Re: [PATCH 1/3 -mm] generic-ipi: fix hotplug_cfd()

[Li Zefan <lizf@cn.fujitsu.com>]

07:27, Andrew Morton wrote:
> On Wed, 29 Jul 2009 15:53:13 +0800
> Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> wrote:
> 
>> Use CONFIG_HOTPLUG_CPU, not CONFIG_CPU_HOTPLUG
>>
>> Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
>> ---
>>  kernel/smp.c |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/kernel/smp.c b/kernel/smp.c
>> index bf9f18b..1b5fd2e 100644
>> --- a/kernel/smp.c
>> +++ b/kernel/smp.c
>> @@ -54,7 +54,7 @@ hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu)
>>  			return NOTIFY_BAD;
>>  		break;
>>  
>> -#ifdef CONFIG_CPU_HOTPLUG
>> +#ifdef CONFIG_HOTPLUG_CPU
>>  	case CPU_UP_CANCELED:
>>  	case CPU_UP_CANCELED_FROZEN:
>>  
> 
> Dammit, that mistake is easy to make.  We should have used #if from day
> one, not #ifdef.  Oh well.
> 
> What's the impact of this bug?  Do we think the fix should be present
> in 2.6.31?  2.6.30.x?
> 

When hot-unpluging a cpu, it will leak memory allocated at cpu hotplug,
but only if CPUMASK_OFFSTACK=y, which is default to n, and I guess no
distro turns it on? All that said, I agree to add this patch to
-stable.

The bug is introduced by:

| commit 8969a5ede0f9e17da4b943712429aef2c9bcd82b
| Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
| Date:   Wed Feb 25 13:59:47 2009 +0100
|
|    generic-ipi: remove kmalloc()

So this patch can be applied to 2.6.29..2.6.31