Opened 10 years ago

Last modified 15 months ago

#3868 assigned bug

Hang on shutdown/reboot (synaptics related?)

Reported by: jackburton Owned by: nobody
Priority: normal Milestone: R1
Component: System Version: R1/pre-alpha1
Keywords: shutdown Cc: umccullough
Blocked By: Blocking:
Has a Patch: no Platform: All

Description

My laptop hangs on shutdown/reboot, if acpi is not enabled. The last thing I see on screen is "Waiting for other processes to quit", then it hangs there. No "reboot now". With ACPI Enabled, instead, it displays the "reboot now" button (it should shutdown or reboot automatically, but this should be a different bug, I guess). This started happening only since some time (a month or so).

The laptop is a Fujitsu Siemens amilo LI2735, centrino based. hrev30622.

Change History (25)

comment:1 by michael.weirauch, 10 years ago

Can confirm this. Happened yesterday with ACPI enabled. Can't recall if it always happens as my system usually freezes during operation and I regularly don't restart ;)

comment:2 by axeld, 10 years ago

The problematic change should be the one from Olivier which kills all apps before rebooting. Obviously, there is one (or more) components in your system that doesn't like to be freed.

Can you remove drivers to find the candidate?

in reply to:  2 comment:3 by jackburton, 10 years ago

Replying to axeld:

The problematic change should be the one from Olivier which kills all apps before rebooting. Obviously, there is one (or more) components in your system that doesn't like to be freed.

Yeah, that's what I was thinking, too. It's weird it doesn't happen with ACPI enabled, though.

Can you remove drivers to find the candidate?

Yes, of course. Will do that this evening.

comment:4 by michael.weirauch, 10 years ago

Oops, misread this... ("not enabled) Will retest.

comment:5 by michael.weirauch, 10 years ago

False alarm, though my initial comment still applies that I had the hang described with ACPI enabled. Perhaps accidentially.

Just tested:

  • with ACPI: reboot fine
  • w/o ACPI: reboot fine

Sorry for disrupting.

comment:6 by jackburton, 10 years ago

Tried removing hda (audio), rtl82xx (network) and the problem is still there. Will try removing more stuff.

comment:7 by jackburton, 10 years ago

Looks like the problematic driver is the touchpad one. If I comment out the whole synaptics_close() implementation, the laptop reboots correctly. Now I'll try to find what exactly hangs.

comment:8 by axeld, 10 years ago

That's not even too surprising considering that the reboot is usually triggered via the keyboard controller. Eventually, the driver needs to reset the controller when it's closed.

comment:9 by Coldfirex, 9 years ago

Is this still an issue in recent builds?

comment:10 by jackburton, 9 years ago

Will try disabling ACPI later today

comment:11 by umccullough, 9 years ago

Cc: umccullough added

comment:12 by tekk, 8 years ago

have the same problem, haven't tried messing with acpi

as a temporary fix you can use shut down and then power on manually, that doesn't seem to be a problem

Last edited 8 years ago by tekk (previous) (diff)

comment:13 by scottmc, 8 years ago

Can you recheck this with a recent Haiku build? It may have been fixed recently.

comment:14 by umccullough, 8 years ago

I have two separate machines that fail to shutdown with r1a3 still. They are both laptops with trackpads, so I have a feeling they're both related to jackburton's comment:7

I haven't had time to test that hypothesis.

comment:15 by scottmc, 8 years ago

Blocking: 7665 added

comment:16 by jackburton, 7 years ago

Still with us in alpha4.1. Any advice on how to debug this, given that I can't enter KDL when the machine is freezed?

comment:17 by korli, 7 years ago

No serial output on this laptop?

in reply to:  17 comment:18 by jackburton, 7 years ago

Replying to korli:

No serial output on this laptop?

Nope. I guess I could get an USB-Serial converter where I work, and another pc with a serial port to interact with it, but is it supported ?

comment:19 by axeld, 7 years ago

Things like these are pretty hard to debug. I would try to remove as much drivers as possible (including legacy things like PS/2), and see if the problem persists. But even if you manage to identify the one driver causing the problem (if there is any), it may not actually cause it :-)

Well, good luck, anyway :)

I had a hard time trying to debug a similar issue on my old SiS 630 based laptop. In the end I think the problem could only be avoided by a very particular shutdown sequence (concerning when to turn off interrupts, and how to talk to the keyboard controller to do the reset).

comment:20 by jackburton, 7 years ago

With recent revisions, the problem has changed a bit: I now get a KDL on shutdown (seems related to ACPI). Should I change the ticket description or should I open a new one ?

comment:21 by tqh, 7 years ago

I'd prefer a new one as I think this one may not be ACPI-related.

comment:22 by pulkomandy, 5 years ago

Blocking: 7665 removed

Not a boot failure...

comment:23 by axeld, 2 years ago

Owner: changed from axeld to nobody
Status: newassigned

comment:24 by tqh, 20 months ago

Keywords: shutdown added
Summary: Hang on shutdown/rebootHang on shutdown/reboot (synaptics related?)

Not sure if this one should be closed.

comment:25 by cocobean, 15 months ago

hrev52017 tested. This is on specific hardware (I haven't reproduced it on any test hardware yet, but known issue on a user's (vidrec) Dell-related system (but shutdown works on his other test system). I've seen it happen on Win/Linux/BSD so not Haiku-specific in nature. If any process hangs, I usually kill it in the dialog box or go back to review apps/processes which require 'kill team' tactics.

I can 'suggest to closing this ticket' and tracking on the wiki on 'specific hardware' with this issue - to consolidate and isolate.

Note: See TracTickets for help on using tickets.