rust_binder: schedule NodeDeath outside of node_refs lock

There's no reason to hold the node_refs lock while scheduling the
NodeDeath to the thread todo list, so don't. The call to set_cleared()
is kept under the lock so that the state update is kept atomic.

Reviewed-by: Matthew Maurer <mmaurer@google.com>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://patch.msgid.link/20260615-binder-noderefs-spin-v3-3-3235f5a3e0a0@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

1 files changed, 4 insertions, 1 deletions

diff --git a/drivers/android/binder/process.rs b/drivers/android/binder/process.rs
index 73a2582c4d9c..7b214dc51586 100644
--- a/drivers/android/binder/process.rs
+++ b/drivers/android/binder/process.rs
@@ -1289,7 +1289,10 @@ impl Process {
 
         // Update state and determine if we need to queue a work item. We only need to do it when
         // the node is not dead or if the user already completed the death notification.
-        if death.set_cleared(false) {
+        let should_schedule = death.set_cleared(false);
+        drop(refs);
+
+        if should_schedule {
             if let Some(death) = ListArc::try_from_arc_or_drop(death) {
                 let _ = thread.push_work_if_looper(death);
             }