powerpc/pseries/cmm: fix managed page counts when migrating pages between zones

Commit 63341ab037 (virtio-balloon: fix managed page counts when migrating
pages between zones) fixed a long existing BUG in the virtio-balloon
driver when pages would get migrated between zones.  I did not try to
reproduce on powerpc, but looking at the code, the same should apply to
powerpc/cmm ever since it started using the balloon compaction
infrastructure (luckily just recently).

In case we have to migrate a ballon page to a newpage of another zone, the
managed page count of both zones is wrong. Paired with memory offlining
(which will adjust the managed page count), we can trigger kernel crashes
and all kinds of different symptoms.

Fix it by properly adjusting the managed page count when migrating if
the zone changed.

We'll temporarily modify the totalram page count. If this ever becomes a
problem, we can fine tune by providing helpers that don't touch
the totalram pages (e.g., adjust_zone_managed_page_count()).

Fixes: fe030c9b85 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191216103058.4958-1-david@redhat.com
This commit is contained in:
David Hildenbrand 2019-12-16 11:30:58 +01:00 committed by Michael Ellerman
parent 0601546f23
commit e352f576d3

View File

@ -539,6 +539,16 @@ static int cmm_migratepage(struct balloon_dev_info *b_dev_info,
/* balloon page list reference */
get_page(newpage);
/*
* When we migrate a page to a different zone, we have to fixup the
* count of both involved zones as we adjusted the managed page count
* when inflating.
*/
if (page_zone(page) != page_zone(newpage)) {
adjust_managed_page_count(page, 1);
adjust_managed_page_count(newpage, -1);
}
spin_lock_irqsave(&b_dev_info->pages_lock, flags);
balloon_page_insert(b_dev_info, newpage);
balloon_page_delete(page);