Bug 213380

Summary: Clock going slow on kernel-xen
Product: [Fedora] Fedora Reporter: Ahmed Kamal <email.ahmedkamal>
Component: kernel-xenAssignee: Xen Maintainance List <xen-maint>
Status: CLOSED WONTFIX QA Contact: Virtualization Bugs <virt-bugs>
Severity: high Docs Contact:
Priority: medium    
Version: 6CC: bstein, gcosta
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: 2008-02-26 23:37:38 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:
Attachments:
Description Flags
dmesg output on non-xen kernel
none
dmesg output on xen kernel
none
var/log/messages after booting my xen kernel none

Description Ahmed Kamal 2006-11-01 08:00:55 UTC
Description of problem:
Booting the Xen kernel, everything seemed to work very slow, even the animations
on the startup screen. Working on the system for half an hour, I noticed its
clock was late for 15 minutes! The clock seemed to be going slow! I did a quick
test of "sleep 30" which should sleep for 30sec, instead it returned after
80seconds!!! So, the system is working about 3 times slower

Version-Release number of selected component (if applicable):
kernel-xen-2.6.18-1.2798.fc6

How reproducible:
100% reproducible

Steps to Reproduce:
1. In a terminal type "sleep 30"
2. Measure the actual sleeping time 
3.
  
Actual results:
80sec

Expected results:
30sec

Additional info:
This is a Toshiba A105-S361 laptop. The non Xen kernel works fine

Comment 1 Glauber Costa 2006-11-17 13:05:51 UTC
I cannot reproduce it on any of my systems. As it leaves me completely clueless
(and most probably any other), it's a good idea to post your dmesg output for
both boots, (xen and non-xen), as well as observing if there are any suspicious
messages delivered while the system is running.

Thanks

Comment 2 Ahmed Kamal 2006-11-24 18:33:13 UTC
Created attachment 142086 [details]
dmesg output on non-xen kernel

Comment 3 Ahmed Kamal 2006-11-24 18:35:03 UTC
Created attachment 142087 [details]
dmesg output on xen kernel

Comment 4 Ahmed Kamal 2006-11-24 18:36:05 UTC
Created attachment 142088 [details]
var/log/messages after booting my xen kernel

Comment 5 Ahmed Kamal 2006-11-24 18:40:10 UTC
ok I added the needed attachements. Please do let me know, if you need further
information (or devel testing on my hardware).
I took a look, and the interesting parts seem to be apic and the "Calibrating
Delay" line! ... Lots of other little differences as well! Didnt think they
would be that different on the same hardware.

Anyway, the problem still presists on the newer kernel (2.6.18-1.2849.fc6)


Comment 6 Glauber Costa 2006-11-27 16:24:30 UTC
Yes, this discrepancy between the xen and non-xen kernel is mostly normal. 
Next step is trying to find out what among it causes your problem ;-)

One theory is that bios may be changing the cpu frequency for energy saving, and
the xen hypervisor is not getting aware of it. Can you please check this, by
completely disabling it at bios level ? 


Comment 7 Ahmed Kamal 2006-11-28 19:46:04 UTC
Well, my BIOS had "CPU Frequency dynamically switchable" enabled, I disabled
that, and made it "CPU Frequency Always High".
Was so sure that was it, but unfortunately it didn't help!

Not that I know anything about that, but the line "Calibrating delay using timer
specific routine" seems too interesting and the numbers are different

Comment 8 Ahmed Kamal 2006-12-16 11:52:53 UTC
Update: Booting the Xen kernel with acpi=off, fixes the problem!!! It seems acpi
was changing the CPU frequency without telling the clocking system!
This way, it seems to me this problem is not really tied very much to my hardware!
I'd be happy to provide any testing you'd require, because a laptop without acpi
is not really useful


Comment 9 Red Hat Bugzilla 2007-07-25 01:34:51 UTC
change QA contact

Comment 10 Chris Lalancette 2008-02-26 23:37:38 UTC
This report targets FC6, which is now end-of-life.

Please re-test against Fedora 7 or later, and if the issue persists, open a new bug.

Thanks