On 21.10.25 22:52, Michael S. Tsirkin wrote:

On Tue, Oct 21, 2025 at 02:59:12PM +0200, David Hildenbrand wrote:

quoted

In order to remove the dependency on the page lock for balloon
pages, we need a lock that is independent of the page.

It's crucial that we can handle the scenario where balloon deflation
(clearing page->private) can race with page isolation (using
page->private to obtain the balloon_dev_info where the lock currently
resides).

The current lock in balloon_dev_info is therefore not suitable.

Fortunately, we never really have more than a single balloon device
per VM, so we can just keep it simple and use a static lock to protect
all balloon devices.

Based on this change we will remove the dependency on the page lock
next.

Signed-off-by: David Hildenbrand <redacted>
---
  include/linux/balloon_compaction.h |  6 ++---
  mm/balloon_compaction.c            | 36 +++++++++++++++++-------------
  2 files changed, 23 insertions(+), 19 deletions(-)

diff --git a/include/linux/balloon_compaction.h b/include/linux/balloon_compaction.h
index 3109d3c43d306..e2d9eb40e1fbb 100644
--- a/include/linux/balloon_compaction.h
+++ b/include/linux/balloon_compaction.h

@@ -21,10 +21,10 @@
   *   i. Setting the PG_movable_ops flag and page->private with the following
   *	lock order
   *	    +-page_lock(page);
- *	      +--spin_lock_irq(&b_dev_info->pages_lock);
+ *	      +--spin_lock_irq(&balloon_pages_lock);
   *
   *  ii. isolation or dequeueing procedure must remove the page from balloon
- *      device page list under b_dev_info->pages_lock.
+ *      device page list under &balloon_pages_lock

Using &balloon_pages_lock with an & is kinda weird here.

Indeed, fixed, thanks!

-- 
Cheers

David / dhildenb