Re: [PATCH v2 14/47] mm/huge_memory: fix zapping migration entries for device-private THPs

[David Hildenbrand (Arm)]

On 10/29/26 23:30, Gregory Price (Meta) wrote:
> On 2026-10-29 21:33 +0100, David Hildenbrand (Arm) wrote:
>> From: Chris Mason <mason@xxxxxxxxxx>
>>
>> A PMD migration entry can refer to a device-private folio.
>> zap_huge_pmd_folio() checks the folio type instead of the entry type,
>> so it treats such a migration entry as a device-private entry.
>>
>> Installing the migration entry already removed the reverse mapping and
>> dropped the mapping reference. Zapping it must not repeat those
>> operations. Concurrent MADV_DONTNEED during migration back to system
>> memory can trigger this.
>>
>> Check the PMD entry type using pmd_is_device_private_entry().
>>
>> Fix generated by an LLM and sent by Chris Mason in private.
>>
>> Fixes: 368076f52ebe ("mm/huge_memory: add device-private THP support to PMD operations")
>> Cc: stable@xxxxxxxxxxxxxxx
>> Signed-off-by: Chris Mason <mason@xxxxxxxxxx>
>> [ david: simplify description ]
>> Signed-off-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
> 
> One note below, but it's not a holdup.
> 
> Reviewed-by: Gregory Price (Meta) <gourry@xxxxxxxxxx>
> 
>> +	const bool is_device_private = !is_present && pmd_is_device_private_entry(pmdval);
> 
> pmd_is_device_private_entry() already returns false for a present
> pmdval - so !is_present is redundant.  If there was a reason to open
> code this, then keep it, but otherwise you could drop it.

Ah, softleaf_from_pmd() takes care of that. Yes, let me drop the is_present check.

(I'll double check that both are guaranteed to be consistent; but otherwise we
might have a bug already)

-- 
Cheers,

David