Re: [PATCH 2/3] mm/rmap: could be partially_mapped only after no entire map

[David Hildenbrand <david@redhat.com>]

Hi,

please don't top-post.

On 16.08.25 12:05, Giorgi Tchankvetadze wrote:
> Since the compiler can prove the expression will always be false in one
> branch, there's no need to execute that computation at all in that
> branch, right?

It's mostly a matter of making the code easier to read. The way we have 
it right now, we calculate partially_mapped in all cases just before the 
end of the case.

For example, partially_mapped is irrelevant for !folio_test_anon() and 
!folio_test_large(), but we rely on the compiler to figure that out and 
optimize the code accordingly.

Having something like

if (folio_test_large(folio) && folio_test_anon(folio) &&
     !folio_test_partially_mapped(folio)) {
	switch (level) {
	case PGTABLE_LEVEL_PTE:
		if (IS_ENABLED(CONFIG_NO_PAGE_MAPCOUNT)) {
			...
			break;
		}
		partially_mapped = ...
	case PGTABLE_LEVEL_PMD:
		...
	}
	if (partially_mapped)
		deferred_split_folio(folio, true);
}

Could be done, but it's arguably harder to read.

> 
> On 8/16/2025 1:16 PM, David Hildenbrand wrote:
>> The compiler can figure out that it can move the whole statement to the
>> "if (last)" branch because it knows that nr == 0 otherwise and
>> partially_mapped == 0 already.
> 


-- 
Cheers

David / dhildenb