Re: [PATCH] mm/page_alloc: make calling prep_compound_head more reliable

[Miaohe Lin]

On 2022/6/15 20:42, Matthew Wilcox wrote:
> On Wed, Jun 15, 2022 at 03:44:06PM +0800, Miaohe Lin wrote:
>>> We definitely don't need the unlikely here.
>>
>> Could you please give me a more detailed explanation? IIUC, the above if condition
>> will only meet at a probability of 1/512. So unlikely tells the compiler to do some
>> optimization around it. Or am I miss something?
> 
> Only add unlikely() when the compiler can't figure out for itself that
> it's unlikely.  You should also check the generated code and/or
> benchmark the results to be sure that it's actually an improvement.
> Using unlikely() needs to be backed up with more than just a feeling.

I see. Many thanks for clarifying. Will keep it in mind. :)

> 
> .
>