Re: [PATCH alt 13/25] mm/fbatch: no lru_add_drain() to collect_longterm_unpinnable_folios()

[David Hildenbrand (Arm)]

On 8/27/26 11:16, Hugh Dickins wrote:
> On Mon, 24 Aug 2026, David Hildenbrand (Arm) wrote:
>> On 8/24/26 16:55, Hugh Dickins wrote:
>>> collect_longterm_unpinnable_folios() has no use for lru_add_drain() nor
>>> lru_add_drain_all(), now that the per-cpu fbatch references are gone.
>>> So remove the recently added lru_cache_drain_for_folio().
>>>
>>> Signed-off-by: Hugh Dickins <hughd@xxxxxxxxxx>
>>> ---
>>> Alternate version for use once David Hildenbrand's two commits
>>> here have been pulled from mm.git into linux.git for 7.3-rc1.
>>
>> For 7.4 we expect another user of the helper in KVM guest_memfd code (the
>> primary reason why we introduce the helper), so depending on when this series is
>> ready it will just remove that user as well, or we'll need some coordination
>> between trees.
> 
> Yes, I could supply an alternate alternate version, providing an empty
> stub for lru_cache_drain_for_folio(), so Ackerley's and my series can
> proceed (or not) independently. Later. But at present your patch or
> patches seem to be in both mm.git and Ackerley's series (17/46),
> but not yet in Linus's tree.

They are in mm-stable, so they should go upstream this week. Anyhow, that's all
stuff that's easily resolved later.

-- 
Cheers,

David