Re: [PATCH v5 1/3] mm/page_alloc: Optimize free_contig_range()

[Vlastimil Babka (SUSE)]

On 4/1/26 11:59, David Hildenbrand (Arm) wrote:
>>> +
>>> +	for (i = 0; i < nr_pages; i++) {
>>> +		bool can_free = true;
>>> +
>>> +		/*
>>> +		 * Contiguous PFNs might not have contiguous "struct pages"
>>> +		 * in some kernel configs: page++ across a section boundary
>>> +		 * is undefined. Use pfn_to_page() for each PFN.
>>> +		 */
>>> +		page = pfn_to_page(pfn + i);
>> 
>> Hm ideally we'd have some pfn+page iterator thingy that would just do a
>> page++ on configs where it's contiguous and this more expensive operation
>> otherwise. Wonder why we don't have it yet. But that's for a possible
>> followup, not required now.
> 
> pfn_to_page() is on relevant configs close to just a "page + i". Not
> entirely, but I am not sure if the micro-gain would really be worth it.
> 
> E.g., on CONFIG_SPARSEMEM_VMEMMAP
> 
> #define __pfn_to_page(pfn)	vmemmap + (pfn)

Oh I see. Yeah agree not worth it. The compiler can probably turn it to
page++ already then.

> 
> Acked-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
>