Re: [PATCH v3] mm/migrate_device: Try to handle swapcache pages

["Mika Penttilä" <mpenttil@redhat.com>]

Hi,

On 6.6.2023 10.46, David Hildenbrand wrote:
> On 06.06.23 09:44, David Hildenbrand wrote:
>> On 06.06.23 07:01, mpenttil@redhat.com wrote:
>>> From: Mika Penttilä <mpenttil@redhat.com>
>>>
>>> Migrating file pages and swapcache pages into device memory is not 
>>> supported.
>>> The decision is done based on page_mapping(). For now, swapcache 
>>> pages are not migrated.
>>>
>>> Things can however be improved, for swapcache pages. Try to get rid 
>>> of the swap cache,
>>> and if successful, go ahead as with other anonymous pages.
>>>
>>> Cc: Alistair Popple <apopple@nvidia.com>
>>> Cc: John Hubbard <jhubbard@nvidia.com>
>>> Cc: Ralph Campbell <rcampbell@nvidia.com>
>>> Cc: "Huang, Ying" <ying.huang@intel.com>
>>> Reviewed-by: "Huang, Ying" <ying.huang@intel.com>
>>> Reviewed-by: Alistair Popple <apopple@nvidia.com>
>>> Signed-off-by: Mika Penttilä <mpenttil@redhat.com>
>>> ---
>>>
>>> v3:
>>>     - Adjust comments
>>>     - Add Reviewed-bys
>>>
>>> v2:
>>>     - use folio_test_anon() (Huang, Ying)
>>>
>>>    mm/migrate_device.c | 22 ++++++++++++++++------
>>>    1 file changed, 16 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/mm/migrate_device.c b/mm/migrate_device.c
>>> index d30c9de60b0d..f76ebccfe067 100644
>>> --- a/mm/migrate_device.c
>>> +++ b/mm/migrate_device.c
>>> @@ -747,13 +747,23 @@ static void __migrate_device_pages(unsigned 
>>> long *src_pfns,
>>>            if (is_device_private_page(newpage) ||
>>>                is_device_coherent_page(newpage)) {
>>> -            /*
>>> -             * For now only support anonymous memory migrating to
>>> -             * device private or coherent memory.
>>> -             */
>>> +
>>>                if (mapping) {
>>> -                src_pfns[i] &= ~MIGRATE_PFN_MIGRATE;
>>> -                continue;
>>> +                struct folio *folio;
>>> +
>>> +                folio = page_folio(page);
>>> +
>>> +                /*
>>> +                 * For now only support anonymous memory migrating to
>>> +                 * device private or coherent memory.
>>> +                 *
>>> +                 * Try to get rid of swap cache if possible.
>>> +                 *
>>> +                 */
>>> +                if (!folio_test_anon(folio) || 
>>> !folio_free_swap(folio)) {
>>> +                    src_pfns[i] &= ~MIGRATE_PFN_MIGRATE;
>>> +                    continue;
>>
>> I'm pretty sure the folio has to be locked in order to call
>> folio_free_swap().
>>
> 
> Oh, staring at the bigger context, I assume we locked the folios via 
> migrate_device_range(), correct?
> 

Yes either that, or when migrating via virtual addresses 
migrate_vma_collect_pmd() has trylock_page() and we migrate only pages 
which succeed locking.

--Mika