Bug 149183

Summary: powersave fails on syslog-ng
Product: [openSUSE] SUSE Linux 10.1 Reporter: Randy Dunlap <rdunlap>
Component: KernelAssignee: Pavel Machek <pavel>
Status: RESOLVED WORKSFORME QA Contact: E-mail List <qa-bugs>
Severity: Minor    
Priority: P5 - None    
Version: Beta 2   
Target Milestone: ---   
Hardware: i586   
OS: SLES 10   
Whiteboard:
Found By: Third Party Developer/Partner Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Randy Dunlap 2006-02-08 19:24:33 UTC 
syslog-ng is not stopped, then suspend fails, system comes back to life.
This is the default SMP installed kernel:
Linux version 2.6.16-rc1-git3-4-smp (geeko@buildhost) (gcc version 4.1.0 20060109 (prerelease) (SUSE Linux)) #1 SMP Tue Jan 24 16:43:20 UTC 2006

syslog during this time:
Freezing cpus ...
Breaking affinity for irq 57
Breaking affinity for irq 73
Breaking affinity for irq 105
Breaking affinity for irq 137
Breaking affinity for irq 145
Breaking affinity for irq 169
Breaking affinity for irq 201
Breaking affinity for irq 217
CPU 1 is now offline
migration_cost=2000
CPU1 is down
Stopping tasks: ================================================================================
 stopping tasks failed (1 tasks remaining)
Restarting tasks...<6> Strange, syslog-ng not stopped
 done
Thawing cpus ...
Booting processor 1/1 eip 3000
Initializing CPU#1
Comment 1 Randy Dunlap 2006-02-08 19:31:05 UTC 
Hm, powersave -U works on the second attempt (no reboots).
Did syslog-ng get killed but at the wrong point in time?
Comment 2 Greg Kroah-Hartman 2006-02-09 01:23:58 UTC 
Pavel, suspend issue?
Comment 3 Pavel Machek 2006-02-09 09:14:29 UTC 
Is it random failure or does it systematically break at first try and work on second try?

What hardware is that? (Just curious :-).
Comment 4 Forgotten User ZhJd0F0L3x 2006-02-09 13:47:23 UTC 
Holger did some tests on SMP (dualcore) so maybe he has seen something similar...
Comment 5 Randy Dunlap 2006-02-09 17:17:37 UTC 
I've seen it more than one time but it's not consistently repeatable.
IOW, powersave -U often (usually) succeeds.
Hardware is a Yonah (Core Duo) development board.
Comment 6 Pavel Machek 2006-02-10 20:18:43 UTC 
There are some improvements in recent -mm kernels. You could try if they help, but unless problem is reproducible I'd rather leave it as-is.
Comment 7 Pavel Machek 2006-03-03 20:18:34 UTC 
Randy, ping? Should I just mark it WORKSFORME or WONTFIX?
Comment 8 Pavel Machek 2006-03-27 09:39:44 UTC 
Randy is not responding, and I think this one is fixed.