On 14.08.19 17:41, David Hildenbrand wrote:
> Commit a9cd410a3d29 ("mm/page_alloc.c: memory hotplug: free pages as higher
> order") assumed that any PFN we get via memory resources is aligned to
> to MAX_ORDER - 1, I am not convinced that is always true. Let's play safe,
> check the alignment and fallback to single pages.
>
> Cc: Arun KS <arunks@xxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Oscar Salvador <osalvador@xxxxxxx>
> Cc: Michal Hocko <mhocko@xxxxxxxx>
> Cc: Pavel Tatashin <pasha.tatashin@xxxxxxxxxx>
> Cc: Dan Williams <dan.j.williams@xxxxxxxxx>
> Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
> ---
> mm/memory_hotplug.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 63b1775f7cf8..f245fb50ba7f 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -646,6 +646,9 @@ static int online_pages_range(unsigned long start_pfn, unsigned long nr_pages,
> */
> for (pfn = start_pfn; pfn < end_pfn; pfn += 1ul << order) {
> order = min(MAX_ORDER - 1, get_order(PFN_PHYS(end_pfn - pfn)));
> + /* __free_pages_core() wants pfns to be aligned to the order */
> + if (unlikely(!IS_ALIGNED(pfn, 1ul << order)))
> + order = 0;
> (*online_page_callback)(pfn_to_page(pfn), order);
> }
>
>
@Michal, if you insist, we can drop this patch. "break first and fix
later" is not part of my DNA :)
--
Thanks,
David / dhildenb
