Re: [PATCH] Resurect proper handling of maxcpus= kernel option

[Max Krasnyansky <maxk@qualcomm.com>]

Ingo Molnar wrote:
> * Max Krasnyansky <maxk@qualcomm.com> wrote:
> 
>>> This will need some test time on 32-bit as that is where this 
>>> represents a material change. ( albeit what matters most is the 
>>> maxcpus=1 distinction - and for that nosmp can be used as well to 
>>> turn off multi-cpu support altogether. So we could do this in 
>>> v2.6.27 as well. )
>> So far we got a couple of reports that it works as expected on 32 
>> (both laptop and server/desktop).
> 
> Yes, but the usecase i'm worried about is when say maxcpus=1 was used to 
> _prevent_ an SMP bootup - because the system would not work otherwise.
> 
> i guess we want to tickle those systems anyway as that case is not 
> supposed to happen (and it can always be totally disabled via nosmp or 
> noapic).
> 
> So i'm not against your fix/change per se, i just wanted to highlight 
> that it has some impact on existing uses of maxcpus that is outside of 
> your cpu-hotplug usecase.

I see what you mean. I think it's fairly safe though since we do not actually
do much for the cpus that are not going to be brought online. Mainly just
setting cpu_*_map and initializing per cpu areas. If something is broken in
there we'd probably want to fix that asap anyway. And like you said nosmp does
the job too.

Max