Re: 2.6.30: hibernation/swsusp stuck in disable_nonboot_cpus

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Thursday 11 June 2009, Johannes Stezenbach wrote:
> Hi,
> 
> on my work machine susend-to-disk is broken in 2.6.30.  Last
> known working kernel was 2.6.29.1. I'm running a 32bit kernel.
> CPU is an AMD Athlon(tm) Dual Core Processor 4850e.
> Board is a Gigabyte GA-MA78GM-S2H rev 1.1.
> 
> - usually it stops on suspend after the message
>   "Suspending console(s) (use no_console_suspend to debug)"
> 
> - I booted with no_console_suspend using serial console
>   as recommended in Documentation/power/basic-pm-debugging.txt.
> 
>   No joy, apparently the serial driver is disabled during suspend
>   and no interesting output on serial console...
> 
> - no_console_suspend on VESA fb console worked better, last msg is
>   "Disabling non-boot CPUs ..."
> 
> - I stripped down my config, and after serveral retries
>   it _sometimes_ gets past the "Disabling non-boot CPUs ...".
>   Then the last lines are (penciled from screen):
> 
>       Enabling non-boot CPUs ...
>       SMP alternatives: switching to SMP code
>       Booting processor 1 APIC 0x1 ip 0x6000
>       Initializing CPU#1
>       Stuck ??
>       Error taking CPU#1 up: -5
> 
>   I also got past this point once, but then it hung at the
>   same spot on resume.
> 
> - I tried    echo 0 > /sys/devices/system/cpu/cpu1/online
> 
>      CPU 1 is now offline
>      SMP alternatives: switching to UP code
>      CPU0 attaching NULL sched-domain.
>      CPU1 attaching NULL sched-domain.
>      CPU0 attaching NULL sched-domain.
>      PM: Removing info for No Bus:msr1
>      PM: Removing info for No Bus:cpu1
> 
> - and     echo 1 > /sys/devices/system/cpu/cpu1/online
> 
>      PM: Adding info for No Bus:msr1
>      PM: Adding info for No Bus:cpu1
>      SMP alternatives: switching to SMP code
>      Booting processor 1 APIC 0x1 ip 0x6000
>      Initializing CPU#1
>      Calibrating delay using timer specific routine.. 5022.45 BogoMIPS (lpj=10044915)
>      CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
>      CPU: L2 Cache: 512K (64 bytes/line)
>      CPU: Physical Processor ID: 0
>      CPU: Processor Core ID: 1
>      Intel machine check architecture supported.
>      Intel machine check reporting enabled on CPU#1.
>      CPU1: AMD Athlon(tm) Dual Core Processor 4850e stepping 02
>      CPU0 attaching NULL sched-domain.
>      Switched to high resolution mode on CPU 1
>      CPU0 attaching sched-domain:
>       domain 0: span 0-1 level CPU
>        groups: 0 1
>      CPU1 attaching sched-domain:
>       domain 0: span 0-1 level CPU
>        groups: 1 0
>      Warning: Processor Platform Limit event detected, but not handled.
>      Consider compiling CPUfreq support into your kernel.
> 
> 
> Kernel config, dmesg and lspci below.  Is there anyting I could
> try besides git bisect?

Perhaps.  Please check if the kernel where the following commit is the head:

commit 2ed8d2b3a81bdbb0418301628ccdb008ac9f40b7
Author: Rafael J. Wysocki <rjw@sisk.pl>
Date:   Mon Mar 16 22:34:06 2009 +0100

    PM: Rework handling of interrupts during suspend-resume

is also broken.

Best,
Rafael