When migrating a balloon page, we first deflate the old page to then inflate the new page.
However, if inflating the new page succeeded, we effectively deflated the old page, reducing the balloon size.
In that case, the migration actually worked: similar to migrating+ immediately deflating the new page. The old page will be freed back to the buddy.
Right now, the core will leave the page be marked as isolated (as we returned an error). When later trying to putback that page, we will run into the WARN_ON_ONCE() in balloon_page_putback().
That handling was changed in commit 3544c4faccb8 ("mm/balloon_compaction: stop using __ClearPageMovable()"); before that change, we would have tolerated that way of handling it.
To fix it, let's just return 0 in that case, making the core effectively just clear the "isolated" flag + freeing it back to the buddy as if the migration succeeded. Note that the new page will also get freed when the core puts the last reference.
Note that this also makes it all be more consistent: we will no longer unisolate the page in the balloon driver while keeping it marked as being isolated in migration core.
This was found by code inspection.
Fixes: 3544c4faccb8 ("mm/balloon_compaction: stop using __ClearPageMovable()") Cc: stable@vger.kernel.org Cc: Andrew Morton akpm@linux-foundation.org Cc: Jerrin Shaji George jerrin.shaji-george@broadcom.com Cc: Broadcom internal kernel review list bcm-kernel-feedback-list@broadcom.com Cc: Arnd Bergmann arnd@arndb.de Cc: Greg Kroah-Hartman gregkh@linuxfoundation.org Signed-off-by: David Hildenbrand david@redhat.com ---
I have no easy way to test this, and I assume it happens very very rarely (inflation during migration failing).
I would prefer this to go through the MM-tree, as I have some follow-up balloon_compaction reworks also mess with this code.
--- drivers/misc/vmw_balloon.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/drivers/misc/vmw_balloon.c b/drivers/misc/vmw_balloon.c index 6df51ee8db621..cc1d18b3df5ca 100644 --- a/drivers/misc/vmw_balloon.c +++ b/drivers/misc/vmw_balloon.c @@ -1737,7 +1737,7 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, { unsigned long status, flags; struct vmballoon *b; - int ret; + int ret = 0;
b = container_of(b_dev_info, struct vmballoon, b_dev_info);
@@ -1796,17 +1796,15 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, * A failure happened. While we can deflate the page we just * inflated, this deflation can also encounter an error. Instead * we will decrease the size of the balloon to reflect the - * change and report failure. + * change. */ atomic64_dec(&b->size); - ret = -EBUSY; } else { /* * Success. Take a reference for the page, and we will add it to * the list after acquiring the lock. */ get_page(newpage); - ret = 0; }
/* Update the balloon list under the @pages_lock */ @@ -1817,7 +1815,7 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, * If we succeed just insert it to the list and update the statistics * under the lock. */ - if (!ret) { + if (status == VMW_BALLOON_SUCCESS) { balloon_page_insert(&b->b_dev_info, newpage); __count_vm_event(BALLOON_MIGRATE); }
base-commit: 1c58c31dc83e39f4790ae778626b6b8b59bc0db8