the current full syslogs. If you need me to I can go back to a known working version and then get the syslogs from that.
Attachments (3)
Change History (11)
by , 2 years ago
Attachment: | new_broken_syslogs.7z added |
---|
comment:1 by , 2 years ago
comment:2 by , 2 years ago
Oh and there's also hrev56293 which changed the touchpad code quite a lot, so that's more likely to be related.
comment:3 by , 2 years ago
i attached two new files. i used pkgman to sync to previous releases and found that 56295 is where the mouse has the disconnect bug introduced.
comment:4 by , 2 years ago
Ok so in the new log we can see this:
2570 KERN: Mouse device exiting, Operation timed out
Indeed it seems something did not work as expected with the move of movementmake to userspace in hrev56293. However I don't really see how this could happen by looking at the code?
follow-up: 7 comment:6 by , 2 years ago
Owner: | set to |
---|---|
Status: | new → in-progress |
comment:7 by , 2 years ago
Replying to waddlesplash:
hrev56301 works! No mouse issues. Awesome! Thank you.
comment:8 by , 2 years ago
Milestone: | Unscheduled → R1/beta4 |
---|---|
Resolution: | → fixed |
Status: | in-progress → closed |
Thanks for reporting and testing :)
I suspected hrev56268 (PS2 keyboard reset changes) but that's just before the range you listed, so, if that range is correct, that's not it.
My next guess, but that is really a guess, is that hrev56274 (kernel/scheduler: Reset the invoke_scheduler flag on reschedule) allows for some race condition with the interrupts in the PS2 code somehow?