Bug 14343

Summary:	very long pause before the power is turned off with clocksource=tsc
Product:	Sisyphus	Reporter:	Ivan Zakharyaschev <imz>
Component:	SysVinit	Assignee:	Dmitry V. Levin <ldv>
Status:	CLOSED WONTFIX	QA Contact:	qa-sisyphus
Severity:	minor
Priority:	P2	CC:	mike, rider
Version:	unstable
Hardware:	all
OS:	Linux

Description

Ivan Zakharyaschev 2008-02-06 18:47:55 MSK

SysVinit-2.86-alt1

I discovered this while experimenting because of the bug with system clock #14258.

$ rpm -q kernel-image-wks-smp-2.6.18-alt1 -i
Name        : kernel-image-wks-smp         Relocations: (not relocateable)
Version     : 2.6.18                            Vendor: ALT Linux Team
Release     : alt1                          Build Date: Срд 13 Дек 2006 02:28:15
Install date: Срд 21 Фев 2007 02:43:51      Build Host: bee5.hasher.altlinux.org
Group       : System/Kernel and hardware    Source RPM:
kernel-image-wks-smp-2.6.18-alt1.src.rpm
Size        : 44948311                         License: GPL
Packager    : Konstantin A. Lepikhov <lakostis@altlinux.ru>
<....>
Steps to Reproduce:
1. boot 2.6.24-wks-smp-alt1 with clocksource=tsc
2. poweroff
3.
Actual Results:  
At the end, it writes "The system is halted" and nothing happens for 2 min 50
sec. Only then the power is tuned off.

Expected Results:  
Normally, the pause is about 3 sec (e.g., on 2.6.24-wks-smp-alt1 with default
options).

Comment 1

Ivan Zakharyaschev 2008-02-06 18:56:38 MSK

A similar effect of clocksource=tsc on the init sequence (a long pause):
https://bugzilla.altlinux.org/show_bug.cgi?id=14345 .

Comment 2

Ivan Zakharyaschev 2008-02-07 05:06:20 MSK

I also experienced this delay on 2.6.24-std-def-alt2 with clocksource=tsc. The
delay on "The system is halted" was about 1 min, but there was also a delay
after the message "Asking all remaining processes to terminate", which lasted
about 3 min.

(With this option, 2.6.24-std-def-alt2 also has the bad behavior of the system
clock: https://bugzilla.altlinux.org/show_bug.cgi?id=14258#c18 ).


Perhaps, this is not a bug and an expected behavior under this kernel parameter.
But I wonder why other programs work without delays then.

Comment 3

Michael Shigorin 2010-08-25 08:48:10 MSD

Ivan, is it still there?

Comment 4

Anton Farygin 2025-02-26 12:09:44 MSK

не акутально