Bug 34374

Summary:	hwclock cannot access the cmos clock
Product:	[Retired] Red Hat Linux	Reporter:	Steve Chang <spammenot>
Component:	util-linux	Assignee:	Elliot Lee <sopwith>
Status:	CLOSED DEFERRED	QA Contact:	David Lawrence <dkl>
Severity:	high	Docs Contact:
Priority:	medium
Version:	7.2	CC:	jorton
Target Milestone:	---
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2002-01-17 13:31:20 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Steve Chang 2001-04-02 20:20:27 UTC

From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0; COM+ 
1.0.2204)


On two out of three of our Tyan S2510 and four of eight SuperMicro P6DGE 
dual PIII motherboards, hwclock hangs when run, including at boot. Ctrl-C 
will abort this step, but then it gets hung up indefinitely again when 
starting up the system logger. Ctrl-\ will abort this step and allows the 
system to boot completely, but the system clock is, of course, not 
initialized. 

Reproducible: Sometimes
Steps to Reproduce:
1. attempt to boot.

Comment 1 S P Arif Sahari Wibowo 2001-07-03 00:11:18 UTC

I second that.

I have a system with Asus A7V133 w/audio, Athlon 850 Mhz, Diamond S540 (Savage4).
Honestly I think it might be hardware problem, but the CMOS clock seems fine
from the BIOS menu. Anybody have an idea on what is the cause?
It was Ok for few days, and then this problem started. I got around it by
renaming hwclock and put an empty file there.

Comment 2 Elliot Lee 2001-07-18 17:25:17 UTC

Hmm, sounds almost like a kernel+hardware problem (since I believe /dev/rtc is
used for clock access...). Not sure how to investigate yet, though.

Comment 3 Joe Orton 2002-01-17 13:15:50 UTC

I have a similar A7V-based system which exhibits this problem too (using a stock
7.2 install, and also using the errata kernel). The hwclock invocation in
init.d/halt also hangs. A workaround for both problems is to add

CLOCKFLAGS="--directisa" 

to /etc/sysconfig/hwclock

once you've managed to boot.  According to http://freshmeat.net/releases/41231/,
there is a version of hwclock which also works around this problem:

"This release avoids the hang (usually a boot time hang) on machines with broken
rtc drivers by timing out after 2 seconds waiting for the clock interrupt, then
falling back to a non-interrupt clock access method."

Comment 4 Joe Orton 2002-01-17 13:20:48 UTC

Additional info is that I never had any problems running hwclock (without
--directisa) using a 2.2 kernel on the same machine.

Comment 5 Joe Orton 2002-01-17 13:31:14 UTC

(Oops, that flag needs to be added to /etc/sysconfig/clock for the workaround)

Comment 6 Elliot Lee 2002-01-29 22:20:11 UTC

I'll suggest to the util-linux maintainer to incorporate the forked hwclock -
otherwise, I'm not going to touch hwclock for fear of blowing up on a lot of
other systems (e.g. architectures that don't know what an ISA bus is :).

The workaround is to boot with init=/bin/sh the first time and hack up
/etc/sysconfig/hwclock.

Comment 7 Elliot Lee 2002-03-27 15:40:10 UTC

*** Bug 55683 has been marked as a duplicate of this bug. ***

Comment 8 Elliot Lee 2002-03-27 15:40:45 UTC

*** Bug 62094 has been marked as a duplicate of this bug. ***

Comment 9 Joe Orton 2002-10-02 09:18:50 UTC

This appears to be fixed in 8.0.

Comment 10 Joe Orton 2002-10-14 18:00:10 UTC

Or maybe it's not fixed. I enabled automatic power-on (at a given time) in my
BIOS, and now hwclock hangs again (repeatably) without the --directisa flag.
I'll try disabling automatic power-on and see if I can boot without passing
--directisa to hwclock.