On Tue, Jul 01, 2025 at 02:32:54PM +0200, David Hildenbrand wrote:

On 01.07.25 13:03, Lorenzo Stoakes wrote:

quoted

On Mon, Jun 30, 2025 at 02:59:58PM +0200, David Hildenbrand wrote:

quoted

Currently, we only support migration of individual movable_ops pages, so
we can not run into that.

Reviewed-by: Zi Yan <ziy@nvidia.com>
Signed-off-by: David Hildenbrand <redacted>

Seems sensible, so:

Reviewed-by: Lorenzo Stoakes <redacted>

Maybe worth adding a VM_WARN_ON_ONCE() just in case? Or do you think not worth it?

Not for now I think. Whoever wants to support compound pages has to fixup a
bunch of other stuff first, before running into that one here.

So a full audit of all paths that handle page_has_movable_ops() is required
either way.

IIRC there was an RFC series last year [1] that adds support for
order > 0 pages in zsmalloc.

Cc'ing Barry and Tangquan in case it's still on their TODO list... 

[1] https://lore.kernel.org/linux-mm/20241121222521.83458-2-21cnbao@gmail.com (local)

-- 
Cheers,
Harry / Hyeonggon