Re: [KVM timekeeping 13/35] Perform hardware_enable in CPU_STARTING callback

[Jan Kiszka <jan.kiszka@siemens.com>]

Zachary Amsden wrote:
> On 08/27/2010 06:32 AM, Jan Kiszka wrote:
>> Zachary Amsden wrote:
>>    
>>> The CPU_STARTING callback was added upstream with the intention
>>> of being used for KVM, specifically for the hardware enablement
>>> that must be done before we can run in hardware virt.  It had
>>> bugs on the x86_64 architecture at the time, where it was called
>>> after CPU_ONLINE.  The arches have since merged and the bug is
>>> gone.
>>>      
>> What bugs are you referring to, or since which kernel version is
>> CPU_STARTING usable for KVM? I need to encode this into kvm-kmod.
>>    
> 
> Prior to the x86_64 / i386 merge, CPU_STARTING didn't work the same way 
> / exist in the x86_64 code... most of this is historical guesswork.  At 
> some point, the 32/64 versions of the code in smpboot.c got merged and 
> now it does.
> 
> Binary searching around my tree shows this timeframe:
> 
> 2.6.11? - 2.6.23 : silver age ; i386 and x86_64 merge underway
>     |
> 2.6.24 : bronze age ; i386 and x86_64 deprecated
>     |
> 2.6.26 : iron age; smpboot_32.c / smpboot_64.c merge
>       \
>       2.6.28 : CPU_STARTING exists and first used
> 
> /me scratches head wondering how this affects operation on older kernels....

I basically need to revert your patch on host kernels without
CPU_STARTING and also on those where it might be broken. So I will set
the barrier to 2.6.28 then.

Thanks,
Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux