Re: [PATCH 1/2] mm: vmscan: remove folio_test_private() check in pageout()

[Baolin Wang <baolin.wang@linux.alibaba.com>]

On 2025/9/12 16:24, David Hildenbrand wrote:
> On 12.09.25 10:24, David Hildenbrand wrote:
>> On 12.09.25 05:45, Baolin Wang wrote:
>>> Currently, we no longer attempt to write back filesystem folios in 
>>> pageout(),
>>> and only tmpfs/shmem folios and anonymous swapcache folios can be 
>>> written back.
>>
>> Can you point me at the code where that is fenced off?

Please see the following check in pageout():

if (!shmem_mapping(mapping) && !folio_test_anon(folio))
	return PAGE_ACTIVATE;

>> I can spot a folio_is_file_lru() check before we call it, but the
>> description tells me that there are indeed ways we could still pass that
>> check for file-lru folios if we are kswapd.

Yes, but this also needs further cleanup, as kswapd also cannot reclaim 
filesystem dirty folios in pageout(). I plan to continue optimizing 
dirty file folios in isolate_lru_folios() to avoid some unnecessary scans.

>>> Moreover, tmpfs/shmem and swapcache folios do not use the PG_private 
>>> flag,
>>> which means no fs-private private data is used. Therefore, we can 
>>> remove the
>>> redundant folio_test_private() checks and related buffer_head release 
>>> logic.
>>
>> If that's indeed the case, do we still need the folio_test_private()
>> check in is_page_cache_freeable()?
> 
> Ah, that's patch #2 :)
>