Re: [PATCH] mm/page_alloc: add one PCP list for THP

[Mel Gorman <mgorman@techsingularity.net>]

On Wed, Jun 19, 2024 at 08:54:32PM +0800, yangge1116@126.com wrote:
> From: yangge <yangge1116@126.com>
> 
> Since commit 5d0a661d808f ("mm/page_alloc: use only one PCP list for
> THP-sized allocations") no longer differentiates the migration type
> of pages in THP-sized PCP list, it's possible that non-movable
> allocation requests may get a CMA page from the list, in some cases,
> it's not acceptable.
> 
> If a large number of CMA memory are configured in system (for
> example, the CMA memory accounts for 50% of the system memory),
> starting a virtual machine with device passthrough will get stuck.
> During starting the virtual machine, it will call
> pin_user_pages_remote(..., FOLL_LONGTERM, ...) to pin memory. Normally
> if a page is present and in CMA area, pin_user_pages_remote() will
> migrate the page from CMA area to non-CMA area because of
> FOLL_LONGTERM flag. But if non-movable allocation requests return
> CMA memory, migrate_longterm_unpinnable_pages() will migrate a CMA
> page to another CMA page, which will fail to pass the check in
> check_and_migrate_movable_pages() and cause migration endless.
> Call trace:
> pin_user_pages_remote
> --__gup_longterm_locked // endless loops in this function
> ----_get_user_pages_locked
> ----check_and_migrate_movable_pages
> ------migrate_longterm_unpinnable_pages
> --------alloc_migration_target
> 
> This problem will also have a negative impact on CMA itself. For
> example, when CMA is borrowed by THP, and we need to reclaim it
> through cma_alloc() or dma_alloc_coherent(), we must move those
> pages out to ensure CMA's users can retrieve that contigous memory.
> Currently, CMA's memory is occupied by non-movable pages, meaning
> we can't relocate them. As a result, cma_alloc() is more likely to
> fail.
> 
> To fix the problem above, we add one PCP list for THP, which will
> not introduce a new cacheline for struct per_cpu_pages. THP will
> have 2 PCP lists, one PCP list is used by MOVABLE allocation, and
> the other PCP list is used by UNMOVABLE allocation. MOVABLE
> allocation contains GPF_MOVABLE, and UNMOVABLE allocation contains
> GFP_UNMOVABLE and GFP_RECLAIMABLE.
> 
> Fixes: 5d0a661d808f ("mm/page_alloc: use only one PCP list for THP-sized allocations")
> Signed-off-by: yangge <yangge1116@126.com>

Too late to be relevant but

Acked-by: Mel Gorman <mgorman@techsingularity.net>

It would have been preferred if the comment stated why GFP_UNMOVABLE is
needed because the original assumption that THP would mostly be MOVABLE
allocations did not age well but git blame is enough. Maybe one day
GFP_RECLAIMABLE will also be added to the list. However, I suspect the
bigger problem later will be multiple THP sizes becoming more common that
are not necessarily fitting within PAGE_ALLOC_COSTLY_ORDER and per-cpu
not being enough to mitigate zone->lock contention in general. That's
beyond the scope of this patch.

Thanks.

-- 
Mel Gorman
SUSE Labs