Re: [GIT RFC] percpu: use dynamic percpu allocator as the default percpu allocator

[Tejun Heo <tj@kernel.org>]

Hello,

Martin Schwidefsky wrote:
> On Thu, 2 Apr 2009 11:24:18 +0400
> Ivan Kokshaysky <ink@jurassic.park.msu.ru> wrote:
> 
>> On the other hand, some tricks with DEFINE_PER_CPU() are indeed possible -
>> for instance, using weak references we could force the compiler to
>> generate proper addressing mode. So DEFINE_PER_CPU(int, foo) in module
>> would expand to something like this:
>>
>> 	extern int per_cpu__foo;
>> 	asm(".weakref per_cpu__foo, per_cpu_mod__foo");
>> 	__attribute__((__section__(".data.percpu"))) int per_cpu_mod__foo
>>
>> The main problem is that our DEFINE_PER_CPU() macro consists of more
>> than one definition, so it won't be possible to specify both storage class
>> and initializer with it.
>>
>> If it's acceptable to change the semantics from
>>
>> static DEFINE_PER_CPU(int, foo) = 1
>>
>> to
>>
>> DEFINE_PER_CPU(static, int, foo) = 1
>>
>> then we're ok.
>>
>> Or maybe just add STATIC_DEFINE_PER_CPU_xx() variants?
> 
> That is what I'm after as well. Just drop the "static" from the
> DEFINE_PER_CPU statement found inside modules and it works.
> 
> My experiments with the weak and visibility attribute failed because
> the static storage class specifier together with the attribute either
> causes a compile error or static just overrides the attribute.

Can STATIC_DEFINE_PER_CPU() be made to work?  It's not pretty but if
that's the only sensible way to reach uniform static/dynamic handling,
I suppose we can ignore the slight ugliness.

Rusty, Ingo, what do you guys think?

Thanks.

-- 
tejun